A Two-stage Case-Control Study of EGFR Polymorphisms and Breast Cancer Risk

The incidence of breast cancer has steadily increased with an estimated 1 million new diagnoses and 400,000 deaths per year worldwide (1). Only a small fraction of breast cancer cases can be explained by rare, highly penetrant mutations in genes such as BRCA1 and BRCA2, suggesting that common genetic variants with low-penetrance may contribute to the susceptibility to breast cancer in a large number of cases (2).

The EGFR gene is overexpressed in several carcinomas (3), and it has been shown that a change at the transcriptional level of the gene may be associated with epidermal growth factor receptor overexpression in breast cancer cells (4). Several polymorphisms in the EGFR gene have been documented in the public databases, although the functional effects of these polymorphisms have not yet been fully clarified. We evaluated 50 single nucleotide polymorphisms (SNP) in the EGFR gene in relation to breast cancer risk in a two-stage population-based case-control study conducted among Chinese women.

A two-stage case-control study design was used to comprehensively evaluate EGFR polymorphisms in relation to breast cancer risk. Promising associations identified in stage I were validated in stage II with an independent study population. Study subjects were participants of the Shanghai Breast Cancer Study, a population-based case-control study among women in urban Shanghai that has been described previously (5, 6).

Of eligible participants, 1,459 (91.1%) cases and 1,556 (90.3%) controls from stage I, and 1,989 cases (83.7%) and 1,989 controls (70.4%) from stage II completed in-person interviews. In stage I, 1,193 cases (81.8%) and 1,310 controls (84.2%) donated blood samples, with 1,062 cases and 1,069 controls successfully genotyped. In stage II, 1,932 (97.1%) cases and 1,857 (93.4%) controls donated either blood or exfoliated buccal cell samples, with 1,925 cases and 1,847 controls successfully genotyped. Our study obtained Institutional Review Board approval from the relevant review boards in both China and United States, and informed consent was obtained from all subjects.

Han Chinese data from the HapMap Project (7) were used to select Haplotype tagging SNPs using the Tagger program (8). SNPs captured showed an r² of ≥0.90, a minor allele frequency of ≥0.05, and were located in the EGFR or within the 5-kb region flanking the gene. Known or potentially functional SNPs were forced into the Haplotype tagging SNP selection process. SNP selection was completed in December, 2005. Fifty-three SNPs were selected, but the design of the assay for two SNPs (rs4947972 and rs13234622) failed. They were replaced by SNPs rs3735064 and rs6964705, respectively, which were in high LD with the original SNPs. These SNPs were genotyped in subjects included in stage I. Two SNPs (rs1357973 and rs2740763) showed unreliable genotyping data and were excluded from data analyses. Twelve SNPs that showed an association with breast cancer risk at a P value of ≤0.05 in stage I were chosen for validation in stage II. However, two SNPs (rs11766798 and rs2241055) showed an r² of ≥0.80 with other SNPs and were not evaluated further. Therefore, 10 SNPs were selected for stage II validation.

The genotyping assay for stage I was done through the Targeted Genotyping System (Affymetrix) using an advanced Molecular Inversion Probe method (9) as part of a large-scale genotyping effort. Blinded duplicated quality control samples (n = 39) were included with the stage I genotyping. Of the 51 SNPs analyzed, 47 SNPs showed a consistency rate of 100%, whereas the other 4 SNPs showed consistency rates of 98.0%.

In stage II, 10 SNPs were genotyped using the TaqMan allelic discrimination assay in the ABI PRISM 7900 Sequence Detection System (Applied Biosystems). Duplicated quality control samples showed average consistency rate of 98.4%, ranging from 95% to 100%.

Hardy-Weinberg equilibrium was tested by comparing the observed and expected genotype frequencies (χ² test) for the cases and controls separately. Odds ratios and corresponding 95% confidence intervals (CI) were determined by logistic regression analyses using additive, dominant, and recessive models. Covariates considered included age, age at menarche, age at first live birth among parous women, age at menopause among postmenopausal women, number of pregnancies, menopausal status, family history of breast cancer, and leisure physical activity. Tests for trend were conducted by coding the number of variant alleles (0, 1, and 2) and reporting the P value for the β coefficient from the logistic regression models.

The distribution of selected demographic and major risk factors for breast cancer in the Shanghai Breast Cancer Study have been previously reported (6). Participants in the two stages were comparable with the exceptions of age and regular physical activity. Cases tended to have earlier age at menarche, older age at menopause, older age at first live birth, higher proportion of first-degree relatives with breast cancer, prior history of fibroadenomas, higher body mass index or waist-to-hip ratio, and less regular physical activity (data not shown).

Descriptive information, regarding the 50 SNPs evaluated in our study, is shown in Table 1. Ten SNPs were found to be significantly associated with an altered risk of breast cancer in stage I (P ≤ 0.05) in at least 1 of the 3 genetic models evaluated (Table 2). However, these associations were not replicated in stage II at a significance level of ≤0.05. Although the variant allele of rs9642391 (C/G) was associated with decreased breast cancer risk in stage II, the direction of association was opposite that observed in stage I. There was no evidence of a significant interaction between any of the 10 SNPs and hormone-related factors on breast cancer risk (data not shown).

In a two-stage case-control study, we did not find any evidence for an apparent association of common genetic polymorphisms in EGFR with breast cancer risk. Our study is the first to comprehensively evaluate common genetic variants in the EGFR gene on breast cancer susceptibility. Our study had sufficient power of >90% in each stage to detect an odds ratio of ≥1.3 (assuming dominant effect, with minor allele frequency of 0.05 and α = 0.05).

Recently, several studies have implicated EGFR SNPs in carcinogenesis. Liu et al. (10) reported that the −216 G/T (rs712829) polymorphism, located in a Sp1 recognition site of the EGFR promoter, increases promoter activity by 30%. Additionally, this polymorphism was associated with a significantly increased risk of glioblastoma in a European Caucasian population (11). However, this −216 G/T polymorphism was not included in this study because the frequency of the minor allele was low (<5%) in Han Chinese (7). Choi et al. (12) reported that the 181946C>T (rs2293347) polymorphism was associated with an altered lung cancer risk in Koreans. In our breast cancer study, this polymorphism was not significantly associated with breast cancer risk.

In addition to SNPs, a (CA)n dinucleotide polymorphism in an intron of EGFR has been shown to affect the basal transcriptional activity of the EGFR gene (13). Brandt et al. (14) reported that subjects having alleles with 19 or more CA repeats showed an increased risk of breast cancer risk in young women. For lung cancer, a case-control study observed an inverse relationship with (CA)n repeat polymorphism in a Caucasian population (15), whereas no association was seen in a Korean population (16). This polymorphism, however, was not included in our study.

Functional significance of a number of genetic variants in the EGFR gene and interethnic differences of these polymorphisms has not been fully elucidated. None of the 10 SNPs selected from stage I were replicated in a stage II, suggesting that the significant associations observed in stage I were likely due to chance finding. Our study underscores the importance of a two-stage study to reduce type I errors in reporting study results. We cannot rule out the possibility that other types of genetic variants that were not evaluated in our study may be associated with breast cancer risk. Future studies are warranted to confirm our findings in other ethnic populations.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank the participants and research staff of the Shanghai Breast Cancer Study for their contributions and commitment to this project, Regina Courtney and Qing Wang for genotyping, and Brandy Sue Venuti for clerical support in the preparation of this manuscript. Stage 1 genotyping was conducted at the Vanderbilt Microarray Shared Resource, which is supported by the Vanderbilt Ingram Cancer Center (P30 CA68485).