| dc.description.abstract | One of the well-established risk factors for breast cancer (BC) is high level of endogenous estrogen, which could be influenced by both genetic and non-genetic factors. Few studies have investigated estrogen related gene-environment interactions, and most of the available ones were extremely underpowered given the available sample size and the design of single-genetic marker analyses. The main objective of this study is to evaluate whether the genetically predicted expression levels of estrogen synthesis, metabolism and bioavailability pathway genes may modulate the association of adiposity, hormone replacement therapy (HRT) use, oral contraceptive intake, and menstrual and reproductive factors with BC risk. Gene expression levels were predicted for each individual in the Breast Cancer Association Consortium using prediction models we built using data from the Genotype-Tissue Expression based on breast, liver, ovary, adipose and cross tissue, separately. A total of 97,694 cases and 87,560 controls were included for the interaction analyses, which were assessed one at a time using Wald test with adjustment for study site or country and the top ten principal components. A total of 24 genes showed a significant interaction with non-genetic factors at P < 0.05 level. If using P < 0.0005 as the cutoff, the expression level of AKR1B15 was observed to be a potential effect modifier for the association between HRT use and BC risk (P-interact = 0.003, β = 0.095 in subcutaneous adipose tissue, and P-interact = 0.004, β = 0.347 in cross tissue). This interaction was attenuated and only reached board line significance if analyses were restricted to post-menopausal women (P-interact = 0.07, β = 0.070 in subcutaneous adipose tissue, and P-interact = 0.05, β = 0.289 in cross tissue). A significant interaction was also observed between HSD17B11 and age at the first full time pregnancy (P-interact = 0.001, β = -0.010 in subcutaneous adipose tissue). This interaction, however, was only observed among post-menopausal women (P-interact = 6.87*10-5, β = -0.016). Our results provide preliminary support that estrogen related genes interact with non-genetic exposure on BC risk, which warrants further investigation. | |
