Genetic Variants in the 53BP1 Gene and Skin Cancer Risk

basal cell carcinoma

squamous cell carcinoma

single nucleotide polymorphism

The conserved p53-binding protein 1 (53BP1) was initially identified as a nuclear protein that interacts with the DNA-binding domain of tumor suppressor p53 and enhances p53-mediated transcription activation (). The interaction region of 53BP1, BRCT (BRCA1 C-terminus) repeats, are present in several proteins involved in DNA repair and cell cycle control, suggesting a direct role of 53BP1 in the cellular response to DNA damage and maintenance of genomic stability (; ). Recently, studies suggested that 53BP1 is a central mediator of the DNA damage checkpoint signaling () and it directly participates in the repair for DNA double-strand breaks (). 53BP1-deficient mice exhibit growth retardation, high-radiation sensitivity, and tumor development features that are indicative of a defective DNA damage response (). 53BP1 has been shown to constitutively have an important role in the etiology of human cancer (). It is plausible that sequence variation in the regulatory and coding regions of the 53BP1 gene might affect its transcription and protein structure thus, its biological function in checkpoint signaling and DNA repair, leads to susceptibility to cancers.

In this study, we hypothesized that that common genetic variants of 53BP1 were associated with risk of skin cancer. We comprehensively surveyed common genetic variation of the 53BP1 gene using two complementary approaches, including putatively functional single nucleotide polymorphisms (SNPs) and choosing tagging-SNPs in the 53BP1 gene locus (introns and exons as well as 20kb upstream and 20kb downstream of the coding region). We performed a skin cancer case-control study of Caucasians nested within the Nurses' Health Study to evaluate whether these common genetic variants are associated with the risks of non-melanoma skin cancers (squamous cell carcinoma (SCC) and basal cell carcinoma (BCC)) along with melanoma risk. We further investigated the association of these genetic variants with pigmentary phenotypes (hair color, skin color, tanning ability, and the number of moles). The nested case-control study consisted of 218 incident melanoma cases, 285 incident SCC cases, 300 incident BCC cases, and 870 age-matched controls. A detailed description of the selection and characteristics of cases and controls was published previously ().

We included four putatively functional SNPs that have been identified in the promoter and coding regions of 53BP1 (http://egp.gs.washington.edu): rs1869258, rs560191, rs689647, and rs2602141. Using genotype data from the 90 CEU samples in the HapMap project, we further selected four tagging-SNPs by the Tagger program (r²>0.8): rs3862138, rs17782975, rs2242069, and rs999047 (Supplementary Figure S1 online and Supplementary Table S1 online). We genotyped the eight SNPs by the 5′ nuclease assay (TaqMan) in 384-well format, using the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA). The distributions of genotypes for the eight SNPs were in Hardy–Weinberg equilibrium among controls. Because the three SNPs (rs1869258, rs560191, and rs2602141) are in high linkage disequilibrium (each pair-wise r²>0.9, Supplementary Table S2 online), we anticipated that these three SNPs would show similar effects on pigmentary phenotypes and skin cancer risk.

We evaluated the main effect of the selected polymorphisms across the three types of skin cancer using unconditional logistic regression (Table 1). In the analyses controlling for the age, we found that three SNPs (rs689647, rs2242069, and rs999047) were consistently associated with a significantly decreased risk of BCC (additive odds ratio and 95% confidence interval: 0.69 (0.50–0.95), 0.68 (0.51–0.90), and 0.66 (0.48–0.91), respectively). These three SNPs are in moderate linkage disequilibrium (pair-wise r² between 0.27 and 0.70, Supplementary Table S2 online). The non-synonymous SNP, rs689647 (G412S), has been investigated in a case-control study of breast cancer but found no association with breast cancer risk (). We did not observe significant association between three putatively functional SNPs (rs1869258, rs560191, and rs2602141) and any type of skin cancer, which was consistent with previous studies (; ; ). We further evaluated haplotypes based on the four SNPs (rs689647, rs2602141, rs2242069, and rs999047) that modified the risk of the cancer subtypes (Table 1). We found that the variant allele of rs689647 was often inherited together with that of the other three SNPs and formed a haplotype in 11% of the population. This haplotype was significantly associated with a decreased risk of BCC (odds ratio, 0.66; 95% confidence interval, 0.45–0.96), suggesting a protective effect on BCC compared with the most frequent haplotype carrying all wild-type alleles in approximately 60% of the population. However, the haplotype carrying only the variant allele of rs2242066 or rs999047 was not significantly associated with BCC risk as suggested in single-SNP analysis, though the latter showed a non-significant decreased risk of BCC. Of note, the last haplotype carrying only the variant allele of rs999047 was associated with a non-significant increased risk of melanoma (odds ratio, 1.97; 95% confidence interval, 0.96–4.07), which was stronger than the association of single-SNP analysis (odds ratio, 1.23; 95% confidence interval, 0.90–1.68). These associations did not change substantially after additional adjustment for potential risk factors including pigmentary phenotypes, family history of skin cancer, geographic region, accumulative sun exposure, sunburns, and sunlamp use or tanning salon attendance. We did not observe any statistically significant interaction between pigmentary phenotypes and genetic variants on skin cancer risk.

We evaluated the associations between the selected SNPs and pigmentary phenotypes among controls using linear regression models (Table 2). We found the three highly correlated putatively functional SNPs (rs1869258, rs560191, and rs2602141) were associated with lighter hair color (P for trend, 0.04, 0.07, and 0.03, respectively) and more moles (P for trend, 0.02, 0.003, and 0.003, respectively). We also found one intronic SNP in 5′ terminus, rs17782975, was significantly associated with more moles (P for trend, 0.01). No statistically significant associations between genetic variants and skin color or tanning ability were observed. To our knowledge, this is the first report that 53BP1 is associated with pigmentary phenotypes in Caucasians, although this gene has shown clear evidence of adaptive skin pigmentation selection in Africans (). Two other genes involved in DNA repair, p53 and MDM2, have been reported to be associated with pigmentary phenotypes (; ).

In summary, we systematically evaluated the common genetic variants of the 53BP1 gene and the risk of skin cancer. We found three SNPs (rs689647, rs2242069, and rs999047) were significantly associated with a decreased risk of BCC. These findings may lead to further replication and functional studies that will elucidate the underlying mechanisms of skin cancer development associated with these genetic variants.