The approval of this research by the Ethical Committee of Guangzhou First People's Hospital (Guangzhou Medical University, China) was provided prior to the commencement of the project. All patients recruited in the present study provided written informed consent.

The present study collected 20 pairs of PCa tissues and adjacent normal tissues following radical prostatectomy at Guangzhou First People's Hospital. The tumor tissue sections were frozen in liquid nitrogen. The tissue microarrays (TMAs; PR807c; Alenabio) contained 50 primary PCa tissues, 10 normal prostate tissues and 20 benign prostatic hyperplasia (BPH) tissues along with detailed follow-up data for further immunohistochemical staining. No previous treatment had been performed on the patients with PCa. Additional investigations were conducted using The Cancer Genome Atlas (TCGA) database, which contains 499 human primary PCa tumors with clinicopathological information, to complement our TMA results. The TCGA dataset was downloaded from the cBioPortal for Cancer Genomics. Biochemical recurrence (BCR) survival and disease-free survival was calculated (11–13). BCR was defined as postoperative serum prostate-specific antigen (PSA) >0.2 ng/ml. Disease-free survival was defined as the time to the first evidence of loco-regional or distant clinical recurrence after the initial surgery.

PC-3, DU145 and LNCaP cell lines (PCa) and RWPE-1 (normal prostate) were purchased from American Type Culture Collection (ATCC). Cells were cultured in a humidified incubator at 37°C with 5% CO₂ according to the protocol outlined in a previous study (11–13). Human premicroRNA Expression Construct Lenti-miR-505 [PMIRH505PA-1; System Biosciences (SBI), LLC] was used as the pMIRNA1 lentivectors to express miR-505 precursor (pre-505). Scramble control hairpin in pCDH-CMV-MCS-EF1α-copGFP (CD511B-1) was designed as a negative control (pre-NC; cat. no. PMIRH000PA-1; SBI). With pPACKH1 Packaging Plasmid Mix (cat. no. LV500A-1; SBI), pre-505/pre-NC were transfected into 293TN cells (SBI) in order to package the construct. By using the LentiConcentin Virus Precipitation Solution (cat. no. LV810A-1; SBI), the virus particles were collected according to the manufacturer's protocol. Following transfection, PC-3 and LNCaP cells were isolated and then seeded.

NRCAM expression plasmid (pCMV-NRCAM) was obtained from Biogot Technology Co., Ltd. The cells transfected with pCMV-NRCAM or pCMV (empty vector) were used as a corresponding control. The cells were isolated for functional analyses at 48 h following transfection.

Gene expression profiles of miR-505-overexpressing LNCaP cells and its corresponding control cells were conducted and normalized. The experiments were peformed according to the manufacturer's protocol as outlined in a previous study (12). Briefly, total RNA from the transfected cells was amplified, labeled and purified. Then array hybridization, wash and scan were determined. Gene Spring Software (Agilent Technologies, Inc.) was used to analyze the data, and genes with a fold change greater than two folds (P<0.05) were selected.

As previously described (11–13), total miRNA of PCa cells and tissues was isolated with the miRNA Isolation Kit (BioTek China), and total RNA was extracted with the RNeasy mini kit (Qiagen GmbH) when detecting the miRNA and mRNA expression, respectively. The cDNA was synthesized using an RT-qPCR Detection kit (GeneCopoeia). The primer sequences were obtained from Thermo Fisher Scientific, Inc. and are presented in Table I. RT-qPCR analysis was performed and the relative changes of miR-505 and the predictive targets were normalized to the levels of RNU6B RNA or 18S rRNA. The 2^−ΔΔCq values were calculated to the relative expression (23).

Immunohistochemistry was performed on the tissue sections of TMAs of the patients with PCa to gauge the expression levels of NRCAM. Following deparaffinization and dehydration, sections were processed to reveal antigens using a microwave oven. Briefly, the sections were blocked and then incubated with primary antibody against NRCAM (1:100, cat. no. ab87427; Abcam). The slides were subsequently incubated with anti-rabbit secondary antibody and visualized by DakoCytomation Liquid DAB plus Substrate Chromogen System (DakoCytomation).

Two pathologists, who were blind to patient data independently evaluated the TMA samples. The immunohistochemistry score was calculated by the sum of the staining intensity and the fraction of positive tumor cells as previously described (11–13).

The potential targets of miR-505 were determined by three different predicting programs, incuding TargetScan (24), miRanda (25), and miRWalk (26). The miRanda-miRSVR scores, TargetScan context score or miRWalk scores were defined to aggregate per gene and miR-505. The intersection of miRanda (score <-0.5) and TargetScan (contextscore <-0.2) and miRWalk (score >0.80) were used to predict the targets of miR-505.

As previously described (11–13), luciferase vectors were designed to express the wild-type (WT) or mutant (MUT) 3′-UTR NRCAM sequences. The PCa cells were co-transfected with WT luciferase vector and miR-505 mimic and the corresponding controls. At 24 h post-transfection, the fluorescence reader (Promega Corporation) was employed to calculate the luciferase signals.

As previously described (11–13), a CCK-8 assay was performed to monitor the proliferative ability of PCa cells following transfection. Briefly, cells were incubated with CCK-8 (Beyotime Institute of Biotechnology) at 37°C containing 5% CO₂ for 4 h. The proliferative ability of PC-3/LNCaP cells was determined at 4, 24, 48 and 72 h.

This assay was performed according to the protocol outlined in previous studies (11–13). When LNCaP and PC-3 cells reached confluence, a linear scratch wound was induced in monolayers with a sterile pipette tip. After washing, the migrated cells through the scratch wound were observed and calculated at 0 and 48 h.

As previously described (11–13), the CytoSelect Cell Invasion Kit supplied by Cell Biolabs was used to detect invasive abilities of PCa cells according to the manufacturer's protocol. Briefly, PC-3/LNCaP cells were collected, and suspended in serum-free DMEM. These cells were located on the Transwell inserts coated with Matrigel (BD Biosciences), while the normal medium was used as an attractant. Following a 48-h incubation, 4% paraformaldehyde was used to fix the membranes at room temperature for 15 min, which were then stained with 0.1% crystal violet at room temperature for 10 min. Finally, the number of cells that had invaded through the membranes were counted in 16 randomly-selected fields of view under an optical microscope (Olympus Corporation) using ×100 and ×400 magnifications.

As previously described (11–13), the present study trypsinized PC-3/LNCaP cells following transfection according to the manufacturer's protocol. Briefly, the cells were washed with phosphate-buffered saline (PBS), stained with propidium iodide (Sigma-Aldrich; Merck KGaA) and incubated for 30 min at room temperature. Finally, flow cytometry (BD Biosciences) was conducted to obtain propidium iodide signals.

The experiment was performed according to the manufacturer's protocol (11–13). Briefly, cells were collected, washed, stained with Annexin V and propidium iodide, and then subjected to the FACScan flow cytometer (BD Biosciences). BD FACSuite™ software (BD Biosciences) was conducted for further calculation.

The results are presented as the mean ± SD and were analyzed using SPSS version 17.0 for Windows (SPSS, Inc.). The unpaired Student's t-test or Mann-Whitney U test were conducted to compare the two groups. One-way analysis of variance (ANOVA) or two-way ANOVA with Bonferroni's post hoc test were also used when comparing more than two groups. To determine the associations between NRCAM and clinicopathological features, Fisher's exact test or Pearson χ² test were performed. Kaplan-Meier method and Cox proportional hazards regression model were conducted for survival estimation. P<0.05 was considered to indicate a statistically significant difference.

The present study, aimed to validate the expression of miR-505 with RT-qPCR in PCa cells and tissues, and the RT-qPCR results demonstrated that miR-505 expression was significantly decreased in PCa tissues when compared with normal tissues (P=0.037; Fig. 1A). The data also revealed that the expression levels of miR-505 were also significantly reduced in PCa cells (P<0.001; Fig. 1B).

The present study then constructed stable LNCaP and PC-3 cell lines expressing miR-505 following lentiviral transduction, and the success of this was confirmed via RT-qPCR (P<0.05; Fig. 2A and B). CCK-8 assays confirmed that the proliferative activities of miR-505-overexpressing LNCaP and PC-3 cells was significantly inhibited when compared with control cells at 24, 48 and 72 h following transfection (all groups P<0.05; Fig. 2C). In addition, flow cytometric analysis indicated that miR-505 could induce cell apoptosis (P<0.01; Fig. 2D and E), which resulted in significant decrease in the number of PCa cells in the G2+S phase (P< 0.05; Fig. 2F and G). Transwell assays revealed that miR-505 suppressed the invasion of PCa cells (P<0.001; Fig. 3A), and wound healing assays revealed that miR-505 also markedly inhibited the migration of PCa cells (P<0.01; Fig. 3B).

Gene expression profiles of miR-505-overexpressing LNCaP cells and its corresponding control cells were compared. From the microarray data, 406 differentially expressed genes (235 genes upregulated and 171 genes downregulated) were identified. It is well established that miRNAs negatively modulate mRNA stability and translation by directly binding to the 3′UTR of specific target genes (4). Therefore, the 171 downregulated genes detected were further analyzed using bioinformatics tools.

The potential targets of miR-505 were determined by three different types of predicting software. In combination with our microarray data of miR-505-overexpressing PCa cells, eight candidate targets were selected, including LPL, NOV, IRF6, SOX6, CACNA2D3, NRCAM, AMOT, and GREM1, which were downregulated in LNCaP cells of miR-505 overexpression (Fig. 4A-C, P<0.05 vs. pre-NC; P<0.01 vs. pre-NC). RT-qPCR confirmed that AMOT, CACNA2D3, NRCAM, IRF6, GREM1 and LPL were significantly decreased in PCa cells (LNCaP and PC-3) with miR-505 overexpression (Fig. 4D, P<0.01 vs. pre-NC). As the oncogenic role of NRCAM had been observed in several malignancies (18–21), NRCAM was selected for subsequent investigation.

In addition, the luciferase reporter assay demonstrated that miR-505 interacts directly with the 3′UTR of NRCAM (Fig. 5A and B, P<0.05). Furthermore, RT-qPCR indicated that restoration of miR-505 attenuated the NRCAM protein expression in PC-3 and LNCaP cells (Fig. 5C, P<0.01 vs. pre-NC).

To reveal whether the inhibitory effect of miR-505 in PCa functioned by mediating NRCAM, pCDNA3.1(+)-Vectors expressing NRCAM were constructed. RT-qPCR revealed that the tumor-suppressive role of miR-505 on NRCAM protein levels in PCa cells could be rescued by NRCAM overexpression (Fig. 5C, P<0.01 vs. pre-NC). In addition, NRCAM stimulation antagonized the inhibitory role of proliferation (Fig. 6A and B), migration (Fig. 6C and E), and invasion (Fig. 6D and F) of LNCaP and PC-3 cells, which were induced by miR-505 upregulation.

To analyze the protein expression of NRCMA in PCa, a TMA was used consisting of 50 PCa tissues, 20 benign hyperplasia tissues, and 10 normal tissues. The immunohistochemical staining revealed that NRCAM was primarily expressed in the cytoplasm of the PCa cells (Fig. 7A, C-F). However, faint NRCAM staining was also identified in the majority of BPH and normal tissues (Fig. 7A, G and H). The present study demonstrated that NRCAM expression was significantly upregulated in PCa cells when compared with benign prostate epithelium (Fig.7B, P<0.05). In addition, the PCa patients with advanced stage exhibited an increased NRCAM expression when compared with lower stage tumors (Fig. 7B, P=0.017).

The present study also performed investigations using The Cancer Genome Atlas (TCGA), which contains data from 499 primary human tumors with PCa-specific mortality and other clinicopathological information. As presented in Table II, the TMA datasets indicated that high NRCAM expression was correlated with a higher Gleason score (P=0.032) and advanced pathological stage (P=0.010), data from TCGA also revealed NRCAM was strongly associated with lymph node metastasis (P<0.001) and Gleason score (P=0.013).

Patients with a low miR-505 expression had a significantly worse BCR after radical prostatectomy as analyzed by the Kaplan-Meier method (P=0.009; Fig. 8A); however, a significant trend for disease-free survival was not observed (P=0.081; Fig. 8B). In addition, high levels of NRCAM were significantly correlated with unfavorable BCR-free survival (P=0.001; Fig. 8C) and disease-free survival (P=0.011; Fig. 8D), respectively. The combined use of these two biomarkers in prognosis is worthy of further investigation. Collectively, these results indicated that patients with PCa with a low expression of miR-505, and a high expression of NRCAM exhibited poor BCR-free survival (P<0.001; Fig. 8E) and disease-free survival (P<0.01; Fig. 8F) when compared with contrasting groups of patients. The multivariate model revealed that a high expression of NRCAM (Hazard Ratio 3.63; P=0.016) may serve as an independent prognostic factor for unfavorable disease-free survival (Table III).