Approval for the present study was obtained from the Research Ethics Committee of Guangzhou First People's Hospital Affiliated to Guangzhou Medical University (Guangzhou, China). For miRNA extraction and immunohistochemistry, 98 tumor tissue samples were obtained from PCa patients who underwent radical prostatectomy at Guangzhou First People's Hospital between January 2002 and August 2012. In addition, 20 benign prostate hyperplasia (BPH) specimens obtained by transurethral resection of the prostate were collected for subsequent experiments. Patients with PCa who received radiotherapy or hormonal treatment prior to surgery were excluded. PCa cases were classified by the World Health Organization criteria (21) and staged according to the tumor node metastasis classification (22) and the Gleason grading system (23). The detailed characteristics of these patients are presented in Table I. BCR was defined as a postoperative serum prostate-specific antigen (PSA) level of ≥0.2 ng/ml.

The DU145 prostate cancer cell line (American Type Culture Collection, Rockville, MD, USA) was cultured in RPMI 1640 medium (GE Healthcare Life Sciences, Logan, UT, USA), supplemented with fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and 100 U/ml penicillin and 100 µg/ml streptomycin (Invitrogen; Thermo Fisher Scientific, Inc.) and maintained at 37°C in an atmosphere of 5% CO₂.

For transient transfection, 0.5×10⁵ DU145 cells were plated in 24 well plates, 12 h prior to transfection. The pGCMV/EGFP/Neo vector (catalog no., C05001) with overexpression of miR-30c was purchased from Shanghai GenePharma Co., Ltd. (Shanghai, China). The blank vector was used as negative control. DU145 human prostate carcinoma cells were transiently transfected with mature miR-30c or control (miR-NC) using Invitrogen™ Lipofectamine 2000 reagent (Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. Following the transfection of cells, BCL9 and several Wnt pathway downstream targets were detected by western blot analysis.

Following the standard protocol, total proteins were extracted from cultured cells using 200 µl M-PER Mammalian Protein Extraction Reagent (Thermo Fisher Scientific, Inc.). Protein concentration was measured using a bicinchoninic acid protein assay kit (Thermo Fisher Scientific, Inc.). Total protein concentration was calculated by measuring the absorbance at a wavelength of 260 nm (NanoDrop 2000c Spectrophotometer; Thermo Fisher Scientific Inc., Wilmington, DE, USA). Next, the lysates were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (NuPAGE™ Novex 4–12% Bis-Tris Gel; catalog no., NP0322BOX; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and blotted onto polyvinylidene fluoride (PVDF) membranes (catalog no., IPFL00010; EMD Millipore, Billerica, MA, USA). The PVDF membranes were blocked with 5% skim milk in phosphate-buffered saline (PBS)-Tween 20 and probed with rabbit polyclonal IgG antibodies against human BCL9 (catalog no., ab37305; dilution, 1:200; Abcam, Cambridge, MA, USA), SOX9 (catalog no., sc-20095; dilution, 1:200; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), c-Myc (catalog no., sc-764; dilution, 1:200; Santa Cruz Biotechnology) or CD44 (catalog no., ab41478; dilution, 1:1,000; Abcam), or a mouse monoclonal IgG₁ antibody against human β-actin (catalog no., sc-47778; dilution, 1:100; Santa Cruz Biotechnology) overnight at 4°C. The membranes were then washed with Tris-Buffered Saline with Tween 20 and incubated with alkaline phosphatase-conjugated goat anti-rabbit (catalog no., SAB3700852; dilution, 1:2,000; Sigma-Aldrich, St. Louis, MO, USA) or anti-mouse IgG secondary antibody (catalog no., A2179; dilution, 1:2,000; Sigma-Aldrich) for 1 h at room temperature. Signals were visualized using SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher Scientific, Inc.) and the STORM 860 Molecular Imager (Amersham Biosciences, Uppsala, Sweden).

The putative miR-30c complementary site in the 3′-UTR of BCL9 mRNA, or mutant sequence, were cloned into the pGL3 luciferase reporter vector (Promega Corporation, Madison, WI, USA) (Fig. 1A). Cultured DU145 cells were cotransfected with 10–24 ng Firefly reporter wild-type BCL9 constructs or the mutant vector (3′-UTR-BCL9wt FLuci or 3′-UTR-BCL9mut FLuci vectors; Promega Corporation), 3 ng Renilla reporter plasmid pGL3 and 30 nM miR-30c mimic or NC mimic. Following 48 h of transfection, PCa cells were harvested, and reporter assays were performed using a Dual-Glo® Luciferase Assay System (Promega Corporation) based on the manufacturer's protocol. The results of the relative reporter activity were normalized to the activity of the Renilla luciferase second reporter (internal control), according to the manufacturer's protocol. The experiment was conducted in triplicate.

BCL9 expression was detected by immunohistochemistry assays performed on formalin-fixed, paraffin-embedded slides of PCa and BPH tissues. The tissues were cut into 5 µm-thick sections. Using a Dako EnVision system (Dako Diagnostics AG, Zug, Switzerland), the slides were deparaffinized with xylene and rehydrated for further hematoxylin and eosin and immunohistochemical staining. Following proteolytic digestion (Trypsin Enzymatic Antigen Retrieval Solution; catalog no., ab970; Abcam) and peroxidase blocking with hydrogen peroxide blocking reagent (catalog no., ab94666; Abcam) of tissue slides, the slides were incubated overnight at 4°C with the primary antibody against BCL9 protein (ab37305; Abcam) at a dilution of 1:150. After washing, the staining was visualized with a peroxidase-labeled polymer (EnVision; Dako, Glostrup, Denamark) and DAB substrate-chromogen system (Dako) using an Olympus AX70 microscope (Olympus Corporation, Tokyo, Japan).

The stained slides were scored independently by two experienced pathologists in a blinded manner. If any discrepant scores were generated, the pathologists simultaneously re-examined the slide to achieve a consensus score. The percentages of positively staining cells exhibiting immunoreactivity in the cell nucleus and cytoplasm in 10 representative microscopic fields were calculated and a score of 0–4 was assigned, as follows: 0, 0%; 1, 1–25%; 2, 26–50%; 3, 51–75%; or 4, 76–100%. Meanwhile, the staining intensity of the cells was calculated and scored as follows: 0, no staining; 1, weakly positive; 2, moderately positive; or 3, strongly positive. The sum of the two scores was calculated to determine a final staining score. Tumor specimens with an overall score of ≥3 were considered to be positive.

This assay was conducted as previously described (8). miRNA was extracted from 98 PCa frozen tissues and transfected DU145 cells using an miRNA Fast Extraction Kit (BioTeke Corporation, Beijing, China). The primers specific for miR-30c and the internal control, RNU6B, were purchased from Thermo Fisher Scientific, Inc. The primer sequences were as follows: Forward, 5′-TGTGTTTTTATTGTTTTTGTTGTCCCA-3′ and reverse, 5′-GGGACAGAACAGGTTAATGGGAA-3′ for miR-30c; and forward, 5′-CTCGCTTCGGCAGCACA-3′ and reverse, 5′-AACGCTTCACGAATTTGCGT-3′ for RNU6B. cDNA synthesis and amplification were performed according to the instructions of the All-in-One™ miRNA qRT-PCR detection kit (GeneCopoeia, Inc., Guangzhou, China) with the specific primers. PCR was performed using the MyiQ2 Two-Color Real-Time PCR Detection System (Bio-Rad Laboratories, Inc.) under the following conditions: Initial denaturation step at 95°C for 5 min, followed by 40 cycles at 95°C for 10 seconds and 60°C for 20 sec, with a final extension step at 72°C for 20 sec. All assays were conducted in triplicate. Relative quantification of miR-30c was performed using IQ5 Standard Edition Optical System version 2.0 software (Bio-Rad Laboratories, Inc.) and the comparative quantification cycle (Cq) method (24), and normalized to the expression of RNU6B. For the analysis of the possible correlations between miR-30c expression levels and BCR, the 98 PCa patients were divided into a negative and a positive expression group according to median relative expression of miR-30c.

All statistical analyses were performed using SPSS version 17.0 for Windows (SPSS, Inc., Chicago, IL, USA). Continuous variables were compared using a Students t-test. The associations between miR-30c and BCL9 expression were assessed by Spearman's rank correlation coefficient. Associations between BCL9 and clinicopathological variables were evaluated by Fisher's exact or Pearsons χ² tests. Kaplan-Meier survival curves were generated to evaluate the effect of miR-30c and BCL9 expression levels on survival rate. A Cox proportional hazards regression model was used to establish independent factors associated with BCR. P<0.05 was considered to indicate a statistically significant difference.

The present study first investigated whether BCL9 is modulated by miR-30c in PCa cells. DU145 cells were transiently transfected with miR-30c vector or blank vector (miR-NC). Subsequent RT-qPCR analysis confirmed that the miR-30c expression level was significantly increased in cells transfected with miR-30c compared with those transfected with miR-NC (P<0.001) (Fig. 1B). The cells were then used for further assays. Western blot analysis revealed that ectopic expression of miR-30c was associated with a significant reduction in the expression of BCL9 protein (P=0.007; Fig. 1C and D). Consistent with the role of BCL9 as a transcriptional coactivator of the Wnt signaling pathway, proteins of multiple target genes of the Wnt/β-catenin pathway, including c-Myc, CD44 and SOX9, were observed to be substantially downregulated in miR-30c-transfected DU145 cells compared with their expression in miR-NC-transfected cells (all P<0.001; Fig. 1C) (17,25).

To further determine whether miR-30c may interact directly with BCL9 protein, a luciferase reporter assay was conducted. Using the databases TargetScan (http://targetscan.org/), PicTar (http://pictar.mdc-berlin.de/) and miRDB (http://www.mirdb.org/miRDB/), the 3′-UTR of BCL9 containing a sequence motif matching with the ‘seed’ sequence of miR-30c was identified (Fig. 1A). The wild-type and mutant BCL9 3′-UTR reporter vectors (Fig. 1A) were cotransfected into DU145 cells, together with miR-30c or miR-NC. The luciferase activity of wild-type, but not mutant, BCL9 3′-UTR reporters was significantly downregulated when cotransfected with miR-30c compared with that of miR-NC-transfected cells, confirming that BCL9 is a direct target of miR-30c (P=0.004; Fig. 1E).

Immunohistochemical analyses were conducted on tissue samples from a cohort comprising 98 PCa and 20 BPH cases, using an antibody specific for BCL9. Staining was predominantly distributed in the nuclei of PCa cells (Fig. 2A–C). In some sections of BPH tissues, very faint BCL9 staining was observed when compared to PCa tissues, and was considered in the negative range (Fig. 2D). The positive expression rate of BCL9 in the tissues from PCa patients (52/98; 53.1%) was significantly higher than that in normal prostate tissues (5/20; 25.0%) (P=0.022).

To further investigate the association between BCL9 expression and miR-30c levels, RT-qPCR was conducted to quantify the expression of miR-30c in the samples from the 98 PCa patients. A non-parametric Spearman's rank correlation coefficient (r_S) analysis was then conducted using the cases with both miR-30c and BCL9 protein quantification data. The results indicated that BCL9 and miR-30c expression across these cases was significantly inversely correlated (r_S=0.38; P<0.001).

The possible correlation between BCL9 and clinicopathological features was further investigated in 98 PCa patients. As shown in Table II, increased expression levels of BCL9 were frequently observed in PCa patients with a high Gleason score (P=0.016) and BCR (P=0.020). However, a statistically significant correlation was not identified between BCL9 expression and other features, including preoperative PSA levels (P=0.326), pathological stage (P=0.073) and surgical margin (P=0.111).

Our previous data indicated the involvement of miR-30c in PCa progression and suggested its potential role as an independent predictor of BCR in PCa (8). In the present study, the combined utility of these two biomarkers was investigated with regard to predicting BCR in patients who had undergone radical prostatectomy. A Kaplan-Meier analysis revealed that increased BCL9 expression (P=0.037) and reduced miR-30c expression (P=0.023) were significant predictors of shorter BCR-free survival time (Fig. 3A and B). Furthermore, the group of PCa patients with miR-30c-negative and BCL9-positive expression had a significantly lower BCR-free survival relative to patients with a miR-30c-positive and BCL9-negative combined expression status (P=0.010) (Fig. 3C) or to those with any other combined expression status (P=0.001) (Fig. 3D).

When a univariate Cox proportional hazards regression model was applied, the pathological stage (P=0.001), Gleason score (P<0.001), PSA level (P=0.001) and surgical margin (P=0.003) were identified to be significant predictors of BCR. Furthermore, a high hazard ratio (HR) for BCR was observed in patients with miR-30c-negative and BCL9-positive expression (HR, 5.79; 95% confidence interval, 1.28–26.19; P=0.023) (Table III). On multivariate analysis, miR-30c-negative and BCL9-positive expression (HR, 5.08; P=0.048) and a higher Gleason score (HR, 6.89; P=0.006) were revealed to be independent prognostic factors for poor BCR-free survival (Table III).