The pigmentation of skin, hair and eyes is genetically controlled and varies significantly among and within human populations. Some aspects of this variation are clearly due to selection that optimizes benefits from solar radiation. However, one cost of this adaptation is the well established link between pigmentation and skin cancer risk in Caucasians. While some features of human pigmentation genetics are established, it has been clear that many genes affecting pigmentation remain to be discovered.

Over 100 genes controlling coat color pigmentation in mice have been identified, roughly half of which have human homologues. About 30 of these genes have been associated with rare human pigmentation disorders. Until recently only six genes were known to harbor relatively common polymorphisms that affect ‘normal’ human pigmentation. These genes include the melanocortin receptor 1 (MC1R), its antagonist the agouti signaling protein (ASIP), tyrosinase (TYR), the P-gene mutated in oculocutaneous albinism type II (OCA2), and ion exchange proteins of the solute carrier family SLC24A5, and SLC45A2 (MATP). That situation has changed dramatically within the last several months with the publication of five papers that double the list of highly polymorphic pigmentation loci, provide new details on specific pigmentation effects of some previously known loci, and ascribe cancer risk to a subset of the loci (Sulem et al., 2007, Sulem et al., 2008, Gudbjartsson et al., 2008, Han et al., 2008, Brown et al., 2008).

The new information comes from genome wide association studies (GWAS) in which single nucleotide polymorphisms (SNPs) distributed across the genome are examined in thousands of individuals to determine if any are associated with specific phenotypic characteristics. A positive finding of association for a particular SNP does not mean that that sequence variant causes the phenotype, just that the causal variation is somewhere nearby in the genome. In aggregate, the studies assessed skin, hair, and eye color, tanning ability, freckling, and risks for melanoma and basal cell carcinoma in populations of European ancestry living in Australia, Iceland and Northern Europe. The phenotypes and the associated loci are summarized in the Table. There is encouraging consistency in the results since previously known pigment genes were rediscovered and novel players co-discovered among the different studies. Not all studies found all associations due to assessment of different phenotypes and use of different sets of SNPs.

A total of five novel loci, most close to or within attractive candidates for the causal gene, were found, two of which (SLC24A4 and IRF4) being identified in two independent studies. SLC24A4 will be familiar to those who follow this field, as its close relative SLC24A5 was previously discovered to be involved in pigmentation in zebra fish and humans (Lamason et al., 2005). By contrast, the interferon regulatory factor IRF4 or MUM1 is a new and somewhat unexpected player, as it was hitherto mostly known to be involved in the inflammatory response in hematopoietic cells. However, IRF4/MUM1 has been demonstrated to be strongly expressed in melanocytic neoplasms (Sundram et al., 2003).

Two of the studies investigated loci associated with melanoma and basal cell carcinoma (Gudbjartsson et al., 2008; Brown et al., 2008). Both identified a region of chromosome 20q11.22 near the gene ASIP as affecting risk for both types of malignancies, but they disagreed on whether ASIP itself was the likely causative gene. Resolution of this disagreement, possibly due to the use of somewhat different sets of SNPs, may reveal additional details about the complexity of the relationship of this genomic region to skin cancer risk (Brown et al., 2008). ASIP is an attractive candidate because of its interaction with MC1R, whose variants are well established risk factors for melanoma and non-melanoma skin cancer. ASIP inhibits the interaction between MC1R and α-melanocyte stimulating hormone. This suggests that the risk-associated ASIP variant would exhibit a gain of function, which would phenocopy the effects of loss-of-function MC1R alleles. The possible role of ASIP in melanoma risk may have been missed by prior studies, which analyzed the ancestral major allele in which the two-SNP haplotype linked to skin cancer arose. The common polymorphism in tyrosinase (R402Q), also was also found to be strongly associated with increased risk for melanoma and basal cell carcinoma, but another polymorphism in this gene was not. A weaker association with melanoma risk, but not risk for basal cell carcinoma, was found for a SNP at the TYRP1 locus.

It is noteworthy that the only associations with skin cancer risk that were discovered involved some, but not all, loci that also affected pigmentation. Since these studies (Gudbjartsson et al., 2008; Brown et al., 2008) employed hundreds of thousands of SNPs selected across the genome without regard to specific gene functions, they could have found other associations if they were strong enough. This suggests that the increased risk primarily involves increased sensitivity to UV-mediated carcinogenesis secondary to decreased pigmentation. The increased risks for each of the three pigmentation loci are relatively small, odds ratios on the order of 1.2 to 1.7, compared with the low frequency/high penetrance germ line alterations in classical tumor suppressor genes such as CDKN2A and CDK4 that are responsible for familial melanoma. These latter genes are involved in signaling checkpoints that restrict proliferation and do not affect pigmentation.

Gene neighborhoodChromHair colora,b,cEye colora,b,cFrecklinga,bTanning abilitya,b,cSkin colorcMelanomad,eBCCd
  1. aSulem et al. (2007). bSulem et al. (2008). cHan et al. (2008). dGudbjartsson et al. (2008). eBrown et al. (2008).

ASIP20q11.22+b +b+b +d,e+d
TYR S192Y11q14.3  +a    
TYR R402Q11q14.3+a+a +a +d+d
TYRP19p23+b+b   +d 
MC1R16q24.3+a,c +a+a   
OCA2 (P-protein)/HERC214q13.1+a,c+a     
KITL (SCF)12q21.33+a      
SLC45A2 (MATP)5p13.3+c  +c+c  

The studies discussed here demonstrate the value of the GWAS approach for revealing the genetic basis of pigmentation and skin cancer phenotypes, and encourage subsequent more sophisticated analyses of genetic variation. For example, the present studies grouped all melanoma into one category, whereas it has now been clearly established that at least several types can be distinguished based on their constellation of somatic genetic changes and clinicopathological features (Viros et al., 2008). For example, variants of MC1R have been shown to be associated with melanomas which have somatically acquired BRAF mutations (Landi et al., 2006). Thus searching for associations with defined cancer subtypes may reveal hits at loci that have been obscured in these initial studies.

An indication that more carefully distinguishing cancer ‘phenotypes’ might be productive comes from the fact that the odds ratios for melanoma risk contributed by ASIP, TYR, and TYRP1 increased when the analysis was restricted to younger patients. As BRAF-mutant melanomas have also been shown to be more frequent in younger patients (Viros et al., 2008), it will be interesting to learn if the risk associations for some or all of these three loci are concentrated on this melanoma type. More generally, as most of the variants in the pigmentation genes have different effects on skin, eye, and hair color it is conceivable that they also play different roles in neoplasms derived from different melanocytic populations, which are subject to differing environmental stresses such as UV exposure. One can expect that the combination of functional, pathway-centric research with more sophisticated studies of normal genetic variation will lead to rapid progress in understanding the complexity of pigmentation phenotypes, and skin cancer risk and etiology.


  1. Top of page
  2. References
  • Lamason, R.L., Mohideen, M.A., Mest, J.R. et al. (2005). SLC24A5, a putative cation exchanger, affects pigmentation in zebrafish and humans. Science 310, 17821786.
  • Landi, M.T., Bauer, J., Pfeiffer, R.M., Elder, D.E., Hulley, B., Minghetti, P., Calista, D., Pinkel, D., Kanetsky, P.A., and Bastian, B.C. (2006). MC1R germline variants confer risk for BRAF-mutant melanoma. Science 313, 512520.
  • Sulem, P., Gudbjartsson, D.F., Stacey, S.N. et al. (2007). Genetic determinants of hair, eye and skin pigmentation in Europeans. Nat. Genet. 39, 14431452.
  • Sundram, U., Harvell, J.D., Rouse, R.V., and Natkunam, Y. (2003). Expression of the B-cell proliferation marker MUM1 by melanocytic lesions and comparison with S100, gp100 (HMB45), and MelanA. Modern Pathol. 16, 802810.
  • Viros, A., Fridlyand, J., Bauer, J., Lasithiotakis, K., Garbe, C., Pinkel, D., and Bastian, B.C. (2008). Improving melanoma classification by integrating genetic and morphologic features. PLoS Med. 5, e120.