MicroRNA Signature in Melanoma: Biomarkers and Therapeutic Targets

Melanoma is the utmost fatal kind of skin neoplasms. Molecular changes occurring during the pathogenic processes of initiation and progression of melanoma are diverse and include activating mutations in BRAF and NRAS genes, hyper-activation of PI3K/AKT pathway, inactivation of p53 and alterations in CDK4/CDKN2A axis. Moreover, several miRNAs have been identified to be implicated in the biology of melanoma through modulation of expression of genes being involved in these pathways. In the current review, we provide a summary of the bulk of information about the role of miRNAs in the pathobiology of melanoma, their possible application as biomarkers and their emerging role as therapeutic targets for this kind of skin cancer.


INTRODUCTION
Arising from unrestrained proliferation of melanocytes, melanoma is the utmost fatal kind of skin neoplasm (1). Though melanoma encompasses less than 5% of all skin cancers, it accounts for most of skin neoplasms mortalities (2). When the cancer is diagnosed in early stages, surgical resection of the tumor is the appropriate therapeutic options for enhancement of survival of patients. Yet, based on the metastatic potential of melanoma, surgery is not satisfactory in advanced stages of melanoma (3). Although the mortality rate of primary melanoma is about 11%, metastatic melanoma has a poor prognosis resulting from inefficiency of conventional therapies (4,5). Meanwhile, novel therapeutic option might offer efficient methods for these patients. For instance, immunotherapeutic approaches such as administration of Anti-PD1 (nivolumab, pembrolizumab) alone, or the combination of anti-PD1 with anti-cytotoxic T lymphocyteassociated protein 4 (CTLA4) ipilimumab has raised the survival of patients who suffer from advanced stages of melanoma (6).
Targeted therapies, like combinations of BRAF inhibitors (Dabrafenib) and MEK inhibitors (vemurafenib) are also frequently used on BRAFV600E mutant melanomas. Superficial spreading, nodular, lentigo maligna and acral lentiginous melanomas represent the main types of melanoma with the first one being the most frequent type (4). Ultraviolet radiation and melanocytic nevi are two main risk factors for development of this kind of skin cancer (4). Molecular changes occurring during the pathogenic processes of initiation and progression of melanoma are diverse and include activating mutations in BRAF and NRAS genes, hyper-activation of PI3K/AKT pathway, inactivation of p53 and alterations in CDK4/CDKN2A axis (4). In addition, several studies have shown the critical role of microRNAs (miRNAs) both in the initiation and in the progression of melanoma (7). These transcripts have sizes around 22 nucleotides and are generated through a multi-step process from DNA sequences into primary, precursor and mature miRNAs, respectively. As a general rule, they regulate gene expression through binding with complementary sequences in the 3′ untranslated region (3′ UTR) of mRNAs and subsequently lead to degradation and suppression of translation of the target transcript. Less frequently, they interact with the 5′ UTR, coding or promoter regions (8). Moreover, there are some reports of activation of translation of certain genes by miRNAs in some situations. For instance, let-7 family of miRNAs can induce translation when cell cycle is arrested in spite of their inhibitory effects on translation during cell proliferation (9). Therefore, miRNAs are regarded as important mediators of gene expression. Besides, their presence in extracellular vesicles provides them the opportunity to module communication between various cells (8). In the current paper, we summarize the bulk of information about the role of miRNAs in the pathobiology of melanoma, their possible application as biomarkers and their emerging role as therapeutic targets for this kind of skin cancer.

DYSREGULATED MIRNAS IN MELANOMA
Expression pattern of miRNAs in melanoma cell lines and clinical specimens has been assessed by both high throughput and candidate gene approaches. An example of the former types of studies is the study conducted by Zhang et al. (10). They reported DNA copy number changes in miRNA coding genes in the majority of the assessed melanoma samples. Notably, miRNA copy alterations have been correlated with miRNA expression. Moreover, they reported copy number alterations in genes contributing in the biogenesis or function of miRNAs in tumor samples (10). Through a microarray-based technique, Aksenenko et al. have identified differential expression of 143 miRNAs between melanoma samples and adjacent skin tissues. Among the dysregulated miRNAs has been the up-regulated miRNA hsa-miR-146a-5p which has been predicted to be associated with Toll-like receptor, NF-kB and ErB pathways. Moreover, this miRNA has been shown to target one of the most recurrently mutated genes in melanoma i.e., the NRAS gene (11).
miRNA also affect activity of melanoma-related signaling pathways. Figure 1 depicts the functional association between two miRNAs and AKT and NF-kB signaling pathways.
Expression profiling of miRNAs in melanocytes and melanoma cells originated from primary or metastatic melanoma cells has provided valuable data about the role of miRNAs in each phase of cancer development. A panel of miRNAs including miR-133a, miR-199b, miR-453, miR-520f, miR-521, and miR-551b has been found consistently upregulated in the course of cancer development from melanocytes to primary cancerous cell and from primary to metastatic melanomas. On the other hand, miR-190 had the opposite trend during this course. Furthermore, expressions of miR-126, miR-29c, miR-506, miR-507, and miR-520d* have been found to be increased during the early progression of melanoma and have been decreased in the metastatic phase. Two other miRNAs including miR-489 and miR-527 had the opposite pattern of expression (15). Levati et al. have demonstrated up-regulation of miR-17-5p, miR-18a, miR-20a, and miR-92a while down-regulation of miR-146a, miR-146b and miR-155 in most of assessed melanoma cell lines compared with melanocytes (16).
Other studies have reported dysregulation of several other miRNAs in the melanoma samples. Among up-regulated miRNAs are miR-221 and miR-222 which induce malignant features through decreasing expression of c-KIT receptor and p27Kip. Both miRNAs promote epithelial-mesenchymal transition (17,18). Moreover, expression of miR-210 has been demonstrated to be elevated in several cancer types including melanoma. Its expression has been correlated with metastatic potential of melanoma tumors. Up-regulation of miR-210 in cancer cell lines facilitates evasion from hypoxia-induced cell cycle arrest and partly upturned the hypoxic gene expression profile. This miRNA has been revealed to target a known MYC antagonist namely MNT. Therefore, miR-210 has been shown to modulate the hypoxia response in cancer cells via regulating an important transcriptional suppressor of the MYC-MAX axis (19). In an attempt to detect the miRNAs that are regulated by BRAFV600E mutation via the ERK pathway, Vitiello et al. have conducted RNA sequencing on A375 cell line and a vemurafenib-resistant clone. Their experiments have led to identification of miR-204 and miR-211 as the utmost overexpressed miRNAs by vemurafenib. In spite of belonging to an identical miRNA family, miR-204 and miR-211 have distinguishing characteristics. miR-204 is regulated by STAT3 and its transcript levels are increased in amelanotic melanoma cells, where it functions as a mediator of antimigratory effects of vemurafenib by modulating expression of AP1S2. On the contrary, miR-211, as a direct target of MITF, is over-expressed in melanotic melanoma cells. miR-211 regulates expression of EDEM1 and subsequently weakens the destruction of Tyrosinase. Thus, miR-211 is a facilitator of pro-pigmentation function of vemurafenib (20). Table 1 displays the list of over-expressed miRNAs in melanoma.
Numerous tumor suppressor miRNAs have been downregulated in melanoma samples. For instance, while miR-34a is constantly detected in normal melanocytes, it is not expressed in uveal melanoma cells. Forced over-expression of this miRNA in uveal melanoma cells remarkably diminishes their growth and migratory abilities. Mechanistically, this miRNA inhibits expression of c-Met protein and decreases the levels of phosphorylated Akt and cell cycle-related proteins (83). Besides, miR-34b, miR-34c, and miR-199a* have been shown to down-regulate MET expression, suppressing the invasive growth features in the melanoma cells (84). Furthermore, expressions of the let-7 miRNAs have been shown to be decreased in primary melanomas compared with benign nevus samples. Forced up-regulation of let-7b in melanoma cells has led to significant decrease in the expression of cyclins D1, D3, and A, and CDK4. The functional interaction between let-7b and cyclin D1 has been verified through in vitro experiments (85). The inhibitory effect of let-7a on expression of integrin beta 3 has been verified in another study (86). In addition, functional studies have shown the role of miR-155 in the suppression of proliferation of a number of melanoma cell lines and induction of apoptosis in these cells (16). Table 2 lists the down-regulated miRNAs in melanoma.

DIAGNOSTIC/PROGNOSTIC MIRNAS IN MELANOMA
Hanniford et al. have introduced a miRNA panel consisting of miR-150-5p, miR-15b-5p, miR-16-5p, and miR-374b-3p whose expression levels could predict the possibility of brain metastasis of melanoma tumors along with clinical stage. Moreover, Kaplan-Meier analysis showed the significance of this miRNA panel in determination of brain-metastasis-free and overall survival of patients with melanoma (273). Stark et al. have assessed expression levels of 17 miRNAs in both melanoma tissues and serum samples of these patients compared with cancer-free individuals. Expression levels of these miRNAs in melanoma samples have been shown to predict stage, recurrence, and survival of patients. Notably, serum expression of a seven-miRNA panel could distinguish melanoma patients from control subjects with 93% sensitivity and more than 82% specificity if at least 4 miRNAs were expressed. Based on the superiority of this miRNA panel above the conventional serological biomarkers for melanoma, it has been suggested as a tool for monitoring disease course in early metastatic melanoma cases to identify relapse after tumor excision or adjuvant therapy (23). Worley et al. have used a high throughput technique to identify the miRNAs whose expression profile could predict the metastatic potential of uveal melanomas. Their approach led to identification of let-7b and miR-199a as the most robust discriminators. Notably, expression profile of six miRNAs could differentiate low and high risk groups with optimal sensitivity and specificity values (274). Table 3 shows the role of miRNAs in the prediction of prognosis of melanoma using Kaplan-Meier or Cox regression analyses.     Receiver operating characteristic (ROC) curves have been used to assess the diagnostic or prognostic values of miRNAs in melanoma. Based on the area under curve (AUC) values, several miRNAs can be suggested as appropriate biomarkers for this kind of cancer. In the field of miRNA application in melanoma diagnosis, these curves depict the diagnostic capability of expression level of a miRNA as a binary classifier system for detection of melanoma cases as its discrimination threshold is            Reduces proliferation, migration and invasion and induces apoptosis in (199) (Continued)     changed. In other words, these curves are generated by plotting the true positive rate against the false positive rate at different threshold points. Notably, serum expression levels of several miRNAs have high sensitivity and specificity values for differentiating between melanoma patients and healthy subjects or between metastatic and non-metastatic melanomas. Tables 4, 5 list the miRNAs whose application as diagnostic or prognostic markers has been evaluated using ROC curve analysis, respectively.  It can be an independent prognostic factor for metastasis. (175) miR-191 32 lymph node metastases Its low expression associated with poor melanoma-specific survival. -- miR-193b 32 lymph node metastases Its high associated with poor melanoma-specific survival. -- hsa-miR-211-5p hsa-miR-514a-3p hsa-miR-508-3p hsa-miR-509-3-5p hsa-miR-513c-5p UM dataset of miRNA expression profiles was obtained from the UCSC Xena Browser Their high expression were associated with poor OS. -- (Continued) Its serum level is predictor of patients OS.
Its serum level is an independent prognostic biomarker for OS.
miR-23b 114 primary melanoma tissues and ANTs Its low expression is associated with short 3-year survival in melanoma patients. -- miR-216a-5p 86 uveal melanoma tissues Its low expression is associated with poor DFS and OS. --

IMPLICATIONS OF MIRNAS IN THE TREATMENT OF MELANOMA
miRNAs are implicated in the therapeutic effects of several anticancer agents. For instance, Genistein, the isoflavone extracted from soybean, has been shown to suppress proliferation of human uveal melanoma cells possibly through modulating expression of miR-27a and its target gene ZBTB10 (291). miRNAs are also involved in conferring resistance to immunotherapeutic modalities. For instance, expression of miR-222 has been shown to be higher in melanoma samples obtained from patients who did not respond to ipilimumab compared with those benefitting from this option (292). Mechanistically, the ADAR1/miR-222/ICAM1 axis has been reported to be involved in this process (292). Other miRNAs such as miR-488-3p, miR-195 and miR-211 participate in the regulation of response to the chemotherapeutic agent cisplatin (130,162,163) Application of miRNAs in the therapeutic settings is limited by target specificity issues (293). However, some miRNAs are currently being tested in some diseases. Among these therapeutic modalities are miR-122/miravirsen and miR-92/MRG 110 which have been manufactured by Roche/Santaris and Regulus Therapeutics, respectively (293).

ASSOCIATION BETWEEN POLYMORPHISMS WITHIN MIRNAS AND RISK OF MELANOMA
Theoretically, polymorphisms with miRNA coding genes can alter their expression or function. Although such polymorphisms are predicted to influence the risk of different cancers such as melanoma, this field has not been vastly explored. Few studies have assessed association between a certain polymorphism within miR-146a namely the rs2910164 G/C and melanoma risk. In spite of the proposed role for allele C of this polymorphism in conferring risk of melanoma (294,295), cell line studies have shown that G allele confers high proliferative capacity to melanoma cells (296). Table 6 summarizes the results of these studies.

DISCUSSION
Dysregulation of miRNAs in melanoma samples and cell line have been reported by several studies. The functional consequences of such dysregulation on cell behavior have also been appraised. However, the underlying mechanism of such dysregulation is not clarified completely. Copy number variations in miRNA-coding genes or genes associated with the biogenesis or function of miRNAs may be responsible for the observed dysregulation of miRNAs in melanoma and other types of cancers (10). Moreover, the role of epigenetic factors in this process should not be ignored. For instance, CpG methylation of the miR-34a promoter has been suggested as an underlying mechanism for down-regulation of this miRNA in primary melanoma samples and melanoma cell lines (297). Another possible mediators of miRNA dysregulation in the melanoma are melanoma-inducing transcription factors such as MITF whose role in the expression of a number of miRNAs has been verified (298). As several miRNAs are implicated in the modulation of skin response to ultraviolet radiation (299), this environmental carcinogen might also affect expression of miRNAs which are involved in the melanomagenesis. Mechanistically, several melanoma-associated miRNAs function upstream or downstream of known oncogenes in melanoma. For instance, miR-137 and miR-182 are among miRNAs that target MITF oncogene (54,300). Moreover, expressions of several miRNAs such as a number of let-7 family members, miR-221/222, miR-17-92 and miR-106-363 clusters, miR-29, miR-146a, miR-148b, and miR-125b have been shown to be modulated by MITF (298). Moreover, several miRNAs such as miR-7, miR-23a and miR-596 have functional interactions with MAPK/ERK and PI3K/PTEN/Akt signaling pathways in the context of melanoma. A number of miRNAs such as miR-378, miR-10b, miR-25, miR-485-5p, miR-708, miR-136, miR-488-5p, miR-29a, miR-22 and miR-140-5p have interactions with Wnt/b-catenin pathway. Finally, miR-21, miR-7-5p, miR-23b, miR-145-5p, miR-9, miR-29a, miR-377 and miR-140-5p interacts with NF-kB signaling in the context of melanoma development. Thus, a number of miRNAs provide functional links between cancer-related pathways in this context. miRNAs have functions both in the paternal cell in which they are produced as well as in the adjoining cells. These transcripts can modulate characteristics of adjacent melanoma cells or directly affect tumor niche by modifying extracellular matrix and function of resident cells in this environment including fibroblasts and endothelial or immune cells. This activity of miRNAs potentiates them as contributors of melanoma metastatic potential through affecting intravasation of cancer cells into vessels, viability of tumor cells in the circulation, their leakage in the target tissues, and establishment of the pre-metastatic milieu in remote organs (301).
Several miRNAs have been shown to differentiate melanoma patients from healthy subjects or distinguish between metastatic and non-metastatic melanoma patients. The prognostic assays founded on miRNAs signature can enhance the efficacy of conventional staging systems in predicting patients' prognosis and their management in the clinical settings in the terms of choosing adjuvant therapies or clinical trial enrolment. Therefore, these miRNAs are potential biomarkers for this kind of skin cancer.
Numerous miRNAs have been dysregulated in tumor samples or peripheral blood of patients with melanoma. Such dysregulation can be used as biomarker for early detection of melanoma or follow-up of patients after initial treatments to uncover any possible tumor recurrence. Blood-based biomarkers are expected to substitute invasive methods of cancer diagnosis  (275) in future. Based on the heterogeneous pattern of miRNAs expression in tumor samples and the varied expressions among affected individuals, multi-miRNA panels are more promising in the diagnostic approaches compared with individual miRNAs. Finally, miRNAs might be implicated in the anti-cancer effects of a number of therapeutic agents including both chemical and herbal medicines. Evidence for supporting this idea has come from several studies including a study which revealed the role of miR-27a in mediating the anti-proliferative effects of Genistein in human uveal melanoma cells (291). Moreover, the observed upregulation of miR-222 in melanoma samples obtained from patients who did not respond to ipilimumab compared with those benefitting from this option (292) implies its contribution in resistance to this agent. Therefore, miRNAs are promising targets for modulation of response of melanoma cells to a wide range of therapeutic options.

PERSPECTIVES AND FUTURE DIRECTIONS
Assessment of expression pattern of miRNAs in cohorts of melanoma patients from different ethnicities and uncovering their association with genetic polymorphisms would facilitate design of prognostic/diagnostic panels. The relationship between aberrant miRNA profile and response to therapeutic regimens should be unraveled. Such kinds of approaches pave the way for design of personalized methods of treatment of melanoma. Therapeutic targeting of miRNAs can influence melanoma course and enhance sensitivity to both conventional therapies and immunotherapeutic approaches. Yet, safety and bioavailability issues remained to be solved before implementation of these techniques in the clinical settings.