MicroRNA-466 inhibits tumor growth and bone metastasis in prostate cancer by direct regulation of osteogenic transcription factor RUNX2

MicroRNAs (miRNAs) have emerged as key players in cancer progression and metastatic initiation yet their importance in regulating prostate cancer (PCa) metastasis to bone has begun to be appreciated. We employed multimodal strategy based on in-house PCa clinical samples, publicly available TCGA cohorts, a panel of cell lines, in silico analyses, and a series of in vitro and in vivo assays to investigate the role of miR-466 in PCa. Expression analyses revealed that miR-466 is under-expressed in PCa compared to normal tissues. Reconstitution of miR-466 in metastatic PCa cell lines impaired their oncogenic functions such as cell proliferation, migration/invasion and induced cell cycle arrest, and apoptosis compared to control miRNA. Conversely, attenuation of miR-466 in normal prostate cells induced tumorigenic characteristics. miR-466 suppressed PCa growth and metastasis through direct targeting of bone-related transcription factor RUNX2. Overexpression of miR-466 caused a marked downregulation of integrated network of RUNX2 target genes such as osteopontin, osteocalcin, ANGPTs, MMP11 including Fyn, pAkt, FAK and vimentin that are known to be involved in migration, invasion, angiogenesis, EMT and metastasis. Xenograft models indicate that miR-466 inhibits primary orthotopic tumor growth and spontaneous metastasis to bone. Receiver operating curve and Kaplan–Meier analyses show that miR-466 expression can discriminate between malignant and normal prostate tissues; and can predict biochemical relapse. In conclusion, our data strongly suggests miR-466-mediated attenuation of RUNX2 as a novel therapeutic approach to regulate PCa growth, particularly metastasis to bone. This study is the first report documenting the anti-bone metastatic role and clinical significance of miR-466 in prostate cancer.

Prostate cancer is the second most leading cause of cancer related deaths among American men. 1 According to recent data, it is estimated that 220, 800 newly diagnosed prostate cancer cases and 27 540 deaths will occur in 2015. 1 The 5 year relative survival rate of early stage prostate cancer is 499% while that of advanced metastatic disease is only 28%. 1 Metastases often occur with no prior indication of tumor invasiveness. 2 A major challenge for treatment of advanced metastatic disease is the lack of understanding of the molecular mechanisms underlying the propensity of prostate cancer to metastasize to other organs, particularly the bone. A number of transcription factors have been identified that play key roles in promoting oncogenesis, tumor growth, metastasis and tissue destruction.
Runt-domain containing protein RUNX2 (also called Osf2/ Cbfa1, AML-3 or Pebp2αA) is a member of the RUNX family of genes. It is a lineage-specific transcription factor with crucial roles in both bone biology and carcinogenesis. [3][4][5] High levels of RUNX2 in metastatic prostate cancer emphasize the significance of this master skeletal transcription factor in potentiating tumor cell progression and metastatic bone disease. [6][7][8] The oncogenic potential of RUNX2 was first documented T-cell lymphoma. 9,10 Since then, a number of independent studies have reported involvement of RUNX2 at early and late stages of tumor progression in various cancers. 6,7,[11][12][13][14][15] RUNX2 was found to be up-regulated and a predictor of metastasis in prostate cancer. 16 Mechanistic studies indicated that RUNX2 accelerated prostate cancer aggressiveness through promotion of cadherin switching, invasion toward collagen I and Akt activation. 16 In addition, attenuation of RUNX2 expression in a PC3 subline led to increased adhesion to fibronectin. 7 Various reports have convincingly shown that the severity of prostate and mammary cancer cell osteolysis is correlated with levels of RUNX2 expression following injection of neoplastic cells in the medullary cavity of the tibia. 7,8,14 These findings suggest that RUNX2 can imbue neoplastic cells with the capacity to breakdown the architecture of surrounding tissues, potentially releasing tumor supporting factors indicating that RUNX2 may facilitate a bone-mimetic program that helps cells adapt and thrive in this foreign environment. RUNX2 also directly induced genes associated with angiogenesis, invasiveness, metastasis and stimulated epithelial-to-mesenchymal transition of primary tumors. [15][16][17] Taken together these studies provide convincing evidence for RUNX2 as a potentially important factor in prostate cancer development and metastasis.
Targeting RUNX2 in prostate cancer could interrupt an integrated network of gene expression required to maintain tumor growth and bone metastasis. Here we report use of miR-466 as a potentially novel avenue to regulate RUNX2 and its downstream genes, thereby inhibiting prostate cancer growth and bone metastasis.
MicroRNAs (miRNA) are small (20-25 nucleotides) evolutionarily conserved non-coding RNAs that negatively regulate transcript levels through sequence-dependent recognition mechanisms. 18 Over the past several years it has become clear that alteration in the expression of miRNA genes contribute to the pathogenesis of many human malignancies. Carcinogenesis involves multiple genetic and epigenetic events, yet the organizing principles underlying their choreography are poorly understood. miRNA deregulation is an important component of this landscape given the oncogenic and tumor-suppressive functions of miRNAs. 18 Since malignant cells have dysregulated expression of miRNAs, which in turn control or are controlled by the dysregulation of multiple protein-coding oncogenes or tumor suppressor genes, these miRNAs may be important for the development of miRNAbased therapies. 18 In addition miRNAs have great potential as diagnostic and prognostic biomarkers. Owing to their tissue specificity, miRNAs have become useful tools for defining the origin of tumors in poorly differentiated cancers. 19 With the advent of miRNA expression profiles, significant efforts have been made to correlate miRNA expression with tumor prognosis. [20][21][22][23] These reports even suggest that expression profiling of miRNAs may be a more accurate method of classifying cancer subtypes than using the expression profiles of protein-coding genes. 22,24 Various studies have established miRNA expression patterns as potential biomarkers for diagnosis, prognosis, personalized therapy, disease management and clinical outcome in various cancers. [25][26][27][28][29][30] To date, 28 645 miRNAs have been identified (http://www.mirbase.org) although their role in disease pathogenesis is not well documented.
Here for the first time we show that miR-466 inhibits prostate cancer bone metastasis. miR-466 is significantly underexpressed in a panel of prostate cancer cell lines and clinical tissues compared to normal. Restoration of miR-466 expression in highly metastatic PC3 and Du145 prostate cancer cell lines impaired proliferation, migration, invasion and induced cell cycle arrest and apoptosis in vitro. Importantly, miR-466 has a profound inhibitory effect on prostate cancer growth and bone metastasis in vivo. We further demonstrated that miR-466 partially exerted these effects by directly targeting runt-related RUNX2, a master osteogenic transcription factor and attenuated its downstream target genes that are known to mediate prostate cancer bone metastasis. In addition, our data indicates that miR-466 is potentially a clinically significant prostate cancer biomarker.
Results miR-466 expression is significantly downregulated in prostate cancer. In a previous study, 31 we performed preliminary screening to identify differentially expressed miRNAs in prostate cancer cell lines compared to a nonmalignant cell line. A set of miRNAs, miR-466, -205, -203, -23b and -34b were found to be significantly downregulated in prostate cancer cells compared to a non-malignant cell line. We validated miR-466 data by miRNA-quantitative real-time PCR (miR qRT-PCR) analysis. The results confirmed that miR-466 was significantly downregulated in a panel of prostate cancer cell lines compared to normal RWPE1 cells ( Figure 1a). Further, miR-466 expression was analyzed in an experimental cohort of laser-captured micro-dissected (LCM) matched patient samples from the San Francisco VA Medical Center (VAMC cohort) and attenuation of miR-466 expression was confirmed (Figure 1b). The expression of miR-466 was downregulated in 85% of PCa tissue samples compared to matched normal tissues (Po0.0001; Figure 1b). Suppression of miR-466 in the experimental cohort was validated by analyzing an independent cohort of prostate adenocarcinoma available publicly at The Cancer Genome Atlas (TCGA) data base (Po0.001; Figure 1c). These results indicate a putative tumor suppressor role for miR-466 in prostate cancer. We reason that tumor suppressor miRNAs would be expressed at low levels in tumors but more highly expressed in normal tissues. MiR-466 fulfills this criterion, exhibiting high expression in normal prostate tissues and a non-malignant cell line and low or silenced expression in prostate cancer samples and cell lines.  miR-466 directly represses osteogenic transcription factor RUNX2. We next sought to determine the underlying molecular mechanism of miR-466 mediated inhibition of tumor growth and bone metastasis. Thus, we utilized in silico computational algorithms to identify miR-466 target genes involved in these processes. Two different miRNA databases (microrna.org; mirdb.org) identified three complimentary miR-466 binding sites in the 3′UTR of RUNX2 (Figure 5a). Ectopic expression of miR-466 significantly attenuated RUNX2 protein levels, suggesting a functional role in controlling protein translation (Figure 5b). Luciferase reporter assays validated that miR-466 directly targets the wild-type 3′ UTR of RUNX2 as co-transfection of the miR-466 along with wild-type RUNX2 3′UTR significantly repressed relative luciferase activity (Figure 5c) in PC3 and Du145 cells. No effect was observed with cont-miR or miR-466 cells transfected with a non-specific 3′UTR control vector ( Figure 5c).   Figure 6a). RWPE1 cells transfected with anti-miR-466 showed a considerable increase in cell proliferation (Figure 6b), migration and invasion (Figures 6c-f Figure 7a). To analyze the correlation with Gleason score, we divided the samples into low o7, medium 7 and high 47 Gleason score groups. In the low group (o7), lower expression of miR-466 was found in 76% cases, whereas 24% cases had high   Figure 7a). In the PSA recurrence patients, 100% of cases had decreased miR-466 expression (Po0.0001; Figure 7a). These results reveal that number of cases with low-miR-466 expression increases from low Gleason grade, low pathological stage to high Gleason grade and high pathological stage. Interestingly, all PSA recurrence patients had low-miR-466 expression suggesting that miR-466 has clinical significance in prostate cancer.
Receiver operating curve (ROC) analyses were performed to evaluate the ability of miR-466 expression to discriminate between normal and tumor tissues. An area under the ROC curve (AUC) of 0.915 (Po0.0001; 95% CI = 0.858-0.954; Figure 7b) was obtained suggesting that miR-466 expression can discriminate between malignant and non-malignant tissues and can be used as a diagnostic marker for PCa. As 100% of patients with biochemical relapse showed low-miR-466 expression, we performed Kaplan-Meier analysis to determine the ability of miR-466 to predict patient recurrence-free survival. Indeed, high miR-466 expression was predictive of better recurrence-free patient survival (P = 0.02) compared to low-miR-466 expression (Figure 7c). Collectively, these results suggest that miR-466 has potential to be a diagnostic and prognostic marker for predicting prostate cancer biochemical recurrence though addition of more samples may strengthen these results.

Discussion
Bone metastases are a major problem in the evolution of prostate cancer. Considerable effort has been devoted to map the requirements of bone lesions in prostate tumors. 32,33 Experimental evidence has demonstrated that RUNX2, a transcription factor essential for osteogenesis, is a key regulator of bone metastasis that becomes highly activated in prostate cancer cells that metastasize to bone. 34,35 RUNX2 has been found to be associated with the osteomimetic properties of bone metastatic prostate cancer cells via transcription of genes implicated in osteoblastic lesions including osteocalcin, osteopontin, VEGF and matrix metalloproteins (MMPs) and so on. 34,35 Thus, therapeutic targeting of RUNX2 may interrupt an integrated network of gene expression that trigger bone metastases and thus deregulate cell survival and migration pathways of invading prostate cancer cells. However, at present transcription factors such as RUNX2 are not considered to be drug targets. 33 Here, we show that miR-466-mediated attenuation of RUNX2 may be a novel therapeutic approach to regulate tumor growth and bone metastasis in prostate cancer. Indeed, reconstitution of miR-466 in metastatic prostate cancer cells significantly attenuated RUNX2 protein levels, suggesting a functional role of miR-466 in controlling RUNX2 protein translation.
miRNAs have emerged as key players in cancer progression and metastatic initiation yet the importance of miRNAs in regulating prostate cancer bone metastasis has just begun to be appreciated. 36 In this study, we have identified miR-466 as a suppressor of prostate cancer growth and metastasis through direct targeting of the bone-related transcription factor RUNX2, which is difficult to target by conventional pharmacologic approaches. 33,37 Our in vivo xenograft studies provide evidence that miR-466 can inhibit primary orthotopic tumor growth and spontaneous metastasis to bone. Expression analyses based on our in-house VAMC patient cohort or the publicly available TCGA data cohort revealed that miR-466 is under-expressed in prostate cancer compared to normal tissues. Conversely RUNX2 has been detected in human prostate cancer tissues but not in normal prostate. 6,34 In this study, we have identified a tumor suppressor role for miR-466 in prostate cancer. Reconstitution of miR-466 in PC3 and Du145 metastatic prostate cancer cell lines inhibited in vitro oncogenic properties such as cell proliferation, migration and invasion compared to controls along with induction of G0/G1 cell cycle arrest and apoptosis in both cancer cell lines. MiRNAs control a wide range of biological functions and may act as tumor suppressors or oncogenes. 22 Alteration of their expression plays a critical role in tumorigenesis and cancer progression. 21,38 Our study corroborates other reports that show miRNAs are linked to the biologic activities of RUNX2 in different cancer types. 37 Recent studies have demonstrated the potential of miRNAs in the intervention of breast cancer progression, metastatic bone disease and other bone pathologies. 37,39 Interestingly, our study shows that the role of miR-466 in the control of prostate cancer progression and metastasis may be partially attributed to decreased cancer cell proliferation, cell cycle arrest and induced apoptosis, which features the beneficial effects of the miR-466 in regulating prostate cancer progression. However, our study highlights a direct role for miR-466 in reducing the bone metastatic potential of prostate cancer cells. This is evidenced by the fact that reconstitution of miR-466 caused a marked downregulation of RUNX2 target genes such as osteopontin, osteocalcin, ANGPT1, ANGPT4, MMP11 including Fyn, pAkt, FAK and vimentin that are involved in migration, invasion, angiogenesis, EMT and metastasis. [15][16][17]34,35 Thus, miR-466 orchestrates the functional activities of RUNX2 by its convergent action on RUNX2 and its downstream target gene network to control prostate cancer bone metastasis. Furthermore, our study also defines the biological relevance of miR-466 in prostate cancer as indicated by induction of pro-cancerous characteristics in non-malignant RWPE1 cells after inhibition of endogenous expression of miR-466. Thus, our data strongly suggests a miR-466 mediated attenuation of RUNX2 may be a novel approach to prevent bone metastatic disease in prostate cancer ( Figure 8). miRNAs have great potential as diagnostic and prognostic biomarkers. [20][21][22][23][24][25][26][27][28][29][30] Owing to their tissue specificity, miRNAs have become useful tools for defining the origin of tumors in poorly differentiated cancers 19 and these reports even suggest that the expression profiling of miRNAs may be a more accurate method of classifying cancer subtype than using the expression profiles of protein-coding genes. 22,24 In this study, we found that miR-466 expression can significantly (Po0.0001) distinguish malignant from normal tissues indicating the powerful diagnostic potential of miR-466 in prostate cancer. In addition, miR-466 expression was predictive of recurrence-free survival (Po0.02) such that patients with higher miR-466 expression levels had better recurrence-free survival. Correlation analyses showed that low-miR-466 expression positively correlated with high pathological T (P = 0.0001), Gleason grade (Po0.001) and PSA failure (Po0.0001). Therefore, our data reveals the clinical significance of miR-466 in prostate cancer, although it would benefit from a larger sample cohort.
In conclusion, for the first time we provide evidence that: (i) miR-466 is an under-expressed metastasis suppressor miRNA in PCa; (ii) miR-466 is biologically relevant and has biomarker potential in PCa; (iii) miR-466 directly regulates RUNX2, a key regulator of bone metastasis; (iv) miR-466 overexpression interrupts RUNX2 integrated network of genes required to maintain PCa growth and bone metastasis. RUNX2, a key transcription factor, is of pivotal significance in the pathogenesis and progression of PCa bone metastases. Therefore, restoration of miR-466 to regulate an integrated network of RUNX2 and its target genes may represent a promising approach to suppress prostate cancer bone metastases.

Materials and Methods
Cell culture, plasmids and probes/primers. Human prostate cancer cell lines PC3, Du145, LNCaP, MDaPCa2b and a non-malignant cell line (RWPE1) were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA) and grown according to ATCC protocol. These human-derived cell lines were authenticated by DNA short-tandem repeat analysis by ATCC. The experiments with cell lines were performed within 6 months of their procurement/resuscitation. Plasmids pEZX-MT01 miRNA 3′UTR target expression clones for RUNX2 and miRNA Target clone control vector for pEZX-MT01 were purchased from GeneCopoeia (Rockville, MD, USA). TaqMan probes including mimics, inhibitors and negative controls for hsa-miR-466 (miR-466) were purchased from Applied Biosystems (Foster City, CA, USA).
Tissue samples, LCM, qRT-PCR and validation of expression in an independent TCGA data cohort. Tissue samples from radical prostectomy (n = 96; 48 pairs) were obtained from the Veterans Affairs Medical Center, San Francisco, CA, USA. A board certified pathologist identified the cancer and normal regions on H&E stained prostate cancer tissue slides. LCM was performed by using AutoPix System (Arcturus) and Arcturus CapSure Macro LCM Caps (Applied Biosystems) following the manufacturer's instructions to isolate pure epithelial cells from normal and cancer areas of prostate tissues. Total RNA was extracted from micro-dissected FFPE tissues or cultured cells using a miRNeasy FFPE Kit (Qiagen, Valencia, CA, USA) or RNeasy mini kit (Qiagen), respectively, following the manufacturer's instructions. Total RNA was assayed for mature miRNAs using the TaqMan MicroRNA Assays in accordance with the manufacturer's instructions (Applied Biosystems). All RT reactions were run in a 7500 Fast Real Time PCR System (Applied Biosystems). Relative expression was calculated using comparative Ct. Expression of miR-466 was also validated by analyzing an independent prostate adenocarcinoma cohort publicly available at TCGA data portal now available as genomic data commons data portal (https://gdc-portal.nci.nih.gov/).  In vivo orthotopic and intracardiac mouse models. Bioluminescent PC3M-Luc-C6 cells (1 × 10 6 cells) constitutively expressing cont-miR or miR-466 were implanted into the posterior prostatic lobe or into the left cardiac ventricle of athymic nude mice (nu/nu; 6-8 weeks old; Harlan Lab., IN, USA) followed by regular monitoring of tumor growth and metastatic dissemination with a live animal bioimaging system (IVIS-PerkinElmer, San Jose, CA, USA). There were ten mice in each group for orthotopic experiment, whereas 8 mice per group for intracardiac implantation. Bioluminescence signal intensity was quantified using the Living Image software (IVIS-PerkinElmer). Orthotopic experiments were performed at PerkinElmer research facility at San Jose, CA, USA.
Antibodies and immunoblot assays. Immunoblotting was performed as described previously. 40 In brief, protein was isolated from 70 to 80% confluent cultured cells using RIPA Extraction Reagent (Pierce Biotechnology, Rockfield, IL, USA) following the manufacturer's directions. Equal amounts of protein were resolved on 4-20% SDS polyacrylamide gels and transferred to nitrocellulose membrane. The resulting blots were blocked with 5% non-fat dry milk and probed with antibodies. All antibodies were obtained from Cell Signalling Technology Inc. Denver, MA, USA except RUNX2 and GAPDH which were purchased from Santa Cruz Biotech. Blots were visualized using Western blotting luminal reagent (sc-2048; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA).
Luciferase assays. Complimentary miR-466 binding sites from two different miRNA databases (microrna.org; mirdb.org) are shown in Figure 5. The RUNX2 wild-type 3′UTR and non-specific scrambled sequence 3′UTR control vectors were purchased from GeneCopoeia and named RUNX2-3′UTR vector and Control-Vector, respectively. The complete sequence of vectors is available at http://www.genecopoeia.com/. For reporter assays, cells were transiently transfected with wild-type or control reporter plasmids and miR-466 or negative miR (control). Firefly luciferase activities were measured using the Dual Luciferase Assay (Promega) 18 h after transfection and the results were normalized with Renilla luciferase. Each reporter plasmid was transfected at least three times (on different days) and each sample was assayed in triplicate.
Statistical analysis. Statistical analyses were performed with GraphPad Prism 5 (La Jolla, CA, USA) and MedCalc version 10.3.2 (medcalc.org; Acacialaan, Ostend, Belgium). All quantified data represents an average of at least triplicate samples or as indicated. Error bars represent S.D.M. or as indicated. All tests were performed two tailed and P-values o0.05 were considered statistically significant. ROC were calculated to determine the potential of miR-466 to discriminate between malignant and non-malignant samples. χ 2 tests were performed to determine the correlation between miR-466 expression and the clinicopathological characteristics. For recurrence-free survival analysis Kaplan-Meier curves (log-rank tests) were also performed.