SPECIFIC SKIN CANCERS
Genetic Predisposition to Melanoma
Melanoma is a relatively uncommon cancer, with an incidence in most of Northern Europe of around 10 per 100,000 per annum (Parkin et al. 2001). In many countries, there is a sex difference in incidence so that in some, particularly in the UK, it has been more common in women, whereas in hot countries it is more common in men. The incidence has increased markedly this century in white people in most Western countries and in Australia and New Zealand, which have the highest rates in the world, but melanoma is fortunately rare in Asian or black skin (Eide and Weinstock 2005). The disparity in incidence between white and pigmented skins gives the first clue that genetics play a major role in determining susceptibility to a form of cancer for which the major environmental determinant is sun exposure (Jones et al. 1999).
The commonest type of melanoma is the superficial spreading type, which has the appearance of a mole progressively changing in shape, size, and color. The melanomas commonly have irregularly distributed hues of brown, grey, black, or red. These Tumours are most frequent on the lower leg in women and on the trunk in males. The prognosis is determined by the Breslow thickness in millimeters, which is the thickness measured by the histopathologist from the granular layer of the skin to the deepest part of the Tumours, as well as sex, age, Tumours site, and other histological features such as the presence of ulceration (Balch et al. 2009). Earlier diagnosis with thinner Tumours is associated with a better prognosis (Rees 2003; Barsh 1996; Elwood et al. 1990; Bahmer et al. 1990; Balch et al. 2001; Anderson and Badzioch 1991; Augustsson et al. 1990; Hayward 2003).
Epidemiological studies have established that the most potent phenotypic risk factor identified to date is the presence of numerous or clinically atypical nevi or moles (Fig. 10.1). The presence of multiple nevi, some of which are clinically atypical, is called the atypical mole syndrome (AMS) phenotype (otherwise known as the dysplastic nevus syndrome (Tucker et al. 2002) or the familial atypical mole and multiple melanoma syndrome FAMMM (Bergman et al. 1992)). It was thought originally that this phenotype was indicative of the inheritance of high-penetrance melanoma susceptibility genes, but it is now recognized that the AMS is relatively common, for example, it was seen in 2 % of the general UK population in one study (Newton et al. 1993), and 18 % of healthy individuals had at least one clinically diagnosed atypical nevus in a Swedish study (Augustsson et al. 1991). The odds ratio for melanoma in individuals with the AMS but no family history of melanoma is in the order of 10 relative to those who have very few nevi at all (Bataille et al. 1996). The absolute risk to such people is therefore moderate. Twin studies have shown that nevi are predominantly genetically determined (Easton et al. 1991; Wachsmuth et al. 2001) and it is hypothesized that the AMS in the absence of a family history is indicative of genetic susceptibility to melanoma with some contribution from sun exposure (particularly sunny holidays). In recent times some of the associated genes have been found using genome-wide association studies (GWAS). Single nucleotide polymorphisms (SNPs) associated with nevus number and melanoma risk were in the TERT gene (Rafnar et al. 2009; Barrett et al. 2011), near to the locus coding for CDKN2A on chromosome 9 (Bishop et al. 2009) and on chromosome 22 in a gene known as PLA2G6 (Bishop et al. 2009). TERT codes for telomerase, a polymerase which maintains the telomere end (or cap) which protects the chromosomes from end-to-end fusion, nucleolytic decay, and atypical recombination (Nan et al. 2011), so this observation is consistent with a previous study in which patients with more nevi had longer telomeres (Bataille et al. 2007). The biology which underpins the second two associations (on chromosome 9p and PLA2G6) is not yet fully understood. Finally, there is some evidence of a role for the gene coding for interferon regulatory factor 4 (IRF4) and nevus number and melanoma risk, but the nature of that relationship is apparently complex (Duffy et al. 2010), with age-dependent effects on nevus number. Even when all these genes are considered together, however, only about 3 % of the variation in nevus number is currently explained (unpublished Leeds in-house data) (Cannon-Albright et al. 1990; Auroy et al. 2001; van der Velden et al. 2001).
Fig. 1 Familial atypical mole-melanoma syndrome: a melanoma arising from a dysplastic or atypical nevus; note the irregular borders and pigmentation (Courtesy of Julia A. Newton-Bishop)
Other phenotypic risk factors relate to the presence of fair skin and hair, freckles, and a reported susceptibility to sunburn often referred to as people’s “skin type” (Gandini et al. 2005a, b). A major genetic determinant of this phenotype (and a genetic risk factor for skin cancer) is the inheritance of common polymorphisms of the melanocortin receptor 1 (MC1R) gene, which codes for variation in the melanocortin receptor to which MSH binds and which result in modulation of the ratio of eumelanin (black pigment) to pheomelanin (red/yellow pigment) in the skin and hair. The association was first reported by Valverde et al. (1996) but has been explored extensively by other groups working on sporadic melanoma (Raimondi et al. 2008) and familial melanoma (Demenais et al. 2010).
As genetically determined phenotypes associated with a tendency to burn in the sun have been unequivocally identified as melanoma risk factors, it is not surprising that there is also an increased risk associated with environmental factors, such as a history of severe sunburn and sunbathing (Gandini et al. 2005a, b; Chang et al. 2009).
Additional pigment genes underlying phenotypic variation in skin and melanoma susceptibility have also been explored in GWAS, and these include SLC45A2 (Barrett et al. 2011), the IRF4 locus, the gene coding for tyrosinase (TYR) (Gudbjartsson et al. 2008; Barrett et al. 2011), and the agouti signaling protein locus ASIP (Barrett et al. 2011; Gudbjartsson et al. 2008) which is an agonist of the MC1R and therefore functions in the same pathway as variants in MC1R having effects on melanin synthesis. Using what is known about the variants of MC1R, it is apparent that we have in all now identified susceptibility genes which “explain” 30 % of the variance in skin pigmentation in white-skinned people taking part in the GWAS (unpublished data).
GWAS have identified a number of additional melanoma susceptibility genes which are not associated with either “at-risk” phenotype (fair skin or increased nevus number) which opens up new and interesting biological pathways to better understand melanomagenesis. These include ARNT (Macgregor et al. 2011) two genes involved in DNA repair, PARP1 and ATM, CASP8 (caspase 8 which is an apoptosis-related gene), CCND1, and MX2 (Barrett et al. 2011).
Clustering of many cases of melanoma predominantly occurs in families in which there appears to be susceptibility to melanoma alone.
There are very rare families demonstrating susceptibility to both cutaneous melanoma and ocular melanoma. Scandinavian families have been reported in which susceptibility to ocular and skin melanoma was linked to a gene on 9p21 (Jonsson et al. 2005) and more recently, an inherited mutation in the BAP-1 gene on chromosome 3p was reported in an ocular melanoma patient (Harbour et al. 2010).
Low numbers of cases of melanoma have also been reported in hereditary non-polyposis colorectal cancer (Lynch syndrome). In one recent series, 9 family members developed melanoma from 8 families in a series of 60 Lynch syndrome families (Ponti et al. 2008). There is a reported increased risk of melanoma in the Li–Fraumeni syndrome and inherited retinoblastoma. In inherited retinoblastoma many of the melanomas have been reported to occur in the radiation field. There is a clear increased risk in patients with xeroderma pigmentosum in whom DNA repair defects are causal. Inheritance of BRCA 2 mutations also appears to increase the risk of melanoma moderately (Liede et al. 2004).
Congenital anomalies such as neurocutaneous melanosis (Arunkumar et al. 2001; De Andrade et al. 2004) and giant congenital hairy nevi also predispose to melanoma.