High number of melanocytic nevi is the most important known risk factor for cutaneous melanoma. We conducted a genome-wide association study for nevus count using 297,108 tagSNPs in 1,524 twins and validated our results in an independent cohort of 4,107 subjects. We
NIH-PA Author ManuscriptNIH-PA Author Manuscript NIH-PA Author Manuscript identified strongly associated variants in MTAP, a gene adjacent to the familial melanoma susceptibility locus CDKN2A on 9p21 (rs4636294, P = 3.4 × 10 -15 ). We further identified PLA2G6 on 22q13.1 (rs2284063, P = 3.4 × 10 -8 ). Both loci were also associated with melanoma risk in 3,131 melanoma cases from two independent studies (odds-ratios 1.23 at rs10757257 and rs132985). About one subject in 11 is homozygous for the variant at both loci with twice the number of nevi compared to those homozygous for the protective alleles, and double the risk for melanoma. These data provide the first evidence for common melanoma alleles whose effects are mediated through nevus number.The worldwide incidence of cutaneous melanoma in European descent populations has risen rapidly over the past 30 years, more than for any other malignancy [1]. The World Health Organization estimates 132,000 new cases of melanoma per year globally. Approximately 10% of malignant melanoma cases occur in multiplex families, with the CDKN2A locus in 9p21 accounting for susceptibility in 20-57% of all melanoma families [2]. However, remaining cases of familial melanoma do not carry mutations in CDKN2A and additional tumor suppressor genes at 9p21 and elsewhere have been proposed [3]. Large numbers of nevi are the strongest known risk factor for melanoma [4] and in many familial cases susceptibility to melanoma is associated with an excess of nevi. Linkage to 9p21 has been found in two genome-wide analyses looking at nevus count [6,7] suggesting that shared genetic factors might be involved in melanoma susceptibility and nevogenesis. Understanding the genetic bases of nevus formation is therefore an important step in understanding melanoma etiology.To search for loci involved in nevus count, we carried out a genome-wide association study (GWAS) in 1,524 healthy adult female twins from the TwinsUK registry using 297,108 SNPs with minor allele frequency (MAF) > 1% on Illumina HumanHap 300k duo chips (See Methods). Total body nevus count was evaluated by trained research nurses and defined as the sum of all nevi at least 2 mm in diameter. Comparison of the observed and expected distributions (Supplementary Figure 1) showed no evidence for inflation of the test statistics (inflation factor = 1.006) and highlighted 46 SNPs having P values < 10 -4 against the expected 29. We selected twelve SNPs for follow up that clustered in either the 9p21 or 22q13 chromosomal regions ( Figure 1). We replicated these SNP associations in an independent sample of 4,107 adolescent subjects of European ancestry from the Brisbane Twin Nevus Study (BTNS) for whom nevus count had been evaluated using a similar protocol (See Methods...