Abstract A84: Reduced LINE-1 methylation in peripheral blood mononuclear cells is associated with risk of melanoma in families segregating CDKN2A germline mutations

Abstract

Abstract Background: CDKN2A is a major susceptibility gene for familial melanoma, but its incomplete penetrance suggests that other factors may modify the effect of this gene in melanoma-prone families. Epigenetic changes involving reduced levels of global DNA methylation in blood have been associated with genomic instability and cancer risk. In addition, genetic background has the potential to influence epigenetic processes through cis- or trans-regulatory variation. Therefore, we hypothesize that the CDKN2A carrier mutation state may be an important factor influencing global DNA methylation in predisposed individuals and that aberrant methylation may be an indicator of modified risk in individuals from melanoma-prone families segregating CDKN2A germline mutations. Methods: We measured the overall level of genomic DNA methylation using bisulfite pyrosequencing at four CpG sites (CpG1, CpG2, CpG3 and CpG4) of the Long Interspersed Nucleotide Element-1 (LINE-1) sequences in peripheral blood mononuclear cells (PBMCs) from 42 CDKN2A mutation positive cases (20 males and 22 females) and 73 controls (29 CDKN2A carrier controls [9 males and 20 females] and 44 non-carrier controls [19 males and 25 females]) in 26 families with CDKN2A mutations. LINE-1 methylation was reported as the average (of all 4 CpGs) and as site-specific percentage methylation. Odds' ratios (OR) and P-values were obtained by comparing melanoma cases to controls using unconditional logistic regression with adjustment for family correlation (using the generalized estimating equations [GEE] approach), age, sex, number of moles and with melanoma status as the outcome variable. Percentage LINE-1 methylation was categorized using tertiles with the highest tertile as the reference. Results: Male gender was significantly associated with higher average LINE-1 methylation (P=<0.001) and methylation at all CpG sites (P=<0.03). In contrast, age, solar injury, dysplastic nevi (DN) and number of moles were not associated with average LINE-1 methylation. Although mean levels of LINE-1 methylation were similar in melanoma cases, carrier controls, and non-carrier controls, after adjusting for family correlation, age, sex and number of moles, low average LINE-1 methylation (tertile 1) was significantly associated with increased risk of melanoma (OR=3.87, 95% confidence interval [CI]: 1.11–13.5, P=0.033, compared to all controls; OR=1.83, 95% CI: 1.09–11.7, P=0.035, compared to carrier controls; and OR=3.43, 95% CI: 1.17–10.0, P= 0.024, compared to non-carrier controls). This association was observed for most CpG sites with the most significant association seen for site 3 and risk of melanoma compared to all controls (OR=5.18, 95% CI: 1.54–17.4, P=0.008). We observed no significant association between average or CpG site-specific LINE-1 methylation and CDKN2A mutation status among controls although the comparison was limited by small sample size. Conclusions: Melanoma cases from families segregating CDKN2A germline mutations have significantly lower LINE-1 global methylation in their PBMCs compared to controls. Changes in LINE-1 methylation in PBMCs from CDKN2A mutation positive cases may reflect a potential genetic background influence on global DNA methylation in these individuals. Further studies are needed to clarify these preliminary observations. Citation Information: Cancer Prev Res 2011;4(10 Suppl):A84.