Multi-Variant Pathway Association Analysis Reveals the Importance of Genetic Determinants of Estrogen Metabolism in Breast and Endometrial Cancer Susceptibility

Despite the central role of estrogen exposure in breast and endometrial cancer development and numerous studies of genes in the estrogen metabolic pathway, polymorphisms within the pathway have not been consistently associated with these cancers. We posit that this is due to the complexity of multiple weak genetic effects within the metabolic pathway that can only be effectively detected through multi-variant analysis. We conducted a comprehensive association analysis of the estrogen metabolic pathway by interrogating 239 tagSNPs within 35 genes of the pathway in three tumor samples. The discovery sample consisted of 1,596 breast cancer cases, 719 endometrial cancer cases, and 1,730 controls from Sweden; and the validation sample included 2,245 breast cancer cases and 1,287 controls from Finland. We performed admixture maximum likelihood (AML)-based global tests to evaluate the cumulative effect from multiple SNPs within the whole metabolic pathway and three sub-pathways for androgen synthesis, androgen-to-estrogen conversion, and estrogen removal. In the discovery sample, although no single polymorphism was significant after correction for multiple testing, the pathway-based AML global test suggested association with both breast (rho(global) = 0.034) and endometrial (rho(global) = 0.052) cancers. Further testing revealed the association to be focused on polymorphisms within the androgen-to-estrogen conversion sub-pathway, for both breast (rho(global) = 0.008) and endometrial cancer (rho(global) = 0.014). The sub-pathway association was validated in the Finnish sample of breast cancer (rho(global) = 0.015). Further tumor subtype analysis demonstrated that the association of the androgen-to-estrogen conversion sub-pathway was confined to postmenopausal women with sporadic estrogen receptor positive tumors (rho(global) = 0.0003). Gene-based AML analysis suggested CYP19A1 and UGT2B4 to be the major players within the sub-pathway. Our study indicates that the composite