Upregulation of miR-130b Contributes to Risk of Poor Prognosis and Racial Disparity in African-American Prostate Cancer

Prostate cancer is among the top estimated diagnosed cancers and second leading cause of cancer-related deaths in U.S. men (SEER Cancer Stat Facts 2019). African-American (AA) men have higher incidence and significantly higher prostate cancer mortality rates than European-American (EA) men (1). A number of groups have found that AA patients have malignant transformation and also greater prostate tumor volumes in radical prostatectomies as compared with similarly staged EA patients (2, 3). While some of the differences in prostate cancer survival or mortality can be attributed to socioeconomic status (SES; refs. 4, 5), some studies have shown less or inconsistent associations between SES and prostate cancer mortality/incidence (6, 7). Another study found that Crude Cox models showed AA men had higher biochemical recurrence (BCR) risk that persisted after adjustment for covariates including SES although AA men were more likely to have lower SES than white men (8). A recent study also reported that there are no interactions between race and SES for BCR. AA men were found to have 10% to 11% increased BCR risk while SES was unrelated to BCR (9). These studies indicate that biological background accounts for a significant portion of prostate cancer disparity with regard to mortality, incidence, and progression in AA men compared with EA men. However, further investigations are required to uncover the mechanisms underlying abnormal gene regulation and racial disparity. Also, there is an unmet need to develop prognostic biomarkers that contribute to AA prostate cancer health disparity.

miRNAs are a small noncoding RNAs that regulate gene expression either by regulating stability or by translational degradation of their cognate mRNAs in multiple cancers (10). miRNAs are potentially powerful diagnostic tools due to their stability in frozen biopsies, formalin-fixed paraffin-embedded (FFPE) tissue samples, blood, and exosomes (11, 12). Unlike studies of racial disparity in relation to protein-coding genes, the role of miRNAs has not been fully explored in prostate cancer.

Dysregulation of miR-130b is known in various human cancers. In hepatocellular carcinoma (HCC), PROM1⁺ and PROM1⁻ cells from human HCC clinical specimens and cell lines identified overexpression of miR-130b in CD133⁺ tumor-initiating cells. Increased miR-130b also reduced TP53INP1 that is a known miR-130b target (13). In prostate cancer, it has been reported that castration-resistant prostate cancers had c-Met activation that increases miR-130b levels, inhibits androgen receptor expression, promotes cancer spreading and resistance to hormone ablation therapy (14). Integrated public database analysis also supports that miR-130b is an oncogenic miRNA. Most published reports concerning miR-130b (27 of 31) reveal upregulation of the miR-130b cluster in prostate cancer that is correlated with adverse outcomes such as tumor malignancy and T-stage in prostate cancer (15). Despite increasing evidence for miR-130b in tumorigenesis, the clinical significance and functional role of miR-130b in prostate cancer racial disparity remains unexplored.

In this study, we sought to elucidate the role of miR-130b in the genetics of prostate cancer health disparity in AA patients. In addition, we examined the biomarker potential of miR-130b to predict prognosis of AA patients with prostate cancer. We identified miR-130b as a significant contributor to prostate cancer racial disparity in AA men. Our results show that (i) miR-130b is significantly overexpressed in AA compared with EA tissue samples and cell lines; (ii) miR-130b has prognostic biomarker potential in AA patients with prostate cancer; (iii) attenuation of miR-130b expression in AA cell lines results in suppression of tumorigenic characteristics as evidenced by repressed cell proliferation, migration, invasion, and induction of cell-cycle arrest; (iv) downregulation of miR-130b modulates essential biological pathways in AA prostate cancer cell lines; (v) FHIT and CDKN1B were confirmed to be direct functional targets of miR-130b. Thus, this study shows that miR-130b plays important role in prostate cancer racial disparity. Also, it may be a risk factor of poor prognosis in patients with AA. Thus, regulation of miR-130b may offer a preventative and therapeutic approach for AA patients with prostate cancer.

Patient studies were conducted as per the “Declaration of Helsinki” ethical guidelines. Informed written consent was obtained from the patients involved in this study. The study was approved by the Institutional Review Board (IRB) on Human Research Protection Program (IRB approval #: 11-06755, ref. #: 225293) at the San Francisco Veterans Affairs Medical Center (San Francisco, CA) and University of California San Francisco (San Francisco, CA).

All cell lines were acquired from the ATCC and were authenticated using DNA short -tandem repeat profiling by ATCC. EA-derived cell lines DU-145, PC3, and LNCaP were maintained in RPMI1640 (Thermo Fisher Scientific) medium according to the manufacturer's protocol. AA cell line, MDA-PCa-2B, was maintained in BRFF-HPC1 (Athena ES, Thermo Fisher Scientific) medium with poly-l-lysine–coated plates according to the manufacturer's protocol.

Clinical formalin fixed paraffin embedded (FFPE) tissue samples from AA (n = 36) and EA (n = 57) were obtained from the Urology Tissue Bank of the San Francisco Veteran Affairs Medical Center (SFVAMC cohort). Clinical data are given in Supplementary Table S1 and S2 and the number of samples for which information was not available are indicated in each variable in same tables. After admission, all patients were followed until their deaths and recorded in the Computerized Patient Record System (CPRS) of the Department of Veterans Affairs. Follow-up was performed by phone or outpatient visits. Overall survival time endpoints were defined from the time of surgery until time of the death or last follow-up. RNA extraction, cDNA synthesis, and TaqMan miRNA assays (Thermo Fisher Scientific) were performed as described previously (16).

The expression levels of mRNAs and miRNAs in 455 (59 AA and 396 EA) prostate cancer patient samples (41–78 years, 61.0 ± 6.7 (mean ± SD) years; Gleason score, 7.5 ± 0.9 (mean ± SD)) were obtained from the Cancer Genome Atlas (TCGA) data portal (https://tcga-data.nci.nih.gov/tcga/) and the Cancer Genetic Ancestry Atlas (TCGAA; ref. 17).

Real-time reverse transcription–PCR (RT–PCR) was carried out using a Quant Studio 7 PCR System, TaqMan Universal PCR Master Mix, TaqMan Reverse Transcription kit, TaqMan miRNA assays (Thermo Fisher Scientific; TaqMan MicroRNA Assays-hsa-miR-130b; catalog no. 000456) according to the manufacturer's instructions. Anti-miR-130b was measured by Custom TaqMan Small RNA Assay (Thermo Fisher Scientific; catalog no. 4440418; designed from antisense of mature miR-130b; 3′-ATGCCCTTTCATCATTGCACTG-5′). The expression levels of miRNA were determined by normalizing with RNU48 (Thermo Fisher Scientific; catalog no. 1006). A QuantiFast SYBR Green PCR Kit was also utilized for gene expression analysis of miR-130b targets. The primers used for SYBR Green–based qPCR analyses are listed in Supplementary Table S3.

A total of 7.5 μg of psPAX2 and 2.5 μg of pMD2.g plasmids were used to pack 10 μg of constructed pLenti-III-mir-Off viral vector (ABM). The filtered medium containing vector was used immediately for infection of prostate cancer cell lines for three days. Then, stable cells were selected with 5 μg/mL puromycin for 1–4 weeks. Antisense miR-130b expression levels were confirmed by qPCR using a custom TaqMan small RNA assay kit (Thermo Fisher Scientific) according to the manufacturer's protocol.

Cell pellets were stained with PI/RNase Staining Buffer (BD Biosciences). Cells were analyzed with a BD FACSVerse (BD Biosciences). All analyses were performed in triplicate, and 10,000 gated events/sample were counted. Cell-cycle–related genes were referred from KEGG PATHWAY Database (https://www.kegg.jp/kegg/pathway.html).

Wound-healing assays were performed with CytoSelect (Cell Biolabs. Inc.) 24-well Wound Healing Assay kit following the manufacturer's instructions. The wound regions calculated by ImageJ (MRI Wound Healing Tool).

For colony-forming assays, anti-miR-130b/control stable MDA-PCa-2b and DU-145 cells were plated onto 6-well plates (MDA; 2,000 cells, DU; 200 cells). After 10 days for cell adherence, colonies of cells were fixed and stained with crystal violet reagent. ImageJ plug-in, “ColonyArea” was used to calculate the occupied areas of the cell colony as described in a published report (18).

Expression profiling of 84 genes involved in cancer was performed utilizing human RT2 Profiler PCR Array (Qiagen, catalog no. PAHS-502Z). Using RNAs isolated from anti-miR-130b stable DU-145 and MDA-PCa-2b cells were applied to the qPCR-arrays. Fold-change calculation and heatmap analysis were done using RT2 Profiler PCR Array Data Analysis (Qiagen) according to manufacturer's protocol.

Over-representative analysis was carried out as described in a published report (19) by using the Max Planck Institute, ConsensusPathDB (CPDB; http://cpdb.molgen.mpg.de). Detailed methodology is described in the documentation of CPDB.

Western blot analyses were performed as described previously (16). The antibodies used were specific for FHIT (Millipore Sigma; 1:1,000 catalog no. 71-9000), CDKN1B (p27 Kip1, Cell Signaling Technology; 1:1,000, catalog no. 2552), CDKN1A [p21 Waf1/Cip1 (12D1)], Cell Signaling Technology; 1:1,000, catalog no. 2947), β-actin (Cell Signaling Technology; 1;1,000, catalog no. 3700), and GAPDH (Santa Cruz Biotechnology; 1:1,000, catalog no. sc-32233).

Cell viability assay was performed using CellTiter-Glo Luminescent Cell Viability Assay (Promega, catalog no. G7570) as described previously (16). Cell viability was measured every 24 hours for four days using established anti-miR-130b and control stable prostate cancer cells following the manufacturer's instructions.

Cell invasion was carried out in chambers coated with Matrigel. Cells were seeded in serum-free medium in the top chamber inserts and cells were allowed to invade into the bottom chambers, which contained medium supplemented with 10% FBS. Inserts were then stained, and the invaded cells were photographed using a light microscope. Cell stain solution was then dissolved with methanol and absorbance was measured at 540 nm using a SpectraMax plate reader according to manufacturer's instructions.

IHC staining was performed with 15 [14 prostate cancer and 1 benign prostate hyperplasia (BPH) tissues] AA and 10 (9 PC and 1 BPH) EA specimens. The Vision UltraVision Detection System (Thermo Fisher Scientific) was used to stain FHIT proteins (Thermo Fisher Scientific, catalog no. 71-9000). The intensity of the staining was classified following Allred scoring (Harvey and colleagues 1999). Negative staining was “0,” low staining “1,” and high staining a “2–3” score. For correlation of FHIT staining with death of patients, we divided the staining intensity into negative group (0 staining) and positive group (1–3 staining score).

Luciferase reporter vectors were constructed by ligation with annealed custom oligonucleotides containing the putative target binding sites of 3′-UTR into pmiR-GLO reporter vector (Promega, catalog no. E1330; Supplementary Table S4). Anti-miR-130b–stable MDA-PCa-2b cells were transfected with 0.1 ng of pmiR-GLO vector containing the 3′-UTR sequences and luciferase activity was measured 48 hours after transfection using a Dual-Luciferase Reporter Assay System (Promega) as per manufacturer's instructions. Relative luciferase activity was calculated by normalizing to Renilla luminescence.

miR-130b overexpression was carried out with mirVana miRNA mimic hsa-miR-130b-3p (Thermo Fisher Scientific; catalog no. 4464066, Product ID MC10777) and mirVana miRNA mimic, Negative Control #1 (Thermo Fisher Scientific; catalog no. 4464058). jetPRIME (Polyplus, catalog no. 114-15) reagent was used for transient transfection of miR-130b mimic and negative control #1.

Statistical analyses were conducted with GraphPad Prism 8 and EZR (20). All quantified data represent the average of three or more independent experiments performed at different times or as indicated. Error bars show SEM of independent experiments. All tests were performed two-tailed and P values <0.05 were considered statistically significant. TCGA gene and miRNA sequencing data were normalized by quintile method to eliminate the bias of the differences between individual patients. miRNA sequencing data were analyzed using Mann–Whitney U test. Youden index was calculated to generate a cut-off value to determine miR-130b high or low expression groups with the SFVAMC EA and AA samples (Supplementary Fig. S1). The endpoints were defined from time of surgery until time of the death, or last follow-up. This cut-off value was used for Kaplan–Meier and uni-/multivariate analyses. Kaplan–Meier analyses were used for overall survival curves and P values were calculated with the log-rank test. Uni- and multivariate analyses were computed by the Cox proportional hazards model. The Pearson r correlation was computed to compare the correlation between miR-130b and cell-cycle–related genes in the TCGA cohort.

For the initial screening, we analyzed the expression of a panel of miRNAs in AA compared with EA prostate cancer samples. We observed significant differential expression profiles in our in-house cohort and in the published literature in AA compared with EA prostate cancer samples for miR-24 (16), miR-34b (21), miR-96, and miR-130b (22). In this study, we investigated the role of miR-130b in prostate cancer racial disparity. First, we confirmed overexpression of miR-130b in prostate cancer tissue samples from AA patient cohorts compared with EA patient samples. Higher expression levels of miR-130b were observed in AA (n = 36) compared with EA patients (n = 57; Fig. 1A; clinicopathologic feature summary and its correlation with miR-130b expression in Supplementary Table S1 and S2, respectively). We further divided samples into low- [Gleason Score (GS)<7(4+3)] and high-grade [GS>7(4+3)] groups and observed that miR-130b expression was consistently higher in AA compared with EA samples in both groups (Fig. 1B). We validated the expression profiles in an independent sample cohort from TCGA PRAD and TCGAA (AA, n = 59; EA, n = 396) datasets (Fig. 1C). A trend in miR-130b overexpression was observed in total (Fig. 1C) and low-grade samples (Fig. 1D, left panel) in AA compared to EA patients. However, significant miR-130b over-expression was observed in high-grade AA compared with EA samples (Fig. 1D, right). We also examined miR-130b expression in EA compared with AA cell lines. miR-130b was underexpressed in EA relative to AA cell lines (Fig. 1E). These results confirm that miR-130b is overexpressed and can act as a contributive factor in aggressive AA prostate cancer.

To determine the clinical significance of miR-130b in AA prostate cancer, we analyzed the correlation between miR-130b expression and overall survival time (OS) in the SFVAMC patient cohort. Kaplan–Meier survival curve analysis showed that higher miR-130b is expression significantly correlated (P = 0.003) with poor OS in AA patients (Fig. 1F). For EA patients, miR-130b expression was not associated with overall survival (Fig. 1G). Univariate (P = 0.0076) and multivariate (P = 0.0078) analyses revealed miR-130b expression to be an independent risk factor linked to poor prognosis in the SFVAMC AA prostate cancer patient cohort (Table 1, top). However, this association was not observed in EA patient cohort (Table 1, bottom). These results confirm that miR-130b overexpression is a risk factor for poor prognosis in AA patients with prostate cancer.

To determine the functional significance of inhibition of miR-130b expression, DU-145, LNCaP (EA cell lines), and MDA-PCa-2b (AA cell line) were transfected with either anti-miR-130b or control lentiviral vector. After establishing anti-miR-130b stable clones in all cell lines, we confirmed anti-miR-130b expression levels by Custom TaqMan qPCR analysis (Fig. 2A). The efficiency of anti-miR-130b in all cell line–stable clones was similar (Fig. 2A). Next, these stable anti-miR-130b cells were used for a series of in vitro assays. miR-130b inhibition reduced cell proliferation exclusively in MDA-PCa-2b (Fig. 2B) in comparison with controls, whereas no such effect was observed in the DU-145 and LNCaP cell lines (Fig. 2B). Similarly, colony formation assay revealed that MDA-PCa-2b cells stably expressing anti-miR-130b resulted in significantly less colony area (P = 0.0008) compared with control (Fig. 2C, top). This effect was not observed in DU-145–stable cells (Fig. 2C, bottom).

FACS cell-cycle analysis showed no change in DU-145 and LNCaP cells between miR-130b knockdown and control cell (Fig. 2D). However, G₂–M arrest was induced in stable MDA-PCa-2b anti-miR-130b cells (controls 15% vs. anti-miR-130b cells 24.4%, P = 0.001; Fig. 2D). Wound-healing migration and chemotactic invasion assays revealed that miR-130b inhibition decreased migration (P = 0.0084; Fig. 2E, top) and invasion (P = 0.032; Fig. 2F, bottom) in stable anti-miR-130b MDA-PCa-2b cells. These effects were not observed in DU-145–stable cells (Fig. 2E, bottom; Fig. 2F, bottom).

Because miR-130b inhibition increased G₂–M arrest in AA prostate cancer cells, we analyzed the correlation between miR-130b and cell-cycle–related gene expression by Pearson r correlation in TCGA prostate cancer samples. We used R package, “Hmisc” (23) to calculate the matrix of Pearson r correlation for all possible pairs of patient gene expression levels (Supplementary Fig. S2). We arranged the r value in descending order. Figure 3A lists the 9 genes that showed the highest negative correlation with miR-130b expression. These r values are based on all possible pairs of patients' gene expression. Next, we determined the correlation of the five topmost negatively correlated genes in AA patients individually (Fig. 3B). The same genes were then correlated in EA patients individually (Fig. 3C). On the basis of these computational analyses, we observed that CDKN1B showed the most inverse correlation (r = 0.26; P = 0.04) with miR-130b expression only in AA patients. However, other genes were more or less similar in both races and hence may not be race related. We also confirmed CDKN1B upregulation by qPCR and Western blot analyses (Fig. 3D and E). To confirm that CDKN1B is direct target of miR-130b, we performed luciferase reporter assays. Two out of three reporter vectors harboring the miR-130b site showed that luciferase activities were upregulated in miR-130b–stable knockdown AA cells compared with deletion vector-transfected cells (Fig. 3F).

We conducted gene expression array analysis using a commercial assay containing oncogenic and tumor suppressive gene sets and compared gene expression patterns in anti-miR-130b–stable cell lines to determine which biological pathways are involved in prostate cancer racial disparity (Fig. 4A). These gene sets were curated by the manufacturer and represent oncogenic and tumor suppressor genes on the qPCR array plate. We randomly selected three up- and downregulated genes from the array data to validate by qRT-PCR analysis. The qRT-PCR results were consistent with the array data and showed that the expression of these genes is modulated by miR-130b inhibition in MDA-PCa-2b cells (Fig. 4B). Upregulated genes were compared with predicted and validated miRNA targets by miRwalk, miRDB, TargetScan, and miRTarBase. Approximately 62% of upregulated genes were predicted by at least one database (Fig. 4C; 26 of 42). The genes that we validated by qRT-PCR analysis such as IGF2R, CDKN2A, and CDKN1A were overlapped in miRWalk and TargetScan (Fig. 4C). The other two, IGF2R and SH3PXD2A, were validated by miRTarBase (Fig. 4C). According to the over-representative analysis with CPDB, 22 pathways were changed with miR-130b inhibition in MDA-PCa-2b cells compared with that of DU-145 and represent the top 10 pathways summarized in Fig. 4D and Supplementary Table S5. In agreement with our functional assays, miR-130b significantly affected central cancer–related gene pathways, “Pathway in Cancer” (P = 0.0002), “p53 signaling pathway” (P = 0.002), and “miRNA in Cancer” (P = 0.002; Supplementary Table S5). Focusing on single genes, we found that 16 genes were upregulated more than four times and five genes were decreased by half in MDA-PCa-2b cells. However, these same genes were less affected in DU-145 compared with MDA-PCa-2b cells (Supplementary Table S6).

On the basis of the tumor-suppressive function of miR-130b inhibition, we focused our search for tumor suppressor genes. Fragile Histidine Triad (FHIT) gene was the fourth ranked significantly altered gene in our qPCR array results (Fig. 4A). We confirmed the array results by qRT-PCR (Fig. 5A). This gene is a tumor suppressor and is located at a fragile region in the human genome (24–26). We observed significant upregulation of FHIT protein expression that was specific to anti-miR-130b–stable MDA-PCa-2b cells as compared with DU-145 or LNCaP stable prostate cancer cells (Fig. 5B). An in silico algorithm (IntaRNA) showed two predicted miR-130b–binding sites in the FHIT 3′UTR. We inserted both binding sites (Fig. 5C) into pmiR-Glo (FHIT 3′UTR#1; FHIT 3′UTR#2) and transfected them into anti-miR-130b stable MDA-PCa-2b cells. The luciferase reporter activity confirmed the FHIT gene to be a direct target of miR-130b in MDA-PCa-2b cells (Fig. 5D). Among the two binding sites, FHIT 3′UTR#2 was responsible for miR-130b binding as we observed that the luciferase activity of this binding site was high in anti-miR-130b–stable MDA-PCa-2b cells. However, the FHIT 3′UTR#1 site did not show any changes (Fig. 5D). Rescuing miR-130b expression using miR-130b mimic in anti-miR-130b stable MDA-PCa-2b cells resulted in FHIT downregulation (Fig. 5E). miR-130b expression levels after miR-130b mimic transfection are shown in Supplementary Fig. S3A. We also randomly selected other miR-130b target genes from the validated or predicted miR-130b target gene list for qRT-PCR analysis and all these target genes were upregulated in the anti-miR-130b–stable MDA-PCa-2b cells compared with controls (Fig. 5F). TP53INP2 (13) and PTEN (27, 28) reported miR-130b targets, were also upregulated in our anti-miR-130b–stable cells. FOXO1, KDM2A, and BTG1 that are predicted miR-130b targets by in silico algorithm miRWalk were also confirmed to be upregulated in our anti-miR-130b–stable clones (Fig. 5F). PCR array data showed another cell-cycle regulator, CDKN1A as differentially expressed in anti-miR-130b–stable MDA-PCa-2b cells compared with DU-145 cells. However, the results of protein expression by Western blots were inconsistent (Supplementary Fig. S3B). Hence, we did not purse CDKN1A further. Similarly, RUNX3 and ESR were upregulated in anti-miR-130b–stable MDA cells in array data and these genes are predicted as miR-130b target in four prediction databases. Nevertheless, qPCR could not detect ESR expression in both cells. RUNX3 mRNA levels were increased in both DU-145 and MDA-PCa-2b anti-miR-130b–stable cells (Supplementary Fig. S3C) and hence appeared not to be race related. RUNX3 has been reported as a miR-130b target in gastric and ovarian cancers (29, 30). These data reveal that among all the PCR array data–predicted targets, only FHIT is a direct functional target of miR-130b relevant to AA race–related prostate cancer.

We sought to determine the correlation of FHIT with AA and EA PC patient survival. We randomly selected a small sample number for IHC staining (AA = 14, EA = 10) from the same SFVAMC cohort that we used for expression analysis in Fig. 1A and B. FHIT staining was observed mostly in the cytoplasm and inside nuclei of glandular epithelial cells. BPH samples from AA and EA tissues showed positive FHIT expression levels (2–3; Fig. 5I, bottom). We classified FHIT expression levels into groups, 0 = negative, 1, 2–3 = positive by Allred scoring, and performed Fisher exact test to determine correlation of FHIT expression with patient death (Fig. 5G). We observed a significant correlation between negative FHIT expression and AA patient death (Fig. 5G) as there were more deceased patients with negative FHIT expression compared with AA patients still alive (Fig. 5G). Furthermore, we correlated FHIT protein expression in the same cohort in both AA and EA groups. IHC results indicated that miR-130b expression levels show an inverse trend with that of FHIT expression (Fig. 5H). In EA, miR-130b expression was similar in negative and positive FHIT–stained samples (Fig. 5I). However, this needs to be confirmed in a large tissue cohort. Nevertheless, these data show that loss of FHIT is a risk factor in AA patients.

In this study, for the first time, we identified miR-130b to be a significant contributor to PC racial disparity in AA men. miR-130b expression levels were higher in AA prostate cancer tissues in both TCGA and SFVAMC cohorts as well as in the AA-derived cell line, MDA-PCa-2b when compared with EA prostate cancer tissues and cells. miR-130b (chr22: 21653304-21653385) forms a cluster with miR-301b (chr22: 21652981–21653058; <10 kb apart) and these two miRNAs harbor the same seed sequence. We compared the expression of miR-301b between AA and EA tissues in TCGA. There was no difference in expression of this miRNA between the two races. (Supplementary Fig. S4A). This indicates that miR-130b may be the dominant miRNA than miR-301b that is related to racial disparity in AA patients. According to the UCSC genome browser, chromatin modification at the suspected miR-130b promoter region is more accessible to binding transcription factors in prostate cancer cells compared with normal prostate epithelial cells (PrEC; ref. 15). Another study, not related to race, has reported that c-Met upregulated miR-130b expression and contributed to cancer progression and castration resistance through downregulation of AR expression (14). However, we observed that the MET mRNA level is slightly lower in AA compared with EA prostate cancer patients according to the TCGA cohort (Supplementary Fig. S4B). Hence it might not be the reason for miR-130b overexpression in AA patients. Meanwhile, the cell viability results in this report are consistent with ours. Taken together, miR-130b over-expression may be the result of open chromatin status in human prostate cancer leading to abnormal overexpression of miR-130b rather than MET-driven regulation. A previously published study also showed that miR-130b tended to be upregulated in a small sample of AA patients with prostate cancer (22). However, our study is the first study to conclusively demonstrate the involvement of miR-130b in the AA prostate cancer health disparity. Clinically, we observed that elevated miR-130b expression positively correlated with poor overall survival of AA patients. Multivariate analysis also demonstrated that miR-130b independently serves as a prognostic marker for overall survival with increased significance (P = 0.008) compared with PSA levels (P = 0.05) for AA patients. Currently, clinicians are seeking noninvasive prognostic and diagnostic biomarkers for early prostate cancer detection. In nephroblastoma and hepatocellular cancer, high expression of miR-130b has been detected in patient serum and showed high diagnostic potential (31, 32). The expression level of serum miR-130b was related to chemosensitivity and chemo-resistance in patients with colorectal cancer (33). In the future, we plan to analyze the potential for miR-130b to be used as a noninvasive biomarker in prostate cancer racial disparity.

Another analysis showed that inhibition of miR-130b caused a drastic change in 42 genes and differential gene expression patterns related to human cancer pathways in MDA-PCa-2b cells compared with DU-145 cells. Given that miR-130b caused different responses in a race-specific manner, there may be race-specific mutations within the miRNA-binding site or differences in the basal expression level of miRNA target genes between the two races. As the qPCR array and its validation showed, this study supports the latter hypothesis, taking into account that most of the genes with altered expression in the array were in anti-miR-130b stable MDA-PCa-2b cells, whereas the majority of genes were not affected in anti-miR-130b–stable DU-145 cells. Basal gene expression differences between DU-145 and MDA-PCa-2b control cells might generate distinct biological pathway activation by miR-130b inhibition in the two different cells.

We found that FHIT and CDKN1B, known as haplo-insufficient tumor suppressors (24, 34, 35), were significantly upregulated in MDA-PCa-2b cells upon miR-130b inhibition and luciferase reporter assay showed that both genes are a direct miR-130b target. A previous study found that stimulation of CDKN1B caused G₂ arrest by repressing cyclin D3 (36). In non–small cell lung cancer (NSCLC), SIRT1 silencing suppresses proliferation and induces senescence in a CDKN1B (p27Kip1)-dependent manner (37). In regard to the clinical relevance of CDKN1B, less CDKN1B protein expression was an independent predictor of adverse outcomes for patients with local prostate cancer (38). CDKN1B expression levels also inversely correlated with BCR-free survival in those patients with prostate cancer who underwent radical prostatectomy with pathologic organ–confined disease (39). Furthermore, SKP2 gene negatively regulates CDKN1B. However, in AA patients, 46.3% showed lower expression of CDKN1B independent of SKP2 (40). On the basis of this report, we infer that CDKN1B might be subdominantly regulated by miR-130b in AA patients. Our analysis selected five cell-cycle–related genes that were correlated with prostate cancer disparity in AA patients. Among them, CDKN1B, a key cell-cycle gene, was inversely correlated with miR-130b expression in TCGA AA patients while EA samples did not show any correlation. Consistent with this result, miR-130b inhibition in AA prostate cancer cells caused upregulation of CDKN1B expression. These results suggest that the cell-cycle pathway may be important prostate cancer racial disparity in AA patients with prostate cancer. Further, FHIT is a tumor-suppressor gene which resides at a fragile chromosome site, FRA3B, and its genetic alterations have been observed in many types of human cancers (41). In human carcinoma, aberrant FHIT transcripts have been frequently observed in lung, esophageal, and gastrointestinal carcinomas (42, 43). Deficiency of FHIT contributed to tumorigenesis by N-nitrosomethylbenzylamine in murine gastric cancer (24, 26). In prostate cancer, downregulation and hypermethylation of FHIT (44, 45) along with its therapeutic intervention potential in human prostate cancer has been reported (44). We validated FHIT regulation via miR-130b by inhibition or rescuing of miR-130b expression. Our study suggested that miR-130b–mediated downregulation of FHIT induces malignant tumors in AA compared with EA patients. It was reported that FHIT upregulated miR-30c and negatively controlled metastasis by modulating EMT genes in lung cancer (46). Similarly, FHIT downregulation caused by miR-130b may mediate prostate cancer metastasis in AA patients with prostate cancer. This is supported by the fact that suppression of miR-130b specifically inhibited tumorigenic characteristics such as cell migration and invasion in AA cells compared with EA cells. Thus, this is the first report indicating that FHIT may be involved in prostate cancer racial disparity.

This study documents the clinical significance of miR-130b in AA prostate cancer. Overexpression of miR-130b was observed to contribute to poor prognosis in AA patients. Understanding the relationships between miRNA expression and downstream regulation of mRNA targets in the AA population is important to comprehend the biological basis for racial health disparity in prostate cancer. Inhibition of miR-130b and rescue of FHIT and CDKN1B, may provide a novel therapeutic approach for the treatment of prostate cancer in AA patients. Furthermore, miR-130b might be a potential prognostic biomarker to predict mortality risk in AA patients with prostate cancer.

No potential conflicts of interest were disclosed.

Conception and design: Y. Hashimoto, M. Shiina, T. Kato, S. Majid, R. Dahiya

Development of methodology: Y. Hashimoto, M. Shiina, S. Maekawa, G. Deng, S. Majid, R. Dahiya

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Hashimoto, M. Shiina, P. Dasgupta, P. Kulkarni, R.K. Wong, Y. Tanaka, V. Shahryari, Z.L. Tabatabai, S. Majid, R. Dahiya

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Hashimoto, M. Shiina, Y. Tanaka, S. Saini, S. Majid, R. Dahiya

Writing, review, and/or revision of the manuscript: Y. Hashimoto, M. Shiina, T. Kato, S. Majid, R. Dahiya

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Shiina, Y. Tanaka, S. Yamamura, G. Deng, S. Majid, R. Dahiya

Study supervision: Y. Tanaka, S. Majid, R. Dahiya

We thank Dr. Roger Erickson for his support and assistance with the preparation of the manuscript.

This work was supported by the NCI at the NIH through grant numbers U01CA184966. Dr. Dahiya is the recipient of a Senior Research Career Scientist Award (# 1K6BX004473) from the Department of Veterans Affairs.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.