Hormone metabolism genes and mammographic density in Singapore Chinese women

Abstract

Background

Female steroid hormone levels and exogenous hormone use influence breast cancer risk. We investigated the association between genetic variation in the hormone metabolism and signaling pathway and mammographic density (MD), a strong predictor of breast cancer risk.

Methods

We genotyped 161 SNPs in 15 hormone metabolism pathway gene regions and evaluated MD in 2,038 Singapore Chinese women. Linear regression analysis was used to investigate SNP-MD association. An overall pathway summary was obtained using the adaptive ranked truncated product test.

Results

We did not find any of the individually tested SNPs to be associated with MD after a multiple testing correction. There was no evidence of an overall effect on MD of genetic variation in the hormone metabolism pathway.

Conclusions

In this cross-sectional study, genetic variation in hormone metabolism pathway was not associated with MD in Singapore Chinese women.

Impact

Consistent with existing data from Caucasian populations, polymorphisms in hormone pathway genes are not likely to be strong predictors of MD in Asian women.

Introduction

Mammographic density (MD) is a relative measure of the amount of epithelium and stroma in the breast, and one of the strongest known predictors of breast cancer risk (1). Endogenous sex steroid hormone levels and exogenous hormone therapy (HT) use have been associated with increased breast cancer risk (2). To date, 18 studies, conducted mainly among Caucasian populations, have described the association between SNPs in one or more hormone pathway genes (reviewed in (3)) and MD and generally reported no associations. In the current study, we investigated the associations between SNPs in hormone metabolism pathway genes and MD among participants of the Singapore Chinese Health Study (SCHS), a population-based prospective study initiated in 1993 in Singapore (4).

Methods

Study participants

Details of the Mammography Subcohort who were women enrolled in both the SCHS and the Singapore Breast Screening Project (SBSP), have been described (4). In brief, we identified 3,777 women common to the SBSP and SCHS databases through a computer linkage, and successfully retrieved mammograms of 3,702 women (98%). Of these, DNA samples were collected from 2,164 women (1,848 blood, 316 buccal) (4). MD was assessed by one of the authors (GU) (4). The Institutional Review Boards at the National University of Singapore, the Singapore National Cancer Center, the University of Southern California and the University of Minnesota had approved this study.

We selected tagging SNPs in the hormone metabolism pathway genes (see Table 1), from 20kb upstream of 5′ untranslated region (UTR) to 10kb downstream of 3′ UTR of each gene. We tagged all common SNPs (minor allele frequency ≥5%) found among non-Hispanic white or Chinese populations, with r²≥0.80. This selection was done using the Snagger software and a custom database of the Hapmap CEU data (release 24) merged with unique SNPs in the Affymetrix 500K panel as well as the Hapmap CHB data release 24 (4).

Table 1.

Number of genotyped SNPs in each hormone pathway gene, and the gene-level and pathway-level summary P-values

	Number of genotyped SNPs	Tagging SNPs selected	Tagging coverage (%)* when using tagging approach; List of selected SNPs when not using tagging approach	Gene-level/Pathway-level summary P-value
Pathway-level summary P-value				0.70
Gene-level summary P-values
AKR1C4	1	No	rs17134592 (Leu311Val)	0.79
AR	2	No	rs5918757, rs1204038	0.82
COMT	29	Yes	75%	0.14
CYP1A1/CYP1A2	13	Yes	81%	0.18
CYP1B1	16	Yes	78%	0.69
CYP17A1	1	No	rs743572	0.72
CYP19A1	2	No	rs727479, rs749292	0.54
ESR1	11	No	rs2077647, rs2295190, rs9479130, rs728524, rs2250122, rs12681, rs3798577, rs1801132, rs9340799, rs1062577, rs3798758	0.49
ESR2	28	Yes	90%	0.18
HSD3B1/HSD3B2	4	No	rs6428830, rs6428828, rs6686779, rs10802107	0.12
HSD17B1	5	Yes	57%	0.37
PGR	32	Yes	86%	0.19
SHBG	10	No	rs6259, rs2955617, rs858518, rs858521, rs858524, rs1624085, rs9898876, rs9913778, rs1642796, rs1619016	0.36
SRD5A2	1	No	rs523349	0.58
SULT1A1/SULT1A2	6	Yes	68%	0.40

^*Hapmap SNPs (Chinese population; release 27) in each gene region captured by the genotyped tagSNPs with minimum pairwise r²≥0.80. Each gene region covers from 20kb upstream of the start of each gene to 10kb downstream of the end of each gene. Full list of tagging SNPs are available upon request.

^¶Tagging SNPs were selected using Snagger (see ref (4) for further description and citation of original paper).

Genotyping of 2,164 samples was performed using the Illumina Golden Gate Assay (Illumina, Inc., San Diego, CA) at the University of Southern California Epigenome Core Facility. We excluded 126 samples whose genotyping success rates were less than 85%. Genotyping concordance based on 42 random duplicate samples was >99.9%. We excluded SNPs with minor allele frequency (MAF) <0.001 or which depart significantly from Hardy-Weinberg equilibrium (P<0.01).

Statistical Analysis

We used linear regression to examine the association between SNPs and MD, adjusting for age at mammogram, body mass index (BMI) (kg/m²) at mammogram, and dialect group (Cantonese, Hokkien). Additional adjustment for other breast cancer risk factors, including parity, menopausal status, and HT use, did not materially change the results; these risk factors were not included in the final model. The linear regression models were based on additive genetic models, in which the regression coefficients are estimates of the difference in MD per copy of the minor allele of a given polymorphism. We tested whether the association between SNPs and MD is modified by established determinants of MD by introducing product terms and conducting Wald tests, adjusting for age, BMI, and dialect. Analyses were conducted using SAS 9.2 (SAS Inc., NC). All P values are two sided.

To summarize the overall statistical significance of the entire set of SNPs under investigation, we applied the adaptive ranked truncated product (ARTP) test of Yu et al. (5).

Results

We genotyped 161 SNPs in 15 hormone metabolism pathway gene regions (Table 1). We did not observe a significant association between SNPs in the hormone metabolism genes and MD in Singapore Chinese women. The overall ARTP test for the set of SNPs investigated was not significant (P=0.70; Table 1). For single SNP analyses, only rs4680 (COMT) had a P value less than 0.01, but this association was not statistically significant after Bonferroni adjustment. There was no evidence of effect modification by parity, menopausal status, or BMI (<25kg/m², ≥ 25kg/m²); however, the sample size for nulliparous (n=133) or premenopausal women (n=216) was limited.

Discussion

Consistent with published data (3), our results do not support hormone-pathway SNPs as genetic determinants of MD in Chinese women. To our knowledge, this is the first study to investigate hormone metabolism pathway genes in association with MD in a large population-based study of Asians.

For SNPs with MAF of 0.2 (average MAF of all tested SNPs), we had 80% power to detect a 3.2% difference in MD per minor allele, with a Bonferroni-corrected type I error rate of 5%. One limitation of the current study is that we were not able to evaluate effect modification by HT (6) due to the very low prevalence of HT use in this population. Another limitation is the incomplete tagging of the selected genes. Nonetheless, the list of hormone-pathway genes and the number of investigated SNPs in the current study is by far the largest among all published studies in Asian population, which examined 2–25 SNPs (3, 7).