1535 Background: While about 10% of cutaneous melanoma (CM) clusters in families, known high-risk loci explain not more than 40% of expected inherited risk. Besides the most frequently mutated genes in FM (e.g. CDKN2A), it is estimated that the remaining 60% of FM susceptibility is due to the interaction of environment with specific pools of rare known loci and yet unknown high-risk genes. In our study, we report the discoveries of novel germline genetic risk factors in FM in a recently developed FM cohort at New York University Langone Medical Center (NYULMC) consisting of CM and multiple primary melanomas (MPM) of Ashkenazi Jewish (AJ) and non-AJ European ancestries. Methods: As part of an ongoing ascertainment of FM at NYULMC, we assessed the status of CDKN2A mutations using Sanger sequencing, examining the coding regions of 47 AJ FM families and 81 non-AJ FM kindreds. In high-risk mutation-negative families, we applied whole-exome sequencing (WXS) and an innovative hot-spot mutational analysis of non-coding regions to identify novel high-risk loci associated with FM susceptibility. Results: We found that frequencies of CDKN2A deleterious mutations in our FM cohort (13%) are comparable with observations from previous studies. We have also identified a specific CDKN2A coding mutation in FM kindreds of AJ ancestry, which is particularly interesting as CDKN2A mutations in AJ cohorts have been sparsely reported in prior studies. The WXS/targeted non-coding sequencing of mutation-negative families identified putatively deleterious mutations in regulatory regions in the vicinity of several novel loci, including SMAD4 and PAX8, co-segregating in FM kindreds. Conclusions: Our unique FM ascertainment, including > 50% AJ kindreds, provides an excellent platform for mapping high-risk genetic susceptibility in FM. Novel deleterious mutations identified in non-coding regulatory regions of SMAD4 and PAX8 genes, some with increased frequency in AJ families, suggest a need for a more thorough investigation of the non-coding genome using a founder FM population, as we propose here. As our ongoing ascertainment expands, we are pursuing validation of our observations through comprehensive sequencing efforts.
