UROLOGICAL ONCOLOGY Potential Prognostic Role for SPOP, DAXX, RARRES1, and LAMP2 as an Autophagy Related Genes in Prostate Cancer Leila Jamali1, Afshin Moradi2,3 , Maziar Ganji1, Mohsen Ayati4, Behrang Kazeminezhad5, Zahra Fazeli Attar1, Hamid Ghaedi1, Seyyed Mohammad Hossein Ghaderian6,1*, Morteza Fallah-Karkan7,8, Arash Ranjbar 7 Purpose: Autophagy plays a critical role in PCa development. DAXX has a potent pro-survival effect by enhanc- ing cell growth in PCa via suppression of autophagy. Here, we depicted a network governed by DAXX and SPOP by which the autophagy pathway is suppressed through the ubiquitination and modulation of key cellular signaling pathways mediators including LAMP2 and RARRES1. Materials and Methods: Through network-based bioinformatics approaches, the expression levels of DAXX, RARRES1, LAMP2, and SPOP genes was assessed in 50 PCa tissues and 50 normal adjacent from the same sam- ple as well as 50 benign prostatic hyperplasia (BPH) tissues by quantitative RT-PCR. The normal adjacent tissues were taken from regions more than 5mm away from the bulk of those tumor tissues with clearly distinct margins. RNA extraction, cDNA synthesis and Real-time Quantitative RT-PCR were done for assessment of gene expres- sion. To evaluate the primary gene network centered on autophagy pathway, according to the Query-dependent weighting algorithm, these two networks were integrated with Cytoscape 3.4 software. Results: We found that in PCa tissues the DAXX expression level was significantly increased (P < 0.001) and the expressions of SPOP, RARRES1, and LAMP2 were significantly down-regulated, when compared to both control groups including normal adjacent and BPH tissues. Moreover, significant correlations were observed between ex- pression levels of all four genes. Additionally, ROC curve analysis revealed that LAMP2 had the most sensitivity and specificity. Conclusion: These findings suggest that the contribution of SPOP, DAXX, RARRES1, and LAMP2 together could be a putative regulatory element acting as a prognostic signature and therapeutic target in PCa. Keywords: prostate cancer; autophagy; gene regulatory network; SPOP; DAXX; RARRES1; LAMP2 INTRODUCTION Prostate cancer (PCa) is the most common malig-nancy in men(1-5). Currently, clinicopathological features including Gleason score (GS), staging, and prostate-specific antigen (PSA) are conventional prog- nostic markers(6,7) and utilized for clinical decision mak- ing. Nonetheless, they are insufficiently accurate to dis- criminate the indolent tumors from aggressive ones due to heterogeneous genetic background of PCa(8-10), which emphasizes the inevitable necessity of identifying novel molecular biomarkers and high-risk individuals(11-13). Autophagy is a survival-promoting pathway that plays the key role of eliminating damaged cellular compart- ments and aggregated proteins in lysosomes(14). Auto- phagy also can serve as a tumor suppressor via prevent- 1Department of Medical Genetics, School of Medicine, Shahid Beheshti University of Medical Sciences, Tehran, Iran. 2Cancer Research Center, Shohada-e-Tajrish Hospital, Faculty of Medicine, Shahid Beheshti University of Medical Sciences, Tehran, Iran. 3Infertility and Reproductive Health Research Center, Shohada-e-Tajrish Hospital, Faculty of Medicine, Shahid Beheshti University of Medical Sciences, Tehran, Iran. 4Uro-oncology Research Center, Tehran University of Medical Sciences, Tehran, Iran. 5Department of pathology, Shahid Beheshti University of Medical Sciences, Tehran, Iran. 6Cellular and Molecular Biology Research Center Shahid Beheshti University of Medical Sciences, Tehran, Iran 7 Urology resident, Shohada-e-Tajrish hospital, Shahid Beheshti Medical University, Tehran, Iran. 8 Laser Application in Medical Science Research Center, Shohada-e-Tajrish Hospital, Faculty of Medicine, Shahid Beheshti University of Medical Sciences, Tehran, Iran. *Correspondence: Department of Medical Genetics, Shahid Beheshti University of Medical Sciences. Velenjak Blvd. Kodakyar Close, Tehran, IRAN. Post Code: 1985717443; Tele Fax: + 98 21 23872572. sghaderian@sbmu.ac.ir Received November 2018 & Accepted February 2019 ing the damaged proteins accumulation. Despite various attempts, role of autophagy and its precise function in PCa, remains unclear(15). Various studies have indicated the crucial roles of speckle-type POZ protein (SPOP) and death-domain associated protein (DAXX) in cell apoptosis, prolifer- ation and that their dysregulated expression may con- tribute to autophagy pathways in tumorigenesis(16,17). Extensive genomic documents have considered SPOP as a tumor suppressive role via degradation of onco- genic substrates in malignant prostate cells(18). DAXX, as a transcriptional repressor, in co-operation with other transcription elements participates in regulating DAPK1/3 tumor suppressor protein kinases(19), which are associated with autophagy(20,21). As another element relevant to autophagy, retinoic acid Urology Journal/Vol 17 No. 2/ March-April 2020/ pp. 156-163. [DOI: 10.22037/uj.v0i0.4935] receptor responder 1 (RARRES1) is a new retinoic inducible gene. Expression of RARRES1 is known to activate the autophagy pathways. Furthermore, dysreg- ulation in RARRES1 was indicated to associate with malignant transformations and tumor progression (1,22). Retinoic acid receptors (RARs) signaling controls the activation of lysosome-associated membrane glyco- protein 2 (LAMP2)(23). LAMP2 encodes a single-span lysosomal membrane protein that is involved in the lys- osomal stability. Nonetheless, apart from preserving the structural integrity of lysosomal membranes, a critical role has been proposed for LAMP2 in lysosomal func- tion and autophagy in the context of cancer(24). In this study, first we aimed to construct a network for candidate genes participating in autophagy pathway based on literatures to clarify the role of these candidate genes in relation with diverse interactions in cellular networks. We hypothesized these interactions may po- tentially affect the individual behavioral of each target gene in a context-dependent manner. Altogether, here we aim to develop network enrichment for SPOP, DAXX, RARRES1, and LAMP2 genes and to examine their expression levels in PCa tissues in comparison to normal adjacent and BPH tissues. MATERIALS AND METHODS Patients and tissue samples After institutional ethical committee approval (ethics code:IR.SBMU.MSP.REC.1396.286), this case-control study was conducted on a series of 189 paraffin-em- bedded prostate tissue samples. The exclusion criteria included the blocks with poor histopathological quality, patients who had history of chemoradiotherapy before surgery, presence of other malignancies in patients, and samples with technical problems in tissue processing. Subjects aged from 55 to 79 years in PCa and 59 to 79 years in Benign Prostatic Hyperplasia (BPH) were included in study. Finally leaving a total of 150 sam- ples for the final analysis. Our collection included 50 PCa and 50 normal adjacent tissues from the same sam- ple in addition to 50 BPH tissues. Considering the fact Table 1. The sequences of primers used in this study. Gene Primer sequences Primer length Product length B2M F: AGATGAGTATGCCTGCCGTG 20 105 R: GCGGCATCTTCAAACCTCCA 20 RARRES1 F: CTAGTGTGAGGCAGTGGAAAAC 22 110 R: GACCAAGTGAATGCGACAGG 20 LAMP2 F: ATGGCTCCGTTTTCAGCATTG 21 106 R: GCTCCAGACACTGAAACAGTC 21 SPOP F: TACCCTCTTCTGCGAGGTGA 20 129 R: CGGGAATTCTCCCACAGTCC 20 DAXX F: GACTATAGGCCAGGCGTTGA 20 144 R: CTCGCCCTCCTCACTTTTGT 20 a: Kruskal-Wallis Test, b: ANOVA. c: Independent sample T test, d: Mann-Whitney U. Note: Normality was checked using Shapiro-Wilk Test. Table 2. The association of four candidate genes with clinicopathological characteristics gathered from PCa patients. a: Kruskal-Wallis Test, b: ANOVA. c: Independent sample T test, d: Mann-Whitney U. Note: Normality was checked using Shapiro-Wilk Test. clinicopathological Features DAXX p- value RARRES1 p- value LAMP2 p- value SPOP p- value N Mea n SD Media n IQR 25 IQR 75 N Mean SD Media n IQR 25 IQR 75 N Mea n SD Media n IQR 25 IQR 75 N Mea n SD Media n IQR 25 IQR 75 Age -1.2675 - 4.458 6 .3736 -3.1850 - 5.142 6 -.9699 -5.2275 - 6.685 5 - 1.978 3 -4.0925 - 6.293 9 - 2.689 8 -1.9525 - 3.280 7 -.7380 -4.0200 - 5.014 1 - 3.258 5 -3.4775 - 4.321 7 - 2.726 3 -4.5075 - 5.177 3 - 3.357 7 -4.6050 - 7.275 5 - 2.030 5 -3.4800 - 5.918 8 - 1.725 2 -1.4800 - 6.246 3 .9703 -2.1150 -.4071 -3.98 GS -1.16 -1.99 1.02 -3.39 -5.25 -3.18 -3.43 -4.76 -1.50 -3.89 -5.93 -2.46 -3.73 -4.27 -1.75 -3.13 -5.00 -1.24 -3.41 -4.61 -1.12 -2.91 -6.08 -2.12 -1.92 -4.61 -1.14 -4.69 -5.73 -2.81 -4.01 -6.23 -1.76 -3.98 -6.97 -1.69 Stage -2.71 -4.29 -1.28 -3.39 -5.00 -2.76 -3.43 -4.76 -1.76 -2.91 -5.93 -2.12 -1.56 -1.75 -.14 -3.48 -6.85 -.52 -2.11 -3.53 -1.36 -2.46 -3.70 -2.12 -1.74 -2.96 -.47 -3.91 -5.38 -2.64 -5.75 -6.42 -4.79 -6.94 -7.79 -3.71 -3.28 -7.58 1.02 -.44 -3.66 2.78 -1.31 -1.50 -1.12 -2.76 -6.13 .61 -3.09 -5.92 -1.14 -5.32 -5.90 -2.81 -3.24 -5.06 -1.48 -3.70 -6.47 -1.69 -1.91 -2.62 -1.53 -3.99 -5.42 -2.43 -6.47 -7.20 -4.06 -5.95 -7.35 -4.51 Perineurial invasion -2.11 -4.17 -1.46 -4.30 -5.00 -3.13 -3.22 -3.53 -1.50 -3.18 -6.08 -2.12 -1.83 -5.10 -1.06 -3.43 -5.73 -2.29 -4.39 -6.23 -1.89 -3.93 -6.47 -2.11 PIN -1.62 -4.29 .42 -3.66 -4.69 -3.13 -1.76 -4.95 -1.36 -3.70 -5.80 -1.41 -1.95 -4.39 -1.15 -3.95 -5.56 -2.50 -3.53 -5.41 -1.91 -3.79 -6.61 -2.12 Circumferenti al Margins & Capsule -2.11 -4.61 -1.14 -3.66 -5.32 -2.31 -3.53 -5.40 -1.74 -3.98 -6.47 -2.12 -1.83 -2.31 -.65 -4.20 -6.24 -3.37 -2.74 -4.88 -1.12 -3.08 -4.51 -1.90 Vol 17 No 02 March-April 2020 157 SPOP, DAXX, RARRES1, and LAMP2 Genes in Prostate Cancer-Jamali et al. that although normal adjacent tissue is normal based on pathological feature, it is abnormal pertaining to molec- ular changes and there is a potential of tumor progres- sion to adjacent tissue cells over time, so we utilized BPH as secondary external controls. The adjacent nor- mal tissues were taken from regions more than 5 mm away from the bulk of those tumor tissues with clearly distinct margins. One-mm tissue cores were obtained from all samples after processing and were transferred to the RNase-free microtubes for RNA extraction. All samples were diagnosed by an expert pathologist to de- termine and confirm the differentiation status of adeno- carcinoma tissue and dysplasia degree of the adenoma tissues. Evaluation of tumor differentiation was based on the architectural and glandular differentiation as well as tumor nuclear features. Tumor grade was figured out based on GS system according to the 2014 International Society of Urological Pathology (ISUP) Consensus on Gleason Scoring of Prostatic Carcinoma(25). Other pa- rameters such as PT category, prostatic intraepithelial neoplasia (PIN), perineural invasion, and serum PSA levels were included too. All these samples were gathered from February 2014 to October 2016 from Mehr and Shohada-e-Tajrish Hos- pitals, Tehran, Iran. The study protocol was approved by the Institutional Review Board of Shahid Beheshti University of Medical Sciences. Bioinformatic analysis We designed a network for candidate genes, gathered from literature review, and their cellular interactions with other genes in various signaling pathways were created. The gene Regulatory network drawing was ac- complished by STRING 1.1 software with confidence cutoff value of 0.1 for the interactions. As the next step, another gene network was plotted with GeneMANIA prediction server by applying unregularized algorithm. To evaluate the primary gene network centered on au- tophagy pathway, according to the Query-dependent weighting algorithm, these two networks were integrat- ed with Cytoscape 3.4 software, validated Regulatory Interactions Network Analysis software(26-28). We ap- plied the Reactome FI 5.2 package to analyze and en- rich the gene network and interactions. RNA extraction and cDNA synthesis Total RNA extraction was carried out from all samples using Formalin-fixed, paraffin-embedded (FFPE) RNA Purification Kit (Cat. 25300; NORGENE, Canada) according to the manufacture’s protocol. The stand- ard de-crosslinking and column purification as well as DNase I treatment were performed to remove proteins and other cellular components in addition to genomic DNA prior to cDNA synthesis. The RNA quality and quantity were measured using agarose gel electropho- resis and Nanodrop 2000 (Thermo Fisher Scientific, Wilmington, DE, USA), respectively. Samples with sufficient yield (>100 ng) and A260/A280 ratio be- tween 1.8–2 were used for single-strand cDNA synthe- sis. Briefly, 500 ng of total RNA from each sample was subjected to reverse transcription for target gene. cDNA was synthesized utilizing Prime Script II reverse tran- scriptase (TaKaRa, Japan) by the following method: at 37 °C for 15 minutes (reverse transcription) followed by 85 °C 5 second (inactivation of reverse transcriptase with heat treatment). Real-time Quantitative RT-PCR (qRT-PCR) The cDNA was diluted 1:10 in nuclease-free water. Re- al-time PCR was performed in duplicate in a LightCy- cler96 instrument (Roche Diagnostics). SYBR Premix Ex Taq II (TaKaRa, Japan) was used for detection of gene expressions. The SYBR qRT-PCR was: 30 sec in- cubation at 95 °C followed by 40, two step cycles of amplifications consists of 95 °C for 15 sec and 60°C for one min. The formation of PCR products was con- firmed through melting curves. Primer sequences were designed by Allele ID software version 7.0 for Win- dows (Premier Biosoft International, Palo Alto, CA, USA). Exon-exon junction method and NCBI Primer Blast were applied (Table 1). Table 1. The sequences of primers used in this study. Each run had a negative control (without cDNA tem- Urological Oncology 158 SPOP, DAXX, RARRES1, and LAMP2 Genes in Prostate Cancer-Jamali et al. Table 3. The correlations between expression levels of candidate genes in PCa tissues. Correlations DAXX RARRES1 LAMP2 SPOP DAXX Correlation Coefficient (r) 1.000 0.320 0.338 0.220 Sig. (2-tailed) . 0.003 ** 0.001 ** 0.029 * RARRES1 Correlation Coefficient (r) 1.000 0.373 0.351 Sig. (2-tailed) . > 0.0001 ** 0.001 ** LAMP2 Correlation Coefficient (r) 1.000 0.535 Sig. (2-tailed) . > 0.0001 ** SPOP Correlation Coefficient (r) 1.000 Sig. (2-tailed) . * Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed). Gene Age PSA GS Stage R P-value R P-value R P-value R P-value DAXX -0.18 0.22 0.033 0.84 -0.14 0.36 -0.026 0.86 RARRES1 -0.07 0.61 0.17 0.3 -0.08 0.57 -0.12 0.42 LAMP2 0.22 0.14 -0.31 0.05 -0.07 0.61 -0.14 0.37 SPOP 0.25 0.09 -0.23 0.14 -0.01 0.92 -0.13 0.38 Table 4. The correlations between gene expression levels and clinicopathological characteristics plate) to check any possible contamination. The relative quantification of expression changes was calculated after normalization to Beta 2Microglobuline (B2M) ex- pression according to previous study (29). Data were evaluated using the comparative cycle threshold (CT). Statistical analysis Relative quantification of mRNA expression was eval- uated using the comparative cycle threshold (CT) meth- od. The expression of samples was normalized to the expression of B2M and fold change was calculated, us- ing the 2−∆∆Ct method. All statistical analyses were conducted on SPSS sta- tistical software version 20 (SPSS Inc., Chicago IL, USA). P values < 0.05 were considered statistically sig- nificant. Mean normalized gene expression ± SD was calculated from independent experiments. For com- parisons between PCa and normal adjacent tissues the normality of the response variables was checked using Shapiro-Wilk statistical test. If normality is accepted, the t test is used; otherwise the Mann-Whitney test or Wilcoxon was utilized for independent non-parametric and dependent variables, respectively. Furthermore, to investigate the correlations, given the normality test of the data the Pearson correlation coefficient for paramet- ric values and Spearman's rank for nonparametric data utilized. Also, to evaluate statistically significant differ- ences between two or more groups of an independent variable one-way ANOVA and Kruskal-Wallis test was used respectively, for parametric and nonparametric data. In addition, authors represented the receiver operating characteristic (ROC) curve to determine the sensitivity and specificity of gene expression levels as diagnostic markers for PCa. ROC was calculated to determine the potential of genes to discriminate between malignant and non-malignant samples. Ethics The ethical committee of Shohada-e-Tajrish Hospital approved this study and permitted us to review patients’ medical data. The personal data of the subjects were not disclosed and the principles of patient secrecy were ob- served. RESULTS Clinicopathological characteristics The mean age and PSA level were 65.49 ± 6.28 years and 10.34 ± 11.10 ng/mL in cancerous sample. The sub- jects with BPH also had the mean age of 66.35 ± 4.55 with 4.34 ± 1.09 ng/mL PSA level. Table 2 illustrates the associations between gene expression levels and clinicopathological characteristics, no significant asso- ciation was found. Table 2. The association of four candidate genes with clinicopathological characteristics gathered from PCa patients. Bioinformatic network analysis The results from network enrichment of STRING and GeneMANIA softwares in terms of statistical signifi- cance (False Discovery Rate < 0.25) revealed the inter- Figure 1. Four network clustering modules based on bioinformatics analysis, shown with different colors. Figure 2. Receiver operating characteristic (ROC) curve showing the area under the curves for discriminating between malignant, BPH tissues by DAXX, RARRES1, LAMP2, and SPOP genes. Vol 17 No 02 March-April 2020 159 SPOP, DAXX, RARRES1, and LAMP2 Genes in Prostate Cancer-Jamali et al. action of candidate genes with other genes in relation to biological and molecular pathways. During functional categorization analysis, four main modules were identi- fied in gene network, each with up to 11 genes (Figure 1). First module included “Co-regulation of androgen receptor (AR) activity” [P = 1.27E-05] and “Pathways in cancer” [P = 3.20E-03] (comprising SPOP, AR, GLI3, RNF4, PDX1, GLI2, SHE, TMF1, SNURF, ATRX, CUL3 genes). The module of “Class I PI3K sig- naling” [P = 4.33E-04] and “Apoptosis” [P = 6.53E-03] (including DAXX, MAP3K5, HIPK1, UBE21, ETS1, SLC2A4, KDM1A, HSPB1, RASSF1, KIF5B genes) was also demonstrated in this network. Moreover, the modules of “Lysosome” [P = 5.92E-03] and “Phago- some” [P = 3.87E-03] (including RARRES1, LAMP2, IMPDH1, LXN) as well as “Retinoic acid receptors-me- diated signaling” [P = 1.18E-07] and “RNA polymer- ase II transcription” [P = 1.40E-08] (comprising genes such as of ESR1, PPARG, NR3C1, KLK3, NCOA2, NCOA1, RXRA, PPARA, RARA) were implicated in this network. The genes in the modules were mainly lo- cated in the nucleus, cytoplasm, and lysosome. Figure 1. The network clustering modules found in this study. Four modules, shown with different colors, are displayed based on bioinformatics analysis. Relative expression levels of SPOP, DAXX, RAR- RES1, and LAMP2 By the 2−∆∆Ct method a statistically significant de- crease in SPOP mRNA expression level of malignant prostate tissues was observed in both normal adja- cent and BPH tissues (both at P < 0.001). As well as, DAXX expression level in PCa tissues was significant- ly up-regulated in comparison with both controls (both at P < 0.001). The expression of RARRES1 gene was significantly down-regulated in PCa group in compar- ison to groups of normal adjacent and BPH samples (P < 0.001, P < 0.011, respectively). Finally expression of LAMP2 in PCa samples showed a total significant de- creased level compared with normal adjacent and BPH groups (both at P < 0.001). Correlation analysis The assessment of correlations between expression levels of our four candidate genes in PCa tissues was accomplished (Table 3). In this regards, there were significant correlations between DAXX and expression levels of RARRES1 (r = 0.320, P = 0.003), LAMP2 (r = 0.001, P = 0.338), and SPOP (r = 0.220, P = 0.029). The expression of RARRES1 demonstrated significant correlations with LAMP2 (r = 0.373, P < 0.0001) and SPOP (r = 0.351, P = 0.001). In addition, the correlation of LAMP2 and SPOP (r = 0.535, P < 0.0001) statisti- cally significant. Table 3. The correlations between expression levels of candidate genes in PCa tissues. The analysis of correlations between gene expression levels and clinicopathological characteristics such as age, PSA, GS, and Stage of disease in PCa tissues was carried out. LAMP2 expression was significantly corre- lated with PSA (r = -0.31, P = 0.05). However, the other correlation related to four candidate genes did not reach a significance level (Table 4). Table 4. The correlations between gene expression lev- els and clinicopathological characteristics ROC curve analysis The predictive value of gene expressions for discrim- inating between malignant and non-malignant tissues was investigated by constructing an ROC curve (Fig- ure 2). Critical cut-off values of significantly different RARRES1, LAMP2, SPOP, and DAXX levels were determined. The area under the curve (AUC), sensitiv- ity, and specificity for RARRES1 were 0.659, 72.1%, and 53.5%, respectively. AUC for LAMP2 expression showed the most predictive power, 0.884 (with sen- sitivity of 90.7% and specificity of 62.8%). As well, the AUC for expression levels of SPOP (sensitivity of 86.0% and specificity of 60.5%) and DAXX (sensitiv- ity of 86.0% and specificity of 60.5%) as predictors of malignancy in prostate tissue was 0.809 and 0.837, re- spectively (Table 5). Figure 2. Receiver operating characteristic (ROC) curve showing the area under the curves (AUC) for discriminating between malignant and non-malignant prostate tissues by DAXX, RARRES1, LAMP2, and SPOP genes. DISCUSSION Gene-module level analysis has emerged as a novel design principle in biological systems. This type of evaluations aims to explain biological network design and system behavior in development of diseases, via highlighting the modules of genes instead of individ- ual genes(30,31). In the current study, first we enriched the candidate genes and directed biological pathways, by bioinformatics approaches, and their gene network as well as interactions with other genes was designed. Then, expression analysis of SPOP, DAXX, RAR- RES1, and LAMP2 was assessed by qRT-PCR in 50 PCa tissues, compared with 50 BPH and 50 normal ad- jacent prostatic tissues. Puto et al. described DAXX as a transcription factor that serves its suppressive role by recruiting DNA me- thyl transferases and histone deacetylases in PCa cell lines(16). Here, by developing a network we depicted that DAXX mediates its key role in cellular process by interactions with SPOP and other factors in pathways related to lysosome, RARs-mediated signaling, PI3K signaling, and AR activity. SPOP encodes an E3 ubiq- uitin ligase component acting through degradation of several regulators of cell proliferation and apoptosis including DAXX in a cancer context(32). Our findings confirms these concepts by showing the inverse rela- tionship between expression levels of DAXX (up-reg- ulated) and SPOP (down-regulated) and further points to their connection and vital roles as autophagy-related Urological Oncology 160 Table 5. The results obtained from ROC curve analysis of four candidate genes. Gene Cut-off point AUC 95% Confidence Interval %Sensitivity %Specificity P-value DAXX -6.302 0.837 [0.751-0.924] 86 60.5 < 0.0001 RARRES1 -2.557 0.659 [0.542-0.776] 72 53.5 0.011 LAMP2 -3.065 0.884 [0.815-0.953] 90.7 62.8 < 0.0001 SPOP -2.437 0.809 [0.715-0.904] 86 60.5 < 0.0001 SPOP, DAXX, RARRES1, and LAMP2 Genes in Prostate Cancer-Jamali et al. genes in PCa. These results are consistent with those of Ju et al. speculating that SPOP selectively suppresses the PCa through stability regulation of cyclin E1 in PCa cell lines. Substantially, expression of cyclin E1 rescues the tumor formation, proliferation, and migration of PCa cells(33). The results from exome sequencing of 112 PCa and normal tissue pairs revealed that SPOP gene had the most frequently recurrent mutations influencing its expression level(34). Noticeably, Dysregulated levels of SPOP may accordingly serve as a specific hallmark in early detection of PCa carcinogenesis(35), which can bring us to precise understanding of molecular mecha- nism and its clinical applications for targeted PCa thera- pies(36). Evidence also suggests that strong expression of DAXX correlates with high GS and elevated cell pro- liferation index, exhibiting its potential prognosticator role in PCa outcomes(37). By considering the network enrichment and correlation outcomes, our study represents for the first the inter- actions of RARRES1 and LAMP2 between themselves and in network of SPOP and DAXX genes, altogether emphasizing a network developed from different mod- ules (Figure 1). Literature introduces RARRES1 as a putative tumor suppressor gene that negatively regu- lates the cell proliferation, while less is known about responsible mechanisms(38). Particularly, it has been emphasized that RARRES1 is involved in autophagy induction(1). Our network enrichment and qRT-PCR findings supported this hypothesis by revealing the sig- nificant reduction of RARRES1 in tumor samples rel- ative to both controls including adjacent and BPH tis- sues. Based on the obtained gene network, this reduced level might be indirectly related to the increased levels of DAXX, offering their inverse relationship in PCa in our study. We found a total reduction in LAMP2 expression lev- el in tumor tissues. Evidence exist that this decreased level can trigger lysosomal membrane permeabilization and subsequently sensitize cells to the lysosomal path- way of cell death(39). Noteworthy, we found that ino- sine monophosphate dehydrogenase 1 (IMPDH1), as a mediator, interacts with LAMP2 in autophagy module of our network. This suggests that DAXX may be able to down-regulate the LAMP2 expression through IMP- DH1 in the depicted gene network. On the other hand, DAXX can adjust the expression of autophagy-related genes such as RARRES1 and LAMP2 by interactions with AR and retinoic acid receptor alpha (RARA) (Fig- ure 1). Importantly, the ROC curve analysis showed the good predictive value (AUC above 0.8) for DAXX, SPOP, and LAMP2 gene expressions in discrimination of ma- lignant and non-malignant tissues in PCa, among which LAMP2 had the most sensitivity and specificity. Again, the integrated picture of ROC curve for all four candi- date genes, in addition to the significant correlations ob- served between their expression levels, potentially sug- gests the putative clinical application of this network. In this kind of study, which different molecular tech- niques can yield different results, it is recommended to study greater sample size with the same methods to in- crease the accuracy of the results. As the next step forward, we suggest the study of a more extended gene network from each module to en- rich our knowledge of reciprocal interactions between these genes. In line, investigation of their expression al- terations in protein level, as well as considering a larger sample size of PCa and BPH tissues, could be of great value for future. CONCLUSIONS In conclusion, given the importance of autophagy in PCa tumorigenesis, these findings not only indicate the complicated cellular networks and context-dependent manner of autophagy induction, also suggest that the contribution of SPOP, DAXX, RARRES1, and LAMP2 together could be a putative regulatory element acting as a prognostic signature and therapeutic target in PCa. ACKNOWLEDGMENTS We would like to thank all staff of Urology and Pathol- ogy department staff in Shohada-e-Tajrish hospital. We would like to show our appreciation towards Dr. Naser Rakhshani and Dr. Arman Morakabatif, for them inval- uable help throughout this study. CONFLICT OF INTEREST None declared by the authors. REFERENCES 1. Roy A, Ramalinga M, Kim OJ, et al. Multiple roles of RARRES1 in prostate cancer: Autophagy induction and angiogenesis inhibition. PloS one. 2017;12:e0180344. 2. Taeb J, Asgari M, Abolhasani M, Farajollahi MM, Madjd Z. Expression of prostate stem cell antigen (PSCA) in prostate cancer: a tissue microarray study of Iranian patients. Pathol Res Pract. 2014;210:18-23. 3. Abedi A-r, Fallah-Karkan M, Allameh F, Ranjbar A, Shadmehr A. Incidental prostate cancer: a 10-year review of a tertiary center, Tehran, Iran. Res Rep Urol. 2018;10:1-6. 4. Javanmard B, Yousefi M, Yaghoobi M, et al. Ureteral Reimplantation or Percutaneous Nephrostomy: Which One Is Better in Management of Complete Ureteral Obstruction Due to Advanced Prostate Cancer?. Int J Cancer Manag. 2017;10: e6074. doi: 10.5812/ ijcm.6074. 5. Karkan MF, Razzaghi MR, Javanmard B, Tayyebiazar A, Ghiasy S, Montazeri S. Holmium: YAG Laser Incision of Bladder Neck Contracture Following Radical Retropubic Prostatectomy. Nephro-Urol Mon. Online ahead of Print ; In Press(In Press):e88677. doi: 10.5812/numonthly.88677. 6. Quinn DI, Henshall SM, Haynes A-M, et al. Prognostic significance of pathologic features in localized prostate cancer treated with radical prostatectomy: implications for staging systems and predictive models. J Clin Oncol. 2001;19:3692-3705. 7. Ghiasy S, Abedi AR, Moradi A, et al. Is active surveillance an appropriate approach to manage prostate cancer patients with Gleason Score 3+ 3 who met the criteria for active surveillance? Turk J Urol. 2018:1-4. DOI: Vol 17 No 02 March-April 2020 161 SPOP, DAXX, RARRES1, and LAMP2 Genes in Prostate Cancer-Jamali et al. 10.5152/tud.2018.72920. 8. Katz B, Reis ST, Viana NI, et al. Comprehensive study of gene and microRNA expression related to epithelial-mesenchymal transition in prostate cancer. PloS one. 2014;9:e113700. 9. Saffari M, Ghaderian SMH, Omrani MD, Afsharpad M, Shankaie K, Samadaian N. The Association of miR-let 7b and miR-548 with PTEN in Prostate Cancer. Urol J. 2018; 8. doi: 10.22037/uj.v0i0.4564. [Epub ahead of print] 10. Vaezjalali M, Azimi H, Hosseini SM, Taghavi A, Goudarzi H. Different Strains of BK Polyomavirus: VP1 Sequences in a Group of Iranian Prostate Cancer Patients. Urol j. 2018;15:44-8. doi: 10.22037/uj.v0i0.3833. 11. Nodouzi V, Nowroozi M, Hashemi M, Javadi G, Mahdian R. Concurrent down-regulation of PTEN and NKX3. 1 expression in Iranian patients with prostate cancer. International braz j urol. 2015;41:898-905. 12. Taheri M, Noroozi R, Rakhshan A, Ghanbari M, Omrani MD, Ghafouri-Fard S. IL-6 Genomic Variants and Risk of Prostate Cancer. Urol J. 2018; 21. doi: 10.22037/uj.v0i0.4543. [Epub ahead of print] 13. Wisniewski T, Zyromska A, Makarewicz R, Zekanowska E. Osteopontin And Angiogenic Factors As New Biomarkers Of Prostate Cancer. Urol J. 2018; 1. doi: 10.22037/ uj.v0i0.4282. [Epub ahead of print] 14. Mishra P, Dauphinee AN, Ward C, Sarkar S, Gunawardena AH, Manjithaya R. Discovery of pan autophagy inhibitors through a high- throughput screen highlights macroautophagy as an evolutionarily conserved process across 3 eukaryotic kingdoms. Autophagy. 2017;13:1556-72. 15. Draz H, Goldberg A, Sanderson T, Safe S. Diindolylmethane and its halogenated analogues induce autophagy in human prostate cancer cells via induction of the astrocyte-elevated gene-1 (AEG-1). [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 3298. doi:10.1158/1538- 7445.AM2017-3298 16. Puto LA, Benner C, Hunter T. The DAXX co-repressor is directly recruited to active regulatory elements genome-wide to regulate autophagy programs in a model of human prostate cancer. Oncoscience. 2015;2: 362–72. 17. Blattner M, Liu D, Robinson BD, et al. SPOP mutation drives prostate tumorigenesis in vivo through coordinate regulation of PI3K/ mTOR and AR signaling. Cancer cell. 2017;31(3):436-451. 18. Mani R-S. The emerging role of speckle-type POZ protein (SPOP) in cancer development. Drug Discov Today. 2014;19:1498-1502. 19. Puto LA, Reed JC. Daxx represses RelB target promoters via DNA methyltransferase recruitment and DNA hypermethylation. Genes Dev. 2008;22:998-1010. 20. Tang HW, Wang YB, Wang SL, Wu MH, Lin SY, Chen GC. Atg1‐mediated myosin II activation regulates autophagosome formation during starvation‐induced autophagy. The EMBO journal. 2011;30:636-51. 21. Zalckvar E, Berissi H, Eisenstein M, Kimchi A. Phosphorylation of Beclin 1 by DAP- kinase promotes autophagy by weakening its interactions with Bcl-2 and Bcl-XL. Autophagy. 2009;5:720-22. 22. Taheri M, Noroozi R, Dehghan A, et al. Association study of Retinoic Acid Related Orphan Receptor A (RORA) gene and risk of prostate disorders. Urol J. 2018; 22. doi: 10.22037/uj.v0i0.4373. [Epub ahead of print] 23. Anguiano J, Garner TP, Mahalingam M, Das BC, Gavathiotis E, Cuervo AM. Chemical modulation of chaperone-mediated autophagy by retinoic acid derivatives. Nat Chem Biol. 2013;9:374-82. 24. Morell C, Bort A, Vara-Ciruelos D, et al. Up-regulated expression of LAMP2 and autophagy activity during neuroendocrine differentiation of prostate cancer LNCaP cells. PloS one. 2016;11:e0162977. 25. Epstein JI, Egevad L, Amin MB, Delahunt B, Srigley JR, Humphrey PA. The 2014 International Society of Urological Pathology (ISUP) consensus conference on Gleason grading of prostatic carcinoma. Am J Surg Pathol. 2016;40:244-52. 26. Kutmon M, Kelder T, Mandaviya P, Evelo CT, Coort SL. CyTargetLinker: a cytoscape app to integrate regulatory interactions in network analysis. PloS one. 2013;8:e82160. 27. Carlin DE, Demchak B, Pratt D, Sage E, Ideker T. Network propagation in the cytoscape cyberinfrastructure. PLoS Comput Biol. 2017;13:e1005598. 28. Türei D, Földvári-Nagy L, Fazekas D, et al. Autophagy Regulatory Network—A systems- level bioinformatics resource for studying the mechanism and regulation of autophagy. Autophagy. 2015;11:155-65. 29. Agell L, Hernández S, Nonell L, et al. A 12- gene expression signature is associated with aggressive histological in prostate cancer: SEC14L1 and TCEB1 genes are potential markers of progression. Am J Pathol. 2012;181:1585-94. 30. Wang X, Dalkic E, Wu M, Chan C. Gene module level analysis: identification to networks and dynamics. Curr Opin Biotechnol. 2008;19:482-91. 31. Razzaghi MR, Mazloomfard MM, Malekian S, Razzaghi Z. Association of macrophage inhibitory factor-173 gene polymorphism with Urological Oncology 162 SPOP, DAXX, RARRES1, and LAMP2 Genes in Prostate Cancer-Jamali et al. biological behavior of prostate cancer. Urol J. 2018 Aug 14. doi: 10.22037/uj.v0i0.3968. [Epub ahead of print] 32. Puto LA, Brognard J, Hunter T. Transcriptional repressor DAXX promotes prostate cancer tumorigenicity via suppression of autophagy. Journal of Biological Chemistry. 2015:jbc. M115. 658765. 33. Ju L-G, Zhu Y, Long Q-Y, et al. SPOP suppresses prostate cancer through regulation of CYCLIN E1 stability. Cell Death Differ. 2018:1. oi.org/10.1038/s41418-018-0198-0. 34. Barbieri CE, Baca SC, Lawrence MS, et al. Exome sequencing identifies recurrent SPOP, FOXA1 and MED12 mutations in prostate cancer. Nat Genet. 2012;44:685. 35. Wang H, Barbieri CE, He J, et al. Quantification of mutant SPOP proteins in prostate cancer using mass spectrometry-based targeted proteomics. J Transl Med. 2017;15:175. 36. Boutros PC. The path to routine use of genomic biomarkers in the cancer clinic. Genome Res. 2015;25:1508-13. 37. Tsourlakis MC, Schoop M, Plass C, et al. Overexpression of the chromatin remodeler death-domain–associated protein in prostate cancer is an independent predictor of early prostate-specific antigen recurrence. Hum Pathol.. 2013;44:1789-96. 38. Rivera-Gonzalez GC, Droop AP, Rippon HJ, et al. Retinoic acid and androgen receptors combine to achieve tissue specific control of human prostatic transglutaminase expression: a novel regulatory network with broader significance. Nucleic Acids Res. 2012;40:4825-40. 39. Fehrenbacher N, Bastholm L, Kirkegaard- Sørensen T, et al. Sensitization to the lysosomal cell death pathway by oncogene- induced down-regulation of lysosome- associated membrane proteins 1 and 2. Cancer Res.. 2008;68:6623-33. Vol 17 No 02 March-April 2020 163 SPOP, DAXX, RARRES1, and LAMP2 Genes in Prostate Cancer-Jamali et al.