195Archivio Italiano di Urologia e Andrologia 2016; 88, 3 ORIGINAL PAPER In the search of novel urine biomarkers for the early diagnosis of prostate cancer. Intracellular or secreted proteins as the target group? Where and how to search for possible biomarkers useful in the everyday clinical practice Amalia Katafigioti 1, Ioannis Katafigiotis 1, Stavros Sfoungaristos 2, Christos Alamanis 1, Konstantinos Stravodimos 1, Ioannis Anastasiou 1, Eleni Roumelioti 1, Mordechai Duvdevani 2, Constantinos Constantinides 1 1 1st University Urology Clinic, Laiko Hospital, Athens, Greece; 2 Hebrew Hadassah University Medical Center, Jerusalem, Israel. Objective: To search which category of proteins can be detected in urine in order to examine subsequently its ability to improve our accuracy for the diagnosis of Prostate Cancer (PCa) as biomarkers in clinical useful fluids like urine and serum. Material and method(s): Urine samples of 127 patients were obtained after a vigorous transrectal prostatic massage to both lobes. The patients were considered to have a high risk for PCa according to their PSA (> 4 ng/ml), their digital rec- tal examination (DRE) (positive for suspicious prostatic lesions) or to their abnormal PSA kinetics (PSA velocity (PSAV > 0.75 ng/mL). All patients subsequently were sub- jected to an extended 10-core per prostatic lobe TRUS-b (total 20 prostatic samples). The proteins that were chosen to be detected in the urine samples with Western-blot, as possible biomarkers, were Glutathione peroxidase 3 precur- sor (GPx3), Cofilin-1 (CFL1), Heat shock protein-90β (HSP 90β), Zinc alpha 2-glycoprotein (ZAG) and secreted protein acidic and rich in cysteine (SPARC).These proteins have been detected previously in the prostatic tissue by proteomics proving their discriminative ability between patients with prostate cancer and benign prostatic hyperplasia. Result(s): From the five proteins, only the secreted Zinc alpha 2-glycoprotein was detected in urine showing a prom- ising ability in the improvement of our diagnostic accuracy for the early diagnosis of prostate cancer. Conclusions: From various categories of proteins that have already been detected in the tissue of prostate by pro- teomics, only secreted protein Zinc alpha 2-glycoprotein showed a clear signal in the urine, proving its discriminative potential for the early diagnosis of PCa. KEY WORDS: Prostate cancer; Biomarker; Urine; Zinc alpha 2-glycoprotein. Submitted 14 February 2016; Accepted 8 May 2016 Summary No conflict of interest declared. rise of PCa incidence over the last 20 years has been attributed also to the PSA (2). Nevertheless, PSA is a test specific for the prostate but not specific for cancer and it can be elevated due to benign prostatic hyperplasia (BPH) and prostatitis. While the conventional PSA threshold of 4 ng/ml has been used in the previous years as a recommendation for transrectal ultrasound guided biopsy (TRUS-b), today PSA is considered a continuous parameter with higher levels of the test rising the likeli- hood of PCa (3). PCa can be present despite low PSA val- ues and even aggressive PCa with Gleason score > 7 can be diagnosed with low serum PSA values, precluding the definition of an optimal PSA threshold for the early diag- nosis of PSA and constituting the urgent need of more specific biomarkers for PCa (3, 4). In our study we tried to examine prospectively, various proteins in the urine of patients with suspect of PCa according to their PSA values or their digital rectum examination (DRE), who subsequently were subjected to TRUS-b. Five proteins were meticulously chosen. Four were previously detected by the use of proteomics in the tissue of patients with (BPH) and PCa helping to differ- entiate these two diseases, while the fifth was chosen due to its proven utility as a secreted protein in bladder can- cer (5, 6). At the same time we classified the proteins in two major categories intracellular and secreted. Our goal was to see which category of proteins can be detected in urine in order to examine subsequently its ability to help us improve our diagnostic accuracy for the PCa as bio- marker in clinical useful fluids like urine and serum. MATERIALS AND METHODS Urine samples of 127 patients were obtained after a vig- orous transrectal prostatic massage to both lobes and they were immediately stored to -80˚C. The patients were considered to have high risk for PCa according to their PSA (> 4 ng/ml), their digital rectal examination (DRE) (positive for suspicious prostatic lesions) or to their abnormal PSA kinetics (PSA velocity (PSAV > 0.75 DOI: 10.4081/aiua.2016.3.195 INTRODUCTION Prostate cancer (PCa) constitutes the most common can- cer and the second cause of cancer specific death in men (1). The widespread use of Prostate Specific Antigen (PSA) test has revolutionized PCa diagnosis. In fact the Katafigioti nuova versione_Stesura Seveso 21/09/16 08:51 Pagina 195 Archivio Italiano di Urologia e Andrologia 2016; 88, 3 A. Katafigioti, I. Katafigiotis, S. Sfoungaristos, C. Alamanis, K. Stravodimos, I. Anastasiou, E. Roumelioti, M. Duvdevani, C. Constantinides 196 ng/mL). All patients subsequently were subjected to an extended 10-core per prostatic lobe TRUS-b (total 20 prostatic samples). The specimen was analyzed in a blinded fashion by an experienced uropathologist (K.P.) and the results of the histology were collected. Exclusion criteria were patients with known PCa, PSA > 25 ng/ml, patients taking finasteride or dutasteride, patients who had been previously subjected to a TRUS-b or to an operation for BPH and patients with rectal extirpation. Patients with PSA > 25 ng/ml were excluded due to the fact that as PSA rises, the probability of PCa also increas- es and the goal was to examine specific the patients that could cause diagnostic dilemma to the urologist and how urine biomarkers could help to the differential diagnosis and no these that Pca diagnosis was higher based to the PSA. Nevertheless PSA > 25 ng/ml was a random choice based on the higher probability of PCa. The study was approved by the ethical committee of the Athens University Medical School and LAIKO Hospital and all the patients signed a written consent after a detailed infor- mation. Urine samples-proteins-Western blot The proteins that was chosen to be detected in the urine samples as possible biomarkers were Glutathione peroxi- dase 3 precursor (GPx3), Cofilin-1 (CFL1), Heat shock protein-90β (HSP 90β), Zinc alpha 2-glycoprotein (ZAG) and secreted protein acidic and rich in cysteine (SPARC). GPx3, CFL1, HSP 90β, ZAG, were chosen based on their performance in the tissue with the use of quantitative pro- teomics analysis in a previous study of ours, where all these four proteins showed different levels between BPH tissue (obtained from suprapubic prostatectomy or transurethral prostatectomy) and PCa tissue (obtained from radical retropubic prostatectomy) 5 (Table 1). Specifically, GPx3 was up-regulated (2.19 ± 0.57 fold change) in PCa compared to BPH tissue, CFL1 was also up-regulated (3.29 ± 1.20 fold change) in PCa compared to BPH tissue, HSP 90β was also up-regulated (3.2 ± 0.61 fold change) in PCa compared to BPH tissue, while ZAG was down-regulated (0.43 ± 0.13 fold change) in PCa compared to BPH5 (Table 1). The fifth protein SPARC was detected from another study of our team, where this pro- tein was found to be differentially expressed at statistical- ly significant levels in the secretome of a cell line model for aggressive bladder cancer, and due its performance and its secreted nature was also chosen in order to examine its possible utility in PCa (6) (Table 1). Sample preparation Urine samples were thawed on ice. Trichloroacetic acid (TCA) with the carrier sodium lauroyl sarcosinate (NLS) precipitation was conducted as followed: 0.1% NLS - 7.5% TCA was added to every sample. Samples were vor- texed and incubated at -20°C o/n. Samples were thawed and centrifuged at 10.000g for 10 min at 4°C. Protein pel- let was washed with ice cold Tetrahydrofuran (THF) and centrifuged at 10.000g for 10 min at 4°C. Washing step was repeated one more time. Pellet was resuspended in lysis buffer (7M Urea, 2M Thiourea, 4% CHAPS, 1% DTE, 2% IPG). Bath sonication was performed for 30 min and finally samples were centrifuged at 10.000g for 10 min RT; 3.6% protease inhibitors (Roche) were added to the supernatant and stored at -20°C until used. Western blot analysis Total protein (20 μg) of urine samples were separated by 10% SDS-PAGE under reducing conditions and elec- troblotted to Hybond-ECL nitrocellulose membrane (Amersham Biosciences). After blocking with 5% non-fat dried milk in TBST (20 mM Tris/pH 7.6, 137 mM NaCl, 0.1% Tween 20) for 2 h at room temperature, membranes were washed with TBST and incubated overnight at 4 °C with the primary antibodies, as applicable: mouse antihu- man SPARC (Santa Cruz; dilution 1:500), mouse antihu- man ZAG (Santa Cruz; dilution 1:1000), mouse antihuman cofilin (Santa Cruz; dilution 1:2000), mouse antihuman GPx-3 (Santa Cruz; dilution 1:1000), goat antihuman HSP90b (Santa Cruz; dilution 1:3000). Membranes were then washed with TBST and incubated with antimouse or antigoat HRP-conjugated secondary antibody (Santa Cruz; dilution 1:10 000) for 2 h at room temperature. A final wash with TBST was made and target protein was detect- ed by Enhanced Chemiluminescence (Perkin-Elmer LAS, Inc.) detection system. Films were scanned and images were analyzed using Quantity One software (Bio Rad). RESULTS Sample baseline characteristics are presented in Table 2. Proteins Nature Performance in tissue between BPH Detection in urine Explanation and PCa or other utility GPx3 Secreted Up-regulated Detected in few samples Extensive unspecific binding (2.19 ± 0.57 fold change) in PCa of the antibody CFL1 Intracellular Up-regulated Not detected both in normal Intracellular (3.29 ± 1.20 fold change) in PCa and prostate cancer samples HSP 90β Intracellular Up-regulated Not detected both in normal Albumin masking effects (3.2 ± 0.61 fold change) in PCa and prostate cancer samples SPARC Secreted Expressed at statistically significant Not detected both in normal Intracellular levels in aggressive bladder cancer and prostate cancer samples ZAG Secreted Down-regulated Significant increase in patients Secreted (0.43 ± 0.13 fold change) in PCa with positive histology for PCa Table 1. Proteins of the study. Katafigioti nuova versione_Stesura Seveso 21/09/16 08:51 Pagina 196 GPx-3 A representative image of western blot analysis of urine samples with Anti-GPx-3 (MW: 23 kDa) is depicted in Figure 1. Glutathione peroxidase 3 is a secreted protein that protects cells and enzymes from oxidative damage, by catalyzing the reduction of hydrogen peroxide, lipid peroxides and organic hydroperoxide, by glutathione (7). This protein has been previously found to be up-reg- ulated (2.19 ± 0.57 fold change) in PCa compared to BPH tissue specimens by a quantitative proteomics analysis (5) (Table 1). GPx-3 was detected in a few sam- ples (79, 80, 84). This was mainly attributed due to the extensive unspecific binding of the antibody as shown in Figure 1. Many optimization protocols were applied to reduce the effect of unspecific binding (lower amount of starting material loaded on the gel, different antibody dilutions and incubation times, different blocking times, different enhanced chemiluminescence detection sys- tems, different film exposure times) without any improvement on the results obtained. Cofilin-1 Representative image of western blot analysis of urine samples with Anti-Cofilin 1 (MW: 19-21 kDa) is depict- ed in Figure 2. Cofilin-1 is an intracellular protein that regulates actin cytoskeleton dynamics. It plays a role in the regulation of cell morphology and cytoskeletal organization (8). It has been also found to be up-regu- lated (3.29 ± 1.20 fold change) in PCa compared to BPH tissue specimens by a quantitative proteomics analysis (5) (Table 1.). However, cofilin-1 was not detected in both normal (1, 2, 3, 4) and PCa (6, 15, 19, 21) urine samples as shown in Figure 2. An explanation for this could be the fact that cofilin-1 is an intracellular protein which reduces the chances to be detected in biological fluids such as urine. SPARC Western blot analysis of Anti-SPARC (MW: 43 kDa) in urine samples is depicted in Figure 3. SPARC is a secreted protein that appears to regulate cell growth through inter- actions with the extracellular matrix and cytokines (6). It binds calcium and copper, several types of collagen, albu- min, thrombospondin, PDGF and cell membranes (6). 197Archivio Italiano di Urologia e Andrologia 2016; 88, 3 In the search of novel urine prostate cancer biomarkers Figure 1. Western blot analysis of Glutathione peroxidase 3 precursor (GPx3). Figure 3. Western blot analysis of secreted protein acidic and rich in cysteine (SPARC). Figure 2. Western blot analysis of Cofilin-1 (CFL1), Table 2. Demographics and clinical characteristics. N (%) Age (years), mean (SD) 65.7 (8.7) BMI (kg/m2), mean (SD) 27.5 (3.5) Smoking No 44 (34.6) Ex-smoker 48 (37.8) Yes 35 (27.6) Family history of cancer No 60 (47.2) Yes 67 (52.8) Family history of prostate cancer No 102 (80.3) Yes 25 (19.7) PSA (ng/mL), mean (SD) 9.1 (5.3) PSA (ng/mL) < 4 6 (4.7) 4-10 89 (70.1) > 10 32 (25.2) DRE Negative 68 (54.0) Positive 59 (46.0) Histology No PCa 56 (44.1) High grade PIN 29 (22.8) PCa 42 (33.1) Prostate volume (ml) 45 BMI, body mass index; High grade PIN, prostatic intraepithelial neoplasia. Katafigioti nuova versione_Stesura Seveso 21/09/16 08:51 Pagina 197 Archivio Italiano di Urologia e Andrologia 2016; 88, 3 A. Katafigioti, I. Katafigiotis, S. Sfoungaristos, C. Alamanis, K. Stravodimos, I. Anastasiou, E. Roumelioti, M. Duvdevani, C. Constantinides 198 This protein was found to be differentially expressed at statistically significant levels in the secretome of a cell line model for aggressive bladder cancer (6). SPARC could not be detected in both normal (1, 3, 7, 8) and PCa (6, 14, 19) urine samples as shown in Figure 3. In most cases a band of 60-70 kDa was detected corresponding to the MW of albumin. SPARC binds to albumin which is present in urine samples making SPARC detection quite difficult due to albumin masking effects. HSP 90β Western blot analysis of Anti-HSP 90β (MW: 90 kDa) in urine samples is depicted in Figure 4. HSP 90β is an intra- cellular protein that acts as a molecular chaperone pro- moting the maturation, structural maintenance and prop- er regulation of specific target proteins involved in many functions such as: cell cycle control and signal transduc- tion (9). This protein has been found to be up-regulated (3.2 ± 0.61 fold change) in PCa compared to BPH tissue specimens by quantitative proteomics analysis (5). HSP 90β was not detected in urine samples mainly because of its intracellular location which does not allow for its detec- tion in body fluids such as urine (Figure 4). Statistical analysis (Table 2 and 3) Continuous variables are presented as mean and standard deviation or median and interquartile range (IQR) while qualitative variables are presented as absolute and relative frequencies. The Kruskal-Wallis test was used for the com- parison of the five proteins between different patient groups. All P values reported are two-tailed. Statistical sig- nifi cance was set at 0.05 and analyses were conducted using STATA statistical software (version 9.0). ZAG Western blot analysis of urine samples with Anti-ZAG (MW: 47 kDa) is depicted in Figure 5. ZAG is a secreted protein which stimulates lipolysis in adipocytes and is responsible for the reduction of fat which is associated with some advanced cancers (10, 11). ZAG was found to be down-regulated (0.43 ± 0.13 fold change) in PCa compared to BPH tissue specimens by quantitative pro- teomics analysis (5). Contrary to the downregulation of ZAG in tissues, ZAG in urine showed a significant increase in patients with positive histology for PCa. ZAG levels were significantly increased as PSA was increased (P = 0.005) and in the patients with positive histology for PCa (P = 0.004) (Table 3). ZAG due to its performance showing statistical significant differences between the urine of patients with BPH and PCa it was further ana- lyzed and proved its utility at least in our preliminary results published by our group. Figure 4. Western blot analysis of Heat shock protein-90β (HSP 90β). Figure 5. Western blot analysis of Zinc alpha 2-glycoprotein (ZAG) n. ZAG GPx-3 Cofilin-1 SPARC HSP 90β MEDIAN IQR P Protein that Not detected both Not detected both Not detected both PSA detected only in normal in normal in normal 0-4 0 0.0-0.1 0.005 in a few urine and Prostate and Prostate and Prostate 4-10 1.1 0.5-1.9 samples cancer samples cancer samples cancer samples > 10 1.4 0.8-2.9 Gleason score 4-6 1.2 1-2.9 0.291 7 1.4 1.3-2.8 8-10 1.9 1.5-3.2 Histology Negative (BPH) 1 0.3-1.8 0.004 High grade PIN 0.8 0.4-1.4 Adenocarcinoma 1.4 1.2-3.0 Table 3. Kruskal – Wallis test – Association of five urine proteins with PSA levels, Gleason score and histology. Katafigioti nuova versione_Stesura Seveso 21/09/16 08:51 Pagina 198 DISCUSSION The advent of proteomics made possible the identifica- tion of thousands of proteins among which, a large num- ber of possible novel biomarkers especially for PCa (5, 12). PSA is the best example of a clinical useful bio- marker and is still considered the best biomarker in our disposal for the early diagnosis of PCa. However, although it is organ specific, it is not disease specific and high levels of PSA > 4 ng/ml are not necessarily attrib- uted to PCa, and at the same time PCa of high Gleason Score > 7 can be diagnosed in low PSA levels < 4 ng/ml, even lower than < 1 ng/ml, while 75% of patients are submitted to high number of unnecessary TRUS-b due to their increased levels of PSA particularly those in the zone of 4-10 ng/ml (3, 12, 13). On the other hand the combination of the high diversity of the nature of PCa, with the increase of the TRUS-b due to the use of PSA, has led to the “overdiagnosis” of PCa of low aggressive- ness (Gleason score