Micro-RNA-186-5p inhibition attenuates proliferation, anchorage independent growth and invasion in metastatic prostate cancer cells

Dysregulation of microRNA (miRNA) expression is associated with hallmarks of aggressive tumor phenotypes, e.g., enhanced cell growth, proliferation, invasion, and anchorage independent growth in prostate cancer (PCa). Serum-based miRNA profiling involved 15 men diagnosed with non-metastatic (stage I, III) and metastatic (stage IV) PCa and five age-matched disease-free men using miRNA arrays with select targets confirmed by quantitative real-time PCR (qRT-PCR). The effect of miR-186-5p inhibition or ectopic expression on cellular behavior of PCa cells (i.e., PC-3, MDA-PCa-2b, and LNCaP) involved the use bromodeoxyuridine (BrdU) incorporation, invasion, and colony formation assays. Assessment of the impact of miR-186-5p inhibition or overexpression on selected targets entailed microarray analysis, qRT-PCR, and/or western blots. Statistical evaluation used the modified t-test and ANOVA analysis. MiR-186-5p was upregulated in serum from PCa patients and metastatic PCa cell lines (i.e., PC-3, MDA-PCa-2b, LNCaP) compared to serum from disease-free individuals or a normal prostate epithelial cell line (RWPE1), respectively. Inhibition of miR-186-5p reduced cell proliferation, invasion, and anchorage-independent growth of PC-3 and/or MDA-PCa-2b PCa cells. AKAP12, a tumor suppressor target of miR-186-5p, was upregulated in PC-3 and MDA-PCa-2b cells transfected with a miR-186-5p inhibitor. Conversely, ectopic miR-186-5p expression in HEK 293 T cells decreased AKAP12 expression by 30%. Both pAKT and β-catenin levels were down-regulated in miR-186-5p inhibited PCa cells. Our findings suggest miR-186-5p plays an oncogenic role in PCa. Inhibition of miR-186-5p reduced PCa cell proliferation and invasion as well as increased AKAP12 expression. Future studies should explore whether miR-186-5p may serve as a candidate prognostic indicator and a therapeutic target for the treatment of aggressive prostate cancer.


Background
Prostate cancer (PCa) is the leading cause of nonmelanoma cancer-related mortality in men in the U.S. [1]. Ninety-one percent of PCa cases are treatable among men diagnosed with localized or regional disease as evidenced by a 100% 5-year survival rate [1]. However, the 5-year survival rate drops to 28% for those with metastatic PCa [2]. Although early detection of PCa has improved, better prognostic biomarkers are urgently needed to refine current detection, prognosis, and clinical management strategies for metastatic PCa [3].
MicroRNAs (miRs or miRNAs) are 17-25 nucleotide short non-coding RNAs that may serve as ideal biomarkers of metastatic PCa for several reasons. First, miRs are stably expressed in tumor tissue and related biological fluids, including serum and plasma [4]. Second, miRs regulate the expression of genes involved in tumor spread including cell proliferation, invasion, migration, angiogenesis, and anchorage-independent growth (reviewed in [5][6][7]). Third, dysregulation of miRNA corresponds with aggressive PCa phenotypes including high tumor stage, high Gleason grade, disease recurrence, and biochemical recurrence [8,9]. Lastly, serum miRNA profiles may distinguish between aggressive and non-aggressive PCa [10,11].
The purpose of the current study was to identify differentially expressed miRNAs in serum from PCa patients versus normal controls. In addition, this study sought to characterize the role of miR-186-5p in metastatic PCa cell models. Our findings will aid in the understanding of miR-186-5p's role in metastatic PCa using pre-clinical, human PCa cell lines. Our novel identification of increased miR-186-5p in the serum of metastatic PCa patients and its pro-migration/invasion oncogenic activity in PCa cell lines suggests miR-186-5p may serve as a diagnostic, prognostic and ultimately a therapeutic tool for the effective clinical management of metastatic PCa.

Isolation of miRNAs from serum
Serum (250 μl) obtained from 15 patients and five disease-free individuals was transferred to 1.5 ml nuclease-free tubes. Trizol LS Reagent (1 ml) was added to each sample and shaken for 30 s (secs) at room temperature. Each sample was spiked with cel-miR-39 (2 μl, 1 nM, internal miRNA control) and incubated for 5 min (mins). ACS 98% grade chloroform (200 μl per 1 ml of Trizol) was added to each sample, shaken for 15 s, and incubated again for 5 mins. Next, total RNA was isolated from serum using the miRVana microRNA Isolation kit (Thermo Fisher Scientific, Waltham, MA). miR profiling in serum using Taqman human MicroRNA arrays Expression analysis of 377 miRNAs involved the use of Taqman Array Human MicroRNA Pool A Cards v.2 (Thermo Fisher Scientific) that included three endogenous controls (RNU6, RNU44, RNU48) and a non-human related negative control (ath-miR-159a). Micro-RNA profiling was assessed using the Applied Biosystems 7900 Real Time PCR system (Thermo Fisher Scientific). RNA was reverse transcribed into cDNA in a 7.5 μl reaction using a TaqMan miRNA Reverse Transcription (RT) Kit and MegaPlex RT Primers (10X) (Thermo Fisher Scientific). Diluted pre-amplified RT products in TE (75 μl, 0.1X, pH 8.0) were added to a reaction mix [100 μl of Taqman Universal PCR Master Mix (2X)] and dispensed into arrays. miRNA profiles in serum were normalized to the global median comparative threshold (Ct) value for each array (global Ct median value -target Ct value) using R-programming software. Fold change was calculated with respect to each tumor stage relative to disease-free individuals. After normalization, targets with ≥50% missing Ct values were imputed using k nearest neighbor (kNN) imputation. Differentially expressed serum-based miRNAs in PCa patients were selected for further validation using the following selection criteria: FDR p-value ≤0.05 and fold change ≥1.5.

DNA isolation from cells
RWPE1 cells were grown to 80% confluency in growth media. Genomic DNA was isolated from cells using DNeasy Blood and Tissue kit (Cat# 69504, Qiagen, Valencia, CA). DNA concentrations (260/280 nm) were measured using a Nano Dropper Spectrophotometer.

Human gene expression array
RNA was extracted from transient and stable transfected PC-3 cells with miR-186-5p inhibitor, RWPE1 cells with miR-186-5p mimic, and corresponding scramble (transient) or empty vector controls (stable) in 3 independent experiments. RNA sample purity and integrity were assessed using the Agilent 2100 Bioanalyzer. RNA (250 ng) was serial diluted and transcribed into cDNA and cRNA. Fragmented and biotin-labeled cRNA (12 μg) was subjected to a series of incubation periods and hybridized to Prime View gene microarrays with appropriate poly-A and hybridization controls using the 3'IVT Plus Reagent kit (Cat# 902416, Affymetrix Inc., Santa Clara, CA), according to the manufacturer's instructions. Each array was washed and stained according to array type. The fluidics protocol FS450_0002 was used to analyze each array via Gene chip scanner (Affy.Command console Version 3.3).

Gene selection
Aberrant gene expression associated with miR-186-5p modification was identified via microarray analysis. Genes up-regulated in stable miR-186-5p inhibited PC-3 cells and down-regulated in stable miR-186-5p overexpressing RWPE1 cells were identified as potential miR-186-5p targets (± 1.2 fold change in expression and false discovery rate p-value < 0.05). Next, potential targets were evaluated based on published reports and in silico databases, including MetaCore, Ingenuity, www.TargetScan.org and the www.microrna.org. Final selection of miR-186-5p targets was based on published reports and www.microrna.org database. The aforementioned miR database used PhastCons (positive value ≥0.57) and mirSVR (negative score ≤ − 0.1) scoring methods to determine highly conserved miRNAs [12,34].

BrdU proliferation assay
Cell proliferation in cells was measured using the Cell Proliferation ELISA 5-bromo-2′-deoxyuridine (BrdU) colorimetric kit (#11647229001, Sigma Aldrich, St. Louis, MO). Transfected cells (5 × 10 3 /well) were seeded into a 96-well plate format, incubated at 37°C for 24 h and labeled with BrdU reagent for 24 h. Absorbance readings were taken at 370 nm and 492 nm (reference) using a Biotek Synergy HT plate reader and Gen5 version 1.08 software (BioTrek, Winooski, VT). Experiments included experimental groups with six replicates that were repeated at least three times.

Anchorage-independent growth assay
The influence of ectopic expression and inhibition of miR-186-5p on 2-dimensional colony formation was assessed using an anchorage independent growth assay. In 6-well plates, 0.7% agar-growth media solution (3 ml), prepared with sterile 3.5% agar and 1X phosphate buffered saline (PBS), was added to each well to form a base layer. Transfected cells (10 × 10 3 ) in growth media (3 ml) were gently mixed with 0.7% agar-media solution (3 ml) seeded on top of base layers. Cells in soft agar were incubated at 37°C for 2-3 weeks. Colonies were quantitated at 4X magnification. Experiments were repeated at least three times.

Matrigel invasion assay
The effect of miR-186-5p inhibition on cellular invasion was evaluated by the Boyden chamber assay, as described elsewhere (Albini,A. et al. 1987). Briefly, polyethylene transwell inserts with 8 μm pore size were coated with a final concentration of 2 mg/ml of reduced growth matrigel. Cells (25 × 10 3 ) were suspended in serum-free media containing reduced growth Matrigel and seeded on top of matrigel. Growth media with FBS (600 μl) was added to the lower chamber of each well. After 24 h of incubation (37°C, 5% CO2), non-invading cells on the upper side of the membrane were removed with 1X PBS. Invading cells were fixed in 100% methanol and stained with 0.2% crystal violet. The number of invading cells was counted under a microscope (EVOS) quantified using a 10X magnification. Assays were repeated at least three times.

Statistical analysis
Differences in demographic/clinical data [age, prostate specific antigen (PSA) levels and BMI values] comparing PCa patients and controls were assessed using the Wilcoxon Rank-Sum test. Differential miRNA expression for each tumor stage was adjusted for multiple hypothesis testing (i.e., FDR) relative to non-cancerous controls using ANOVA and modified t-test with the R package limma [35,36]. Differential gene expression was identified in PC-3 and RWPE1 cells using the Partek Genomics Suite 6.6 software (St. Louis, MO), after adjusting for multiple hypothesis testing using the false discovery test (FDR). MicroRNA/mRNA expression and biological assays were evaluated using two-sided unpaired t-tests. (GraphPad 6 Software, Inc., La Jolla, CA). All statistical significance was established using an alpha cut-off value of 0.05 or FDR ≤ 0.05. All statistical analysis was performed using GraphPad 6 Software, Inc., (La Jolla, CA).

Population description
Serum was collected from 15 PCa patients diagnosed with tumor stage I, III, IV and five disease-free patients who selfidentified as men with European ancestry (Additional file 1: Table S1). There was no significant difference in the median age or BMI levels between cases and controls, respectively. Median PSA levels among cases were significantly higher than non-cancerous controls (p = 0.048). Notably PSA levels in controls were higher than the normal range.
According to the clinical data provided for the control biospecimens, each patient was classified as having a normal prostate at the time of serum collection. Although the "control" subjects had high PSA levels, their prostate was designated as disease-free or "normal" based on negative biopsy results. Unfortunately, there was no data on whether the controls had BPH and there was no available clinical follow-up data for the patients designated as controls. Tumor classification for 60 % (n = 9) of the cases were diagnosed with adenocarcinoma and 67% had a smoking history. However, the smoking history status among the controls was not available for this study population. The majority of the PCa patients received at least two types of therapy (73.3%), including hormonal therapy (n = 14, 93. 3%), radiation (n = 8, 53.3%), surgery (n = 7, 46.7%), and chemotherapy (n = 2, 13.3%).

Validation of miR-186-5p in PCa patient serum
MiR-186-5p expression was validated in two independent experiments within the same cohort of serum samples. MiR-186-5p was significantly upregulated in PCa serum relative to disease-free individuals (Fig. 1a). To examine its potential role in PCa, miR-186-5p was selected for further validation and characterization using PCa cell lines and normal prostate epithelial cells.

Discussion
Altered miRNA profiles contribute to dysregulation of gene expression involved in the pathogenesis of metastatic PCa [40]. In the current study, we identified upregulation of miR-186-5p in the serum from PCa patients (tumor stage I, III and IV) and metastatic PCa cell lines (LNCaP, MDA-PCa-2b, PC-3) relative to their respective controls. We also demonstrated inhibition of miR-186-5p reduced PCa cell proliferation, anchorageindependent cell growth, colony formation, and invasion of metastatic PC-3 and/or MDA-PCa-2b cells. Collectively, these findings suggest miR-186-5p plays an oncogenic role in PCa. Indeed, ectopic miR-186-5p expression enhanced anchorage independent growth of LNCaP cells but not PC-3 or MDA-PCa-2b cells.
There are limited reports on the role of miR-186 in PCa [28][29][30][31][32]41]. Commensurate with our serum and in vitro findings, Ambs and colleagues (2008) observed higher miR-186-5p expression in laser microdissected tumor tissue from PCa patients diagnosed with extra-prostatic disease relative to patients with no extra-prostatic disease among European (n = 30) and African American (n = 30) men [28]. In contrast, other studies reported miR-186 was down-regulated in non-microdissected PCa tissue [30,31] and PCa cell lines, i.e., M12, P69, PC-3, Tsu-Pr1, LNCaP, 22Rv1, DU145 [30][31][32]. Collectively, these studies suggest a tumor suppressor role for miR-186 in PCa. However, these authors do not distinguish whether the miR-186 precursor, miR-186-5p or miR-186-3p were responsible for apparent cell effects of miR-186 overexpression in PCa cell lines or tissue. Our report suggests miR-186-5p may have an oncogenic role due to its up-regulation in the serum of PCa patients and tumor cell lines. The discrepancies between our findings and other miR-186 reports may also be attributed to the following: (1) tumor tissue processing or storage methods; (2) selection of "normal prostate" tissue using microdissected versus non-microdissected tissue; (3) types of "control" cell lines used for comparison purposes; (4) degree of specificity of primers used for quantitation of miR-186-3p or miR-186-5p; (5) selection of normalizer for miRNA quantification; and (6) methods of miRNA isolation and detection. For the detection and semiquantitation of miR-186-5p in the current study, we used normal epithelial cell lines (e.g., RWPE1) for comparison purposes, the miRVana miRNA isolation kit, primers specific for miR-186-5p, and U44 for normalization. In the current study, the U44 levels did not vary among normal epithelial and the prostate cancer cell lines (data not shown). The studies that indicated an tumor suppressor role for miR-186 used: (1) primary culture prostate epithelial cells for comparison purposes with no further details about these cell models; (2) snoRNAs U24, or U6 as controls to normalize miRNA levels; (3) varying kits for miRNA isolation (i.e., Trizol, Qiagen miRNeasy mini kit, E. Z.N.A.® miRNA kit); and (4) non-specific rather than specific miRNA primers [30][31][32]. To our knowledge, the current study is the first to demonstrate the up-regulation of miR-186-5p in serum from PCa patients and in metastatic PCa cell lines. For the serum based studies, we used  the miRVana isolation kit and an exogenous miRNA control, namely cel-miR-39 (i.e., C. Elegans 39a nonhuman miRNA). We identified miR-186-5p targets using in silico tools followed by functional assays. Among the established miR-186-5p targets, we selected AKAP12 for further evaluation due to its central role in cell proliferation, colony formation, cell invasion and epithelial mesenchymal transition and tumor growth [39,[52][53][54][55][56]. A previous study demonstrated AKAP12 is a bona fide direct target of miR-186-5p using 3'-UTR-luciferase reporter assay [57]. Down-regulation of AKAP12 suppresses cell proliferation, survival, motility, migration, anchorage independent growth, angiogenesis and invasion in several cancers (reviewed in [52]).
Notably, several cancer phenotypes (i.e., cell proliferation, colony formation, invasion) were attenuated upon miR-186-5p inhibition in metastatic PCa cell lines in the current study. This reduction in aggressive PCa behavior may partially correspond with upregulation of AKAP12. Ectopic expression of AKAP12 is associated with a decrease in cell invasion and anchorage independent growth of mouse PCa cells [54,58]. Furthermore, knockout of AKAP12 in mice resulted in prostatic hyperplasia and dysplastic foci [39], supporting a role of AKAP12 as a tumor suppressor gene.
AKAP12 knockout mice also showed increased expression of pAKT in prostate tissue [39]. pAKT signaling inhibits Glycogen Synthase Kinase-3 (GSK3) activity, which leads to β-catenin nuclear accumulation [38]. Loss of tumor suppressor PTEN, which is common in PCa, leads to activation of PI3K and AKT signaling and subsequent β-catenin phosphorylation [59]. Phosphorylation of β-catenin increases its stability and nuclear localization, which leads to its association with T-cell specific transcription factor/lymphoid enhancer-binding factor 1 family proteins (TCF/LEF-1) in the nucleus. The TCF/LEF-1-βcatenin complex recruits coactivators such as B-cell lymphoma 9 protein (Bcl-9) and cAMP response element binding protein (CREB)-binding protein (CBP) [60], resulting in an increase in the transcription of genes involved in tumorigenesis, angiogenesis, extracellular matrix, and cell cycle progression, including MYC, MMP7, VEGF, and CCND1 [61,62].
In addition, β-catenin is an integral signaling protein for epithelial mesenchymal transition (EMT) and the transcription of genes in the canonical Wnt signaling pathway [38]. Here, we observed inhibition of miR-186-5p resulted in a decrease in β-catenin protein in PC-3 and MDA-PCa-2b cells. We speculate the repression of the aggressive PCa phenotype observed in miR-186 inhibited PCa cells is partially attributed to AKAP12 mediated down-regulation of pAKT, which in turn, may downregulate β-catenin (modeled in Fig. 6). Several reports demonstrate pAKT phosphorylates and stabilizes βcatenin and phosphorylated β-catenin activates AKT expression [31,63]. Consequently, we propose an increase in miR-186-5p expression stimulates AKT signaling via decreasing AKAP12, leading to increased nuclear β-catenin, an EMT mediator. This proposed mechanism may explain the influence of miR-186-5p on cell invasion and anchorage-independent growth in metastatic PCa cell models in the current study. Future studies will validate whether miR-186-5p inhibition down-regulates AKT signaling, allowing GSK3 to promote β-catenin degradation.
We evaluated the strengths, limitations and future directions of the current study. In a pilot study, our lab was the first to demonstrate the up-regulation of miR-18-5p in the serum of patients diagnosed with non-metastic and metastatic prostate cancer when compared to those without prostate cancer. Although the controls on average had elevated PSA, they were designated as prostate cancer-free at the time of the serum collection, following a biopsy. We cannot exclude the possibility that the controls had benign prostatic hyperplasia. If we assume the controls in the (See figure on previous page.) Fig. 5 Tumor suppressor AKAP12 is a target of miR-186-5p in PCa cells. AKAP12 transcript and protein expression were measured using qRT-PCR and western blot analysis, respectively. a Three potential miR-186-5p binding sites in AKAP12 were identified; site 1) mirSVR score: − 0.5641 PhastCons score: 0.6188; site 2) mirSVR score: − 0.7981 PhastCons score: 0.5553; and site 3) mirSVR score: − 0.2640 PhastCons score: 0.6218 (values from www.microrna.org). b AKAP12 transcript expression was significantly lower in metastatic LNCaP (p = 0.0011) and MDA-PCa-2b (p < 0.0001) cells and higher in PC-3 cells relative to normal prostate RWPE1 cells (p = 0.0011). c Stable anti-miR-186-5p expression in PC-3 cells increased AKAP12 transcript expression (p = 0.0013). AKAP12 expression was normalized to GAPDH. d HEK 293 T cells were transfected with scrambled miR control or miR-186-5p mimic. A representative western blot shows ectopic expression of miR-186-5p in HEK 293 T cells decreased AKAP12 protein by 30% 72 h post-transfection (p = 0.0063). All analyses involved at least three independent experiments. e and f Representative western blots of AKAP12, β-catenin and pAKT expression in HEK 293 T (72 h), PC-3 (48 h), and MDA-PCa-2b (24 h) cells post-transfection. Data analyses included three independent experiments for e and two experiments for f. g and h Quantitation of western blots presented as black bar (scramble control) and grey bar (miR-186-5p inhibitor). AKAP12 protein was increased in PC-3 (p = 0.0049) and MDA-PCa-2b (p = 0.0318) cells transiently transfected with miR-186-5p inhibitor. In contrast, β-catenin/ β-actin protein was decreased in PC-3 (p = 0.0434) and MDA-PCa-2b (p = 0.0048) cells transfected with miR-186-5p inhibitor. Data analysis was based on mean ± S.D. of target protein relative to β-actin in the same blot. (*p-value < 0.05, **p-value < 0.007, **** p-value < 0.002) current study had BPH, we would anticipate an even higher fold increase serum-based miR-186-5p levels from prostate cancer patients relative to levels among non-BPH controls. Future studies will confirm whether miRNA-186-5p detected in serum and matched micro-dissected prostate tumor specimen correspond with high tumor grade/stage, metastatic disease, higher risk of biochemical/disease recurrence, or hormone refractory status within large and ethnically diverse patient sub-groups.
Although miR-186-5p was also up-regulated in nonmetastatic and metastatic prostate cancer cell lines, miR-186-5p levels did not vary by androgen receptor status. Interestingly, over-expression of miR-186-5p in PC-3 cell lines did not result in an increase in aggressive prostate cancer cellular behavior. Since the PC-3 cells have the highest expression of miR-186-5p, we believe overexpression of miR-186-5p in PC-3 cells had no further biological effect in these cells. The high baseline levels of miR-186-5p in the PC3 cells may have saturated any biological effects (i.e., cell proliferation, colony formation and cell invasion) and prevented our capacity to detect further aggressive phenotypes in these already transformed cells. However, additional pre-clinical studies are needed to assess whether miR-186-5p overexpression and inhibition alter aggressive tumor behavior in other metastatic PCa cell models (e.g., TSU-Pr1, MDA PCa-2a, VCaP, DuCaP) as well as tumorigenesis in animal models in the presence and absence of androgen stimulation.
To our knowledge, this is the first study to demonstrate AKAP12 as a miR-186-5p target in metastatic PCa cell lines (PC-3, MDA-PCa-2b). Importantly, inhibition of miR-186-5p suppressed three metastatic PCa cell hallmarks, namely proliferation, invasion, and colony growth. We speculate the reduction in cell proliferation, invasion, and anchorage independent growth with anti-miR-186-5p may be attributed, at least in part, to AKAP12's role in pAkt mediated suppression of β-catenin. Following inhibition/overexpression of miR-186-5p, we demonstrated an up-regulation of tumor suppressor AKAP12 as well as a downregulation pAkT and β-catenin in vitro in total cell lysates of PC3 cells. Since increased β-catenin is apparent in many cancers, including prostate cancer [4,5,7], future studies will focus on the evaluation of β-catenin levels in both the cytoplasm and nucleus, following inhibition/ overexpression of miR-186-5p using in vitro cell models. Apparently, the translocation of β-catenin from the cytoplasm to the nucleus is where it mediates the transcription of target genes related to EMT and metastasis. Since Bcatenin mediates its effects on EMT markers.
In our microarray analysis, we observed previously published miR-186 validated targets (AKAP12, ROCK1, PPM1B, and PTTG1) were downregulated as anticipated in RWPE1 stably transfected with miR-186-5p and/or up-regulated in PC-3 cells stably transfected with anti-miR-186. Although hundreds of potential miR-186-5p targets were identified in the miR-186-5p depleted PC-3 cells, we focused on AKAP12, a validated target involved in prostate cancer. We also focused on AKAP12 because it plays a role in many of cell behaviors (e.g., cell proliferation, colony formation, cell invasion, cell motility, EMT, metastasis, cell invasion, cell cycle arrest, and cell death) that were also modified in our miR-186-5p suppressed metastatic cell models. AKAP12 appears to mediate its effects on cell invasion, presumably through the p-Akt/β-catenin the pathway. In future studies, our lab will validate both common and unique targets revealed in the top 30 targets in anti-miR-186-5p transfected PC-3 and miR-186-5p overexpressing prostate cancer cell models cells.

Conclusions
This is the first report that miR-186-5p is upregulated in PCa patient serum and cell lines. We demonstrated inhibition of miR-186-5p inhibited anchorage-independent cell growth and invasion and reduced pAKT and β-catenin levels, while increasing tumor suppressor AKAP12. Overall, our data suggest miR-186-5p may function in an oncogenic capacity and serve as a potential prognostic tool and therapeutic target in PCa. Lastly, future studies will elucidate how the miR-186-5p-AKAP12 axis and other miR-186-5p targets play a role in prostate cancer using in vitro and in vivo models.

Additional files
Additional file 1 Table S1. De-identified demographic and clinicopathological data. Clinical data and serum from 15 men diagnosed with prostate cancer and five disease-free individuals were obtained from the BioServe Biotechnologies Biorepository (Beltsville, MD). Subjects were selfidentified as European American males. There were no significant differences in median age between cases and controls (p = 0.726). Among men diagnosed with prostate cancer, 60% were diagnosed with adenocarcinoma, 66.7% were smokers, and 73.3% received two or more therapies. Relative to controls, cases had higher median serum PSA level (ng/ml) (p = 0.048) and BMI (p = 0.225) values. (DOCX 63 kb) Additional file 2 Table S2. Differentially expressed human miRNAs in serum from PCa patients. Relative to disease-free individuals, human miRNAs were down-regulated (fold change ≤ − 1.5) and up-regulated (fold change ≥1.5) in PCa patients when compared to disease-free individuals based on Taqman Human MicroRNA Array data (p-value ≤0.05). Global normalization of miRNA profiles identified 26 differentially expressed human miRNAs (9 downregulated and 17 up-regulated) in the serum from patients diagnosed with tumor stage I (n = 5), III (n = 5) and IV (n = 5) relative to disease-free individuals (n = 5). miRNAs (miRs-106b-5p, − 186-5p, −302b-3p, − 342-3p, −520e, − 885-5p), highlighted in gray, represent targets that survived multiple hypothesis testing (FDR p-value ≤0.05).