Estimated Melanocortin-1 Receptor (MC1R) Haplotypes

Peter A. Kanetsky, Timothy R. Rebbeck, Tara Friebel, Marianne Berwick, for the GEM Study Group

Background The human melanocortin-1 receptor (MC1R) protein is highly polymorphic with nearly 40 published sequence variants that occur within the one-exon coding region spanning only 951 basepairs. It has been suggested that MC1R coding region variants are inherited independently with the exception of the observed co-segregation of the V92M and T314T variants. To formally address this issue, we estimated MC1R haplotypes among unrelated individuals with cutaneous melanoma collected for a large, international investigation of melanoma etiology. Methods MC1R genotyping was accomplished by direct sequencing of genomic DNA on 311 individuals from Australia (New South Wales and Tasmania), 230 from the United States (Southern California, New Jersey, North Carolina, and Michigan), and 121 from Canada (British Columbia and Ontario). The hapipf and ipf commands in Stata were used to estimate MC1R haplotypes. Results We observed a total of 33 MC1R polymorphisms including seven synonymous, one frameshift, and 25 missense variants. Five hundred fifty-four (83.7%) individuals carried at least one MC1R variant; two hundred thirty-five (35.5%) individuals carried two variants, 74 (11.2%) carried three, and 16 (2.4%) carried four. Disregarding synonymous changes, 539 (81.4%) persons carried at least one variant; two hundred forty-seven (37.3%) carried two, and four carried three (0.6%). Five common estimated haplotypes occurred at a frequency of greater than 5% and accounted for over 86% of all MC1R haplotypes. With the addition of five uncommon haplotypes (estimated at frequency of greater than 1%), explanation of MC1R haplotypes increased to 96%. For these ten haplotypes, only two (—V92M—T314T— and —I155T—T314T—) were multi-variant and both occurred in the context of the same synonymous variant. All remaining estimated haplotypes represented uni-variant carriage (—V60L—, —D84E—, —R142H—, —R151C—, —R160W—, —R163Q—, —D294H—) or carriage of no MC1R variant —consensus—). Several other multi-variant MC1R haplotypes containing two missense variants were estimated, although the expected haplotype frequencies ranged downward from 0.1%. We did not observe significant differences in haplotype frequencies when the data were stratified according to geography (Australia vs. Canada vs. United States). Conclusion Our results support the notion that MC1R variants are indeed inherited individually on separate alleles with few exceptions. Hence, the 251 individuals genotyped here with two or more missense MC1R variants do not carry an allele encoding the wildtype MC1R (although for many variants the effect upon receptor function is unknown or unclear). Because of the rarity of multi-variant haplotypes containing missense variants, knowledge of MC1R haplotypes contribute very little to analyses of associations of MC1R variants and outcomes. These types of analyses reduce simply to models of outcomes and single variants, either alone or in combination.