Abstract 31: The association between GWAS SNPs and mammographic density in Norwegian postmenopausal women stratified by nongenetic factors.

Abstract

Background: Mammographic density (MD) is one of the strongest known breast cancer risk factors. Dense breast tissue is characterized by increased stromal tissue and possibly by increased numbers of breast epithelial cells. Twin studies have suggested that part of the variation in MD is determined by genetic factors, however, it is also clear that non-genetic factors such as hormone use, body weight and reproductive factors play a role. Exactly which genetic factors determine MD is not clear, although several genome-wide association studies (GWAS) conducted over the past few years have suggested a number of variants. We hypothesized that the effect of genetic variants on MD may be strongly modified by non-genetic factors. We therefore evaluated the association between percent MD and 17 single nucleotide polymorphisms (SNPs) identified from several published GWAS of MD/breast cancer by strata defined by non-genetic factors.

Methods: We assessed MD on 2036 postmenopausal women aged 50-69 years who participated in the Norwegian Breast Cancer Screening Program (NBCSP) in 2004 using a computer assisted method (Madena, University of Southern California). We used linear regression (additive model) to determine the association between each SNP and MD, adjusting for age and body mass index (BMI). The postmenopausal women were stratified on postmenopausal hormone therapy (HT) use, parity, BMI, and alcohol intake.

Results: In all postmenopausal women a variant (rs29815829) in the FGFR2 gene and a variant (rs3734805) in the 6q25.1 gene associated significantly with MD (regression coefficient (beta)=-1.42, p=0.0306 and beta=0.74, p=0.0390). The same variants showed significant associations in never HT users with slightly stronger effect estimates, but not in current combined estrogen and progestin therapy users.

We also observed some possible effect modification by parity. In all postmenopausal parous women as well as in parous never HT users, the 6q25.1 and the FGFR2 variants were the most important (lowest p-values). However, in nulliparous never HT users a variant (rs3803662) in the TOX3 gene and a variant (rs1045485) in the CASP8 gene associated significantly with MD (beta=1.41, p=0.0120 and beta= 1.42, p=0.0275). There were no apparent differences between alcohol drinkers or non-drinkers in never HT users, regardless of parity. When examining strata of BMI among never HT users, we found a significant association with MD in the CASP8 gene (rs1746827) in women with a BMI < 25 (beta=0.81, p=0.0507) and in the FGFR2 gene (rs2981582) in women with a BMI >25 <29 (beta=-2.34; p=0.0284).

Conclusion: We found some evidence that the associations between variants in the 6q25.1 gene and in the FGFR2 gene and MD were modified by HT use. Further, our results suggest that the genetic determinants of MD may differ between parous and nulliparous women, and across BMI strata.

Citation Format: Merete Ellingjord-Dale, Isabel dos Santos Silva, Eunjung Lee, Elisabeth Couto, David Van Den Berg, Solveig Hofvind, Tom Grotmol, Giske Ursin. The association between GWAS SNPs and mammographic density in Norwegian postmenopausal women stratified by nongenetic factors. [abstract]. In: Proceedings of the AACR Special Conference on Post-GWAS Horizons in Molecular Epidemiology: Digging Deeper into the Environment; 2012 Nov 11-14; Hollywood, FL. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2012;21(11 Suppl):Abstract nr 31.