Lack of somatic alterations of MC1R in primary melanoma
PRESENT ADDRESS J. A. Curtin, Respiratory Medicine Research Group, North West Lung Centre, South Manchester University Hospital NHS Foundation Trust, University of Manchester, Manchester, UK.
Boris C. Bastian, e-mail: firstname.lastname@example.org
Germline variation of the melanocortin 1 receptor gene (MC1R) is a risk factor for cutaneous melanoma. Recent studies have indicated that the risk is significantly higher for melanomas with somatic BRAF mutations, suggesting that MC1R variants may have a more specific role than their demonstrated effects on skin and hair pigmentation. To address the possibility that MC1R may act like a tumor suppressor gene by creating a permissive condition for melanocytes with specific somatic mutations to proliferate or survive, we analyzed 103 primary melanomas for somatic MC1R mutations and copy number alterations. This cohort included melanomas from skin with and without chronic sun-induced damage, mucosal membranes, and acral skin (palms, soles, and subungual). We did not find somatic mutations or frequent DNA copy number alterations at the MC1R locus, nor any skewed pattern of copy number alterations that would favor one allele type over the other. In conclusion, our findings indicate that MC1R is not a frequent target of somatic alterations in melanoma.
The melanocortin-1 receptor (MC1R) is a G-protein coupled receptor expressed on melanocytes that is a critical element in the tanning response (Chhajlani and Wikberg, 1992). It is stimulated by the release of alpha-melanocyte stimulating hormone (MSH), from keratinocytes in response to UV radiation (Chakraborty et al., 1996; Schauer et al., 1994). The MC1R gene is highly-polymorphic with over 65 variants (Rees, 2004), influencing phenotypic traits such as skin pigmentation and red hair (Naysmith et al., 2004). Variants that result in partial or near-complete loss of function are particularly common in Caucasian populations. Alleles associated with red hair (R-alleles) and non-red hair alleles (r-alleles) increase melanoma risk (Healy et al., 1999; Rees and Healy, 1997; Valverde et al., 1996). The assumption that the increase in melanoma risk is attributed to a decreased ability to produce eumelanin in response to MSH upon UV radiation is contradicted by several observations. The melanoma risk contributed by MC1R variants is independent of their effect on skin pigmentation (Kennedy et al., 2001; Palmer et al., 2000). Furthermore, the increase in risk among the most common types of melanomas in Caucasians occurring on the non-chronically sun damaged skin of the trunk and proximal extremities is mostly restricted to melanomas with somatic mutations in BRAF (Landi et al., 2006). These observations raise the possibility that MC1R may affect melanoma risk by mechanisms reaching beyond its effect on pigmentation, perhaps acting as a tumor suppressor gene by creating a permissive condition for melanocytes with certain somatic mutations to proliferate. For example, prior studies have shown that MC1R signaling induces repair of DNA damage following UV exposure (Bohm et al., 2005; Kadekaro et al., 2005; Robinson and Healy, 2002). A direct role for MC1R in melanocytic transformation would be supported by finding somatic alterations of MC1R in tumor DNA from melanoma. An analysis of 14 cases looking for somatic alterations by Valverde et al. (1996) did not find such changes. We analyzed a larger series of primary melanomas for the presence of somatic alterations of MC1R by re-sequencing the coding region of the gene and analyzing CGH data for copy number changes at the MC1R locus.
Results and discussion
We analyzed the coding region of MC1R in 103 primary melanoma samples including samples from skin with and without chronic sun-induced damage (CSD and non-CSD, respectively), mucosal membranes, and acral (palms, soles, and subungual) sites. At least one MC1R variant was found in 71.8% of our samples (Table 1) with the most common being R163Q, V60L, R151C, V92M, and R160W. Since the tissue samples came from different ethnic populations, our population is not suited to draw conclusions about polymorphism frequency in melanoma subtypes. Normal DNA from adjacent tissue was analyzed from 37 samples, including all samples with uncommon or previously unreported polymorphisms. These samples also included 6 of 15 samples that were homozygous for the R163Q polymorphism, a common non-red hair allele. We did not find evidence for somatic mutations in any of the tumors. We also examined potential interactions between MC1R genotype, and published BRAF mutation status (Curtin et al., 2005, 2006; Maldonado et al., 2003) and noted a similar association between BRAF and MC1R as in prior studies (Landi et al., 2006), where patients with BRAF-mutant melanomas tended to have a higher number of variant MC1R alleles (P = 0.038, Mann–Whitney U).
Table 1. Frequency distribution of MC1R variants
|29insA||r||0 (0.0)||0 (0.0)||0 (0.0)||1 (4.5)||1 (0.9)|
|Gln23Term||R||0 (0)||0 (0.0)||1 (5.6)||0 (0.0)||1 (0.9)|
|V60L||r||1 (2.3)||9 (28.1)||6 (33.3)||4 (18.2)||20 (17.4)|
|V92M||r||9 (20.9)||2 (6.3)||1 (5.6)||2 (9.1)||14 (12.2)|
|I120T||r||2 (4.7)||0 (0.0)||0 (0.0)||0 (0.0)||2 (1.7)|
|R142H||R||1 (2.3)||1 (3.1)||0 (0.0)||0 (0.0)||2 (1.7)|
|R151C||R||2 (4.7)||6 (18.8)||1 (5.6)||8 (36.4)||17 (14.8)|
|I155T||r||0 (0.0)||0 (0.0)||0 (0.0)||1 (4.5)||1 (0.9)|
|R160W||R||2 (4.7)||2 (6.3)||4 (22.2)||5 (22.7)||13 (11.3)|
|R163Q||r||25 (58.1)||12 (37.5)||1 (5.6)||1 (4.5)||39 (33.9)|
|D294H||R||1 (2.3)||0 (0.0)||4 (22.2)||0 (0.0)||5 (4.3)|
|wt/wt|| ||2 (7.1)||17 (45.9)||5 (27.8)||5 (25.0)||29 (28.2)|
|r/wt, R/wt|| ||9 (32.1)||8 (21.6)||8 (44.4)||8 (40.0)||33 (32.0)|
|r/r, R/r, R/R|| ||17 (60.7)||12 (32.4)||5 (27.8)||7 (35.0)||41 (39.8)|
To determine potential DNA copy number changes at the MC1R locus on chromosome 16q24.3, we analyzed CGH data that was available from published studies (Curtin et al., 2005, 2006) for 93 of our cases. Twelve samples showed copy number alterations – five lost and seven gained at least one copy of the MC1R locus (Table 2). We analyzed the sequencing electropherograms from the samples with copy number alterations and compared the peak heights of the consensus and variant alleles. If no copy number changes of the MC1R locus are present, the peaks on the sequence electropherograms of the two alleles are of approximately equal height. In the presence of a copy number loss, the peak height of the allele residing on the lost chromosome is decreased compared to the retained allele. Similarly the peak height of an allele present in multiple copies will be higher compared with an allele present at normal copy number. As the DNA from all samples contain varying amounts of DNA from contaminating stromal cells, the peak of any allele that is lost in the tumor cells is never completely absent. Only two of the five cases with losses of the MC1R region carried one or more MC1R variants which limited this analysis. One of the five samples, case 136, had a R160W substitution, an R-allele known to result in loss of MC1R function (Beaumont et al., 2007). In this sample, the wild-type peak was three times as high as the R-allele peak, indicating that the R-allele, not the wild type allele was deleted in the tumor. By contrast, examination of the sequence electropherogram of the other sample, AM17, indicated a loss of the wild type allele with retention of the variant allele. With the exception of case 136, the deletions of the MC1R region were rather large and ranged from 4.3 to 9.1 megabases, and were not specifically focused on MC1R. This could indicate that genes other than MC1R were the target of these aberrations. In addition, seven melanoma samples had increased copy number of the MC1R region, further indicating that the decrease of MC1R gene dosage may not provide a growth advantage. Interestingly, six were mucosal melanomas and one an acral melanoma, whereas copy number increases of the MC1R region did not occur in melanomas from the CSD or non-CSD group. However, acral and mucosal melanomas have a higher number of chromosomal abnormalities than CSD or non-CSD melanomas so that pattern could have arisen by chance (Curtin et al., 2005). Similar as with the losses, the gains were not restricted to MC1R but extended over a larger region ranging from 4.2 to 5.9 Mbases.
Table 2. Losses and gains of the MC1R locus
| AM209||−0.56382||Mucosal||wt/wt|| ||N/A|
| D196||−0.45103||Acral||wt/wt|| ||N/A|
| AM175||−0.39148||Mucosal||wt/wt|| ||N/A|
| AM17||−0.42007||Acral||r/wt||V92M||V92M:wt (2:1)|
| 136||−0.32501||CSD||R/wt||R160W||R160W:wt (1:3)|
| AM214||0.531057||Mucosal||wt/wt|| ||N/A|
| AM210||0.507865||Mucosal||wt/wt|| ||N/A|
| AM217||0.466146||Mucosal||R/r||R142H, V60L||V60L:R142H (1.5:1)|
| AM112||0.439243||Acral||r/r||R163Q, R163Q||N/A|
| AM146||0.41888||Mucosal||wt/wt|| ||N/A|
In summary, the absence of somatic mutations in MC1R and the lack of a consistent pattern of DNA copy number alterations at the MC1R locus indicate that MC1R is not a frequent target of somatic genetic alterations in melanoma.
We analyzed DNA extracted from a series of archival, paraffin-embedded primary melanomas that in part were previously analyzed and reported (Curtin et al., 2005, 2006; Maldonado et al., 2003). All tumors were invasive and had areas that could be microdissected, in which tumor cells predominated over stromal cells. The melanoma samples were classified into 4 groups based on anatomic location and degree of exposure to sun as indicated by the presence or absence of solar elastosis in the dermis surrounding the melanomas as previously described (Maldonado et al., 2003). In total we examined 103 primary melanomas: from skin with (18) and without (20) chronic sun-induced damage (CSD and non-CSD, respectively), mucosal membranes (37), as well as melanomas from palms, soles, and subungual (acral) sites (28). The study was approved by the Institutional Review Board of the University of California, San Francisco.
The entire 951-bp coding region of MC1R was amplified from tumor DNA in five overlapping segments by PCR and sequenced as described previously (Landi et al., 2006). Sequence alterations in MC1R that were identical to common MC1R alleles found in the germline were considered germline alterations. For all samples in which uncommon or previously unreported sequence alterations were found, we obtained normal DNA by micro- dissecting the uninvolved skin adjacent to the melanoma tissue.
To assess the DNA copy number status of MC1R, we examined the array-based comparative genomic hybridization (CGH) data that was available from prior studies (Curtin et al., 2005, 2006) for 93 of the samples. The MC1R locus was deemed to have an altered number of copies if the clone closest to MC1R at chromosome position 16q24.3 (RP11-7D23) along with either of the two flanking clones RP11-122P17and PAC191P24 (telomere) had a contiguous change in the number of copies. Specifically, a log2 ratio of <−0.30 was defined as a loss and a log2 ratio of >0.30 was defined as a gain, as previously reported (Curtin et al., 2006). BRAF mutation status of the samples was available from prior studies (Curtin et al., 2005; Maldonado et al., 2003). The Mann–Whitney U-test was used to assess interactions between MC1R and BRAF where MC1R genotypes were ranked ascending according to predicted MC1R loss of function as follows, 1 = wt/wt, 2 = r/wt, 3 = R/wt, 4 = rr, 5 = rR and 6 = RR.
We thank Susan Charzan for excellent technical assistance. This article was supported by grants from the National Cancer Institute (PO1 CA025874-25-A1).