Being in the red: a no‐win situation with melanoma

The importance of skin pigmentation in determining the risk for sun-induced skin cancers, including melanoma, is supported by overwhelming clinical evidence. Numerous epidemiological studies have shown that the incidence of melanoma, as well as non-melanoma skin cancers, is considerably higher in individuals with light skin complexion, red hair, and poor tanning ability than in those with dark skin and hair color. This was originally attributed to the shielding effect of melanin, which limits the penetration of ultraviolet rays (UV) through the epidermal layers of the skin, thus reducing solar UV-induced DNA damage. An important determinant of human skin pigmentation is the relative eumelanin and pheomelanin contents, with eumelanin content being seemingly more variable in different pigmentary phenotypes, and correlating directly with visual pigmentation. It is well known that eumelanin is more photoprotective than pheomelanin in human skin. Histological and electron microscopic examination established that in dark skin, melanosomes that are enriched in eumelanin remain intact and persist in the superbasal layers of the epidermis, while in lightly pigmented skin, melanosomes are degraded, leaving only ‘melanin dust’ in the upper epidermal layers. Analysis of purified pheomelanin and eumelanin revealed that pheomelanin is photolabile and can generate reactive oxygen species that can damage DNA, proteins and lipids, while eumelanin is efficient in scavenging reactive oxygen species. However, the exact role of pheomelanin synthesized and deposited within melanocytes and its impact on melanoma formation remained unclear. Genetic studies on mouse coat color led to the identification of the extension locus, which codes for the melanocortin 1 receptor (MC1R), as the main regulator of eumelanin synthesis in follicular melanocytes. Loss-of-function mutation in the mc1r (e/e; recessive yellow) results in yellow coat color. In humans, certain allelic variants of the MC1R are strongly associated with red hair and fair skin, and result in loss of function of the MC1R expressed on melanocytes. Data primarily from our laboratory and that of Bohm et al. described an important and unexpected role for the MC1R, namely enhancement of nucleotide excision repair. We further found that activation of the MC1R reduces oxidative stress and augments antioxidant defenses in melanocytes, and that all these effects on DNA damage were not necessarily linked to increased pigmentation. The activation of DNA repair and antioxidant pathways by the MC1R offers an explanation for how the MC1R acts as a melanoma susceptibility gene. A pending question remained: how do eumelanin and pheomelanin fit in the melanoma puzzle? The study by Mitra et al. provides a direct in vivo evidence for the possible role of pheomelanin in melanoma development. Their manuscript addressed the effect of pheomelanin on melanomagenesis in mice expressing the constitutively active BRAFV600E, the commonly expressed mutation in human melanoma tumors and nevi. However, expression of this mutation is not sufficient for melanomagenesis. In the absence of any UV exposure, recessive yellow (C57BL/6 Mc1re/e, Tyr+/+, BrafCA/CA) mice expressing the Braf mutation had increased melanoma tumor formation as compared to C57BL/6 mice that express wild-type MC1R (Mc1r+/+, Tyr+/+, BrafCA/CA). What is more intriguing is that C57BL/6 Mc1re/e, Tyr+/+, BrafCA/CA developed significantly more spontaneous melanoma tumors than albino mice expressing the Braf mutation, with or without mutant mc1r (Mc1r+/+, Tyrc/c, BrafCA/CA or Mc1re/e, Tyrc/c, BrafCA/CA, respectively). Tumor formation could be inhibited by the Braf inhibitor PLX4720, indicating that constitutively active Braf was the driver for melanomagenesis. Based on these results, it can be concluded that pheomelanin cooperates with mutated Braf to enhance melanoma tumor formation and that pheomelanin has more detrimental effects than the total absence of melanin, as in the albino mice. There is increasing evidence supporting the importance of oxidative DNA damage in melanoma formation, and the concept that melanoma is a tumor driven by reactive oxygen species. In squamous and basal cell carcinomas, UV-signature mutations, such as in the tumor suppressor p53, are more common than in melanoma. Dysplastic nevi that are precursors of melanoma have higher pheomelanin content and express more reactive oxygen radicals than normal human melanocytes. In addition to the MC1R, p16, the product of the highly penetrant melanoma susceptibility gene CDKN2A, has been found to counteract oxidative stress in melanocytes (Jenkins et al., 2011). Moreover, Landi et al. had proposed that the common somatic BRAFV600E, which is mostly characterized by A>T transversion, is caused by increased generation of reactive oxygen radicals. This mutation is significantly more prevalent in melanomas arising on non-chronically sun-exposed than on sun-exposed sites and is markedly increased in individuals with any MC1R variant, with the odds ratio increased from 7.2 in individuals with one MC1R variant to 17 in those with multiple variants who have reduced antioxidant capacity. Recently, one SNP in the glutathione S-transferase gene GSTP1, namely rs 1695 that reduces the detoxification of reactive oxygen species by the enzyme, has been associated with melanoma susceptibility and reported to further increase melanoma risk in the presence of MC1R variant alleles (Ibarrola-Villava et al., 2012). Furthermore, loss of PTEN in melanoma tumors results in increased Akt activity, which leads to the generation of high levels of superoxide and progression to the aggressive vertical growth phase. Based on previous findings that UV increases the generation of reactive oxygen species and oxidative DNA damage in melanocytes and that pheomelanin itself can be a source of oxygen radicals, Mitra et al. compared the levels of two forms of oxidative DNA damage, namely 8,5′-cyclo-2′-deoxyadenosine and 8,5′-cyclo-2′deoxyguanosine, in DNA derived from skin of recessive yellow, versus albino mice (Mc1re/e, Tyrc/c), both expressing mutated Braf. They found higher levels of these two lesions in yellow mice, compared with albino mice. However, the authors did not provide any data on these lesions in C57BL/6 Mc1r+/+, Tyr+/+, BrafCA/CA mice, and thus, it can only be assumed that they were lower than in the recessive yellow counterparts. These results have important implications on the underlying drivers of melanoma formation. Using a genetic and a biochemical approach, the study provides evidence for the role of pheomelanin in increasing oxidative DNA damage and lends further support to the significance of this form of damage in melanomagenesis. The results presented also suggest that loss of MC1R function is detrimental to melanocytes in several ways, (i) inhibition of eumelanin, which is expected to counteract the oxidative effects of pheomelanin, and (ii) deprivation of melanocytes from the beneficial antioxidant effects of α-MSH, which include immediate inhibition of generation of reactive oxygen species, and enhanced activity and expression of antioxidant enzymes. Earlier this year, Noonan et al. reported that UVA irradiation induced melanoma tumor formation in HGF mice via an oxidative process (Noonan et al., 2012). This effect was pigment-dependent, with greater tumor multiplicity in C57BL/6-HGF mice than in their albino counterparts. Spontaneous tumor formation was observed in pigmented, but not in albino HGF mice. The authors argued that melanoma tumor formation was due to endogenous oxidation resulting from dysregulated melanin synthesis in the absence of UV exposure, and that this dysregulation was exacerbated by UVA, known to increase generation of reactive oxygen species. That oxidative damage is the underlying cause of melanoma formation in HGF mice is supported by other reports demonstrating inhibition of melanoma by antioxidants. Recently, it was shown that not only pheomelanin, but also eumelanin was degraded by UVA in human hair, which further confirms the detrimental effect of UVA on pigmentation and that UVA-induced pigmentation does not confer photoprotection (Wakamatsu et al., 2012). The study by Mitra et al. is highly significant. However, to demonstrate the causal role of increased oxidative DNA damage in melanoma formation, it will be crucial to show that tumor formation can be blocked by antioxidants. Also, in humans, there is no clinical evidence supporting a causal role of pheomelanin in melanomagenesis in the absence of UV exposure, since in red-haired individuals melanoma formation is very rare in non-sun-exposed anatomical sites. The findings by Mitra et al. can have an additional or alternative explanation, namely that the oxidative damage observed in recessive yellow mice might be due, at least in part, to reduced levels of glutathione, an important cellular antioxidant. Cysteine is the common precursor of pheomelanin and glutathione. Preferential utilization of cysteine for pheomelanin synthesis is expected to limit the production of glutathione. One needs to be cautious in extrapolating the results obtained using mouse models to melanoma formation in humans. Recessive yellow mice synthesize almost exclusively pheomelanin, while human melanocytes, regardless of the MC1R genotype, synthesize both pheomelanin and eumelanin, the latter of which might counteract the effects of pheomelanin by scavenging reactive oxygen radicals. The ability of the activated MC1R to augment the repair of UV-induced DNA damage and reduce oxidative stress in human melanocytes seems to be independent of pigmentary phenotype, because we observed these effects in albino human melanocytes, and in melanocytes expressing different eumelanin and pheomelanin contents (Kadekaro et al., 2012). Comparison of the responses to UV of cultured human melanocytes with comparable or disparate pheomelanin and eumelanin contents, and expressing either functional or non-functional MC1R, revealed that the main differences were due to the ability to respond to α-MSH and activate the MC1R. This is consistent with epidemiological data that increased melanoma risk conferred by expression of certain MC1R allelic variants is independent of pigmentation, because these variants can be expressed in individuals with olive skin color, who have high eumelanin content. While the studies discussed above support a role of oxidative stress in melanoma formation, the role of melanin, be it pheomelanin or eumelanin, particularly in humans, needs to be further investigated. Mouse models have illuminated many of the pathways that lead to melanoma formation. However, the question whether melanin synthesized in mouse melanocytes is chemically identical, particularly in terms of pheomelanin to eumelanin ratio, to melanin synthesized by human melanocytes is pending. Pheomelanin in human red hair contains more benzothiazole, while mouse yellow hair contains more benzothiazine. As melanin synthesis seems to be aberrant in nevi and in melanoma tumors, it is very timely to compare the chemistry of melanins synthesized constitutively in different skin types, in chronically sun-exposed versus non-exposed skin, and in normal skin versus melanoma tumors or melanoma precursors. Such analysis might lead to the identification of various forms of melanin based on the constituting precursors and provide a decisive explanation for the role of melanins as pro- or antioxidants.