The tissue microarrays (TMAs), including 126 primary PCa tissues and 22 adjacent non-cancerous prostate tissues along with detailed clinical information of the samples, were purchased from Xi'an Alenabio Biotech Co., Ltd. (cat no.: PR752 & PR753). All human tissues were collected under Institutional Review Board (IRB) and Health Insurance Portability and Accountability (HIPPA) approved protocols. Patients known to have undergone chemotherapy and/or radiotherapy prior to tissue isolation were excluded from the study.

Additionally, a cohort from the Memorial Sloan-Kettering Cancer Center (the Taylor cohort) with publicly available data (GEO accession no. GSE21032), including 150 primary PCa tissues and 29 adjacent non-cancerous prostate tissues, was assessed in regards to clinicopathological parameters. Of these cases, 111 of the primary PCa tissues and all of the non-cancerous tissues had available miRNA microarray expression data (22). Biochemical recurrence (BCR) was defined as PSA ≥0.2 ng/ml on two consecutive measurements after radical prostatectomy. BCR-free survival was defined as the time interval between initial surgery and the date of BCR. Overall survival was determined as the time interval between the initial surgery and the date of the last follow-up or patient death.

Animal housing and the experiments in this study were performed in compliance with the guidelines of the Institute for Laboratory Animal Research at Guangzhou Medical University (Guangzhou, China). A total of 32 BALB/c nude mice (4- to 5-week-old males) were purchased from Guangdong Medical Laboratory Animal Center and were housed in wire-top cages (5 mice per cage) with sawdust bedding in an isolated, clean, air-conditioned room at a temperature of 25–26°C and relative humidity of ~50% under a 12-h light/dark cycle.

The human PCa cell lines, LNCaP (cat. no. 63462566) and DU145 (cat. no. 61761869) were purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA) in 2015 and were cultured in RPMI-1640 medium (Hyclone, USA) supplemented with 10% fetal bovine serum (Gibco, USA), 2 mM L-glutamine, and antibiotics. Cell line characterization and passaging for ~3 months were performed prior to purchase. Therefore, we did not carry out reauthentication of the cell lines. ATCC used Short Tandem Repeat (STR) profiling to detect the misidentified, cross-contaminated, or genetically drifted cells. A Promega PowerPlex^® 18D System was used to amplify 17 STR loci plus amelogenin. Human umbilical vein endothelial cells (HUVECs) were obtained from the Cell Bank, Chinese Academy of Sciences (Shanghai, China). All cell lines were maintained at 37°C in a humidified chamber supplemented with 5% CO₂.

The miRNA mimics (miR-195, cat. no. miR10000461-1-5) and negative control miRNA mimics (miR-NC, cat. no. miR01201-1-5) used for transient transfection were designed and synthesized by RiboBio (Guangzhou, China). The PRR11 coding sequence (without the 3′ UTR) was cloned into a pCDNA3.1 (+)-Vector (cat. no. V790-20; Invitrogen, USA), while a blank vector was used as a negative control. Short interfering RNA (siRNA) against PRR11 (si-PRR11) and negative control siRNA with non-specific sequences (si-NC) were synthesized by Sigma-Aldrich (USA). The targeting sequences of the si-PRR11 were as follows: 5′-CUGCAUAACCCAGAGUUUAdTdT-3′ (sense) and 5′-UAAACUCUGGGUUAUGCAGdTdT-3′ (antisense). The sequences of the si-NC were as follows: 5′-UUCUCCGAACGUGUCACGUTT-3′ (sense) and 5′-ACGUGACACGUUCGGAGAATT-3′ (antisense). Cells were transfected with the miRNA mimics, siRNA and pCDNA3.1(+)-PRR11 using Lipofectamine 2000 Reagent (cat. no. 11668019; Invitrogen, USA)according to the manufacturer's protocol. At 48 h after transfection, the cells were used in the cell cycle, CCK-8, RNA extraction and western blotting assays.

Packaging of lentiviral particles was performed as follows: Briefly, three lentivirus expression plasmids containing siRNA against PRR11 were constructed by GeneChem Corporation (Shanghai, China) and were used to infect DU145 and LNCaP cells in the presence of 6 µg/ml Polybrene: PRR11-shRNA1: 5′-TCAGATGGATCTGCGGAAACTTCCTGTCAGATTTCCGCAGATCCATCTGA-3′; PRR11-shRNA2: GGATCTGCGGAAACTGCTTCTTCCTGTCAGAAAGCAGTTTCCGCAGATCC; and PRR11-shRNA3: CCTAGAAGCCCAACTCCAACTTCCTGTCAGATTGGAGTTGGGCTTCTAGG. Infected cells were selected for using puromycin, and the knockdown of PRR11 was confirmed via western blotting using anti-PRR11 antibody (cat. no. HPA023923; Sigma-Aldrich, USA). One of the PRR11-shRNAs was chosen for further experiments.

Total RNA was extracted using TRIzol reagent (cat. no. 15596-018; Life Technologies, Carlsbad, CA, USA) following the manufacturer's instructions, and RNA integrity was checked against an RIN threshold in an Agilent Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA). Qualified total RNA was further purified with an RNeasy Micro kit (cat. no. 74004; Qiagen, GmBH, Germany) and RNase-Free DNase Set (cat. no. 79254; Qiagen). Total RNA was then amplified, labeled and purified using a GeneChip 3′IVT Express Kit (cat. no. 901229; Affymetrix, Santa Clara, CA, USA) following the manufacturer's instructions to obtain biotin-labeled cRNA. For array hybridization, the hybridization procedure was performed using a GeneChip^® Hybridization, Wash and Stain Kit (cat. no. 900720; Affymetrix) in a Hybridization Oven 645 (cat. no. 00-0331-220V) and Fluidics Station 450 (cat. no. 00-0079; Affymetrix) following the manufacturer's instructions. For data acquisition, slides were scanned with a GeneChip^® Scanner 3000 (cat. no. 00-00212; Affymetrix) using Command Console Software 3.1 (Affymetrix) at default settings. Raw data were normalized by the MAS 5.0 algorithm using Gene Spring Software 11.0 (Agilent Technologies).

The online program TargetScan (release 6.2) (23) was used to predict potential target genes for miR-195.

The expression levels of miR-195 and PRR11 mRNA in the PCa cell lines were detected by RT-qPCR analysis according to the protocol in our previous study (24). The oligonucleotide sequences (5′-3′) of the primers used in the present study were as follows: PRR11 (F, CCTGCTAGCTACATTTACA, R, GAATGGTCAAGTCATTTAGC); GAP DH (F, CATGGGTGTGAACCATGAGAAGTA, R, CAGTAGAGGCAGGGATGATGTTCT); hsa-miR-195-5p (cat. no. HmiRQP0283; GeneCopoeia, USA) and U6 (cat. no. HmiRQP9001; GeneCopoeia, USA).

The expression levels of PRR11 protein in the PCa cell lines were detected by western blot analysis according to the protocol in our previous study (24). The antibodies used in the present study were as follows: Anti-PRR11 (polyclonal rabbit, cat. no. HPA023923; Sigma-Aldrich, USA), anti-GAPDH (HRP-conjugated monoclonal mouse, cat. no. KC-5G5; KangChen Bio-Tech, Shanghai, China).

The expression pattern and subcellular localization of PRR11 protein in clinical PCa tissues, and of CD31 and Ki-67 in the subcutaneous tumor xenografts of nude mice, were detected by immunohistochemistry. The specimens were fixed in 10% neutral buffered formalin and subsequently embedded in paraffin. The paraffin-embedded tissues were cut to a thickness of 4 µm and deparaffinized with xylene, and then rehydrated for further peroxidase (DAB) immunohistochemistry staining employing a Dako EnVision System (Dako Diagnostics, Switzerland). For this staining assay, following a brief proteolytic digestion and peroxidase blocking of the tissue slides, the slides were incubated overnight at 4°C with the abovementioned anti-PRR11 antibody at a dilution of 1:600, along with anti-CD31 (rabbit monoclonal antibody; cat. no. ZA-0568; ZSGB-BIO, China) at a dilution of 1:200, and anti-Ki-67 (rabbit monoclonal antibody; cat. no. ZA-0502; ZSGB-BIO, China) at a dilution of 1:200. After washing, peroxidase-labeled polymer and chromogen substrate were used to visualize the staining of the proteins of interest. In each immunohistochemistry run, negative controls were included by omitting the primary antibody.

Following hematoxylin counterstaining, immunostaining was scored by two independent experienced pathologists who were blinded to the clinicopathological data and clinical outcomes. The scores of the two pathologists were compared and any discrepant scores were reviewed through re-examination of the staining by both pathologists to achieve a consensus score. Scores were assigned by evaluating the immunolabeling of tumor cells. The number of positively stained cells in 10 representative microscopic fields was counted, along with the percentage of positive cells. Given the heterogeneity of the target protein staining, tumor specimens were scored in a semi-quantitative manner. The scoring system based on the percentage of immunoreactive tumor cells was as follows: 0 (0–5%), 1 (6–25%), 2 (26–50%), 3 (51–75%) and 4 (>75%). Additionally, staining intensity was visually scored and stratified as follows: 0 (negative), 1 (weak), 2 (moderate) and 3 (strong). Final immunoreactivity scores (IRS) for PRR11 and Ki-67 were then obtained for each sample by multiplying the percentage of positive cells by the intensity score. Vasculature density in the tumor xenografts was also determined from the number of CD31-positive vessels.

For the in vivo tumor formation assays, DU145 or LNCaP cells transfected with lentivirus expression plasmid containing PRR11-shRNA or negative control (scramble) were trypsinized and suspended in PBS. Subsequently, the cells were subcutaneously injected into the right flank of each nude mouse (8 mice per group); DU145 cells were injected with 0.2 ml PBS at a concentration of 2.5×10⁷ cells/ml, while LNCaP cells were injected as a mixture of 0.1 ml PBS at a concentration of 5×10⁸ cells/ml and an equal volume of Matrigel (cat. no. 356234; BD Biosciences). The tumor sizes were measured at 2-day intervals as soon as the tumors were measurable, and the tumor volumes were calculated as follows: V (mm³) = width (mm²) × length (mm)/2. On day 42, all mice in the LNCaP and DU145 groups were sacrificed.

The expression of the target gene of miR-195 was evaluated in LNCaP cells by a luciferase reporter assay. The putative miR-195 complementary site in the 3′ UTR of PRR11 mRNA (NCBI reference sequence: NM_018304.3; 3′ UTR-1: 3490–3496 and 3′ UTR-2: 4827–4833) or a mutant sequence was cloned into a psiCHECK-2 luciferase reporter vector (Promega, Madison, WI, USA). LNCaP cells were co-transfected with 50 nM miR-195 mimic or miR-NC and 0.5 µg of psi-PRR11-3′ UTR-1-WT, psi-PRR11-3′ UTR-2-WT, psi-PRR11-3′ UTR-1-MUT or psi-PRR11-3′ UTR-2-MUT. Cells were collected 48 h after transfection and analyzed with a Dual-Luciferase Reporter Assay System (Promega). The firefly and Renilla luciferase signals were detected with a GloMax fluorescence reader (Promega), and the Renilla luciferase signal was normalized to the firefly luciferase signal.

For cell viability assays, 2×10³ cells were seeded in 96-well plates and cultured for 24, 48 and 72 h. Cells were then incubated with 20 µl of CCK-8 solution (cat. no. C0038; Beyotime, China) for 4 h at 37°C. The absorbance was measured at a wavelength of 495 nm with a spectrophotometer, and data were expressed as means ± SD of three independent experiments.

A total of 200 µl human umbilical vein endothelial cells (HUVECs; 2×10⁴ cells) were seeded in 48-well plates containing 200 µl BD Matrigel Basement Membrane Matrix (cat. no. 356234; BD Biosciences) for 8–12 h at 37°C. LNCaP and DU145 cells transfected with oligonucleotide and/or plasmid were seeded in the upper Transwell chambers (cat. no. 3495; Corning Incorporated, Corning, NY, USA), in which the conditioned medium permeated through the 0.4-µm micropores to the Matrigel, which established a non-contact co-culture system. Images were acquired with a phase-contrast microscope. The numbers of tubes were counted in three individual wells and presented as the mean ± SD.

A flow cytometry assay (kit cat. no. KGA511) was performed to assess the cell cycle distributions of the DU145 and LNCaP cells. At 48 h after cell transfection, attached and suspended cells were harvested with a pipette, washed once with 1 ml PBS, and resuspended in 500 µl PBS containing 50 µg/ml propidium iodide (PI). RNase A (100 µg/ml) and 0.2% Triton X-100 were then added to the cells, which was followed by incubation at 4°C for 30 min in the dark prior to flow cytometric analysis (BD FACSCaliber). Data analysis was performed using ModFit software (Verity Software House, Inc., Topsham, ME, USA.).

Continuous variables are expressed as means ± SD. SPSS version 20.0 for Windows (SPSS, Inc., IL, USA) and SAS 9.1 (SAS Institute, Cary, NC, USA) were used for all statistical analyses, which were performed by two independent biostatisticians. The RT-qPCR and western blot data were analyzed by Wilcoxon signed-rank tests. The Fisher's exact test was used for any 2×2 tables and the Pearson χ² test was used for non-2×2 tables. Mann-Whitney U and Kruskal-Wallis H tests were performed to examine the associations between PRR11 expression and the clinicopathological characteristics of PCa patients in the Taylor cohort. The Kaplan-Meier method was used for survival analysis, and Cox regression analysis was used for univariate and multivariate analyses. Finally, Spearman correlations were calculated for the expression levels of miR-195 and PRR11 in the Taylor cohort. A P-value of <0.05 was considered to indicate a statistically significant difference.

In our previous study, we found that miR-195 suppressed invasion and migration while promoting apoptosis in PCa cells (21). To verify the biological role of miR-195 in angiogenesis and proliferation in vitro, we used LNCaP (P<0.001) and DU145 (P<0.001) PCa cell lines overexpressing miR-195 mimics (miR-195 group) or negative control miR (miR-NC group) via transient transfection. RT-qPCR analysis confirmed that transient transfection of the cell lines was successfully established (Fig. 1A). To investigate whether miR-195 has biological effects on tumor angiogenesis, HUVEC tube formation assays were carried out. The results indicated that the tube formation of HUVECs was strongly inhibited by the conditioned media from DU145 and LNCaP cells overexpressing miR-195 when compared with their negative control (miR-NC) media (Fig. 1B). Notably, the relative tube-forming abilities of miR-195-overexpressing DU145 and LNCaP cells, determined by comparing the number of complete tubes per field, were significantly increased when compared with their miR-NC counterparts (P=0.017 and P=0.005, respectively; Fig. 1C). Moreover, in both the DU145 and LNCaP cells, overexpression of miR-195 strongly inhibited cell proliferation (Day 4: P=0.008 and P=0.006, respectively; Fig. 1D) compared with the controls.

Since miRNAs in most cases only cause modest decreases in protein translation, the miRNA-mediated regulation of proteins with long half-lives may not be detected by measuring steady-state protein levels using standard proteomic quantification (25), which was used in our previous study (21). Therefore, to solve the shortage of the standard proteomic quantification and identify more novel targets of miR-195 in PCa, we performed microarray analysis on the same cells (LNCaP cells overexpressing miR-195 or miR-NC) to detect miR-195-induced changes in mRNA levels. As shown in Fig. 2A, a total of 153 genes differentially regulated with fold changes ≤-1.5 or ≥1.5 (miR-195 vs. miR-NC) were identified in at least two independent experiments. miRNAs typically destabilize post-transcriptional mRNAs. Thus, as expected, among the 153 genes, there were 138 (90.2%) downregulated genes and 15 (9.8%) upregulated genes. We then investigated the molecular roles of the differentially expressed genes regulated by miR-195 by using Ingenuity Pathway Analysis (IPA). Among the pathways enriched for differentially expressed genes, there were a number of tumor-related pathways, including mTOR signaling, EIF-2 signaling, cell cycle signaling, and remodeling of epithelial adherens junctions signaling (Fig. 2B), which all have broad effects on cell behavior. These results suggested that the biological pathways and functions regulated by miR-195 were closely associated with PCa progression.

We additionally used an miRNA target prediction program (TargetScan) to predict the top 10 candidate targets of the downregulated genes identified by microarray. TFEC, ZFHX3, PRR11, GCC2, RPS6KB1 and ENAH were predicted, which demonstrated that miR-195 may bind to the 3′ UTR sequences of these genes (Fig. 3A). We further found that PRR11 and RPS6KB1 were involved in the top 10 enriched pathways (data not shown). RPS6KB1 is a critical component of mTOR signaling and was identified as a target gene of miR-195 in our previous study (21). Meanwhile, the present results identified PRR11 as a key component in cell cycle signaling, implying that PRR11 might also be an important mediator of the biological role of miR-195 in PCa.

We continued to investigate the relationship between miR-195 expression and PRR11 mRNA expression in the Taylor cohort, comprised of mRNA and miRNA expression profiles for 113 primary PCa tissues (22). The results demonstrated that PRR11 expression was negatively correlated with miR-195 expression (r=−0.355, P<0.001; Fig. 3B), which corresponded with the targeting data for miR-195.

More notably, we found that miR-195 downregulation combined with PRR11 upregulation was associated with aggressive clinicopathological features and poor prognosis of PCa patients in the Taylor cohort. As shown in Table I, miR-195 downregulation and PRR11 upregulation were frequently observed in cases presenting with the most aggressive features, including those with higher Gleason scores (P=0.005), higher risk of metastasis (P<0.001) and BCR (P=0.003). By contrast, miR-195 upregulation and PRR11 downregulation often occurred in the most indolent tumors, while other regulation states, including simultaneous upregulation or downregulation of miR-195 and PRR11, could be seen in the patients with intermediate risk of progression. Kaplan-Meier analysis also indicated that miR-195 expression combined with PRR11 expression could significantly stratify patients into three groups based on BCR-free survival time (P=0.001; Fig. 3C). Notably, patients with low miR-195 and high PRR11 expression were likely to have the shortest BCR-free survival time. Additionally, univariate and multivariate Cox regression analysis revealed that miR-195 expression combined with PRR11 expression served as an independent predictor for BCR-free survival (P=0.001 and P=0.034, respectively; Table II). These results indicated that PRR11, in an opposing manner to miR-195, might act as a promotive factor in PCa progression. Thus, PRR11 may be a potential target of miR-195 in PCa.

According to the TargetScan prediction, two putative binding sites for miR-195 were found in the 3′ UTR of PRR11 at 4827–4833 and 3490–3496 bp (Fig. 3D and F). To confirm these predictions, a luciferase reporter assay was carried out in LNCaP cells. In this assay, relative luciferase activity was markedly reduced in cells co-transfected with psi-PRR11-3′ UTR-2-WT luciferase reporter and miR-195 mimic compared with negative control cells (P=0.027; Fig. 3E). In contrast, the expression of the luciferase reporter containing a mutated sequence of the PRR11 fragment (psi-PRR11-3′ UTR-2-MUT) was not affected by co-transfection with hsa-miR-195 mimics. Meanwhile, relative luciferase activity was not significantly changed in cells co-transfected with psi-PRR11-3′ UTR-1-WT luciferase reporter and miR-195 mimic relative to the negative control cells (P>0.05; Fig. 3G). These results indicated that the sequence at 4827–4833 bp in the 3′ UTR of PRR11 was the complementary site for the miR-195 seed region, and further demonstrated that PRR11 is a direct target of miR-195.

The direct targeting of PRR11 by miR-195 was evaluated in the PCa cell lines DU145 and LNCaP. We found that the mRNA and protein levels of PRR11 were significantly decreased by miR-195 overexpression in DU145 and LNCaP cells (both P<0.001; Fig. 4A and B). Subsequently, we knocked down the endogenous expression of PRR11 in DU145 and LNCaP cells with PRR11 siRNA (both P<0.001; Fig. 4A and B). Compared with the negative controls, knockdown of PRR11 in the DU145 and LNCaP cells inhibited the tube formation abilities of HUVECs (P=0.046 and P=0.006, respectively; Fig. 4C and D), suppressed cell cycling (G1 phase: both P=0.045; S phase: both P=0.001; Fig. 4E-G) and suppressed cell proliferation (Day 4: P=0.003 and P=0.005, respectively; Fig. 4H and I). Furthermore, we transfected cells with pCDNA3.1 (+)-vectors expressing PRR11 without its 3′ UTR. As shown in Fig. 4A and B, the endogenous PRR11 expression levels were detected by qRT-PCR and western blot analysis in the DU145 and LNCaP cells transfected with the miR-195 mimics in the presence of PRR11 or vector control for 48 h (P<0.001 and P<0.002, respectively). In turn, restoration of PRR11 expression markedly attenuated the effects of miR-195 on HUVEC tube formation (P=0.008 and P<0.001, respectively; Fig. 4C and D), cell cycling (G1 phase: P=0.033 and P=0.027, respectively; S phase: P=0.041 and P=0.003, respectively; Fig. 4E-G) and cell proliferation (Day 4: P=0.007 and P=0.009, respectively; Fig. 4H and I). Collectively, these findings indicated that miR-195 serves a tumor suppressive role by downregulating PRR11 expression.

The in vitro assays indicated that PRR11 downregulation inhibited the proliferative and angiogenic activities of tumor cells. To evaluate the biological functions of PRR11 in vivo, we stably suppressed PRR11 expression in DU145 and LNCaP cells through transfection with a lentivirus expression plasmid containing siRNA against PRR11. As shown in Fig. 5A and B, we chose the DU145 and LNCaP cells transfected with the third shRNA for subsequent experiments. The PRR11-suppressed cell lines (PRR11-shRNA) and control PCa cell lines (scramble) were subcutaneously injected into the right side of male nude mice (8 mice per group). Compared with the controls, the PCa cells with suppressed PRR11 expression formed significantly smaller tumor nodules (Fig. 5C and D). Additionally, the knockdown of PRR11 markedly reduced the growth of tumor nodules and the weight of tumors in the PRR11-shRNA groups on day 42 when compared with the scramble control groups (Fig. 5E-H). The proliferative and angiogenic abilities of the tumor xenografts were subsequently evaluated via histopathological staining for the proliferative marker Ki67 and pan-endothelial marker CD31. The results showed that in the tumor xenografts of each cell line, downregulation of PRR11 reduced the number of vessels (P=0.003 and P=0.015, respectively; Fig. 5I) and suppressed cell proliferation (P=0.003 and P=0.038, respectively; Fig. 5J).

To investigate the expression pattern of PRR11 in PCa, immunohistochemical staining for PRR11 was employed to detect the expression pattern and subcellular localization of PRR11 protein in 126 primary PCa tissues and 22 adjacent non-cancerous prostate tissues (Fig. 6A). As shown in Fig. 6B-D, we identified PRR11-positive staining in the cytoplasm and cellular membrane of PCa and benign glandular epithelium cells, with evenly distributed staining patterns observed at weak, intermediate and strong intensities. PRR11 was recorded as low if the final IRS was no >4. As shown in Table III, we found that high PRR11 expression frequently occurred in the PCa tissues (P<0.001) and in patients with advanced pathological stage PCa (P=0.008).

Finally, we validated the clinical value of PRR11 in the Taylor cohort. As shown in Table III, PRR11 was significantly upregulated in PCa tissues relative to non-cancerous tissues (P<0.001). More notably, PRR11 overexpression was significantly associated with higher Gleason score (P=0.013), more advanced pathological stage (P=0.012), positive metastasis (P<0.001), positive BCR (P=0.003) and shorter overall survival time (P=0.007) in the PCa patients (Table III). Kaplan-Meier analysis was also conducted to assess the prognostic value of PRR11 expression in human PCa. The data demonstrated that PRR11 expression level could markedly stratify the patients into high risk and low risk groups regarding BCR (P<0.001; Fig. 6E). However, there was no difference between the overall survival times of patients with high and low PRR11 expression (P=0.09; Fig. 6F). Meanwhile, univariate analysis revealed that PRR11 expression (P=0.001), pathological tumor stage (P<0.001) and Gleason score (P<0.001) were significant prognostic factors for BCR-free survival time in patients with PCa (Table IV). Additionally, Cox proportional hazards multivariate analysis also indicated that PRR11 expression level (P=0.018), pathological tumor stage (P=0.003) and Gleason score (P<0.001) were independent predictors of BCR-free survival time in PCa patients (Table IV). Taken together, these findings demonstrated that PRR11 overexpression is associated with aggressive tumor behavior and poor prognosis.