Molecular classification of melanoma
Molecular classification of melanoma
Over the last decade significant molecular data have shed new light on differences between melanoma subtypes. This continues to inform an emerging classification, which integrates epidemiology, clinical and histopathological features with patterns of somatic mutations and mutational mechanisms. This final portion of the melanoma chapter focuses on conceptual information regarding the molecular classification of melanoma and its correlation to UV damage, as well as the progression from melanoma in situ to invasive and metastatic melanoma. Molecular data continue to inform our traditional classification schemes, validating them in some instances and expanding them in others. While traditional classification categories of melanoma are still employed, the molecular characterization of melanocytic neoplasia continues to progress and will have to be continually integrated with the traditional features to establish a classification system that correctly describes distinctive phenotypes but is firmly based in the underlying mechanistic alterations that are in large part caused by genetic alterations.
1355 Molecular classification of melanoma
indicate such a signature.5,6 For example, the most common mutation at BRAF codon 600 is a T to A transition and thus does not match the classical UV signature.7 The fact that these mutations lack a UV signature does not exclude a role for UV radiation in their formation, as they can form as the result of indirect consequences of UV radiation such as the formation of reactive oxygen species. Furthermore, activating an oncogene requires mutations to change very specific amino acid sequences in the protein, whereas abrogating a tumor suppressor protein’s function can occur through numerous possible mutations (missense, nonsense, and frameshift mutations). Therefore, mutations that activate oncogenes tend to cluster in a few critical hotspots, whereas mutations in tumor suppressor genes are more widespread and enriched for damaging mutations that truncate or scramble the DNA sequence. The need for specific growth promoting mutations in oncogenes such as BRAF and NRAS therefore significantly skews the mutations observed in these genes. The causative role for UV radiation in mutagenesis of melanomas on sun-exposed skin is clearly demonstrated by the high burden of UV signature mutations in the tens to hundreds of thousand mutations found in the genomes of cutaneous melanomas and by the observation that mutations in critical genes such as the TERT promoter, CDKN2A, P53, and PTEN are indeed enriched for UV signature mutations.8,9
Melanoma epidemiology Melanomas are most common in Caucasians, where they typically arise on sun-exposed skin. In patients under the age of 50, the highest density is found on the trunk and extremities. In patients older than 50, the highest density of melanomas is found on the head and neck. The high frequency of melanoma and non-melanoma skin cancers in Caucasians is considered to be the result of light skin pigmentation with the consequence of increased UV-induced mutagenesis. Some aspects of skin color variation are likely due to selective advantage during human evolution that optimizes benefits from solar radiation (e.g., lighter skin pigmentation to enhance vitamin D synthesis) in less sunny climates. Association studies have revealed common polymorphisms in pigmentation genes are genetic factors controlling skin color, tanning ability, and freckling, with some variants increasing melanoma risk. These pigmentation genes include the melanocortin receptor one (MC1R), its antagonist the agouti signaling protein (ASIP), tyrosinase (TYR), tyrosinase-related protein (TYPR), the P-gene mutated in oculocutaneous albinism type II (OCA2), ion exchange proteins of the solute carrier family SLC24A5, SLC24A4, and SLC45A2 (MATP), and the interferon regulatory factor IRF4 or MUM1.1
However, not all melanoma subtypes harbor genomic damage attributable to UV radiation. Melanomas arising on the palms and soles, nail beds, and mucosal sites, as well as uveal melanoma, typically lack the high mutation burden with a UV signature found in cutaneous melanomas. This indicates that while UV radiation plays a critical role in the pathogenesis of some melanoma subtypes, it does not in others. This could also explain why acral and mucosal melanomas affect all world populations independent of skin complexion.10
Melanoma is composed of distinct subtypes There is considerable variation in the clinical and histopathological presentation of melanomas that partially depends on the anatomic site in which the tumor arises and how much ultraviolet radiation exposure has occurred. As outlined previously in this chapter, these differences formed the traditional ‘histogenetic’ classification of melanoma as initially proposed in the Sydney Classification.11 There has been considerable controversy on the justification of this classification and opponents have long posited that melanoma is a single disease, with the variation in clinical and pathological features representing secondary phenomena. This belief stemmed both from the lack of difference with regards to prognosis of advanced disease and the inability in the past to predict a different response to therapy between subtypes.12 In addition, there was significant overlap in the defining phenotypic features that hindered unequivocal classification of a considerable portion of primary tumors.13
Decreased pigmentation in Caucasian skin leads to decreased protection from UV radiation and results in increased somatic mutations in skin cells, including melanocytes. However, increased melanoma susceptibility in Caucasians may not be entirely attributed to differences in skin pigmentation, because albinos of African descent do not share the increased incidence of melanomas, while they do suffer more frequently from cutaneous squamous cell carcinoma.2 Melanin, which is present in Caucasian skin and absent in albinos who lack functional tyrosinase, appears to have both protective and deleterious effects for DNA. Mutations from UV radiation continue to accumulate in the skin for hours after removal from UV light.3 These ‘dark’ mutations do not occur in albino mice lacking melanin. In mice with intact melanin production, these mutations were twice as frequent in those with increased pheomelanin (due to MC1R mutation) versus those with high eumelanin, implicating melanin, particularly pheomelanin, as a critical factor in these ‘dark’ mutations. Further evidence is found in transgenic mouse models in which melanoma development after UVA radiation only occurs in black mice with intact melanin production and not in albino mice.4
Considerable progress has now been made in the identification of genetic alterations in melanoma and has provided strong support for the concept of biologically distinct melanoma types. This is consistent with work leading to the discovery and validation of clinically relevant genetic subtypes that have emerged in other malignancies such as lung and colon cancer. This is a decisive trend in oncology that is critical for the implementation of precision cancer medicine. Mutations in specific genes, mutational signatures, and other genetic alterations such as genome-wide chromosomal aberrations correlate with specific clinical and histopathological features of melanoma, providing genetic support for the original concept that melanomas fall into different categories of diseases. New genetic data offer an opportunity to integrate causal genetic alterations with clinical or histopathological disease attributes, as well as predisposing and environmental factors. The emerging melanoma subtypes that integrate clinicopathological features overlap considerably with the original ‘histogenetic’ melanoma types, but offer a more nuanced picture of the categories. This continuity is satisfying and underscores the enduring importance of thoughtful clinicopathological characterization as a basis for initial classification. The extent of UV radiation damage can serve as an initial branch point in the classification of melanoma: melanomas found in skin with high cumulative sun damage (high-CSD), low cumulative sun damage (low-CSD), and melanomas on sites with negligible (e.g., acral) or no sun exposure (e.g., mucosal).
Melanoma and UV radiation The establishment of a causal link between UV radiation and melanoma formation requires several considerations. UV radiation causes a wide spectrum of mutations. Among these UV signature mutations, pyrimidine dimers resulting in C > T or CC > TT transitions are the most common and are due to a direct photochemical reaction with the DNA. However, mutations in common melanoma oncogenes such as BRAF and NRAS typically do not
1356 Melanoma
Melanoma on sun-exposed skin Epidemiological data alone suggest differences between melanomas from sun-exposed skin and melanomas on acral or mucosal sites, the latter of which do not show variation in incidence between races. Additionally, the anatomic site of origin and the cumulative dose of UV radiation are two intricately linked features that are strongly associated with distinctive clinical and histopathological presentations of melanoma. Melanomas on the head and neck tend to occur in older individuals with a peak incidence of about 72 years of age.14 The skin surrounding these melanomas typically shows signs of chronic sun damage, as evidenced by pronounced solar elastosis and other signs of chronic sun-induced damage such as solar lentigines and actinic keratoses.15 Total cumulative UV dose appears to represent a major risk factor for these melanomas. The traditional lentigo maligna variant of melanoma falls into the high-CSD category.
%\0XWDWLRQ6WDWXV
1(,7+(5 %56 0XWDWLRQ
:+27\SH
3LJPHQW
6FDWWHU
1HVWLQJ
&LUFXPVFULSWLRQ
By contrast, melanomas involving the trunk arise earlier in life, with a peak incidence at around age 54, and signs of chronic sun damage (severe solar elastosis) are typically absent.14 The drop in incidence of melanomas on the trunk after age 54 suggests that it is not the cumulative dose of UV radiation but timing of UV exposure that plays a decisive role in these melanomas, with a particular window of vulnerability present earlier in life. Epidemiological studies demonstrate that exposure to UV irradiation in childhood plays an important role in the pathogenesis of melanoma in general.16 Patients with melanomas on the trunk, but not with melanomas on chronically sun-exposed sites such as the face, have more melanocytic nevi, suggesting a common pathogenesis between low-CSD melanomas and nevi.17 This is supported by the fact that low-CSD melanomas more frequently have adjacent precursor nevi than CSD melanomas and that nevi and low-CSD melanomas share a high frequency of BRAF V600E mutations. A functional scale for assessing sun damage has been proposed to grade the degree of solar elastosis involving the dermis.18,19 Recent studies have demonstrated how melanomas progress from such nevi through the acquisition of additional UV-associated mutations.20 The fact that most nevi arise during childhood and adolescence furthers highlights the possibility of a window of vulnerability towards UV radiation early in life for these types of melanocytic neoplasms.21 Genetic data also support the concept of distinct melanoma types based on sun-exposure as discussed below.22