key: cord-293372-saqoft9p
authors: Heffner, Kelley; Hizal, Deniz Baycin; Majewska, Natalia I.; Kumar, Swetha; Dhara, Venkata Gayatri; Zhu, Jie; Bowen, Michael; Hatton, Diane; Yerganian, George; Yerganian, Athena; O’Meally, Robert; Cole, Robert; Betenbaugh, Michael
title: Expanded Chinese hamster organ and cell line proteomics profiling reveals tissue-specific functionalities
date: 2020-09-28
journal: Sci Rep
DOI: 10.1038/s41598-020-72959-8
sha: 
doc_id: 293372
cord_uid: saqoft9p

Chinese hamster ovary (CHO) cells are the predominant production vehicle for biotherapeutics. Quantitative proteomics data were obtained from two CHO cell lines (CHO-S and CHO DG44) and compared with seven Chinese hamster (Cricetulus griseus) tissues (brain, heart, kidney, liver, lung, ovary and spleen) by tandem mass tag (TMT) labeling followed by mass spectrometry, providing a comprehensive hamster tissue and cell line proteomics atlas. Of the 8470 unique proteins identified, high similarity was observed between CHO-S and CHO DG44 and included increases in proteins involved in DNA replication, cell cycle, RNA processing, and chromosome processing. Alternatively, gene ontology and pathway analysis in tissues indicated increased protein intensities related to important tissue functionalities. Proteins enriched in the brain included those involved in acidic amino acid metabolism, Golgi apparatus, and ion and phospholipid transport. The lung showed enrichment in proteins involved in BCAA catabolism, ROS metabolism, vesicle trafficking, and lipid synthesis while the ovary exhibited enrichments in extracellular matrix and adhesion proteins. The heart proteome included vasoconstriction, complement activation, and lipoprotein metabolism enrichments. These detailed comparisons of CHO cell lines and hamster tissues will enhance understanding of the relationship between proteins and tissue function and pinpoint potential pathways of biotechnological relevance for future cell engineering.

Proteomics sample preparation. Samples for proteomics were thawed on ice and lysed in a solution of 2% sodium dodecyl sulfate (SDS) in 500 µL of cell lysis buffer supplemented with 0.1 mM phenylmethylsulfonyl fluoride (PMSF) and 1 mM ethylenediaminetetraacetic acid (EDTA), pH 7-8. Lysates were sonicated x3 times for 60 s at 20% amplitude followed by a 90 s pause. Protein concentration was measured by a bicinchoninic acid (BCA) protein assay. One hundred micrograms of each sample were reduced in 10 mM tris(2-carboxyethyl) phosphine (TCEP), pH 7-8, at 60 °C for 1 h on a shaking platform. Iodoacetamide was added to alkylate the sample to a final concentration of approximately 17 mM and incubated for 30 min in the dark. Next, samples were passed through 10 kDa filters to reduce the SDS concentration as described in the filter-aided sample preparation (FASP) method 23 . During the FASP method, tetra-butyl ammonium bicarbonate (TEABC) was added after the urea washes to increase protein recoverability from the filters. The samples were finally digested using trypsin/LysC enzyme mix (Promega V507A) at an enzyme to substrate ratio of 1:10, overnight at 37 °C on a Scientific RepoRtS | (2020) 10:15841 | https://doi.org/10.1038/s41598-020-72959-8 www.nature.com/scientificreports/ shaking platform. After digestion, peptides were cleaned up by C18 cartridges and labeled with TMT reagents. All TMT labeled samples were combined and vacuum centrifuged to dryness removing the entire liquid.

Labeling. In order to compare protein expression, samples were labeled in duplicate (biological replicates)

using two TMT-10plex labeling kits (ThermoFisher Scientific, Waltham, MA, USA). Triplicates were used for ovary tissue and CHO-S. Each of the 10 reagents has the same nominal mass and chemical structure, so that for each sample a unique reporter mass (126-131 Da) was used to relate protein expression levels. Specifically, we included technical and biological replicates of CHO-S (two samples in TMT 1 and one sample in TMT 2) to aid in comparisons between and within the TMT experiments. All other samples contained only biological replicates, which were randomly assigned to one of the TMT-10plex kits. The TMT labeling design is provided in Table 1 . Following protein digestion, TMT reagents were thawed, and acetonitrile was used to dissolve the reagents. One reagent tube was added to each sample and then incubated at room temperature for 1 h. Hydroxylamine was subsequently added to quench the reaction before the tubes were combined. All TMT labeled samples were combined and vacuum centrifuged to dryness, removing the entire liquid. Each TMT-10plex was subjected to analysis by two-dimensional liquid chromatography tandem MS (2D LC/MS/MS). Protein identifications were made using Proteome Discoverer software with a high confidence cutoff [< 1% false discovery rate (FDR)]. The protein intensities were evaluated by fold change, using the CHO-S cell line technical replicate in each TMT as the basis. Protein accession numbers were mapped to gene symbols using the biological database network (https ://biodb net-abcc.ncicr f.gov) for functional analysis by gene ontology (GO). For GO annotation, gene symbols were mapped to biological processes, using the GO-CHO platform (https ://ebdru p.biosu stain .dtu.dk/gocho ). All programming for the hypergeometric test were calculated in MATLAB version 2015vB (https ://www.mathw orks.com/produ cts/matla b) and RStudio (https ://www.rstud io.com). Enrichment and depletion p-values were calculated using the hygecdf and hygepdf functions in MATLAB. These values were 

This study was undertaken to compare protein expression of various CHO cell lines and hamster tissues, resulting in the most comprehensive multi-tissue analysis of the Cricetulus griseus proteome (Fig. 1A) . This multi-tissue and multi-cell line analysis aims to improve our understanding of the Chinese hamster as the original tissue www.nature.com/scientificreports/ source for CHO cell lines. Additionally, since CHO is the dominant biopharmaceutical production host in biotechnology, this comparison elucidates similarities and differences across cells and tissues. An overview of the proteomics workflow is shown in Fig. 1B . Following MS identification, the protein accession numbers were determined using the annotated Chinese hamster genome 25 . Protein accession numbers were converted to gene symbols for functional analysis 26 . For missing gene symbols, the accession numbers were searched against mouse and human databases using the online database, bioDBnet 26 . In this study, gene ontology (GO) and ingenuity pathway analysis (IPA) were used for functional analyses of the differentially expressed proteins. GO analysis converts gene symbols to their molecular function, cellular component, and biological processes in order to evaluate the relationship between the molecular activities of gene products, location of activity, and pathways comprising the activity of multiple gene products, respectively 27 . In another functional analysis, each gene symbol was mapped to the relevant IPA pathways 28 , suggesting enriched and depleted pathways for comparisons between different tissues and cells. For both GO and IPA, the enrichment and depletion values were determined based on p-values that were calculated via the hypergeometric distribution, with p < 0.05 set for evaluating significance.

Protein identification and total protein intensity differences. PCA comparison. For each sample, the number of unique proteins and peptides identified are listed in Table 2 along with the number of spectra obtained. The complete list of identified proteins and fold change ratios are listed in Supplementary Table 1 and are searchable online through the NIST database (https ://pepti de.nist.gov). Over 6000 proteins were identified in each sample with at least two unique peptides per protein; this corresponded to a total of 8470 unique proteins containing a false discovery rate (FDR) of 1% for protein identification when the TMT samples were combined and the duplicates were removed ( Table 2) . Between the two TMT experiments, there were 4430 common proteins identified in both sets ( Fig. 2A) . A total of 2244 proteins were unique to TMT1 and 1796 proteins were unique to TMT2 ( Fig. 2A) .

Protein intensity fold change ratios were initially evaluated through principal component analysis (PCA) of proteins identified in all samples, as shown in Fig. 2B 29 . The PCA distribution represents differences in protein expression for those proteins identified in all samples and is influenced by proteins highly expressed in one tissue versus another. Not surprisingly, as shown by PCA, the cell samples clustered together and were distinct from all the tissues. For the tissue samples, the expression of proteins in spleen, liver, and heart clustered together. Similarly, protein expression in the lung and kidney clustered together. Interestingly, the first cluster (spleen, liver and heart) comprises tissues with dense connective tissue and capillary systems 30 . In contrast, the second cluster (lung and kidney) is specialized for transport, with tubular systems and thin insterstitium [30] [31] [32] [33] . Ovary and brain were clustered separately from the other organs.

Next, the two replicates from the same tissue were plotted to examine their consistency. Fitting a curve to the scatter plot and calculating the R 2 value assessed the linearity of the two replicates. Shown in Fig. 2C -F are the plots for the tissue replicates with the highest R 2 value for each cluster, specifically lung (from the lung/kidney cluster) and heart (from the heart/kidney/spleen cluster), plus brain and ovary. For the following analyses, we studied these representatives from each of the clusters; additional data on all proteins is provided in Supplementary Table 1 .

To compare between cell lines and tissues, protein intensity was averaged between replicates and plotted in Fig. 3 . The data was log2 transformed in order to ensure a normal distribution for each sample. A student's t-test was performed to determine the likelihood of significant differences between samples from cells and tissues. A summary of the statistical analysis, performed using student's t-test as a means comparison, is shown in Table 3 .

When comparing the cell lines to each other, no significant differences were observed between the CHO-S and CHO DG44 cell lines (Table 2 and Fig. 3A ) at a p-value of 0.05. We identified 178 proteins with significantly higher expression in CHO-S relative to CHO-DG44 and 155 proteins with significantly higher expression in CHO DG44 relative to CHO-S, representing the lowest number of total outliers for any comparison. A comparison between cells and tissues is shown in Fig. 3B through I. For these comparisons, the total number of proteins with low and high expression (fold change < 0.5 or > 2.0, respectively) is approximately 50%. For the cell to tissue comparison, all tissues show a statistically significant difference as compared to either the CHO-S or CHO DG44 cell lines. The proteins expressed at significantly higher levels in cells over tissues include proteins related to DNA replication, transcription, translation, and controlling cell apoptosis as expected to maintain rapid cell growth in exponential culture. Among the most highly expressed proteins in CHO-S and CHO DG44 are DNA-directed RNA polymerase II, eukaryotic translation initiation factor, histone H3.1t, general transcription factor 3C, 60S ribosomal protein, and apoptosis inhibitor 5. www.nature.com/scientificreports/ In comparison to ovary tissue (Fig. 3E ,I respectively, p < 0.01), CHO-S and CHO DG44 cells exhibited differences in expression patterns. This is somewhat surprising considering that CHO cells were derived from a mixture of ovary and the surrounding connective tissue. In Fig. 3E ,I, the proteins colored blue and yellow represent those with higher expression in ovary or CHO cells, respectively. In addition to ovary, statistical differences in protein expression were observed for CHO-S and CHO DG44 cells and to lung tissue (Fig. 3D,H, p < 0.01) . Similar to the comparison against ovary, an increase in upregulated proteins with higher expression was observed in the lung tissue in comparison to the cell lines. This suggests that tissue-specific functions may contribute to differences in expression patterns with cells regardless of the type of tissue.

Next, both CHO cell lines show statistically significant differences in outliers as compared against heart tissue (Fig. 3C ,G, p < 0.01). Similar to the ovary and lung comparison, there are a greater number of proteins with higher expression in the heart tissue when compared to cell lines. Over 50% of the cells in the heart are cardiac fibroblasts, which contributes to the specificity of this organ 34 . In addition, the heart has endothelial, smooth muscle, and pacemaker cells. This high degree of specialization is likely influential for the differences in proteins between CHO and heart. Finally, both CHO-S and CHO DG44 also show a difference in expression www.nature.com/scientificreports/ when compared to the brain, which is also likely related to the high degree of specialization required for brain cells such as neurons and glia (Fig. 3B ,F, p < 0.01). Amongst the tissues, wide differential regulation, both upregulation and downregulation, can be observed when comparing tissues against each other (Fig. 3J-O) . There is a greater difference in terms of the total percentage of outliers in the brain versus heart comparison (~ 62% total outliers for brain and ~ 51% total outliers for heart). Interestingly, only brain and heart tissues were found to have a statistically significant difference between each other (Table 3 ). All other tissue to tissue comparisons were found to have insignificant differences in the percentage of outliers. One reason for the relatively high number of outliers in brain tissue may be attributed to the distinct separation in terms of embryonic development from the other organs. The brain originates from the ectoderm whereas the circulatory system (heart), epithelial layer of the lungs, and ovary develop from the Table 2 ) are provided in the appendix. We also examined some of the top upregulated proteins in each tissue. Hierarchical clustering of protein expression is depicted in the center plot of Fig. 4 in which the color pink represents highly expressed proteins Table 3 . Means comparison using Student's t-test for total percentage of outliers. *p-value < 0.05 indicates that the outlier comparison is statistically significant. Coloring is shown from low (green) to high (pink) abundance. Distinct clusters are shown for brain, lung, heart, and ovary tissues. From top to bottom: the protein functions exhibiting high expression for each tissue are shown relative to other tissues for ovary (A), brain (B), heart (C), and lung (D) plotted using Genesis software. www.nature.com/scientificreports/ for each specific tissue. For each sample, the top 200 upregulated proteins, corresponding to approximately 3% of the total proteins in a tissue, were identified in order to highlight tissue specificity (Supplementary Table 3 ). For example, disintegrin and metalloproteinase domain-containing proteins (ADAM7, 10, 15) were highly upregulated in hamster ovary (Supplementary Table 3 ). Indeed, ADAM 10 can control follicle formation by regulating the recruitment of ovarian follicle supporting cells 36 . Disintegrin would be useful for control of the extracellular matrix as agrees with results from the human ovary-specific proteome 37 . Similarly, highly upregulated brain proteins include amyloid beta A4 protein, neuron navigator 1 (NAV1), calcium and integrin binding protein, and serine/threonine protein kinase (Supplementary Table 3 ). NAV1 is specifically targeted to the nervous system 38 .

In contrast, the heart is a strong muscle that must contract continually and is predominantly composed of cardiomyocytes and fibroblasts. The human heart tissue proteome was found to have 201 upregulated proteins 37 , including retinol dehydrogenase (RDH1) which was also significantly upregulated in our hamster heart proteome (Supplementary Table 3 ). Previous studies have indicated that the knock-down of RDH1 led to abnormal neural crest cell migration and an abnormal heart loop in mutant embryos, signaling its importance in heart tissue 39 . We also observed high expression of proteins including pleckstrin homology domain-containing family F member 1-like, actin-related protein, glutathione S-transferase, protein O-glycosyltransferase, and Ras GTPase-activating protein (Supplementary Table 3) .

We also identified the top 200 upregulated proteins in hamster lung tissue (Supplementary Table 3 ). Examples include branched chain aminotransferase, glycine N-methyltransferase, integrin alpha-6, and vesicle-associated membrane protein, indicating the importance in the lung of key metabolic and membrane process. As a comparison, the human proteome identified 183 genes highly expressed in the lung including similar membrane and secretory proteins 37 . Indeed, the lung is particularly adept at expressing membrane and secreted proteins such as surfactants and solute carrier proteins 40 . Analysis of the most abundant proteins for each tissue or cell line can, at least in some cases, relate to key function and roles of specific tissues and provide targets of opportunity for genetic engineering of CHO cell lines.

IPA pathway analyses. Next, the proteins were annotated with gene symbols in order to perform pathway analysis. Proteins with fold change values of less than 0.5 or greater than 2.0 were used for each comparison in order to determine downregulation and upregulation of pathways in the IPA software. As shown in Fig. 4 , protein intensity hierarchical clustering was used to identify proteins with significantly higher expression in a specific tissue.

Protein functions identified in IPA that are enriched in the brain include G1/S phase arrest, metabolism of acidic amino acids, and organization of the Golgi apparatus, and transport of phospholipids (Fig. 4B) . The Golgi plays a central role in cholesterol and other lipid metabolism; almost 25% of the human body's unesterified cholesterol is present in the brain 41 . Additionally, glycosphingolipids are abundant in the nervous system, and are synthesized in the endoplasmic reticulum and completed in the Golgi apparatus 42 . Acidic amino acids, including glutamate and aspartate, serve important signaling functions in the brain, so an increase in their metabolic activity would be expected, especially since these metabolites are not readily obtained from the diet 43, 44 .

Pathways enriched in the lung include catabolism of branched chain amino acids (BCAAs), metabolism of reactive oxygen species (ROS), trafficking of vesicles, and synthesis of lipids (Fig. 4D) . The lung shows high secretory capacity and thus trafficking of the secretory products through vesicles is important 45 . Furthermore, the lung is the source of surfactants, composed of 90% lipids; thus, lipid synthesis will be an important component of its function 46, 47 . The lungs are also particularly sensitive to hypoxia occurring at high altitude. The lung responds to these hypoxic conditions through signaling, including the release of ROS species to trigger hypoxia-inducible factor (HIF) 48 . Thus, the capacity for ROS metabolism may be critical for lung function and adaptation to changes in oxygen levels in different environments.

Similarly, in the heart, enriched pathways include metabolism of alpha-amino acid, sorting of protein, and translation (Fig. 4C ). Amino acid metabolism was studied in rat heart with amino acid levels rising up to fivefold higher than plasma levels 49 , accompanied by increases in ribosome activity and translation 50 . Finally, protein functions enriched in the ovary include development of extracellular matrix, docking of vesicles, and exocytosis (Fig. 4A) . The matrix of ovary tissue is important for numerous physiological activities, including growth, migration, and differentiation, and the composition changes during ovulation 31 , critical to the fertility process 51 . Further, malfunctions of the matrix are observed during ovarian cancer 52 . High levels of extracellular matrix components in Chinese hamster ovaries are in agreement with our previous research examining the hamster ovary tissue proteome using label-free proteomics approaches 12 .

For GO functional analysis, gene symbols were annotated for biological processes in order to group proteins with similar biological relevance. The biological process GO category was analyzed to determine enrichment and depletion p-values for the cell-to-tissue and tissue-to-tissue comparisons. The appendix lists the top 10 most enriched biological processes for CHO-S (Supplementary Table 4 ) and CHO DG44 (Supplementary Table 5 ) comparisons with tissue. Enrichment is determined by hypergeometric distribution, with a p-value of < 0.05 used for significance.

Biological processes, representing a biological function involving the gene product, complements the pathway analysis in "IPA pathway analyses" shown above 27 . In both CHO-S and CHO DG44, the most common biological processes enriched involve DNA and mRNA processing, and metabolism. Some of the most common enriched processes in CHO-S include mRNA processing and splicing, DNA replication and repair, transcription, cell cycle, chromatin modification, and chromosome condensation.

In comparison, signaling, transport, and adhesion were significantly enriched across the different hamster tissues. Enriched brain biological processes relative to the CHO cell lines include ion transport, axon guidance, www.nature.com/scientificreports/ synaptic transmission, and metabolic process, among others 27 . Ion transport helps to maintain the stability of cerebral function due to the key roles that ions play in currents and synaptic transmission 53 . Abnormal distributions of ions in the brain can lead to defects in neuronal function including seizures and depression. Enriched biological processes in the heart compared to the CHO cell lines show functions related to circulation and heart function such as vasoconstriction, sodium-independent organic ion transport, complement activation, blood coagulation, and lipoprotein metabolism. For example, complement activation involves mannosebinding lectin and complement components, C3, C5, and CD59. The complement cascade can be activated during heart disease or failure, especially in cases of myocardial ischemia and reperfusion 54, 55 . Biological processes that are enriched in the lung as compared to the CHO cell lines include G-protein coupled receptor signaling pathway, innate immune response, vesicle-mediated and transmembrane transport, and signal transduction. Indeed, vesicle transport is an important component of secretory pathway machinery. Unraveling the complexities of lung secretions may yield new insights into ways by which secretion differs in tissues and cell lines. Signaling through G-protein coupled receptors and other pathways is an important proinflammatory response in lungs, which can undergo modifications leading to lung cancers 56, 57 . Not surprisingly, protection of the lungs using the host innate immune response is critical as this organ is exposed to a variety of pathogens including bacterial, fungal, and viral, during breathing 58, 59 .

Finally, enriched biological processes in the ovary highlight differences between the cells and tissue around the region from which CHO cells were derived. Enriched ovarian biological processes include transmembrane transport, protein transport, vesicle-mediated transport, cell adhesion, and cell death. When vesicle production rate was quantified in the ovary, the turnover indicated high vesicle recycling across the endomembrane system 59 , which is consistent with these proteomic results. Cell adhesion is also important to ovary function and follicle maturation through interactions with the extracellular matrix and direct cell-cell contacts. Mutation of the ovarian surface is causative of approximately 90% of malignant ovarian tumors 60,61 .

The results from the Chinese hamster proteome provide new insights into global protein expression across a wide variety of tissues and multiple cell lines. These differences highlight the role of tissues in executing key organ functions which require a specific metabolic processes, such as transport and communication, in comparison to CHO cells, which are focused on replication and gene expression, characteristics useful for rapid growth and the production of biologics. Because of their relevance to biomedicine and the biotechnology industry, we compared the tissue proteome to CHO cell lines and each other in order to identify functional differences in expression across tissues and cell lines. Specifically, we observed enrichment of many physiological pathways in tissues that were not enriched in cells, such as ion, protein, and vesicle transport, signal transduction, and cell adhesion. Often, these differences correlated with specific tissue functions while the activities in cell lines were often correlated with DNA replication, cell cycle, or RNA processing. Furthermore, some of the proteins with high expression in lung, ovary, or other tissues versus CHO, such as vesicle-mediated and protein transport, provide significant opportunities for CHO cell engineering going forward. In this way, the study expands on our CHO and Chinese hamster tissue knowledge base by virtue of establishing an atlas to differentiate proteins across cells and tissue for this critically important biotechnological and biomedical host species. Indeed, this comparison has enabled us to appreciate the changing proteomic landscape across cells and tissues and furthermore to recognize how the expressed proteins from different cell types can represent signatures for some of their key physiological or biotechnological functions.

The complete list of identified proteins and fold change ratios are listed in Supplementary Table 1 

Life cycle analysis of mammalian cells: II. Cells from the Chinese hamster ovary grown in suspension culture

CHO quasispecies-Implications for manufacturing processes

Recombinant antibody production in CHO and NS0 cells: Differences and similarities

Isolation of Chinese hamster cell mutants deficient in dihydrofolate reductase activity

Deletion of the diploid dihydrofolate reductase locus from cultured mammalian cells

The genomic sequence of the Chinese hamster ovary (CHO)-K1 cell line

Genomic landscapes of Chinese hamster ovary cell lines as revealed by the Cricetulus griseus draft genome

Sequencing the CHO DXB11 genome reveals regional variations in genomic stability and haploidy

Construction of BAC-based physical map and analysis of chromosome rearrangement in Chinese hamster ovary cell lines

A reference genome of the Chinese hamster based on a hybrid assembly strategy

Proteomic analysis of Chinese hamster ovary cells

Lessons from the Hamster: Cricetulus griseus tissue and CHO cell line proteome comparison

Elucidation of the CHO super-Ome (CHO-SO) by proteoinformatics

Animal Cell Culture

Integrated miRNA, mRNA and protein expression analysis reveals the role of post-transcriptional regulation in controlling CHO cell growth rate

A deep proteome and transcriptome abundance atlas of 29 healthy human tissues

Initial quantitative proteomic map of 28 mouse tissues using the SILAC mouse

The Chinese hamster as a model for the study of diabetes mellitus

Insulin and glucagon response of the diabetic Chinese hamster in the Asahikawa colony

Effects of neonatal ovarian X-irradiation in the Chinese hamster. I. Correlation between the age of irradiation and the fertility span

Specific aneusomies in Chinese hamster cells at different stages of neoplastic transformation, initiated by nitrosomethylurea

Mice, hamsters, ferrets, monkeys. Which lab animals can help defeat the new coronavirus

Universal sample preparation method for proteome analysis

Genesis: Cluster analysis of microarray data

An evaluation of public genomic references for mapping RNA-Seq data from Chinese hamster ovary cells

The biological database network

Gene ontology: Tool for the unification of biology

Causal analysis approaches in Ingenuity Pathway Analysis

Principal component analysis of proteomics (PCAP) as a tool to direct metabolic engineering

Extracellular matrix of the developing ovarian follicle

Decidual cells in the human ovary at term. I. Incidence, gross anatomy and ultrastructural features of merocrine secretion

Extracellular matrix of ovarian tumors

Mending broken hearts: Cardiac development as a basis for adult heart regeneration and repair

Differentiation of human embryonic stem cells into embryoid bodies compromising the three embryonic germ layers

ADAM10-Notch signaling governs the recruitment of ovarian pregranulosa cells and controls folliculogenesis in mice

Tissue-based map of the human proteome

Mouse neuron navigator 1, a novel microtubule-associated protein involved in neuronal migration

Retinol dehydrogenase, RDH1l, is essential for the heart development and cardiac performance in zebrafish

Site-specific differences in gene expression of secreted proteins in the mouse lung: Comparison of methods to show differences by location

Cholesterol metabolism in the brain

The role of glycosphingolipid metabolism in the developing brain

Dietary amino acids and brain function

Amino acid precursors of monoamine neurotransmitters and some factors influencing their supply to the brain

Mass spectrometry-based secretome analysis of non-small cell lung cancer cell lines

Intermediary metabolism of the lung

The role of lipids in pulmonary surfactant

Lung cell hypoxia: Role of mitochondrial reactive oxygen species signaling in triggering responses

Regulation of protein synthesis in heart muscle. I. Effect of amino acid levels on protein synthesis

Characterization of the human heart mitochondrial proteome

The roles of the ovarian extracellular matrix in fertility

Alterations of the extracellular matrix in ovarian cancer studied by Second Harmonic Generation imaging microscopy

Ion regulation in the brain: implications for pathophysiology

Complement activation in heart diseases. Role of oxidants

Effects of complement activation in the isolated heart. Role of the terminal complement components

Common pathways for activation of proinflammatory gene expression by G protein-coupled receptors in primary lung epithelial and endothelial cells

Pathogenesis of lung cancer signalling pathways: Roadmap for therapies

Innate immunity in the lungs

Estimation of Golgi membrane flow rates in ovary glands of aptenia cordifolia using cytochalasin B

Cell-cell adhesion in the normal ovary and ovarian tumors of epithelial origin; An exception to the rule

Cell biology of human ovarian surface epithelial cells and ovarian carcinogenesis

The authors acknowledge the financial support of AstraZeneca for funding this project. In addition, this material is based upon work supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. DGE-1232825 and Grant No. DGE-1746891. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

The authors declare no competing interests.

Supplementary information is available for this paper at https ://doi.org/10.1038/s4159 8-020-72959 -8.Correspondence and requests for materials should be addressed to M.B.Reprints and permissions information is available at www.nature.com/reprints.Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.