Polymorphisms associated with circulating sex hormone levels in postmenopausal women

Abstract

Background: Reports suggest a relationship between circulating sex hormone levels and breast cancer risk, but genetic association studies have been inconclusive. We investigated the association between levels of sex hormones and single nucleotide polymorphisms (SNPs) in genes coding for the enzymes that regulate them. Methods: We assayed circulating levels of estradiol, testosterone, estrone, androstenedione, 17α-hydroxyprogesterone, and sex hormone-binding globulin (SHBG) in 1975 normal postmenopausal women. Fifteen SNPs in the CYP17, CYP19, EDH17B2, SHBG, COMT, and CYP1B1 genes were genotyped in these postmenopausal women and in a breast cancer case-control study. Associations of genotypes with breast cancer risk were evaluated in the case-control study and with hormone levels in the postmenopausal women using multiple linear regression with assay batch, body mass index, parity, peri- or postmenopausal status, and age band as covariates. Results: CYP19 SNPs (rs10046 and [TCT]+/-) were associated with differences in estradiol level (P = .0006 and P = .0003, respectively) and the estradiol:testosterone ratio (P = .000001 and P = .002). SNP rs10046 explained 1.6% of the variance (r2) in the estradiol:testosterone ratio. SHBG SNPs (5′ untranslated region [5′UTR] g-a and D356N) were associated with both SHBG levels (p<10-6 and P = .005) and the estradiol:SHBG ratio (P = .000008 and P = .01). These SNPs explained 2.4% and 0.6% of the variance in SHBG levels, respectively. SNPs in the other genes were not associated with differences in any hormone levels, and none were statistically significantly associated with breast cancer risk. Conclusion: Genetic variation in CYP19 and SHBG contributes to variance in circulating hormone levels between postmenopausal women, but low r2 values may explain why these genes have given inconclusive results in breast cancer case-control studies. © Oxford University Press 2004, all rights reserved.