MicroRNA-138 negatively regulates the hypoxia-inducible factor 1α to suppress melanoma growth and metastasis

ABSTRACT Melanoma with rapid progression towards metastasis has become the deadliest form of skin cancer. However, the mechanism of melanoma growth and metastasis is still unclear. Here, we found that miRNA-138 was lowly expressed and hypoxia-inducible factor 1α (HIF1α) was highly expressed in patients’ melanoma tissue compared with the paracancerous tissues, and they had a significant negative correlation (r=−0.877, P<0.001). Patients with miRNA-138low/HIF1αhigh signatures were predominant in late stage III/IV of melanoma. Further, bioinformatic analysis demonstrated that miRNA-138 directly targeted HIF1α. We found that the introduction of pre-miRNA-138 sequences to A375 cells reduced HIF1α mRNA expression and suppressed cell proliferation, migration and invasion. Overexpression of miRNA-138 or inhibition of HIF1α significantly suppressed the growth and metastasis of melanoma in vivo. Our study demonstrates the role and clinical relevance of miRNA-138 and HIF1α in melanoma cell growth and metastasis, providing a novel therapeutic target for suppression of melanoma growth and metastasis.


INTRODUCTION
Melanoma, the most common and aggressive form of skin cancer, is highly heterogeneous and can progress rapidly to metastatic disease (Gupta et al., 2005;Luo et al., 2016). However, the mechanism and effective therapies for melanoma remain elusive. Recently, many studies have pointed toward microRNAs (miRNAs) as playing key roles in the development, metastasis and prognosis of melanoma (Deng et al., 2016;Hanniford et al., 2015). miRNAs, 18-25 nucleotides in length non-coding RNAs, act as negative regulators of post-transcriptional gene regulation through directly binding to the 3′-untranslational region (UTR) of their target mRNAs, and promoting translational suppression or their degradation, finally modulating the expression and function of those genes.
In melanoma, HIF1α is well substantiated to promote tumorigenesis (Loftus et al., 2017). Elevated expression of HIF1α correlates with a cell phenotypic transition from proliferation to metastasis (O'Connell et al., 2013;Widmer et al., 2013). Additionally, mouse melanoma tumors exposed to hypoxia show increased tumor growth and metastatic potential (Cheli et al., 2012). Although some studies have shown that miRNA-138 negatively regulates HIF1α to affect the proliferation and metastasis of melanoma  and ovarian cancer (Yeh et al., 2013), it is not clear whether the relationship in human melanoma tissues is established and the therapy for targeting miRNA-138 and HIF1α is effective.
Here, we found that the expression of miRNA-138 was significantly downregulated in both human melanoma tissues (HMTs) and melanoma cell lines. HIF1α has been identified as the direct downstream target of miRNA-138 in melanoma. Further studies showed that exogenous overexpression of miRNA-138 or knockdown HIF1α inhibits melanoma cell growth and metastasis, both in vitro and in vivo. Thus, miRNA-138/HIF1α is a potential target for the clinical diagnosis and treatment of melanoma.

miRNA-138 expression was downregulated in HMTs
Firstly, we quantitatively detected the expression of miRNA-138 and a variety of its related downstream proteins including FOXC1, SOX4, HIF1α, CDK6, E2F2 and E2F3 in HMTs and paracancerous tissues (PTs). The results showed that miRNA-138 was significantly downregulated in HMTs compared with PTs (Fig. 1A). Conversely, HIF1α had higher level than other downstream proteins in HMTs compared with PTs (Fig. 1B). Then, we analyzed the correlation between miRNA-138 and HIF1α using Pearson's correlation coefficient. The results showed that the miRNA-138 level was negatively correlated with the HIF1α mRNA level (Fig. 1C). We further examined the difference of the miRNA-138 and HIF1α expression at different stages of melanoma. The late stages of melanoma tissues had further decreased miRNA-138 and increased HIF1α expression (Fig. 1D). These results suggest that miRNA-138 The above results motivated us to observe whether miRNA-138 is involved in the growth and metastasis of melanoma. We designed loss-of-function and gain-of-function experiments to investigate the role of miRNA-138 in melanoma cells. We firstly observed that primary WM35 cells had low proliferation ( Fig. 2A), migration ( Fig. 2C) and invasion (Fig. 2D). However, proliferation, migration and invasion of miRNA-138 knockdown WM35 cells were increased. Metastatic A375 cells had high proliferation (Fig. 2B), migration ( Fig. 2E) and invasion (Fig. 2F). However, the proliferation, migration and invasion of miRNA-138 overexpressing A375 cells were decreased. Those results suggested that miRNA-138 regulates the proliferation and invasion of melanoma cells.
Then, we further observed the effect of miRNA-138 on apoptosis of melanoma cells. As shown in Fig. 2G, cell apoptosis increased significantly in miRNA-138 overexpressing A375 cells compared with A375 cells. Therefore, miRNA-138 functioned as a suppressor of melanoma occurrence and development. miRNA-138 played a regulatory role for the epithelial-tomesenchymal transition (EMT) , which has an increased incidence of tumor metastasis (Qin et al., 2017). Thus, the effect of miRNA-138 on EMT was further examined. We found that miRNA-138 promoted E-cadherin expression and inhibited Vimentin expression (Fig. 3H left), suggesting that the process of EMT was inhibited by miRNA-138. Importantly, siRNA HIF1α further promoted the effect of miRNA-138 (Fig. 3H, right). These results indicate that functions of HIF1α and miRNA-138 were antagonistic in regulating proliferation, invasion and metastasis of melanoma cells.
Subsequently, the changes of HIF1α mRNA levels under miRNA-138 knockdown or overexpression conditions were undertaken to analyze by quantitative PCR. We found that miRNA-138 significantly reduced HIF1α mRNA levels (Fig. 4C,D). These data indicate that miRNA-138 directly suppresses HIF1α translation, resulting in the decrease of its mRNA levels.

miRNA-138 and HIF1α inhibitor synergistically suppressed melanoma growth and metastasis
To further observe the therapeutic effects of miRNA-138 and HIF1α in vivo, we used miRNA-138 knockdown and overexpression cells to inoculate nude mice. The results showed that miRNA-138 knockdown WM35 tumors grew faster than primary WM35 tumors (Fig. 5A). Conversely, miRNA-138 overexpressing A375 tumors grew very slowly (Fig. 5B). Additionally, inhibition of HIF1α through oral administration of BAY 87-2243 also significantly inhibited the tumor growth (Fig. 5B) and metastasis (Fig. 5C,D). To our surprise, the combination therapy of overexpressing miRNA-138 and inhibition of HIF1α eradicated tumor growth and largely inhibited tumor metastasis, suggesting that the targeted therapy via miRNA-138 and HIF1α was a novel approach to melanoma treatment. The body weight of mice under the above treatment had no significant change (Fig. 5E), suggesting that the use of drugs in mice is reasonable.

DISCUSSION
miRNAs play critical roles in tumor development and prognosis (Deng et al., 2016). The function of miRNA-138 has been extensively studied in other tumors (Jiang et al., 2016;Yeh et al., 2013;Yu et al., 2015), but its role in melanoma tissue remains unclear. Herein, we identified and verified miRNA-138 as an effective tumor suppressor on cell migration, invasion and metastasis in melanoma. Our results also show that the molecular mechanism underlying miRNA-138 action in melanoma inhibits cell proliferation and metastasis by directly targeting degradation of HIF1α.
Tumor cells can adapt and survive under hypoxic conditions and this may be caused by HIF1-induced numerous target genes (Mineo et al., 2016), which mediate coagulation, angiogenesis and metabolism to provide more nutrients and oxygen for tumor tissues (Ebersole et al., 2018). Studies have showed that HIF1α, a subunit of HIF1, can directly promote the expression of EMT-related genes (Jiang et al., 2011;Yang et al., 2008), enhancing tumor cell proliferation and metastasis. Indeed, HIF1α was highly expressed in melanoma tissues (Fig. 1B,D). Decreased HIF1α activity by RNA interference significantly inhibited E-cadherin expression and increased Vimentin expression, and reversed the process of EMT (Fig. 3H), suggesting that HIF1α may be a new target to treat melanoma metastasis. Other studies show that HIF1α activates EMT regulators including SIP1, ZEB1 SLUG, SNAIL or TWIST to promote of cancer cell metastasis (Huang et al., 2009;Lin et al., 2016;Yang et al., 2008;Yoo et al., 2011). In melanoma, the downstream signal regulation mechanism of HIF1α is still under further study by ourselves.
Current therapeutic interventions for highly metastatic melanoma are not satisfactory. Thus, efforts to understand the molecular mechanisms underlying the metastatic process of melanoma are of highly significant clinical importance. Our data from the tumor tissues of melanoma patients from different stages show that the expression of miRNA-138 decreases progressively with the enhancement of metastatic potential and the corresponding HIF1α level increases gradually. Therefore, we have revealed that the functions of miRNA-138 and HIF1α and their correlation are important for the clinical treatment of melanoma. In our study, the combination therapy of overexpressing miRNA-138 and inhibition of HIF1α in vivo showed eradicated tumor growth and metastasis (Fig. 5B,C). These results reveal the mechanism of melanoma invasion and metastasis, and show that combination therapy is effective for treatment of melanoma. These results will motivate us to explore combinations with other immunotherapies (Chang et al., 2017). To summarize, we identified miRNA-138 as a regulator for controlling melanoma cell proliferation, invasion and metastasis, and revealed a novel mechanism by which miRNA-138 negatively regulates the HIF1α to suppress melanoma growth and metastasis. Upregulation of the miRNA-138 may serve as a new therapeutic target for patients with melanoma.

Tissue collection
Human HMTs (n=16) were collected following surgical resection of cutaneous melanomas at the Second Affiliated Hospital of South China University of Technology from September 2017 to July 2018, following a protocol approved by the hospital's ethics committee. This cohort included nine male and seven female patients. The mean age was 59.6 years (range: 32-69 years). Tumor stages were determined according to the consensus of the European Thyroid Association (ETA) risk stratification system and the tumor, node, metastasis (TNM) staging system, respectively. Tissue specimens are diagnosed as cutaneous melanomas by histopathology and cytology.

Cell lines and reagents
WM35 and A375 melanoma cells and HEK293 cells were purchased from the Cell Bank of Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China) and authenticated and tested for contamination. They were cultured in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum (Sijiqing, Hangzhou, China) at 37°C in a humidified atmosphere containing 5% CO 2 .

Lentivirus production and transfection
pre-miRNA-138 sequence and anti-miRNA-138 sequence, and their nonspecific control, were synthesized by GenePharma (Shanghai, China). Those sequences were inserted into the lentivirus vector pLL3.7 (Sigma-Aldrich) to generate a miRNA-138 overexpressing vector (LV-miRNA-138 overexpression) and its non-specific control vector (LV-NC), and a miRNA-138 knockdown vector (LV-miRNA-138 knockdown) and its non-specific control vector (LV-NC knockdown). The miRNA-138 overexpressing stable cells were established by infecting A375 cells with lentiviruses encoding LV-miRNA-138 overexpression and LV-NC followed by selection in 2 mg/ml puromycin. Similarly, the miRNA-138 knockdown stable cells were established by infecting WM35 cells with lentiviruses encoding LV-miRNA-138 knockdown and LV-NC knockdown followed by selection in 2 mg/ml puromycin. The overexpression and knockdown efficiency of miRNA-138 was detected by qPCR.

Cell migration/invasion assay
Migration of A375 cells was assessed using Transwell chamber (Costar, Pleasanta, CA, USA) containing a polycarbonated filter with a pore size of 8 µM. A375 (3×10 4 cells in 200 μl) cells in complete medium were mixed and the cell suspension was added to the upper chamber. The cells were allowed to migrate at 37°C incubation for 24 h and cells remaining on the topside of the transwell membrane were manually removed with a cotton swab. The membrane was washed with PBS. Cells that had migrated to the underside were fixed with 95% ethanol and stained with Crystal Violet for counting under a light microscope. Cell migration was quantified in five random fields for each chamber.

Animal study
To explore the therapeutic effect of targeting miRNA-138, female 6-to 7-week-old mice were subcutaneously (SC) challenged with 3×10 5 LV-NC or LV-miRNA-138 overexpressing A375 cells or 3×10 5 LV-NC or LV-miRNA-138 knockdown WM35 cells. Tumor growth and survival of mice were monitored every day. To explore the therapeutic effect of targeting HIF1α, female 6-to 7-week-old mice were SC challenged with 3×10 5 melanoma cells. BAY 87-2243 were orally administered as a 2-week continuous infusion at 9 mg/kg/day (Schöckel et al., 2015). Mice were monitored for tumor growth every day.

Statistical analysis
Quantitative data were presented as mean±s.e.m. All experiments were performed at least in triplicate. Differences between groups were performed by the SPSS12.0 software using two-tailed Student's t-test or one-way analysis of variance (ANOVA) for experiments with more than two groups. Survival analysis was made with the Kaplan-Meier method. Correlation analysis was performed by using two-tailed Pearson's correlation coefficient. Significant differences were determined at P<0.05.