DNA methylome of familial breast cancer identifies distinct profiles defined by mutation status

It is now understood that epigenetic alterations occur frequently in sporadic breast carcinogenesis, but little is known about the epigenetic alterations associated with familial breast tumors. We performed genome-wide DNA methylation profiling on familial breast cancers (n = 33) to identify patterns of methylation specific to the different mutation groups (BRCA1, BRCA2 and BRCAx) or intrinsic subtypes of breast cancer (basal, luminal A, luminal B, HER2-amplified and normal-like). We used methylated DNA immunoprecipitation (meDIP) on Affymetrix promoter chips to interrogate methylation profiles across 25,500 distinct transcripts. Using a support vector machine classification algorithm, we demonstrated that genome-wide methylation profiles predicted tumor mutation status with estimated error rates of 19% (BRCA1), 31% (BRCA2) and 36% (BRCAx), but did not accurately predict the intrinsic subtypes defined by gene expression with error rates of 43% (basal) and 54% (luminal A). Furthermore, using unsupervised hierarchical clustering we identified a distinct subgroup of BRCAx tumors defined by methylation profiles. Finally, gene expression profiling and the SNP CGH array previously performed on the same samples allowed full integration of methylation, gene expression and copy number datasets. This integrated analysis revealed frequent hypermethylation of genes that also displayed loss of heterozygosity compared with the tumors that were diploid for that gene. We also observed frequent hypermethylation of genes that show copy number gains compared with diploid tumors providing a potential mechanism for expression dosage compensation. Together these data show that methylation profiles for familial breast cancers are defined by the mutation status and distinct from the intrinsic subtypes.