Serum levels of miR-194 have been reported to predict prostate cancer recurrence after surgery, but its functional contributions to this disease have not been studied. Herein, it is demonstrated that miR-194 is a driver of prostate cancer metastasis. Prostate tissue levels of miR-194 were associated with disease aggressiveness and poor outcome. Ectopic delivery of miR-194 stimulated migration, invasion, and epithelial–mesenchymal transition in human prostate cancer cell lines, and stable overexpression of miR-194 enhanced metastasis of intravenous and intraprostatic tumor xenografts. Conversely, inhibition of miR-194 activity suppressed the invasive capacity of prostate cancer cell lines in vitro and in vivo. Mechanistic investigations identified the ubiquitin ligase suppressor of cytokine signaling 2 (SOCS2) as a direct, biologically relevant target of miR-194 in prostate cancer. Low levels of SOCS2 correlated strongly with disease recurrence and metastasis in clinical specimens. SOCS2 downregulation recapitulated miR-194–driven metastatic phenotypes, whereas overexpression of a nontargetable SOCS2 reduced miR-194–stimulated invasion. Targeting of SOCS2 by miR-194 resulted in derepression of the oncogenic kinases FLT3 and JAK2, leading to enhanced ERK and STAT3 signaling. Pharmacologic inhibition of ERK and JAK/STAT pathways reversed miR-194–driven phenotypes. The GATA2 transcription factor was identified as an upstream regulator of miR-194, consistent with a strong concordance between GATA2 and miR-194 levels in clinical specimens. Overall, these results offer new insights into the molecular mechanisms of metastatic progression in prostate cancer. Cancer Res; 77(4); 1021–34. ©2016 AACR.

Prostate cancer is the most common non-skin cancer of men and causes >300,000 deaths worldwide annually (1). The predominant cause of mortality from prostate cancer is metastasis (2), which can be present at the time of diagnosis or develop after failure of primary treatment (surgery and/or radiotherapy). Therefore, both the identification of markers that accurately predict or demarcate metastatic disease at an early stage, as well as the development of strategies that effectively inhibit metastasis, would have a significant impact on patient outcomes. Both of these goals have been hindered by an imprecise understanding of the mechanisms governing dissemination of tumor cells from the prostate.

Metastasis of carcinomas (epithelial-derived cancers) encompasses a complex series of events whereby epithelial tumor cells invade the surrounding stroma, enter blood or lymphatic circulation, arrest at distant anatomic sites, exit the vasculature, and colonize a secondary location through metastatic outgrowth (3). Given the intricacy of the metastatic cascade, it is not surprising that many regulators of metastasis have been identified, often acting in a context-dependent manner. One such class of metastatic regulatory factors are miRNAs, small, noncoding RNA molecules of 21–23nt that regulate gene expression in a sequence-specific manner at both posttranscriptional and epigenetic levels (4). Aberrant miRNA expression is a common feature of many cancers, including prostate cancer (5), and this dysregulation can both promote and inhibit metastasis (6).

In a recent study, we identified miRNAs in the circulation that were predictive of disease recurrence following surgery for primary prostate cancer (7). The expression of one of the miRNAs identified in this earlier work, miR-194, was elevated in metastatic tissues, leading us to speculate that it could be a driver of disease progression. We report herein the elucidation of a novel prometastatic pathway in prostate cancer, in which GATA2-regulated miR-194 targets suppressor of cytokine signaling 2 (SOCS2) and thereby enhances the STAT3 and ERK signaling pathways.

Analysis of miR-194 and SOCS2 expression in published datasets

The Cancer Genome Atlas (TCGA) data (8) were obtained from the web portal (https://tcga-data.nci.nih.gov/tcga/) on October 31, 2015. The TCGA dataset used in this study comprised miRNA and mRNA expression data from 349 and 414 radical prostatectomy (RP) tumor samples, respectively. miRNA expression data from the Memorial Sloan Kettering Cancer Center (MSKCC; New York, NY) cohort (9) were processed as described previously (10); mRNA and clinical data for this cohort were obtained from the cBioPortal for Cancer Genomics (11). The MSKCC dataset used in this study comprised miRNA expression data from 99 RP tumor samples and mRNA expression data from 131 RP tumor samples, 19 metastases, and 29 normal prostate tissue samples. Data from the Grasso cohort (12), comprising 49 RP tumor samples, 27 metastases, and 12 normal prostate tissue samples, were obtained from the Gene Expression Omnibus (GEO; GSE35988). The association between SOCS2 expression in primary tumors (RP samples) and subsequent biochemical recurrence (BCR) and metastasis was also evaluated in a large “Multi-Institutional” cohort, which comprises microarray and clinical data from 1,447 RP tumor samples [from five previously published datasets: the Mayo Clinic (13, 14), Johns Hopkins University (15), Thomas Jefferson University (16), and the Cleveland Clinic (17)]. Microarray processing and normalization of this integrated dataset was performed as described previously (18); raw data are available from the GEO (GSE46691, GSE62116, GSE62667, GSE72291, GSE79956, and GSE79957).

Reagents

ERK1/2 inhibitor (SCH772984) was obtained from Selleckhem (S7101). Pan-JAK inhibitor (JAK1 inhibitor I) was obtained from Calbiochem (420097).

Cell line culture and transfection

PC-3, DU145, 22Rv1, C4-2B, and LNCaP human prostate carcinoma cells were obtained from the ATCC. DU145, C4-2B, LNCaP, and 22Rv1 cells were cultured in RPMI + 10% FBS. PC-3 cells were maintained in RPMI1640 containing 5% FBS. All cell lines underwent verification by short tandem repeat profiling in 2016 by CellBank Australia.

Cells were transfected with 20 nmol/L miRNA mimics (miR-194 or negative control mimic; Shanghai GenePharma), 50 nmol/L locked nucleic acid (LNA) miRNA inhibitors (miR-194 LNA inhibitor or negative control inhibitor; Exiqon), and 20 nmol/L siRNA [SOCS2 siRNA (Thermo Fisher Scientific, 4392420); GATA2 siRNA (Thermo Fisher Scientific, 1299001); or negative control siRNA; Qiagen)] using RNAiMAX Transfection Reagent (Life Technologies), according to the manufacturer's instructions. For plasmid transfections, cells were transfected with Lipofectamine 2000 (Life Technologies), according to the manufacturer's instructions.

Measuring levels of miR-194 in serum

Serum total RNA samples from men with metastatic disease have been described previously (10). During this same study, whole blood samples from men with localized prostate cancer were also collected. All samples were collected with institutional approval from the Research Ethics Board of the British Columbia Cancer Agency. Informed consent was obtained from all participating patients and volunteers. Blood was collected in accordance with the National Institute of Cancer standard operating procedures for serum and plasma processing. Total RNA and TaqMan qRT-PCR analysis of miR-194 (with a preamplification step) was performed as described previously (10).

miRNA in situ hybridization

In situ hybridization (ISH) was performed to determine the patterns of expression of miR-194 in human clinical prostate cancer tissue from unmatched benign and malignant prostate tissues. ISH was performed using an LNA-conjugated miR-194–specific probe from Exiqon according to the manufacturer's instructions (19). Slides were examined with the aid of an Olympus BX50 microscope; three random fields at ×60 magnification were analyzed for each sample.

Cell proliferation assays

Proliferation curves were performed essentially as described previously (20), with some minor modifications. Briefly, cells were seeded at 2 × 105 (PC3) or 3 × 105 (22Rv1) cells/well in 6-well plates and transfected in suspension with miRNA mimic, miRNA inhibitor, or siRNA as described above. Live and dead cells were subsequently quantified at the indicated time points using Trypan blue.

Cellular migration and invasion assays

In vitro scratch wound migration assays and Matrigel invasion assays in prostate cancer cell lines were conducted as described previously (21). For the SOCS2 rescue experiment, cells were cotransfected with miR-194 and a SOCS2 overexpression vector (Origene, SC108265).

qRT-PCR analysis of miRNA expression

Total RNA was extracted from prostate cancer cells using TRIzol, essentially as described previously (22), except that the RNA was precipitated with 2.5 volume of ethanol, 10 mmol/L MgCl2, 0.1 volume of 5 mol/L NaCl, and 20 μg of Glyco-Blue (Life Technologies) overnight at −20°C. Levels of miR-194 and U6 small nuclear RNA were measured by qRT-PCR using TaqMan assays, following the manufacturer's instructions (Life Technologies).

qRT-PCR analysis of mRNA expression

RNA extraction from cells, using TRIzol reagent, and qRT-PCR were done as described previously (22). GAPDH was used for normalization of qRT-PCR data. Primer sequences are available on request.

Western blotting

Protein extraction from cells, using RIPA buffer, and Western blotting were done as described previously (22). Antibodies used in Western blotting were E-cadherin (BD Biosciences, 610182), ERK (Cell Signaling Technology, 9102), phospho-ERK (Cell Signaling Technology, 9101), FLT3 (Cell Signaling Technology, #3462), N-cadherin (Santa Cruz Biotechnology, sc-7939), SOCS2 (Cell Signaling Technology, 2779), STAT3 (Cell Signaling Technology, 9132), phospho-STAT3 (Ser727; Cell Signaling Technology, 9134), ZO-1 (Santa Cruz Biotechnology, sc-10804), and GAPDH (Millipore, MAB374).

Immunofluorescence

22Rv1 cells were transfected with miR-194 mimic or control, as described above, plated onto chamber slides (Lab-Tek, Thermo Fisher Scientific), and stained at day 3. For E-cadherin staining, cells were fixed in 4% paraformaldehyde, permeabilized in 0.1% Triton X-100, and probed with an anti-E-cadherin antibody (1:500; BD Biosciences, 610182). To detect nuclei, cells were costained with 4′-6-diamidino-2-phenylindole (DAPI; Invitrogen). For F-actin staining, fixed and permeabilized cells were incubated with rhodamine phalloidin (Invitrogen) for 10 minutes. Cells were observed on and pictures were taken using a confocal microscope (Leica SP5).

Luciferase assays

The SOCS2 3′ untranslated region (3′UTR) was cloned into psi-CHECK-2 (Promega); details are available on request. To determine whether miR-194 directly targets the SOCS2 3′UTR, PC3 cells were transfected with miR-194 mimic or negative control. The following day, cells were transfected again with 500 ng of the psi-CHECK-2 SOCS2 3′UTR construct using Lipofectamine 2000 reagent. After 2 days, luciferase activity was measured using a Dual Luciferase Reporter Assay (Promega). Firefly luciferase activity was normalized to Renilla luciferase activity.

Generation of miR-194–overexpressing PC3 and LNCaP cells

Luciferase-tagged PC3 and LNCaP cells were a kind gift from Professor Andreas Evdokiou (University of Adelaide, Adelaide, Australia). The lines were generated using the retroviral expression vector SFG-NES-TGL, which gives rise to a single fusion protein encoding herpes simplex virus thymidine kinase, GFP, and firefly luciferase (Luc), as described previously (23). Lentivirus particles designed to overexpress miR-194 were prepared using a standard third-generation packaging system in HEK293T/17 cells after transfection of cells with packaging vector and GFP hsa-miR-194-5p miRNA lentivector (Applied Biological Materials Inc., mh11109) or empty vector control (Applied Biological Materials Inc., m003). For viral transduction experiments, luciferase-tagged PC3 and LNCaP cells were seeded at 1.25 × 107 cells per T75 flask and left overnight to adhere. The next day, cells were transduced with concentrated lentivirus using an MOI of 1 and 6 μg/mL (as final concentration) Polybrene (Sigma) in normal growth media.

Chick chorioallantoic membrane assays

Chick chorioallantoic membrane (CAM) assays were approved by the University of Adelaide Animal Ethics Committee (approval number M-2014_079). CAM assays were carried out essentially as described previously (24), using the in ovo method. PC3 cells (5 × 104) per 15 μL of growth media and LNCaP cells (8 × 104) per 15 μL of growth media were mixed with an equal volume of Matrigel and grafted on top of the CAM of day 11 chick embryos. After 3 days, the CAM implants (containing the cell:Matrigel graft) were removed and fixed in 4% formaldehyde and paraffin embedded. Hematoxylin and eosin staining and pan-cytokeratin IHC were performed as described previously (24). For quantitative analysis of cancer cell invasion into the mesoderm layer, 8 to 12 CAM images from each embryo (6–8 embryos per treatment) were assessed.

Intravenous experimental metastasis model

Intravenous xenograft experiments were approved by the University of Adelaide Animal Ethics Committee (approval number M-2014-180C). Eight-week-old male NOD/SCID mice (Animal Resources Centre, Western Australia) received tail vein injections of 1 × 106 PC3-194 cells in 200 μL of PBS. Noninvasive, whole-body imaging to monitor luciferase-expressing cells in mice was done at the time of injection and then once every week using the IVIS Lumina XRMS Series III Imaging System (PerkinElmer). Mice were injected intraperitoneally with 100 μL d-luciferin (Biosynth; Cas # 115144-35-9) solution at 300 mg/kg body weight and then gas-anesthetized with isoflurane. The photon emission transmitted from mice was captured and quantitated by converting to physical units of surface radiance (photons/sec/cm2/sr) using PerkinElmer Living Image (version 4.5).

Intraprostatic experimental metastasis model

Intraprostatic xenograft experiments were approved by the University of Adelaide Animal Ethics Committee (approval number M-2014-180C). Eight-week-old male NOD/SCID mice (Animal Resources Centre, Western Australia) received intraprostatic injections of 1.5 × 105 PC3-194 cells in 10 μL of PBS. In vivo imaging was carried out as described above. After sacrificing the animals, organs were removed for ex vivo bioluminescence imaging.

Prospective primary prostate cancer cohort

Primary tumor specimens were obtained with written informed consent through the Australian Prostate Cancer BioResource from 44 men who underwent robotic RP at St. Andrew's Hospital (Adelaide, Australia). The study was approved by the University of Adelaide Human Research Ethics Committee (approval number H-2012-016). A subset of these samples (n = 26) was first reported in a previous publication from our group (21); the remainder were collected subsequently. Tissues were homogenized in Qiazol using a Precellys 24 tissue homogenizer (Bertin Technologies) before RNA extraction with miRNeasy Mini Kits (Qiagen). DNase treatment was performed using a TURBO DNase Kit (Ambion) according to the manufacturer's instructions. RNA was quantified using a NanoDrop. Reverse transcription was performed on 400 ng total RNA using the iScript Kit (Bio-Rad Laboratories) according to the manufacturer's instructions. Levels of miR-194 and SOCS2 were not normally distributed; hence, Spearman correlation tests were used to examine relationships between their levels.

Gene set enrichment analysis

Gene set enrichment analysis (GSEA) was performed as described previously (21). The GATA2-regulated gene set used in this study was composed of genes that were downregulated (P < 0.001) in response to GATA knockdown (25).

Statistical analyses

All statistical analyses were carried out using GraphPad Prism (version 5; GraphPad Software). Details of statistical tests used are provided in the figure legends. For Kaplan–Meier analyses, optimal cutoffs for dichotomizing variables were defined as the point with the most significant (Fisher exact test) split using Cutoff Finder (26).

miR-194 enhances metastatic features of prostate cancer cells

In silico analyses of TCGA (8) prostate cancer cohort indicated that miR-194 levels are higher in primary tumors of men who subsequently experienced a new tumor event (Supplementary Fig. S1A) or BCR (Supplementary Fig. S1B). Moreover, miR-194 was elevated in high Gleason score tumors in both the TCGA cohort and an independent dataset from MSKCC (Fig. 1A; ref. 9). Notwithstanding the robust association between miR-194 and Gleason score, multivariate Cox proportional hazards regression revealed that these variables had independent prognostic value (Supplementary Table S1). Assessment of miR-194 expression in an “in-house” localized tumor cohort by ISH supported its association with Gleason grade (Fig. 1B) and demonstrated that it is expressed primarily in malignant epithelial cells (Fig. 1C). Moreover, expanding on our previous work in which miR-194 was identified as a marker of poor outcome following surgery (7), levels of this miRNA were significantly higher in the serum of men with metastatic castration-resistant prostate cancer (n = 18) compared with men with localized (n = 21) disease (P < 0.05; Supplementary Fig. S1C). Collectively, these analyses of multiple clinical cohorts demonstrated that miR-194 expression in prostate cancer is associated with tumor aggressiveness and metastasis.

Figure 1.

miR-194 is associated with disease aggressiveness in prostate tumors. A, Expression level of miR-194 in localized prostate tumors (obtained during RP) according to increasing Gleason score in the TCGA (8) and MSKCC (9) cohorts. Middle line, mean; lines above and below, ±SD. NS, not significant. P values were determined using unpaired two-sided t tests (*, P < 0.05; **, P < 0.01). B, Expression level of miR-194 in localized prostate tumors (obtained during RP) according to increasing Gleason grade in 23 primary tumors, as estimated using ISH. Error bars, SEM. P values were determined using Mann–Whitney U tests (**, P < 0.01). C, Representative ISH images from B demonstrate that miR-194 is expressed in prostatic epithelial cells. U6 staining confirmed the preservation of intact small RNAs in the same case, and a scrambled probe demonstrated specificity. Red arrows, weak staining in more normal glandular structures; black arrows, strong staining in foci of higher tumor grade. Scale bars, 50 μm. Gleason grade, GG.

Figure 1.

miR-194 is associated with disease aggressiveness in prostate tumors. A, Expression level of miR-194 in localized prostate tumors (obtained during RP) according to increasing Gleason score in the TCGA (8) and MSKCC (9) cohorts. Middle line, mean; lines above and below, ±SD. NS, not significant. P values were determined using unpaired two-sided t tests (*, P < 0.05; **, P < 0.01). B, Expression level of miR-194 in localized prostate tumors (obtained during RP) according to increasing Gleason grade in 23 primary tumors, as estimated using ISH. Error bars, SEM. P values were determined using Mann–Whitney U tests (**, P < 0.01). C, Representative ISH images from B demonstrate that miR-194 is expressed in prostatic epithelial cells. U6 staining confirmed the preservation of intact small RNAs in the same case, and a scrambled probe demonstrated specificity. Red arrows, weak staining in more normal glandular structures; black arrows, strong staining in foci of higher tumor grade. Scale bars, 50 μm. Gleason grade, GG.

Close modal

Given the association between miR-194 and tumor aggressiveness, we hypothesized that it may directly enhance prometastatic phenotypes. To test this concept, we modulated miR-194 activity in prostate cancer cell lines by overexpression (transfection of miRNA mimics) or inhibition (transfection of LNA miRNA inhibitors) and assessed cell migration, invasion, and proliferation. Ectopic delivery of miR-194 mimic caused PC3, DU145, and 22Rv1 cells to become more migratory and invasive in nature (Fig. 2A and B). Conversely, suppression of endogenous miR-194 activity with an LNA inhibitor suppressed cell invasion (Fig. 2C). Interestingly, concomitant with enhancing motility and invasiveness, miR-194 attenuated cancer cell proliferation in all models tested (Supplementary Fig. S2).

Figure 2.

miR-194 enhances metastatic features of prostate cancer cells. A, miR-194 promotes migration of PC3, DU145, and 22Rv1 cell lines. Error bars, ± SEM. B and C, miR-194 promotes invasion of PC3, DU145, and 22Rv1 cells. Values for the negative control (NC) were set to 1, and error bars are SEM. P values were determined using unpaired two-sided t tests (*, P < 0.05; ***, P < 0.001). B, Cells were transfected with miR-194 mimic or negative control (NC) mimic. C, Cells were transfected with miR-194 inhibitor or negative control inhibitor. Error bars, SEM. D, miR-194 enhances mesenchymal features of prostate cancer cells. The expression of epithelial (E-cad, ZO-1) and mesenchymal (N-cad) markers was examined by Western blotting following transfection of miR-194 or negative control mimic in PC3 and 22Rv1 cells. GAPDH was used as a loading control. E, Immunofluorescence analysis of E-cad and F-actin expression in 22Rv1 cells transfected with miR-194 or negative control mimic. Nuclei were stained with DAPI.

Figure 2.

miR-194 enhances metastatic features of prostate cancer cells. A, miR-194 promotes migration of PC3, DU145, and 22Rv1 cell lines. Error bars, ± SEM. B and C, miR-194 promotes invasion of PC3, DU145, and 22Rv1 cells. Values for the negative control (NC) were set to 1, and error bars are SEM. P values were determined using unpaired two-sided t tests (*, P < 0.05; ***, P < 0.001). B, Cells were transfected with miR-194 mimic or negative control (NC) mimic. C, Cells were transfected with miR-194 inhibitor or negative control inhibitor. Error bars, SEM. D, miR-194 enhances mesenchymal features of prostate cancer cells. The expression of epithelial (E-cad, ZO-1) and mesenchymal (N-cad) markers was examined by Western blotting following transfection of miR-194 or negative control mimic in PC3 and 22Rv1 cells. GAPDH was used as a loading control. E, Immunofluorescence analysis of E-cad and F-actin expression in 22Rv1 cells transfected with miR-194 or negative control mimic. Nuclei were stained with DAPI.

Close modal

Increased motility and invasiveness of cancer cells, coupled with a decrease in proliferative index, has been associated with epithelial–mesenchymal transition (EMT; ref. 27). EMT is a fundamental aspect of morphogenesis and wound healing during which sessile epithelial cells lose cell adhesion and convert to migratory and invasive mesenchymal cells. Cancer cells can hijack this process to enable migration and invasion of tumor cells into the surrounding stroma and entry/exit from the bloodstream (27). To test whether the promigratory and proinvasive activities of miR-194 were associated with EMT, expression of epithelial [E-cadherin (E-cad) and Zona occludens-1 (ZO-1)] and mesenchymal [N-cadherin (N-cad)] markers was examined by Western blotting. In both PC3 and 22Rv1 cells, delivery of miR-194 mimic resulted in changes reminiscent of EMT, namely loss of E-cad/ZO-1 and gain of N-cad (Fig. 2D). Moreover, the cells lost cohesion and gained a more elongated, mesenchymal-like morphology (Fig. 2E; Supplementary Fig. S3). Immunofluorescence analysis validated the loss of E-cad, and staining for F-actin revealed cytoskeletal rearrangement characteristic of EMT (Fig. 2E).

miR-194 promotes prostate cancer invasion and metastasis in vivo

Our in vitro experiments indicated that miR-194 could enhance prometastatic features of tumor cells. This notion was further investigated using in vivo models of tumor invasion and metastasis. Luciferase-tagged PC3-overexpressing miR-194, termed PC3-194, were generated by lentiviral transduction. We first examined the invasive capacity of these cells in in ovo CAM assays (24), which allow visualization of cell invasion through ectoderm into the mesoderm. Cell/Matrigel grafts were implanted onto the CAM and after 3 days, and invasion was assessed by pan-cytokeratin IHC. As expected, PC3-194 cells exhibited significantly augmented invasive capacity compared with control cells (Supplementary Fig. S4).

Having verified the ability of miR-194 to augment prostate cancer cell invasion, we next turned to experimental metastasis assays in immunocompromised mice. The luciferase-tagged PC3-194 and control (PC3-NC) cells were injected into the tail veins of NOD/SCID mice, which were then assessed by weekly whole-animal bioluminescence imaging. In this assay, overexpression of miR-194 enhanced both the speed and extent of metastatic colonization (Fig. 3A).

Figure 3.

miR-194 promotes invasion and metastasis of prostate cancer in vivo. A, Overexpression of miR-194 promotes metastasis of PC3 cells in tail vein assays. The graph on left represents luciferase intensity as assessed by whole-animal bioluminescent imaging over time in mice injected with control cells (PC3-NC, n = 5) or miR-194–overexpressing cells (PC3-194, n = 5), as assessed by noninvasive bioluminescent imaging. Tumor incidence was 100% for both cell lines. Significance of the differences in bioluminescence was assessed using unpaired two-sided t tests (***, P < 0.001; ****, P < 0.0001). Error bars, ± SEM. Right, images of mice at week 7. B, Overexpression of miR-194 promotes growth and metastasis of PC3 intraprostatic xenografts. The graph on left represents luciferase intensity as assessed by whole-animal bioluminescent imaging over time in mice injected with control cells (PC3-NC, n = 7) or miR-194–overexpressing cells (PC3-194, n = 9), as assessed by noninvasive bioluminescent imaging. Tumor incidence was 77.8% for PC3-NC (7/9) and 90% (9/10) for PC3-194. Significance of the differences in bioluminescence was assessed using unpaired two-sided t tests (*, P < 0.05; ***, P < 0.001). Error bars, ± SEM. Right, images of mice at week 5. C, Representative bioluminescent images of organs obtained from 2 PC3-NC and 2 PC3-194 mice shown in D at the time of sacrifice. K, kidney; L, lung; Li, liver; P, prostate primary tumor; S, spleen. D, Kaplan–Meier analysis showing survival of mice with intraprostatic xenografts. P value was determined using a log-rank test (**, P < 0.01). E, Representative immunohistochemical staining of pan-cytokeratin in CAM invasion assays. PC3 and LNCaP cell/Matrigel grafts (CM) were placed on top of the ectoderm (ET) layer, and cancer cell invasion into the CAM mesoderm (MD) was assessed in day 14 chick embryos. Endoderm, EN. Scale bar, 100 μm. F, Quantitation of CAM invasion assay data. Data were generated from 48 to 60 images from at least 6 chicken embryos per treatment. Data represent the mean percentage of images with invasion into the mesoderm. Error bars, SEM. P value was determined using an unpaired two-sided t test (***, P < 0.001; ****, P < 0.0001).

Figure 3.

miR-194 promotes invasion and metastasis of prostate cancer in vivo. A, Overexpression of miR-194 promotes metastasis of PC3 cells in tail vein assays. The graph on left represents luciferase intensity as assessed by whole-animal bioluminescent imaging over time in mice injected with control cells (PC3-NC, n = 5) or miR-194–overexpressing cells (PC3-194, n = 5), as assessed by noninvasive bioluminescent imaging. Tumor incidence was 100% for both cell lines. Significance of the differences in bioluminescence was assessed using unpaired two-sided t tests (***, P < 0.001; ****, P < 0.0001). Error bars, ± SEM. Right, images of mice at week 7. B, Overexpression of miR-194 promotes growth and metastasis of PC3 intraprostatic xenografts. The graph on left represents luciferase intensity as assessed by whole-animal bioluminescent imaging over time in mice injected with control cells (PC3-NC, n = 7) or miR-194–overexpressing cells (PC3-194, n = 9), as assessed by noninvasive bioluminescent imaging. Tumor incidence was 77.8% for PC3-NC (7/9) and 90% (9/10) for PC3-194. Significance of the differences in bioluminescence was assessed using unpaired two-sided t tests (*, P < 0.05; ***, P < 0.001). Error bars, ± SEM. Right, images of mice at week 5. C, Representative bioluminescent images of organs obtained from 2 PC3-NC and 2 PC3-194 mice shown in D at the time of sacrifice. K, kidney; L, lung; Li, liver; P, prostate primary tumor; S, spleen. D, Kaplan–Meier analysis showing survival of mice with intraprostatic xenografts. P value was determined using a log-rank test (**, P < 0.01). E, Representative immunohistochemical staining of pan-cytokeratin in CAM invasion assays. PC3 and LNCaP cell/Matrigel grafts (CM) were placed on top of the ectoderm (ET) layer, and cancer cell invasion into the CAM mesoderm (MD) was assessed in day 14 chick embryos. Endoderm, EN. Scale bar, 100 μm. F, Quantitation of CAM invasion assay data. Data were generated from 48 to 60 images from at least 6 chicken embryos per treatment. Data represent the mean percentage of images with invasion into the mesoderm. Error bars, SEM. P value was determined using an unpaired two-sided t test (***, P < 0.001; ****, P < 0.0001).

Close modal

Tail vein assays evaluate tumor cell extravasation and metastatic colonization. To determine how overexpression of miR-194 influences spontaneous metastasis, we employed an orthotopic assay in which cells were injected directly into the prostate (i.e., intraprostatic xenografts). Weekly monitoring of the mice revealed that tumor incidence was 77.8% for PC3-NC and 90% for PC3-194 cells, and PC3-194 tumors grew more rapidly (Fig. 3B). Ex vivo imaging of tissues following intraprostatic tumor growth revealed that in 6 of 9 mice, PC3-194 cells spread to major visceral organs, such as kidneys, lungs, livers, and spleens (Fig. 3C). In contrast, PC3-NC cells remained contained within the prostate in all mice (Fig. 3C). The longevity of the mice postinjection further highlighted the aggressiveness of the PC3-194 cells in the orthotopic assay (Fig. 3D).

To assess the potential of targeting miR-194 activity as a therapeutic strategy, we tested whether inhibition of endogenous miR-194 activity reduced invasion of prostate cancer cells in CAM assays. For these experiments, we used unmodified PC3 and LNCaP lines, the latter being chosen because it is amenable to these types of assays. Following transfection of a negative control LNA inhibitor, both cell types invaded through the ectoderm into the mesoderm layer of the CAM, although as expected, PC3 was more aggressive (Fig. 3E and F). Importantly, transfection of cells with miR-194 LNA inhibitor greatly reduced the invasive capacity of both cell lines (Fig. 3E and F).

SOCS2 is a clinically relevant target of miR-194

To elucidate the mechanism(s) by which miR-194 exerts its prometastatic effects, we identified candidate gene targets by intersecting outputs from three distinct prediction algorithms, TargetScan (28), miRanda (29), and PicTar (Supplementary Fig. S5; ref. 30). From the resultant list of 69 genes, we focused on factors that are downregulated in prostate cancer metastasis. One such gene, SOCS2, has a well-conserved miR-194 target site in its 3′UTR (Fig. 4A). In the MSKCC and Grasso (12) clinical cohorts, expression of SOCS2 was decreased in metastatic samples (Fig. 4B). Moreover, low tumoral expression of SOCS2 was associated with BCR in the MSKCC and TCGA cohorts and both BCR and metastasis in a large, integrated “multi-institutional” cohort (Fig. 4C; Supplementary Fig. S6).

Figure 4.

SOCS2 is a direct target of miR-194. A, Schematic showing conservation of putative miR-194 target site in the 3′ UTR of SOCS2. B, Relative expression of SOCS2 in normal prostate tissue, primary prostate tumors (obtained during RP), and metastases (Mets). Two different cohorts, Grasso (12) and MSKCC (9), were analyzed. Middle line, mean; lines above and below, ±SD. P values were determined using unpaired two-sided t tests (****, P < 0.0001). C, Kaplan–Meier curves showing estimated BCR-free or metastasis-free probability following RP in patients with high or low levels of tumor SOCS2 mRNA in the MSKCC (9) and Multi-Institutional cohorts. P values and HRs were determined using log-rank tests. D, Western blot showing SOCS2 levels following transfection with miR-194 mimic or a negative control (NC) in PC3 and 22Rv1 cells. GAPDH was used as a loading control. E, Western blot showing SOCS2 levels following transfection with a miR-194 inhibitor or a negative control inhibitor in 22Rv1 cells. Tubulin was used as a loading control. Normalized levels of SOCS2 are shown beneath the lanes. F, miR-194 directly targets the SOCS2 3′UTR. A Luciferase:SOCS2 3′UTR construct was cotransfected with miR-194 mimic or negative control in PC3 cells. Luciferase activity for “NC” was set to 1; bars represent the average of 6 wells per treatment, and error bars are SEM. A P value was determined using an unpaired t test (*, P < 0.05). G, Negative correlation of miR-194 and SOCS2 in prostate cancer. The graph shows the levels of miR-194 versus SOCS2 in 44 primary tumors, as determined by qRT-PCR. SOCS2 was normalized to GAPDH, whereas miR-194 was normalized to the reference small RNA U6. P and r values were calculated using a Spearman correlation test.

Figure 4.

SOCS2 is a direct target of miR-194. A, Schematic showing conservation of putative miR-194 target site in the 3′ UTR of SOCS2. B, Relative expression of SOCS2 in normal prostate tissue, primary prostate tumors (obtained during RP), and metastases (Mets). Two different cohorts, Grasso (12) and MSKCC (9), were analyzed. Middle line, mean; lines above and below, ±SD. P values were determined using unpaired two-sided t tests (****, P < 0.0001). C, Kaplan–Meier curves showing estimated BCR-free or metastasis-free probability following RP in patients with high or low levels of tumor SOCS2 mRNA in the MSKCC (9) and Multi-Institutional cohorts. P values and HRs were determined using log-rank tests. D, Western blot showing SOCS2 levels following transfection with miR-194 mimic or a negative control (NC) in PC3 and 22Rv1 cells. GAPDH was used as a loading control. E, Western blot showing SOCS2 levels following transfection with a miR-194 inhibitor or a negative control inhibitor in 22Rv1 cells. Tubulin was used as a loading control. Normalized levels of SOCS2 are shown beneath the lanes. F, miR-194 directly targets the SOCS2 3′UTR. A Luciferase:SOCS2 3′UTR construct was cotransfected with miR-194 mimic or negative control in PC3 cells. Luciferase activity for “NC” was set to 1; bars represent the average of 6 wells per treatment, and error bars are SEM. A P value was determined using an unpaired t test (*, P < 0.05). G, Negative correlation of miR-194 and SOCS2 in prostate cancer. The graph shows the levels of miR-194 versus SOCS2 in 44 primary tumors, as determined by qRT-PCR. SOCS2 was normalized to GAPDH, whereas miR-194 was normalized to the reference small RNA U6. P and r values were calculated using a Spearman correlation test.

Close modal

In light of the robust association between reduced levels of SOCS2 and subsequent development of metastatic disease, a finding that would be expected of a biologically relevant miR-194 target, we undertook experiments to validate the functional association between these two factors. SOCS2 protein levels were decreased by transfection of miR-194 mimic and increased by miR-194 inhibition (Fig. 4D and E). SOCS2 mRNA was decreased by exogenous miR-194 (Supplementary Fig. S7) but not to the same extent as SOCS2 protein, suggesting that targeting is composed of both translational repression and transcript degradation. Downregulation of SOCS2 by miR-194 was due to a direct miRNA:mRNA interaction, as miR-194 inhibited the expression of a luciferase reporter gene fused to a fragment of the SOCS2 3′UTR containing the target site (Fig. 4F). Finally, confirming the potential for miR-194 targeting of SOCS2 in vivo, there was a negative relationship between the expressions of these factors in an “in-house” cohort composed of 44 tumor samples (Fig. 4G).

Next, we assessed the relevance of SOCS2 in processes associated with miR-194–driven metastasis. Knockdown of SOCS2 (Fig. 5A) recapitulated miR-194 overexpression in all phenotypes tested, specifically downregulation of SOCS2 reduced proliferation (Supplementary Fig. S8) but enhanced migration (Fig. 5B) and invasion (Fig. 5C). To demonstrate dependence of miR-194 on SOCS2 to mediate its proinvasive ability, we enforced expression of the SOCS2 open reading frame without its 3′UTR in PC3 cells (Fig. 5D). In this rescue assay, cotransfection of the SOCS2 open reading frame overexpression construct (SOCS2-OE) with miR-194 significantly reversed miR-194–mediated invasion (Fig. 5E). These data demonstrate that SOCS2 is a key target through which miR-194 signals to promote metastatic phenotypes in prostate cancer cells.

Figure 5.

SOCS2 is a biologically relevant target of miR-194 and suppresses metastatic features of prostate cancer cells. A, Western blot showing SOCS2 levels in PC3 cells following transfection with two specific siRNAs [siSOCS2 (1) and siSOCS2 (2)] or a negative control (siNC). GAPDH was used as a loading control. B, SOCS2 knockdown (using a pool of the two siRNAs shown in A) promotes migration of PC3 cells. Error bars, ± SEM. C, SOCS2 knockdown (using a pool of the two siRNAs shown in A) promotes the invasion of PC3, DU145, and 22Rv1 cells. P values were determined using unpaired two-sided t tests (**, P < 0.01; ***, P < 0.001; ****, P < 0.0001). Error bars, SEM. D, Western blot showing SOCS2 levels following transfection with a SOCS2 overexpression construct (SOCS2-OE) or a negative control (NC-OE). E, Overexpression of SOCS2 reverses miR-194–induced invasion in PC3 cells. P values were determined using unpaired two-sided t tests (*, P < 0.05; **, P < 0.01). Error bars, SEM.

Figure 5.

SOCS2 is a biologically relevant target of miR-194 and suppresses metastatic features of prostate cancer cells. A, Western blot showing SOCS2 levels in PC3 cells following transfection with two specific siRNAs [siSOCS2 (1) and siSOCS2 (2)] or a negative control (siNC). GAPDH was used as a loading control. B, SOCS2 knockdown (using a pool of the two siRNAs shown in A) promotes migration of PC3 cells. Error bars, ± SEM. C, SOCS2 knockdown (using a pool of the two siRNAs shown in A) promotes the invasion of PC3, DU145, and 22Rv1 cells. P values were determined using unpaired two-sided t tests (**, P < 0.01; ***, P < 0.001; ****, P < 0.0001). Error bars, SEM. D, Western blot showing SOCS2 levels following transfection with a SOCS2 overexpression construct (SOCS2-OE) or a negative control (NC-OE). E, Overexpression of SOCS2 reverses miR-194–induced invasion in PC3 cells. P values were determined using unpaired two-sided t tests (*, P < 0.05; **, P < 0.01). Error bars, SEM.

Close modal

miR-194–mediated suppression of SOCS2 activates STAT3 and ERK signaling pathways

SOCS2 is a member of the SOCS family of ubiquitin E3 ligases, which have pleiotropic roles in normal physiology of the immune and central nervous systems as well as in the pathophysiology of various cancers (31). In the immune system, these functions are at least partly mediated by the ability of SOCS2 to target key tyrosine kinases, such as JAK2 and Fms-like tyrosine kinase 3 (FLT3), for degradation by the ubiquitin–proteasome system (31–33). Thus, we speculated that enhanced activity of JAK2- and FLT3-regulated pathways in response to loss of SOCS2 might be a mechanism by which miR-194 promotes metastasis. Indeed, total levels of both JAK2 and FLT3 were elevated in response to miR-194 overexpression and SOCS2 knockdown in prostate cancer cells (Fig. 6A). Moreover, miR-194 and siSOCS2 both enhanced the phosphorylation of STAT3 and ERK, respective substrates and mediators of JAK2 and FLT3 (Fig. 6B). Activation of STAT3 and ERK by miR-194 was further confirmed by measuring the expression of downstream prometastatic effector genes, including BIRC5, BRF1, CDC25A, ELK1, ELK4, FSCN2, and MYC (Fig. 6C). All of these genes were also positively correlated with miR-194 in clinical samples (Fig. 6D; Supplementary Fig. S9), suggesting that miR-194–mediated enhancement of ERK and STAT3 pathways is physiologically relevant.

Figure 6.

miR-194 activates STAT3 and ERK signaling pathways. A, Western blotting demonstrates that expression of total JAK2 and FLT3 is increased by miR-194 or SOCS2 knockdown in 22Rv1 cells. GAPDH was used as a loading control. B, Western blotting demonstrates that expression of phospho-STAT3 and phospho-ERK is increased by miR-194 or SOCS2 knockdown. GAPDH was used as a loading control. C, Expression of prometastatic downstream effectors of STAT3 and ERK signaling is increased in response to miR-194, as revealed by qRT-PCR. Levels of genes were normalized to the GAPDH. The control treatment [negative control (NC) mimic] was set to 1. Error bars, SEM. P values were determined using unpaired two-sided t tests (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001). D, Correlation of miR-194 and ERK/STAT3 effector genes in 414 prostate tumors from the TCGA cohort. P and r values were calculated using Pearson correlation tests. E, Inhibition of ERK (with SCH772984; left graph) or JAK (with JAK1 inhibitor I; right graph) reverses miR-194–mediated invasion in 22Rv1 cells. Error bars, SEM.

Figure 6.

miR-194 activates STAT3 and ERK signaling pathways. A, Western blotting demonstrates that expression of total JAK2 and FLT3 is increased by miR-194 or SOCS2 knockdown in 22Rv1 cells. GAPDH was used as a loading control. B, Western blotting demonstrates that expression of phospho-STAT3 and phospho-ERK is increased by miR-194 or SOCS2 knockdown. GAPDH was used as a loading control. C, Expression of prometastatic downstream effectors of STAT3 and ERK signaling is increased in response to miR-194, as revealed by qRT-PCR. Levels of genes were normalized to the GAPDH. The control treatment [negative control (NC) mimic] was set to 1. Error bars, SEM. P values were determined using unpaired two-sided t tests (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001). D, Correlation of miR-194 and ERK/STAT3 effector genes in 414 prostate tumors from the TCGA cohort. P and r values were calculated using Pearson correlation tests. E, Inhibition of ERK (with SCH772984; left graph) or JAK (with JAK1 inhibitor I; right graph) reverses miR-194–mediated invasion in 22Rv1 cells. Error bars, SEM.

Close modal

If ERK and STAT3 signaling pathways are important mediators of miR-194′s prometastatic activity, pharmacologic inhibition of these pathways should reverse miR-194–driven phenotypes. To test this hypothesis, we transfected 22Rv1 cells with miR-194 in the presence and absence of an ERK1/2 inhibitor (SCH772984) or a pan-JAK inhibitor (JAK1 inhibitor I) and measured invasion. In the presence of these drugs, the proinvasive activity of miR-194 was completely blocked (Fig. 6E).

GATA2 is an upstream regulator of miR-194

Having established key downstream mediators of miR-194 activity, we wished to identify upstream regulators that may be responsible for overexpression of this miRNA in prostate cancer. Initially, we searched for genes that were expressed concordantly with miR-194 in prostate tumors. GATA2 was one of the most highly positively correlated genes in the TCGA cohort (Fig. 7A). GATA2 is a transcription factor that plays multiple, key roles in prostate cancer growth and metastasis by modulating the androgen receptor (AR) and IGF2 signaling pathways (25, 34). Importantly, miR-194 levels were also closely associated with GATA2 activity in primary prostate tumors, as revealed by GSEA (Fig. 7B). Supporting the clinical association, knockdown of GATA2 by siRNA caused a significant reduction in miR-194 expression (Fig. 7C). Collectively, these data suggest that GATA2 is an upstream regulator of miR-194 in prostate cancer.

Figure 7.

miR-194 is a target of the transcription factor, GATA2. A, Positive correlation of miR-194 and GATA2 in prostate cancer. The graphs show the levels of the two miR-194–encoding transcripts (MIR194-1 and MIR194-2) versus GATA2 in 414 tumors from the TCGA cohort. P and r values were calculated using Pearson correlation tests. RPKM, reads per kilobase per million mapped reads. B, Positive association between miR-194 and a signature of GATA2 transcriptional activity (25), as demonstrated by GSEA, in the TCGA cohort. NES, normalized enrichment score; FWER, family-wise error rate. C, GATA2 knockdown (siGATA2) causes reduced miR-194 expression in 22Rv1 cells, as determined by qRT-PCR. Two different concentrations of siGATA2 were used in the experiment. Levels of miR-194 were normalized to the reference small RNA U6. P value was determined using an unpaired two-sided t test (*, P < 0.05). Error bars, SEM. Knockdown of GATA2 at the two different siRNA concentrations is shown below; GAPDH is used as a loading control. D, Model of miR-194 regulation and prometastatic activity, via SOCS2, in prostate cancer.

Figure 7.

miR-194 is a target of the transcription factor, GATA2. A, Positive correlation of miR-194 and GATA2 in prostate cancer. The graphs show the levels of the two miR-194–encoding transcripts (MIR194-1 and MIR194-2) versus GATA2 in 414 tumors from the TCGA cohort. P and r values were calculated using Pearson correlation tests. RPKM, reads per kilobase per million mapped reads. B, Positive association between miR-194 and a signature of GATA2 transcriptional activity (25), as demonstrated by GSEA, in the TCGA cohort. NES, normalized enrichment score; FWER, family-wise error rate. C, GATA2 knockdown (siGATA2) causes reduced miR-194 expression in 22Rv1 cells, as determined by qRT-PCR. Two different concentrations of siGATA2 were used in the experiment. Levels of miR-194 were normalized to the reference small RNA U6. P value was determined using an unpaired two-sided t test (*, P < 0.05). Error bars, SEM. Knockdown of GATA2 at the two different siRNA concentrations is shown below; GAPDH is used as a loading control. D, Model of miR-194 regulation and prometastatic activity, via SOCS2, in prostate cancer.

Close modal

To assess the relevance of miR-194 in the GATA2 signaling pathway, we tested whether it could rescue a phenotype induced by GATA2 knockdown. Specifically, 22Rv1 cells were cotransfected with siGATA2 and miR-194 mimic (or appropriate negative controls), and invasion assays were carried out. Knockdown of GATA2 potently inhibited invasion of this cell line (Supplementary Fig. S10), supporting an earlier finding from a distinct prostate cancer model system (35). Importantly, cotransfection with miR-194 reversed this phenotype but a negative control mimic did not (Supplementary Fig. S10). This observation is consistent with miR-194 being a mediator of GATA2′s proinvasive activity.

In this study, we have provided insight into the molecular mechanisms underlying prostate cancer metastasis by identifying a new pathway influencing this process. Specifically, by targeting SOCS2, miR-194 coordinately stimulates multiple prometastatic effectors downstream of JAK2/STAT3 and FLT3/ERK. These novel mechanistic findings likely underlie the observation that miR-194 is elevated in the serum of men with metastatic disease (this study) and, as we previously reported (7), in serum of patients with a rapid biochemical recurrence following surgery. A model summarizing these concepts is shown in Fig. 7D.

Providing new insight into miR-194 in cancer is an important outcome of this study because its precise functions have been difficult to resolve, with reports of both tumor-suppressive and oncogenic activity. For example, in models of gastric, colorectal, liver, kidney, breast, and endometrial cancers, miR-194 has variably been shown to suppress tumor growth, invasion, and metastasis, acting at least partly by driving epithelial differentiation and inhibiting EMT (36–40). In contrast, other groups have demonstrated that miR-194 can promote invasion and metastasis of pancreatic and endometrial cancer cells (41, 42) and colorectal cancer angiogenesis and growth (43) and that its expression is elevated in and during progression of esophageal adenocarcinoma (44). Moreover, an elegant study using laser-captured tissues demonstrated that high expression of miR-194 at the invasive front of liver cancers was a prognostic indicator of poor recurrence-free and overall survival in colorectal liver metastases (45). Our demonstration that miR-194 promotes invasion, migration, EMT, and metastasis is consistent with these latter findings. During the preparation of this manuscript, Zhang and colleagues published a study in which miR-194 was shown to target BMP-1 in PC3 cells, leading to decreased expression of matrix metalloproteinases 2 and 9 and subsequent suppression of invasion (46). We cannot currently explain the inconsistency between these findings and our current work. One explanation could simply be the known interlaboratory heterogeneity of cancer cell lines, which can lead to variable results. In our study, to ensure that findings were not predicated on a single model and therefore likely to be robust, we utilized multiple cell line models in in vitro and in vivo assays and interrogated miR-194 and SOCS2 expression in both “in-house” clinical samples and published clinical datasets. Further research on miR-194 in prostate cancer, including identification of its complete “targetome,” will undoubtedly unravel its precise role and resolve this apparent discrepancy.

The dichotomy of miR-194 action, acting as an oncogenic miRNA in some contexts and a tumor suppressor in others, is not uncommon in miRNA biology. This phenomenon is underpinned by the fact that miRNA function in a particular cell or tissue is contingent on the expression of relevant gene targets in that environment. Thus, the context-dependent roles of miR-194 can be attributed to its many cancer-relevant targets, including the EMT factor N-cad (47), the Wnt pathway receptor frizzled-6 (48), the cytoskeletal protein talin2 (49), the SCF E3 ubiquitin ligase component RING box protein1 (RBX1; ref. 50), MAP4K4 (39), the EMT-promoting transcription factor Bmi-1 (51), and the oncogenic transcription factor YAP1 (52). Here, we expand upon miR-194′s target repertoire by identifying SOCS2, which we believe represents a particularly critical mediator of miR-194 action in prostate cancer. SOCS2 is a substrate-recruiting component of E3-ubiquitin ligase complexes that promotes degradation of its targets via the ubiquitin–proteasome pathway (31). In the context of prostate cancer, we demonstrated that SOCS2 downregulates two key kinases, JAK2 and FLT3, thereby constraining STAT3 and ERK signaling. STAT3 and ERK act as key conduit molecules in prostate cancer, integrating various extrinsic signals to drive processes including cell survival, proliferation, suppression of apoptosis, AR activity, and phenotypic plasticity governing EMT and cancer stem cell maintenance (53, 54). By targeting SOCS2, miR-194 can therefore derepress multiple oncogenic and prometastatic signaling pathways.

Like miR-194, the function of SOCS2 in cancer is apparently complex, with reports of both oncogenic and tumor-suppressive activities (31). With this in mind, it is not surprising that recent studies of prostate cancer have produced conflicting data in relation to the function of SOCS2 (55, 56): Whereas one reported SOCS2 to exert growth-promoting effects (55), others demonstrated that SOCS2 inhibits tumor growth in vitro and in vivo and is a negative prognostic indicator postsurgery (56, 57). Our study at least partially resolves these conflicting findings by identifying a dual role for SOCS2 in this disease: It is required for normal cancer cell growth but acts concomitantly to inhibit metastatic characteristics of cancer cells mediated by JAK2 and FLT3. This latter function is reinforced by our observation that reduced expression of SOCS2 in large, contemporary prostate cancer cohorts is strongly associated with increased incidence of biochemical recurrence and metastasis.

In addition to elucidating downstream effectors of miR-194 activity, in this study, we identified GATA2 as an upstream regulator of miR-194 expression. GATA2 has a well-characterized function as an AR coregulator (25, 58), but can also operate independently of the AR signaling axis to promote chemoresistance, tumorigenicity, and metastasis (34, 35). These latter AR-independent functions are at least in part mediated by direct regulation of IGF2 by GATA2, which results in the activation of IGF1R and INSR and a downstream polykinase program (34). Our study expands the multifaceted nature of GATA2 in prostate cancer, providing a novel link between GATA2 and the oncogenic signal transducers STAT3 and ERK, with miR-194 at the nexus (Fig. 7D). Although we cannot rule out interplay between these two mechanisms of action of GATA2, we propose that miR-194 and IGF2 represent distinct downstream targets of GATA2 that nevertheless act in a complementary manner to additively enhance oncogenic signaling pathways and thereby promote prostate cancer progression. Importantly, we believe the GATA2–miR-194 pathway may be generalizable to other tissues and malignancies, as we found evidence for a positive association between miR-194 and GATA2 signaling in breast cancer (Supplementary Fig. S11).

A caveat of our in vitro mechanistic experiments is that the cell lines are all models of advanced, aggressive prostate cancer. However, our analyses of both published and “in-house” clinical samples provided compelling support for (i) an association between disease progression and high or low levels of miR-194 and SOCS2, respectively; (ii) an inverse relationship between miR-194 and SOCS2; and (iii) a positive association between miR-194 and signaling pathways regulated by STAT3, ERK, and GATA2. Therefore, the clinical data collectively suggest that the cell lines utilized herein accurately model miR-194 function in prostate cancer.

Our interest in miR-194 arose from an earlier finding that its levels in serum/plasma, measured immediately prior to RP for localized prostate cancer, were predictive of recurrence after a disease-free period ranging from 1 to 70 months (7). In this previous study, we also noted that tissue expression of miR-194 was higher in metastatic disease, leading us to speculate that its potential as a biomarker was a consequence of its release into the blood from tumor cells with high metastatic potential and/or clinically undetectable micrometastases. This hypothesis is further strengthened by the work shown in this study. More specifically, using large independent clinical cohorts, we show that miR-194 levels in the primary tumor are highly prognostic for prostate cancer metastasis and that circulating miR-194 is higher in men with metastatic compared with localized disease. Collectively, these findings support the idea that measuring circulating miR-194 could be used to detect micrometastases at the time of treatment or predict metastatic recurrence posttreatment. However, we acknowledge that the putative utility of miR-194 as a biomarker requires substantial additional validation.

In summary, our study has identified a new metastasis-promoting miRNA, miR-194, a new metastasis-suppressing factor, SOCS2, and a novel GATA2-driven pathway influencing prostate cancer metastasis. These findings have implications in terms of utilizing miR-194 as a biomarker and warrant further investigation into potential targeting of this miRNA to suppress prostate cancer progression.

S.G. Zhao has ownership interest (including patents) in and has provided expert testimony for GenomeDx Biosciences. E. Davicioni is the president at and has ownership interest (including patents) in GenomeDx Biosciences. F.Y. Feng is the founder of PFS Genomics and is a consultant/advisory board member for Celgene, Dendreon, GenomeDx, Medivation/Astellas, and Sanofi. No potential conflicts of interest were disclosed by the other authors.

Conception and design: R. Das, L.M. Butler, L.A. Selth

Development of methodology: R. Das, S.G. Zhao, H.K. Armstrong, N.A. Lokman, E.D. Williams, L.A. Selth

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Das, I. Denis, Q. Wang, S.L. Townley, A.R. Hanson, H.K. Armstrong, E. Davicioni, R.B. Jenkins, R.J. Karnes, A.E. Ross, R.B. Den, E.A. Klein, K.N. Chi, H.S. Ramshaw, E.D. Williams, W.D. Tilley, L.A. Selth

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R. Das, P.A. Gregory, R.C. Fernandes, I. Denis, Q. Wang, S.L. Townley, S.G. Zhao, H.K. Armstrong, E. Ebrahimie, E. Davicioni, R.B. Jenkins, R.J. Karnes, E.D. Williams, G.J. Goodall, F.Y. Feng, L.M. Butler, W.D. Tilley, L.A. Selth

Writing, review, and/or revision of the manuscript: R. Das, P.A. Gregory, S.G. Zhao, H.K. Armstrong, R.B. Jenkins, R.J. Karnes, A.E. Ross, R.B. Den, E.A. Klein, K.N. Chi, E.D. Williams, A. Zoubeidi, G.J. Goodall, L.M. Butler, W.D. Tilley, L.A. Selth

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R. Das, R.C. Fernandes, I. Denis, A.R. Hanson, M.A. Pickering, R.B. Jenkins, E.D. Williams, F.Y. Feng, W.D. Tilley

Study supervision: P.A. Gregory, L.M. Butler, L.A. Selth

The authors thank Dr. Denise Noonan (University of Adelaide Laboratory Animal Services) for assistance with animal experiments, Dr. Agatha Labrinidis (University of Adelaide Microscopy) for assistance with animal imaging, Suraya Roslan (Centre for Cancer Biology, SA Pathology and University of South Australia) for guidance with xenograft experiments, Dr. Carmela Ricciardelli (University of Adelaide) for guidance with CAM assays, Professor Andreas Evdokiou (University of Adelaide) for providing luciferase-tagged cell lines, and Dr. Margaret Centenera and Swati Irani (University of Adelaide) for assistance with prostate tumor collection. The authors are grateful to study participants, as well as the urologists (particularly Dr. Peter Sutherland), nurses, and histopathologists, who assisted in the recruitment and collection of patient samples and clinical data. Finally, the results published here are in part based on data generated by TCGA, established by the National Cancer Institute and the National Human Genome Research Institute, and we are grateful to the specimen donors and relevant research groups associated with this project.

This work was supported by funding from the National Health and Medical Research Council of Australia (ID 1083961 to L.A. Selth, W.D. Tilley, G.J. Goodall, and P.A. Gregory) and a Prostate Cancer Research Programs Transformative Impact Award from the U.S. Department of Defense (W81XWH-13-2-0093 to W.D. Tilley and L.A. Selth). R. Das was supported by an award from The Hospital Research Foundation. L.A. Selth was previously supported by a Young Investigator Award from the Prostate Cancer Foundation (the Foundation 14 award). P.A. Gregory was supported by a Beat Cancer Project fellowship from the Cancer Council of South Australia. The research programs of L.M. Butler, W.D. Tilley, L.A. Selth, and E.D. Williams were supported by the Movember Foundation and the Prostate Cancer Foundation of Australia through Movember Revolutionary Team Awards.

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.

1.
Ferlay
J
,
Soerjomataram
I
,
Dikshit
R
,
Eser
S
,
Mathers
C
,
Rebelo
M
, et al
Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012
.
Int J Cancer
2015
;
136
:
E359
86
.
2.
Wu
JN
,
Fish
KM
,
Evans
CP
,
Devere White
RW
,
Dall'Era
MA
. 
No improvement noted in overall or cause-specific survival for men presenting with metastatic prostate cancer over a 20-year period
.
Cancer
2014
;
120
:
818
23
.
3.
Valastyan
S
,
Weinberg
RA
. 
Tumor metastasis: molecular insights and evolving paradigms
.
Cell
2011
;
147
:
275
92
.
4.
Brodersen
P
,
Voinnet
O
. 
Revisiting the principles of microRNA target recognition and mode of action
.
Nat Rev Mol Cell Biol
2009
;
10
:
141
8
.
5.
Selth
LA
,
Tilley
WD
,
Butler
LM
. 
Circulating microRNAs: macro-utility as markers of prostate cancer?
Endocr Relat Cancer
2012
;
19
:
R99
R113
.
6.
Bouyssou
JM
,
Manier
S
,
Huynh
D
,
Issa
S
,
Roccaro
AM
,
Ghobrial
IM
. 
Regulation of microRNAs in cancer metastasis
.
Biochim Biophys Acta
2014
;
1845
:
255
65
.
7.
Selth
LA
,
Townley
SL
,
Bert
AG
,
Stricker
PD
,
Sutherland
PD
,
Horvath
LG
, et al
Circulating microRNAs predict biochemical recurrence in prostate cancer patients
.
Br J Cancer
2013
;
109
:
641
50
.
8.
The Cancer Genome Atlas Research Network
. 
The molecular taxonomy of primary prostate cancer
.
Cell
2015
;
163
:
1011
25
.
9.
Taylor
BS
,
Schultz
N
,
Hieronymus
H
,
Gopalan
A
,
Xiao
Y
,
Carver
BS
, et al
Integrative genomic profiling of human prostate cancer
.
Cancer Cell
2010
;
18
:
11
22
.
10.
Selth
LA
,
Townley
S
,
Gillis
JL
,
Ochnik
AM
,
Murti
K
,
Macfarlane
RJ
, et al
Discovery of circulating microRNAs associated with human prostate cancer using a mouse model of disease
.
Int J Cancer
2012
;
131
:
652
61
.
11.
Cerami
E
,
Gao
J
,
Dogrusoz
U
,
Gross
BE
,
Sumer
SO
,
Aksoy
BA
, et al
The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data
.
Cancer Discov
2012
;
2
:
401
4
.
12.
Grasso
CS
,
Wu
YM
,
Robinson
DR
,
Cao
X
,
Dhanasekaran
SM
,
Khan
AP
, et al
The mutational landscape of lethal castration-resistant prostate cancer
.
Nature
2012
;
487
:
239
43
.
13.
Erho
N
,
Crisan
A
,
Vergara
IA
,
Mitra
AP
,
Ghadessi
M
,
Buerki
C
, et al
Discovery and validation of a prostate cancer genomic classifier that predicts early metastasis following radical prostatectomy
.
PLoS One
2013
;
8
:
e66855
.
14.
Karnes
RJ
,
Bergstralh
EJ
,
Davicioni
E
,
Ghadessi
M
,
Buerki
C
,
Mitra
AP
, et al
Validation of a genomic classifier that predicts metastasis following radical prostatectomy in an at risk patient population
.
J Urol
2013
;
190
:
2047
53
.
15.
Benzon
B
,
Zhao
SG
,
Haffner
MC
,
Takhar
M
,
Erho
N
,
Yousefi
K
, et al
Correlation of B7-H3 with androgen receptor, immune pathways and poor outcome in prostate cancer: an expression-based analysis
.
Prostate Cancer Prostatic Dis
2016
Nov 1.
[Epub ahead of print]
.
16.
Den
RB
,
Feng
FY
,
Showalter
TN
,
Mishra
MV
,
Trabulsi
EJ
,
Lallas
CD
, et al
Genomic prostate cancer classifier predicts biochemical failure and metastases in patients after postoperative radiation therapy
.
Int J Radiat Oncol Biol Phys
2014
;
89
:
1038
46
.
17.
Klein
EA
,
Yousefi
K
,
Haddad
Z
,
Choeurng
V
,
Buerki
C
,
Stephenson
AJ
, et al
A genomic classifier improves prediction of metastatic disease within 5 years after surgery in node-negative high-risk prostate cancer patients managed by radical prostatectomy without adjuvant therapy
.
Eur Urol
2015
;
67
:
778
86
.
18.
Prensner
JR
,
Zhao
S
,
Erho
N
,
Schipper
M
,
Iyer
MK
,
Dhanasekaran
SM
, et al
RNA biomarkers associated with metastatic progression in prostate cancer: a multi-institutional high-throughput analysis of SChLAP1
.
Lancet Oncol
2014
;
15
:
1469
80
.
19.
Jorgensen
S
,
Baker
A
,
Moller
S
,
Nielsen
BS
. 
Robust one-day in situ hybridization protocol for detection of microRNAs in paraffin samples using LNA probes
.
Methods
2010
;
52
:
375
81
.
20.
Hickey
TE
,
Irvine
CM
,
Dvinge
H
,
Tarulli
GA
,
Hanson
AR
,
Ryan
NK
, et al
Expression of androgen receptor splice variants in clinical breast cancers
.
Oncotarget
2015
;
6
:
44728
44
.
21.
Selth
LA
,
Das
R
,
Townley
SL
,
Coutinho
I
,
Hanson
AR
,
Centenera
MM
, et al
A ZEB1-miR-375-YAP1 pathway regulates epithelial plasticity in prostate cancer
.
Oncogene
2016
.
22.
Moore
NL
,
Buchanan
G
,
Harris
JM
,
Selth
LA
,
Bianco-Miotto
T
,
Hanson
AR
, et al
An androgen receptor mutation in the MDA-MB-453 cell line model of molecular apocrine breast cancer compromises receptor activity
.
Endoc Relat Cancer
2012
;
19
:
599
613
.
23.
Zinonos
I
,
Labrinidis
A
,
Lee
M
,
Liapis
V
,
Hay
S
,
Ponomarev
V
, et al
Anticancer efficacy of Apo2L/TRAIL is retained in the presence of high and biologically active concentrations of osteoprotegerin in vivo
.
J Bone Miner Res
2011
;
26
:
630
43
.
24.
Lokman
NA
,
Elder
AS
,
Ricciardelli
C
,
Oehler
MK
. 
Chick chorioallantoic membrane (CAM) assay as an in vivo model to study the effect of newly identified molecules on ovarian cancer invasion and metastasis
.
Int J Mol Sci
2012
;
13
:
9959
70
.
25.
He
B
,
Lanz
RB
,
Fiskus
W
,
Geng
C
,
Yi
P
,
Hartig
SM
, et al
GATA2 facilitates steroid receptor coactivator recruitment to the androgen receptor complex
.
Proc Natl Acad Sci U S A
2014
;
111
:
18261
6
.
26.
Budczies
J
,
Klauschen
F
,
Sinn
BV
,
Gyorffy
B
,
Schmitt
WD
,
Darb-Esfahani
S
, et al
Cutoff Finder: a comprehensive and straightforward Web application enabling rapid biomarker cutoff optimization
.
PLoS One
2012
;
7
:
e51862
.
27.
De Craene
B
,
Berx
G
. 
Regulatory networks defining EMT during cancer initiation and progression
.
Nat Rev Cancer
2013
;
13
:
97
110
.
28.
Lewis
BP
,
Burge
CB
,
Bartel
DP
. 
Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets
.
Cell
2005
;
120
:
15
20
.
29.
Enright
AJ
,
John
B
,
Gaul
U
,
Tuschl
T
,
Sander
C
,
Marks
DS
. 
MicroRNA targets in Drosophila
.
Genome Biol
2003
;
5
:
R1
.
30.
Krek
A
,
Grun
D
,
Poy
MN
,
Wolf
R
,
Rosenberg
L
,
Epstein
EJ
, et al
Combinatorial microRNA target predictions
.
Nat Genet
2005
;
37
:
495
500
.
31.
Letellier
E
,
Haan
S
. 
SOCS2: physiological and pathological functions
.
Front Biosci
2016
;
8
:
189
204
.
32.
Kazi
JU
,
Ronnstrand
L
. 
Suppressor of cytokine signaling 2 (SOCS2) associates with FLT3 and negatively regulates downstream signaling
.
Mol Oncol
2013
;
7
:
693
703
.
33.
Sen
B
,
Peng
S
,
Woods
DM
,
Wistuba
I
,
Bell
D
,
El-Naggar
AK
, et al
STAT5A-mediated SOCS2 expression regulates Jak2 and STAT3 activity following c-Src inhibition in head and neck squamous carcinoma
.
Clin Cancer Res
2012
;
18
:
127
39
.
34.
Vidal
SJ
,
Rodriguez-Bravo
V
,
Quinn
SA
,
Rodriguez-Barrueco
R
,
Lujambio
A
,
Williams
E
, et al
A targetable GATA2-IGF2 axis confers aggressiveness in lethal prostate cancer
.
Cancer Cell
2015
;
27
:
223
39
.
35.
Chiang
YT
,
Wang
K
,
Fazli
L
,
Qi
RZ
,
Gleave
ME
,
Collins
CC
, et al
GATA2 as a potential metastasis-driving gene in prostate cancer
.
Oncotarget
2014
;
5
:
451
61
.
36.
Dong
P
,
Kaneuchi
M
,
Watari
H
,
Hamada
J
,
Sudo
S
,
Ju
J
, et al
MicroRNA-194 inhibits epithelial to mesenchymal transition of endometrial cancer cells by targeting oncogene BMI-1
.
Mol Cancer
2011
;
10
:
99
.
37.
Li
Z
,
Ying
X
,
Chen
H
,
Ye
P
,
Shen
Y
,
Pan
W
, et al
MicroRNA-194 inhibits the epithelial-mesenchymal transition in gastric cancer cells by targeting FoxM1
.
Dig Dis Sci
2014
;
59
:
2145
52
.
38.
Meng
Z
,
Fu
X
,
Chen
X
,
Zeng
S
,
Tian
Y
,
Jove
R
, et al
miR-194 is a marker of hepatic epithelial cells and suppresses metastasis of liver cancer cells in mice
.
Hepatology
2010
;
52
:
2148
57
.
39.
Wang
B
,
Shen
ZL
,
Gao
ZD
,
Zhao
G
,
Wang
CY
,
Yang
Y
, et al
MiR-194, commonly repressed in colorectal cancer, suppresses tumor growth by regulating the MAP4K4/c-Jun/MDM2 signaling pathway
.
Cell Cycle
2015
;
14
:
1046
58
.
40.
Wu
X
,
Liu
T
,
Fang
O
,
Leach
LJ
,
Hu
X
,
Luo
Z
. 
miR-194 suppresses metastasis of non-small cell lung cancer through regulating expression of BMP1 and p27(kip1)
.
Oncogene
2014
;
33
:
1506
14
.
41.
Zhang
J
,
Zhao
CY
,
Zhang
SH
,
Yu
DH
,
Chen
Y
,
Liu
QH
, et al
Upregulation of miR-194 contributes to tumor growth and progression in pancreatic ductal adenocarcinoma
.
Oncol Rep
2014
;
31
:
1157
64
.
42.
Chung
TK
,
Lau
TS
,
Cheung
TH
,
Yim
SF
,
Lo
KW
,
Siu
NS
, et al
Dysregulation of microRNA-204 mediates migration and invasion of endometrial cancer by regulating FOXC1
.
Int J Cancer
2012
;
130
:
1036
45
.
43.
Sundaram
P
,
Hultine
S
,
Smith
LM
,
Dews
M
,
Fox
JL
,
Biyashev
D
, et al
p53-responsive miR-194 inhibits thrombospondin-1 and promotes angiogenesis in colon cancers
.
Cancer Res
2011
;
71
:
7490
501
.
44.
Mathe
EA
,
Nguyen
GH
,
Bowman
ED
,
Zhao
Y
,
Budhu
A
,
Schetter
AJ
, et al
MicroRNA expression in squamous cell carcinoma and adenocarcinoma of the esophagus: associations with survival
.
Clin Cancer Res
2009
;
15
:
6192
200
.
45.
Kahlert
C
,
Klupp
F
,
Brand
K
,
Lasitschka
F
,
Diederichs
S
,
Kirchberg
J
, et al
Invasion front-specific expression and prognostic significance of microRNA in colorectal liver metastases
.
Cancer Sci
2011
;
102
:
1799
807
.
46.
Zhang
C
,
Shu
L
,
Kim
H
,
Khor
TO
,
Wu
R
,
Li
W
, et al
Phenethyl isothiocyanate (PEITC) suppresses prostate cancer cell invasion epigenetically through regulating microRNA-194
.
Mol Nutr Food Res
2016
;
60
:
1427
36
.
47.
Han
K
,
Zhao
T
,
Chen
X
,
Bian
N
,
Yang
T
,
Ma
Q
, et al
microRNA-194 suppresses osteosarcoma cell proliferation and metastasis in vitro and in vivo by targeting CDH2 and IGF1R
.
Int J Oncol
2014
;
45
:
1437
49
.
48.
Krutzfeldt
J
,
Rosch
N
,
Hausser
J
,
Manoharan
M
,
Zavolan
M
,
Stoffel
M
. 
MicroRNA-194 is a target of transcription factor 1 (Tcf1, HNF1alpha) in adult liver and controls expression of frizzled-6
.
Hepatology
2012
;
55
:
98
107
.
49.
Le
XF
,
Almeida
MI
,
Mao
W
,
Spizzo
R
,
Rossi
S
,
Nicoloso
MS
, et al
Modulation of microRNA-194 and cell migration by HER2-targeting trastuzumab in breast cancer
.
PLoS One
2012
;
7
:
e41170
.
50.
Chen
X
,
Wang
Y
,
Zang
W
,
Du
Y
,
Li
M
,
Zhao
G
. 
miR-194 targets RBX1 gene to modulate proliferation and migration of gastric cancer cells
.
Tumour Biol
2015
;
36
:
2393
401
.
51.
Huang
D
,
He
X
,
Zou
J
,
Guo
P
,
Jiang
S
,
Lv
N
, et al
Negative regulation of Bmi-1 by AMPK and implication in cancer progression
.
Oncotarget
2016
;
7
:
6188
200
.
52.
Jung
KH
,
McCarthy
RL
,
Zhou
C
,
Uprety
N
,
Barton
MC
,
Beretta
L
. 
MicroRNA-194 regulates hepatocytic differentiation of progenitor cells by targeting YAP1
.
Stem Cells
2016
;
34
:
1284
96
.
53.
Bishop
JL
,
Thaper
D
,
Zoubeidi
A
. 
The multifaceted roles of STAT3 signaling in the progression of prostate cancer
.
Cancers
2014
;
6
:
829
59
.
54.
Rodriguez-Berriguete
G
,
Fraile
B
,
Martinez-Onsurbe
P
,
Olmedilla
G
,
Paniagua
R
,
Royuela
M
. 
MAP kinases and prostate cancer
.
J Signal Transduct
2012
;
2012
:
169170
.
55.
Hoefer
J
,
Kern
J
,
Ofer
P
,
Eder
IE
,
Schafer
G
,
Dietrich
D
, et al
SOCS2 correlates with malignancy and exerts growth-promoting effects in prostate cancer
.
Endocr Relat Cancer
2014
;
21
:
175
87
.
56.
Iglesias-Gato
D
,
Chuan
YC
,
Wikstrom
P
,
Augsten
S
,
Jiang
N
,
Niu
Y
, et al
SOCS2 mediates the cross talk between androgen and growth hormone signaling in prostate cancer
.
Carcinogenesis
2014
;
35
:
24
33
.
57.
Zhu
JG
,
Dai
QS
,
Han
ZD
,
He
HC
,
Mo
RJ
,
Chen
G
, et al
Expression of SOCSs in human prostate cancer and their association in prognosis
.
Mol Cell Biochem
2013
;
381
:
51
9
.
58.
Wu
D
,
Sunkel
B
,
Chen
Z
,
Liu
X
,
Ye
Z
,
Li
Q
, et al
Three-tiered role of the pioneer factor GATA2 in promoting androgen-dependent gene expression in prostate cancer
.
Nucleic Acids Res
2014
;
42
:
3607
22
.