Genomics and Epigenomics in the Molecular Biology of Melanoma—A Prerequisite for Biomarkers Studies

Melanoma is a common and aggressive tumor originating from melanocytes. The increasing incidence of cutaneous melanoma in recent last decades highlights the need for predictive biomarkers studies. Melanoma development is a complex process, involving the interplay of genetic, epigenetic, and environmental factors. Genetic aberrations include BRAF, NRAS, NF1, MAP2K1/MAP2K2, KIT, GNAQ, GNA11, CDKN2A, TERT mutations, and translocations of kinases. Epigenetic alterations involve microRNAs, non-coding RNAs, histones modifications, and abnormal DNA methylations. Genetic aberrations and epigenetic marks are important as biomarkers for the diagnosis, prognosis, and prediction of disease recurrence, and for therapeutic targets. This review summarizes our current knowledge of the genomic and epigenetic changes in melanoma and discusses the latest scientific information.


Introduction
Melanoma is a malignant tumor originating from melanocytes, the cells specialized in producing the melanin pigment. Melanocytes emerge from the neural crest, an embryonic structure consisting of migratory pluripotent cells from which several different cell types originate [1]. In the maturation process, melanocyte progenitors migrate, differentiate, and colonize the skin, hair follicles, uvea, and mucous membranes throughout the body. Accordingly, melanoma can arise in any of these sites, leading to a genetically, histologically, phenotypically, and clinically heterogeneous disease.
In Caucasian populations, the most common type of melanoma is cutaneous melanoma (CM). In recent decades, a continuous increase of CM frequency rates has been observed in Caucasian populations worldwide, making CM the cancer with the most rapidly increasing occurrence [2]. Melanoma is dangerous because it poses a greater risk for metastasizing than other skin cancers. [3].
Cutaneous melanoma comprises four main subtypes: superficial spreading melanoma, lentigo melanoma, acral lentiginous melanoma, and nodular melanoma. Superficial spreading melanoma accounts for about 70% of melanoma cases. It is more frequent in fairerskinned patients, the locations are the trunk and legs, and the excessive UV exposure is a risk factor [4]. Lentigo melanoma targets older patients with sun-damaged skin and the affected skin areas are the face, ears, and neck [5].
Acral lentiginous melanoma is typically found in individuals with darker skin. Clinical presentation is a dark spot on the sole or palms. Subungual melanoma represents a

Genomics in Melanoma and Implications for Biomarkers
Genetic aberrations, concurrently with epigenetic characteristics and the tumor microenvironment (TME), lead to tumor progression that eventually results in uncontrolled cell proliferation, an escape of immune annihilation, and induces a metastatic potential. Discovery of the genomic modifications of melanocytic tumors has remarkably enhanced our understanding of pathogenesis. Thus, we established correlations between the genetic abnormalities and the clinical and pathologic characteristics. The uncontrolled cell proliferation of melanocytes depends on the mutations in some oncogenes, such as MC1R, CDK4, BRAF, CCND1, RAS, NRAS, c-KIT, GNAQ, and GNA11 or in tumor suppressor genes, such as, TP53, BCORL1, PPP2R3B, RASA2, PTEN, and CDKN2A [16]. The main signalling pathways of melanoma are: MAPK (mitogen-activated protein kinase), AKT (protein kinase B) pathway (PI3K/PTEN/AKT), cell-cycle regulation pathway, pigmentation-related pathway (MITF signalling pathways), and p53 pathway [17].
The progression to melanoma involves a minimum of three key mutations, but the number of genetic abnormalities is high in melanoma. Sometimes, the occurrence of mutations is so fast that a precursor lesion exists only for a short time, which is not enough to detect them [18][19][20][21].
The MAPK pathway plays a major role in the regulation of cell proliferation, cell survival, invasion, the angiogenesis process, and metastasis. Mutant Ras proteins play a role in oncogenesis by activation of the downstream cascade without the stimulation of the respective upstream pathway. Furthermore, it has been suggested that the activation of BRAF, NRAS, and PI3K can occur in various stages of melanoma development [22].
Up to 30% of melanomas are linked to melanocytic precursor lesions, including nevi or intermediate melanocytic tumors [23,24]. It is believed that the initial mutations affect the MAPK pathway, leading to the formation of nevi. When no other genomic aberration is present, melanocytes with a MAPK pathway-activating mutation allow for an oncogeneinduced senescence. If benign or intermediate melanocytic tumors develop into further genetic aberrations in cancer-related genes, they are liable to progress and acquire metastatic prospects [25].
The genes from MAPK pathways involved in melanoma pathogenesis are summarized in Table 1.

BRAF
BRAF is a RAF gene. The connection in the signaling pathway is with the small G protein RAS and through MEK1/2 activate ERK1/ERK2 [3].
BRAF mutations represent a frequent event in oncogenesis, mainly in melanoma [28]. The most common mutation (75%) is V600E, when the valine is replaced with glutamic acid in position 600.
BRAF mutations appear in the first stages of melanocytic tumors [29], whereas individual studies indicated that BRAF mutations have an unfavorable prognosis in patients with melanoma [30,31], which has not been definitively shown in other studies [32].
BRAF mutation is implicated in melanoma progression, sustained angiogenesis, tissue invasion, and metastasis, as well as the evasion of the immune response. Multiple studies assessed the prognostic importance of BRAF mutations but their role remains controversial. Most studies found a positive association between BRAF mutation and poor clinical outcome. The analysis performed in the phase III Keynote 054 trial showed that there are differences between the relapse-free survival (RFS) in patients with BRAF mutations compared to those without BRAF mutations only in the group treated with a placebo. There is no difference in patients treated with pembrolizumab [33]. Other studies suggested an opposite correlation between BRAF mutation and patient outcome prediction. For example, Tas and Ertuk investigated the prognostic value of BRAF V600E mutation in 151 stage III patients. The mutation was present in 51% of patients and was associated with better overall survival (OS) and longer disease-free survival [34].
A recent study indicates that extracellular vesicles (EVs) could be a promising source of mutant DNA for BRAF mutation status for evaluating BRAF therapy [35].

NRAS
NRAS is a member of the RAS family of oncogenes and was originally found in oncogenic viruses [36]. NRAS was found to be mutated in melanoma cell lines [37]. NRAS mutations emerge in about 20% of melanomas and are evenly spread among cutaneous, acral, and mucosal melanomas [20,21]. In opposition, variants in HRAS (HRas protooncogene, GTPase) and KRAS (KRAS proto-oncogene, GTPase) are less common: 2% and 3%, respectively, of all melanoma [26]. These mutations are recurrent and they are mutually exclusive with other RAS gene mutations [38].
Mutations in RAS genes generally concern codons 12, 13, and 61. Whereas most aberrations in NRAS involve the glutamine on position 61 (Q61), mutations in KRAS and HRAS are usually found at the glycine 12 and 13 (G12, G13). All three hotspot mutations (Q61, G12, and G13) contribute to GTPase inactivation, resulting in a constitutively functional GTP-bound protein. Activation of the RAS protein signals is generated via multiple oncogenic downstream paths, including activation of the MAPK signaling pathway by the RAS/RAF/MEK/ERK signaling cascade and activation of the PI3K/AKT pathway through PI3K (phosphatidylinositol 3OH-kinase) phosphorylation [37].

NF1
NF1 (neurofibromin 1) is a tumor suppressor gene [39]. It intervenes as an oncogene in different cancers and it was seen to be non-functional in multiple human malignancies, including lung cancers, neuroblastomas, and glioblastomas [40,41]. NF1 encodes the protein neurofibromin, which negatively regulates RAS by hydrolysis of RAS-bound GTP to GDP [42]. Functional inactivation of NF1 contributes to the activation of RAS and its downstream signaling pathways, including the MAPK, PI3K/AKT, and mTOR pathways [43].
It has been observed that NF1 inactivating mutations occur in melanomas without BRAF and NRAS mutations and cause the activation of the MAPK pathway. NF1 mutations are typically inactivating, often truncating mutations or losses, and there are no mutation hotspots. Usually, melanomas occurring in heavily sun-damaged skin or desmoplastic melanomas show a significantly higher rate of NF1 mutations [44][45][46]. NF1 mutations are more frequent in acral and mucosal melanomas [47,48].
Studies have pointed out that NF1 could be associated with resistance to BRAF and MEK inhibitors [49,50]. Nevertheless, NF1-mutated melanomas have been associated with tumors with a high tumor mutation burden (TMB) that respond well to immunotherapy [51,52]. These discoveries demonstrate that the NF1 mutational status could be important for therapeutic decision making.

MAP2K1/MAP2K2 (MEK1/MEK2)
MAP2K1 and MAP2K2 encode the protein kinases MEK1 and MEK2. They function specifically in the MAPK/ERK cascade and determine the activation of MAPK3/ERK1 and MAPK1/ERK2 and further transduction of the signal within the MAPK/ERK cascade [53]. MAP2K1/MAP2K2 mutations have been associated with resistance to RAF and MEK inhibitors [54,55].

KIT
The KIT gene encodes the mast/stem cell growth factor receptor Kit [56]. KIT gene modifications (mutations, copy number variations) are mutually exclusive to BRAF, NRAS, and NF1 mutations. The most involved exons are 11 and 13. The L576P and K642E mutations have a positive response to KIT inhibitors, but this is a temporary answer [57][58][59][60].
Translocations represent another mechanism that activates the MAPK and other oncogenic pathways. They are more frequent in tumors with spitzoid morphology. In these cases, fusion gene results in ALK, ROS1, RET, BRAF, NTRK1, NTRK3, and MET genes [61,62]. These translocations were determined as mutually exclusive of each other and other MAPK-activating mutations, sustaining the pinpointed translocations as being critical driver events.

CDKN2A
CDKN2A was first associated with familial melanoma predisposition [64]. Latter studies recognized CDKN2A as an essential gene in this locus, responsible for controlling the cell cycle [65]. CDKN2A is the most often concerned tumor suppressor gene in sporadic melanoma [66]. CDKN2A losses, frequently biallelic, are seen in 50-80% of sporadic melanoma [20,67]. The most common abnormalities are inactivating mutations and promoter methylation [68].
The CDKN2A gene encodes for p16 and p14ARF. p16 is important in cell cycle control. It acts on CDK 4/6 and blocks the phosphorylation of the retinoblastoma protein (Rb) [69,70]. Rb phosphorylation releases the E2F transcription factor with cell cycle progression from G1 to the S phase [71]. p14ARF inhibits the ubiquitin ligase MDM2, which targets proteasomal degradation of TP53 [70].

PTEN
PTEN is a tumor suppressor gene with inactivating action in the PI3K signaling pathway. The PTEN mainly determines the inhibition of the AKT pathway.
PTEN has been reported to be mutated or nonexistent in up to 70% of melanomas [20]. Epigenetic silencing of PTEN may likewise play a role in melanoma [72]. PTEN loss is more often found in BRAF-mutant than in NRAS-mutant melanoma [67], a finding compatible with PI3K/AKT pathway activation by NRAS but not BRAF mutations, which need an additional hit. PTEN inactivation has been linked to resistance to BRAF inhibitors [73] and immunotherapy [74] in melanoma patients.

TERT Promoter Mutations
Mutations in the telomerase reverse transcriptase (TERT) promoter melanomas are activatied in melanoma [75]. The effects are an increased gene expression, and cell proliferation without senescence or apoptosis [76]. TERT promoter mutations were demonstrated to promote tumorigenesis in two phases, mainly by repairing the shortest telomeres [77]. These mutations are markers of aggressive behavior and also of inferior prognosis [78].

TYRP1
The microarray analysis of melanoma metastasis conducted by Journe found that the gene TYRP1 was associated with shorter survival. TYRP1 expression in the validation group demonstrated a powerful correlation between TYRP1 protein level and distant metastasis-free survival and OS. This discovery indicates TYRP1 as a possible prognostic marker for stage III melanoma patients [79].
Even though nowadays we know that TYRP1 plays an important role, the link with patient survival and how its expression affects cell behaviour is still unclear [14].

ctDNA and CTC
In a recent systematic review, Gandini et al. summarized the studies about the link between ctDNA and survival in over 2000 stage III and IV melanoma patients. They observed that detectable ctDNA before treatment and during follow-up correlated to poorer progression-free survival (PFS) and OS, with no differences by tumour stage or systemic regimen. Even though ctDNA has a high potential as a prognostic biomarker, the standardization of a methodology is necessary before introducing liquid biopsy in clinical practice [80].

Common Variants
A recent meta-analysis genome-wide association study identified 54 significant loci for melanoma and 68 independent SNPs (Single nucleotide polymorphisms) in the meantime. These variants concern genes or loci located in the vicinity of genes involved in pathogenic pathways of melanoma such as DNA repair and telomere length, differentiation of melanocyte and cell adhesion, and immunity [81].
SNPs identified in following genes (or genetic regions located in their proximity) are implied in melanocyte development and differentiation pathway: FOXD3 (rs670318), NOTCH2 (rs2793830), MITF ( rs149617956), NOTCH1, and SOX10 [81]. E-cadherin encoded by the CDH1 gene (cadherin 1) plays a major role in the adhesion between melanocytes and keratinocytes. E-cadherin expression is lost in the stage of melanoma progression. SNP rs4420522 in the CDH1 gene is a risk allele [81].
Common variants in immunity genes associated with melanoma susceptibility have been identified, for example, rs28986343 at the HLA locus and association between rs408825 and MX2 gene. On the other hand, rs1126809 in the TYR gene or rs6059655 in the ASIP gene are protective effects [81].

Copy Number Variations (CNV)
Numerous numerical and structural chromosomal aberrations, unlike benign proliferations that do not have such changes, were discovered in malignant melanomas [82]. Chromosomes most frequently involved are 1,6,7,9,10, and 11. These chromosomal regions 3)] involved in the MAPK pathway, which is consistent with the involvement of this pathway in the pathogenesis of melanoma [82,83].
Copy number variations can be generated by loss or gain of genetic material and these correlate with the stage of tumorigenesis, patient age, and histological type [82,83].
The most frequent anomalies are losses of chromosome 9 (81% of tumours) and chromosome 10 (63% of tumours), with both anomalies appearing in the early stage of tumorigenesis. At the level of chromosome 9, usually, region 9p21 where the CDKN2A gene is located, which encodes cyclin-dependent kinase inhibitor 2A, is concerned. It interacts with CDK4 (cyclin-dependent kinase 4) and CDK6 (cyclin-dependent kinase 6) and acts as a negative regulator of normal cell proliferation [82]. For chromosome 10, the frequently involved regions are 10q21-22 and 10q24-qter [82].
In final stages of melanoma and in metastasis, gains on chromosome 6, 7, 8, and 1q were observed, and 1q chromosome gain is associated with 6p chromosome gain and 10q chromosome losses [82].
Chiu et al identified 249 copy number variations in circulating tumor cells (139 copy number gains and 110 copy number loss) found in more than 50% of the studied cases. A panel of five such CNV proved to have a negative prognostic impact (copy number gains: 1p35.1, 2q14.3, 14q32.33, and copy number loss: 14q32.11, 21q22.3) [84].

Hereditary Melanoma
Hereditary melanomas represent 5-12% of all melanomas [85]. Germline mutations are characterized by high penetrance and are associated with other cancers or other locations [85]. The mechanisms of oncogenesis include: activation of oncogenes, loss of tumor suppressor genes, and chromosomal instability. Most cancer predisposition syndromes are transmitted in an autosomal dominant manner [85][86][87]. Table 2 summarizes the genes and the syndromes with predisposition to cancer and risk for melanoma.

Epigenomics in Melanoma and Implications for Biomarkers
Epigenetic changes relate to gene expression independent of changes in the DNA sequence that persist over several cell divisions [89].
Modifications of the epigenome are involved in cancer initiation, progression, and resistance to antitumor drugs [90]. Epigenetic marks are important as biomarkers for diagnosis, prognosis, predictive for disease recurrence, and therapeutic targets [91,92]. The epigenetic changes are reversible [93,94].
MicroRNAs, non-coding RNAs, histones modifications, and abnormal DNA methylations were associated with the stages of melanoma progression [95].

Histones Modifications
In histone modifications, 3 types of proteins are involved: histone writer proteins, eraser proteins, and reader proteins. The role of the first category is to add different chemical groups to histones through various chemical processes (acetylation, methylation, phosphorylation, ubiquitylation, glycosylation, ADP-ribosylation, carbonylation, and SUMOylation) [96]. Eraser proteins remove chemical groups. The reader proteins modulate gene expression by recruiting transcription factors or transcription repressors [90,97,98]. Chromatin remodelling occurs by modifying histones. These changes produce gene hyperexpression or gene inactivation [90]. The most important histone modifications are histone methylation and histone acetylation.

Methyltransferases and histone demethylases
Methylation of histones occurs most frequently at the level of lysine residues and/or arginine residues [96]. Lysine methylation consists of the addition of one to three methyl groups. The conversion from the unmethylated form to the methylated form is done under the action of methyltransferases (writer) and the reverse demethylation process is carried out by histone demethylases (eraser) [96,99]. Table 3 and Figure 1 summarize methyltransferases and histone demethylases, their action, and their place of action.

Epigenomics in Melanoma and Implications for Biomarkers
Epigenetic changes relate to gene expression independent of changes in the DNA sequence that persist over several cell divisions [89].
Modifications of the epigenome are involved in cancer initiation, progression, and resistance to antitumor drugs [90]. Epigenetic marks are important as biomarkers for diagnosis, prognosis, predictive for disease recurrence, and therapeutic targets [91,92]. The epigenetic changes are reversible [93,94].
MicroRNAs, non-coding RNAs, histones modifications, and abnormal DNA methylations were associated with the stages of melanoma progression [95].

Histones Modifications
In histone modifications, 3 types of proteins are involved: histone writer proteins, eraser proteins, and reader proteins. The role of the first category is to add different chemical groups to histones through various chemical processes (acetylation, methylation, phosphorylation, ubiquitylation, glycosylation, ADP-ribosylation, carbonylation, and SUMOylation) [96]. Eraser proteins remove chemical groups. The reader proteins modulate gene expression by recruiting transcription factors or transcription repressors [90,97,98]. Chromatin remodelling occurs by modifying histones. These changes produce gene hyperexpression or gene inactivation [90]. The most important histone modifications are histone methylation and histone acetylation.

Methyltransferases and histone demethylases
Methylation of histones occurs most frequently at the level of lysine residues and/or arginine residues [96]. Lysine methylation consists of the addition of one to three methyl groups. The conversion from the unmethylated form to the methylated form is done under the action of methyltransferases (writer) and the reverse demethylation process is carried out by histone demethylases (eraser) [96,99]. Table 3 and Figure 1 summarize methyltransferases and histone demethylases, their action, and their place of action.  Histone methylation has two effects: silencing signatures and actively transcribed chromatin. Table 3. Main methyltransferases and histone demethylases and their actions [27,96,100].
The most important methyltransferase with a role in melanoma is EZH2 (Histonelysine N-methyltransferase EZH2) encoded by the EZH2 gene (enhancer of zeste 2 polycomb repressive complex 2 subunit). EZH2 represents the catalytic subunit of Polycomb Repressive Complex 2 (PRC2). PRC2 has a role in (mono-, di-, or tri-) the methylation on H3 lysine 27 and in such a way in the silencing of target genes [100]. In total, 3% of melanomas are associated with activating mutations in EZH2, with a role in melanoma progression [97,102].
Acetylation is done exclusively at the level of lysine residues [94]. The balance of histone acetyl transferases and histone deacetylases determines the level and state of acetylation [97].

Histone Variants
Histone variants contribute to epigenome plasticity [91]. The canonical histones are H2A, H2B, H3, and H4 [97,104]. The histone variants have different sequences and properties and can replace canonical histones. The effect is altered chromatin structures and gene transcription [97,104].

Long Non-Coding RNAs
The majority of the human genome transcript (90%) is not transcribed into proteins and has a role in the regulation of gene expression. Long non-coding RNA (lncRNAs) is represented by DNA sequences that has a length over 200 pb. It modifies the regulation of gene expression by transcription and translation regulation, chromatin changes, RNA changes through editing or splicing or degradation, and miRNA sequestration. At the cellular level, the modification of gene expression translates into the regulation of cell proliferation, differentiation, migration, and invasion [110]. Guo [17].
In more than 90% of melanomas, an lncRNA SAMMSON, which plays an oncogenic role in association with MITF, is present [111].
Several lncRNAs were studied because of the high expression levels in melanoma patients, including SPRY4-IT1, MALAT-1, BANCR, UCA1, HOTAIR, and SNHG8. The levels of UCA1 and MALAT-1 were remarkably more elevated in patients with melanoma compared to healthy controls, and their levels were associated with the stage of the disease [112,113].
The expression of LINC02249 was found to be elevated in cutaneous melanoma. The high expression corresponds to poor OS and disease-specific survival, which is an independent prognostic factor [114].

MicroRNAs
Small ncRNAs includes miRNAs, piwi-interacting RNAs-piRNAs, and small nucleolar RNAs-snoRNAs. Their length is up to 200 nucleotides [115,116]. MiRNA (single stranded DNA molecules 18-25 nucleotides long) is involved in cancer by regulating oncogenes and tumour suppressor genes. The premature miRNAs are exported to the cytoplasm where mature miRNAs are formed. These molecules will bind to the target mRNA because of the complement of the bases. Mechanisms for regulating gene expression involve the degradation of the target mRNA or inhibition of translation into proteins [115,117].
Huber et al. found a set of miRNAs (let-7e, miR-99b, miR-100, miR-125a, miR-125b, miR-146a, miR-146b, and miR-155) implicated in the transformation of monocytes into immunosuppressive MDSCs. MiRNAs such as miR-28 and miR-17-5p seem to interfere with PD-1 and PD-L1 expression at a post-transcriptional level, enabling resistance to immunotherapy [119]. Kanemaru et al. suggested that circulating miR-221-3p could be used as a melanoma biomarker, showing significantly distinct expressions between stage I/IV melanoma patients and healthy controls. Their work explained how miR-221-3p levels reduced after surgical removal of the primary tumor and increased upon disease recurrence, suggesting that circulating miRNA-221-3p could be a new tumor marker. High levels of miRNA-221 have been found in early-stage melanomas compared to healthy individuals. The levels of expression were also found to be proportionate with the stage of the disease [120].
Tang et al. summarized the circRNA involved in the inhibition of apoptosis, hyperproliferation, activation of invasion, migration, carbohydrate metabolism, and metastasis. Some of circRNA have an oncogenic function and activate several physiological processes in melanoma: circ_0084043 and circ-FOXM1 (proliferation, apoptosis, invasion). Other cir-cRNA molecules have suppressive functions: circ_0023988 and circ_0030388 (proliferation, invasion, migration) [121].

Abnormal DNA Methylation
DNA methylation is the biochemical process in which a methyl group is added to a cytosine or adenine at the 5-position of carbon where the DNA base thymine is located. The cytosine is converted to methylcytosine [127,128]. In total, 60% of gene promoters are associated with unmethylated CpG islands [128].
DNA methylation is a mark of suppression of gene expression. The effects are on cell differentiation and cell proliferation. Sigalotti et al. studied the importance of DNA methylation as a prognostic biomarker in stage III melanoma. They assessed the genome-wide methylation profiles from 45 patients. Based on global methylation, the cohort was split into a favorable group, with a median survival of 31.5 months, and an unfavorable group, with a median survival of 10.3 months, with a 5-year overall survival of 41.2 and 0%, respectively. The group identified a 17-gene methylation signature sufficient to distinguish the good prognosis group, characterized by low methylation density. Hypomethylation was a significant predictor of increased OS [129].

Conclusions
Cutaneous melanoma is the result of multiple genomic and epigenomic changes, some of them in close correlation. The negative prognosis of melanoma has raised interest in the discovery of prognostic and predictive biomarkers in order to improve life expectancy.
For these reasons, the approach to the patient must be in relation to genetic and epigenetic changes. Prevention targeted the approaches of modifiable and non-modifiable risk factors. The existence of genetic and epigenetic biomarkers validated in various evolutionary stages would allow for a more precise molecular framing and a targeted approach to the changes present in the patient/group of similar patients. The general objective must be to avoid the metastatic stage.