Human RWPE-1, DU145, LNCaP, 22RV1, PC-3 and VCaP cells were obtained from the American Type Culture Collection (ATCC; American Type Culture Collection, Manassas, VA, USA) and cultured according to the manufacturer's instructions. The C4-2B cell line was purchased from MD Anderson Cancer Center (Houston, TX, USA). RWPE-1 cells were grown in defined keratinocyte-SFM (1X) (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA). LNCaP, 22RV1, C4-2B and PC-3 cells were grown in RPMI-1640 medium supplemented with 10% FBS, while DU145 and VCaP cells were grown in Dulbecco's modified Eagle's medium (all from Invitrogen; Thermo Fisher Scientific, Inc.) supplemented with 10% FBS. All cells were incubated at 37°C in a humidified atmosphere with 5% CO₂.

The 127 archived PCa tissues, including 81 non-bone metastatic PCa tissues and 46 bone metastatic PCa tissues were obtained during surgery or needle biopsy at Jiangmen Central Hospital (Guangzhou, China). Patients were diagnosed based on clinical and pathological evidence, and the specimens were immediately snap-frozen and stored in liquid nitrogen tanks. For the use of these clinical materials for research purposes, prior informed consents from patients and approval from the Institutional Research Ethics Committee were obtained from Jiangmen Central Hospital (Jiangmen, China). The clinicopathological features of the patients are presented in Table I.

Total RNA from tissues or cells was extracted using TRIzol (Life Technologies; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. Messenger RNA (mRNA) and miRNA were reverse-transcribed from total mRNA using the Revert Aid First Strand cDNA Synthesis kit (Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Real-time PCR was performed according to a standard method, as previously described (15). The list of primers used is presented in Table II. Primers for U6 and miR-505-3p were synthesized and purified by Guangzhou RiboBio Co., Ltd. U6 or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an endogenous control for miRNA or mRNA, respectively. Relative fold expression was calculated using the comparative threshold cycle (2^−∆∆Cq) method (16).

The (CAGAC) 12/pGL3 TGF-β/Smad-responsive luciferase reporter plasmid and control plasmids (cat. no. 32177; Clontech; Takara Bio, Inc., Tokyo, Japan) were used to quantitatively assess the transcriptional activity of TGF-β signaling components. The 3′UTR regions of SMAD2 and SMAD3 were PCR-amplified from genomic DNA and cloned into the pmirGLO luciferase reporter vector (Promega Corporation, Madison, WI, USA). The miR-505-3p mimics were obtained from Guangzhou RiboBio Co., Ltd. Transfection of siRNAs and plasmids was performed using Lipofectamine 3000 (Life Technologies; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions.

Migration and invasion were performed using Transwell chambers consisting of 8-mm membrane filter inserts (Corning Incorporated, Corning, NY, USA) coated with or without Matrigel (BD Biosciences, Franklin Lakes, NJ, USA) as previously described (17).

Cells (4×10⁴) were seeded in triplicate in 24-well plates and cultured for 24 h as previously described (18). Cells were transfected with 250 ng pTGF-β/Smad reporter luciferase plasmid, or pmirGLO-SMAD2-3′UTR and -SMAD3-3′UTR luciferase plasmid, plus 5 ng pRL-TK Renilla plasmid (Promega Corporation) using Lipofectamine 3000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. Luciferase and Renilla signals were assessed 36 h after transfection using a Dual-Luciferase Reporter Assay kit (Promega Corporation) according to the manufacturer's protocol.

Cells were co-transfected with HA-Ago2 (cat. no. 10822; Addgene, Inc., Cambridge, MA, USA), followed by HA-Ago2 immunoprecipitation using an HA-antibody (1:1,000; cat. no. H3663; Sigma-Aldrich; Merck KGaA, Darmstadt, Germany), as previously described (19). Real-time PCR analysis of the IP material was used to assess the association of SMAD2 and SMAD3 mRNA with the RISC complex.

Western blotting was performed according to a standard method, as previously described (20). Proteins were visualized using ECL reagents (Pierce; Thermo Fisher Scientific, Inc., Waltham, MA, USA). Antibodies against SMAD2 (cat. no. 5339), SMAD3 (cat. no. 9523), pSMAD2/3 (cat. no. 9510) and SMAD2/3 (cat. no. 8685) were purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA), and p84 (cat. no. PA5-27816) was obtained from Invitrogen; Thermo Fisher Scientific, Inc. All antibodies aforementioned were diluted 1:1,000. The membranes were stripped and reprobed with an anti-α-tubulin antibody (1:5,000; cat. no. ab7291, Abcam, Cambridge, UK) as the loading control.

The miRNA expression levels and clinical profile datasets from PCa patients were downloaded from The Cancer Genome Atlas (TCGA; http://tcga-data.nci.nih.gov/tcga/). The log2 values of miRNAs in each sample were analyzed using Excel 2010 and GraphPad 5.0 software (GraphPad Software, Inc., La Jolla, CA, USA). The miRNA expression levels of all PCa tissues were statistically analyzed using paired t-tests or unpaired t-tests. The miRNA expression levels in each sample were analyzed as previously described (21). Briefly the high and low expression levels of miR-505-3p were stratified by the medium expression level of miR-505-3p in PCa tissues. Gene set enrichment analysis was performed using Molecular Signatures Database (MSigDB) v5.2.

TargetScan (http://www.targetscan.org/vert_71/) and miRanda (http://34.236.212.39/microrna/home.do) datasets were analyzed as previously described respectively (22,23).

All values are presented as the means ± standard deviation (SD). Significant differences were determined using GraphPad 5.0 software. Student's t-test was used to determine significant differences between two groups. The chi-square test was used to analyze the relationship between miR-505-3p expression and clinicopathological characteristics. Survival curves were plotted using the Kaplan Meier method and compared by log-rank test. P<0.05 was considered to indicate a statistically significant difference. All experiments were repeated three times.

To investigate the aberrant miRNA expression between bone metastatic PCa tissues and non-bone metastatic PCa tissues, we analyzed the miRNA sequencing datasets of PCa tissues from TCGA. We found that that there was no significant difference in miR-505-3p expression between 498 PCa tissues and 52 adjacent normal tissues (ANT) (Fig. 1A), or between 52 paired PCa tissues compared with the expression in the matched ANT (Fig. 1B). Notably, miR-505-3p expression was significantly downregulated in bone metastatic PCa tissues compared with the expression in non-bone metastatic PCa tissues (Fig. 1C). The miR-505-3p expression in our 81 non-bone metastatic PCa tissues and 46 bone metastatic PCa tissues was further validated via real-time PCR and the results revealed that miR-505-3p expression was significantly downregulated in bone metastatic PCa tissues compared with the expression in non-bone metastatic PCa tissues (Fig. 1D). The percentage of low miR-505-3p expression was much higher in bone metastatic PCa tissues compared to that in non-bone metastatic PCa tissues (Fig. 1E), which was consistent with the results of TCGA (Fig. 1F). Therefore, these results indicated that downregulation of miR-505-3p was involved in bone metastasis of PCa.

We further investigated the clinical association of miR-505-3p expression with clinicopathological characteristics in PCa patients and found that miR-505-3p expression was downregulated in N1, M1 or Gleason grade (>7) PCa tissues, but not in T3-T4 PCa tissues (Fig. 2A-D). Statistical analysis of PCa tissue samples revealed that miR-505-3p expression was inversely associated with serum PSA levels, Gleason grade, N classification, M classification and bone metastasis status in PCa patients (Table I). Kaplan Meier analysis revealed that miR-505-3p expression levels had no effect on the overall 5-year survival in PCa patients (Fig. 2E); whereas low expression of miR-505-3p was strongly and positively associated with shorter bone metastasis-free survival in PCa patients (Fig. 2F). Thus, our results demonstrated that downregulation of miR-505-3p predicted poor bone metastasis-free survival in PCa patients.

To determine the biological role of miR-505-3p in bone metastasis of PCa, we performed gene set enrichment (GSEA) analysis of miR-505-3p expression against the oncogenic signatures collection of MSigDB, and the results revealed that downregulation of miR-505-3p was significantly and positively associated with metastatic propensity (Fig. 3A-D). Therefore, we further examined whether miR-505-3p has an effect on the invasion and migration of PCa cells. We first examined miR-505-3p expression in normal prostate cells (RWPE-1) and 6 PCa cell lines and found that miR-505-3p expression was differentially downregulated in metastatic cell lines, including lymph node metastatic cell line LNCaP, brain metastatic PCa cells DU145 and bone metastatic PCa cell lines (PC-3, C4-2B and VCaP), particularly in PC-3 and C4-2B, but slightly elevated in primary PCa 22RV1 cells (Fig. 3E). This finding further elucidated the pro-metastatic roles of miR-505-3p downregulation in PCa. According to the expression levels of miR-505-3p shown in Fig. 3E, we upregulated miR-505-3p in bone metastatic PC-3, C4-2B and VCaP cells via transfection with miR-505-3p mimics to further evaluate the effects of miR-505-3p on the invasion and migration of PCa cells (Fig. 3F). As shown in Fig. 3G, miR-505-3p overexpression significantly decreased the invasive and migratory abilities of PCa cells. Therefore, these results indicated that low expression of miR-505-3p was implicated in bone metastasis of PCa cells via promotion of the invasion and migration abilities of PCa cells.

GSEA analysis of miR-505-3p expression revealed that downregulation of miR-505-3p was positively associated with TGF-β signaling activity (Fig. 4A-C). These results indicated that miR-505-3p may negatively regulate the TGF-β signaling pathway, which has been demonstrated to play an important role in bone metastasis of PCa (24). As shown in Fig. 4D, miR-505-3p overexpression reduced the transcriptional activity of the TGF-β/Smad-responsive luciferase reporter in PCa cells. Cellular fractionation and western blot analysis revealed that overexpression of miR-505-3p decreased nuclear accumulation of pSMAD2/3 (Fig. 4E). Real-time PCR analysis revealed that upregulation of miR-505-3p differentially reduced downstream target genes of TGF-β signaling in PCa cells, including IL11, PTHRP and CTGF (Fig. 4F-H). Thus, these results indicated that miR-505-3p inhibited the activity of TGF-β signaling in PCa cells.

Using the publicly available algorithms TargetScan and miRanda, we found that the critical downstream effectors of TGF-β signaling SMAD2 and SMAD3 were potential targets of miR-505-3p (Fig. 5A). Real-time-PCR and western blot analysis revealed that miR-505-3p overexpression reduced the mRNA and protein expression levels of SMAD2 and SMAD3 (Fig. 5B and C). Luciferase assays revealed that miR-505-3p overexpression suppressed the luciferase reporter activity of SMAD2 and SMAD3, but not the expression of SMAD2 and SMAD3 with the 3′UTR mutation within the miR-505-3p-binding seed regions in PCa cells (Fig. 5D-F). Microribonucleoprotein (miRNP) immunoprecipitation (IP) assays revealed a direct association of miR-505-3p with SMAD2 and SMAD3 transcripts (Fig. 5G-I), further elucidating the direct suppressive effects of miR-505-3p on SMAD2 and SMAD3. Collectively, these results demonstrated that miR-505-3p inhibited the activity of TGF-β signaling by targeting SMAD2 and SMAD3in PCa cells.

We further investigated whether SMAD2 and SMAD3 had an effect on the invasion and migration abilities in miR-505-3p-overexpressing PCa cells and found that individual upregulation of SMDA2 and SMAD3 markedly reversed the inhibitory effects of miR-505-3p on the invasion and migration abilities in PCa cells (Fig. 6). These results indicated that miR-505-3p suppresses the invasion and migration abilities by inhibiting SMAD2 and SMAD3 activity in PCa cells.