Published OnlineFirst April 20, 2010; DOI: 10.1158/1055-9965.EPI-09-1250 Cancer Research Article Epidemiology, Biomarkers & Prevention Polymorphisms in the TOX3/LOC643714 Locus and Risk of Breast Cancer in African-American Women Edward A. Ruiz-Narváez1,2, Lynn Rosenberg1,2, Yvette C. Cozier1,2, L. Adrienne Cupples3, Lucile L. Adams-Campbell4, and Julie R. Palmer1,2 Abstract Background: The rs3803662 single nucleotide polymorphism (SNP) in the TOX3/LOC643714 region was identified as a breast cancer susceptibility genetic variant in recent genome-wide association studies of women of European ancestry and has been replicated in other populations of European ancestry. The position of the causal variant tagged by the rs3803662 marker is still unknown. In fact, because the rs3803662 polymorphism is located between the TOX3 and the LOC643714 loci, it is unclear which gene is the one causally related to the risk of breast cancer. Because linkage disequilibrium blocks are smaller in populations of African ancestry, fine- mapping in African ancestry samples might be an effective approach to narrowing the position of the causal variant(s) in the TOX3/LOC643714 locus. Methods: We evaluated a total of 60 tagging SNPs throughout the TOX3/LOC643714 region in a nested case-control study of breast cancer within the Black Women's Health Study, which included 906 cases and 1,111 controls. Results: No significant association was found for the rs3803662 SNP. However, four other SNPs (rs3104746, rs3112562, rs3104793, and rs8046994), all of them located in the LOC643714 gene, were associ- ated with risk of breast cancer. The strongest association was observed for rs3104746: each copy of the A-rs3104746 allele was associated with a 23% higher risk of breast cancer (odds ratios, 1.23; 95% confidence intervals, 1.05-1.44; P = 0.009). Conclusions: Our results confirm the association observed in genome-wide association studies of European ancestry populations. Impact: The results narrow the locus to a smaller linkage disequilibrium block in the LOC643714 gene. Cancer Epidemiol Biomarkers Prev; 19(5); 1320–7. ©2010 AACR. Introduction populations, showed that the risk conferred by the rs3803662 polymorphism was either restricted to or The TOX3/LOC643714 locus on chromosome 16 was more strongly associated with estrogen receptor (ER)– one of the first breast cancer regions to be identified positive tumors compared with ER-negative cancers through genome-wide association studies in popula- (2, 3). Because most of the replication studies have tionsofEuropeanandEastAsianorigin(1).Outof been carried out in populations of European ancestry several single nucleotide polymorphisms (SNP) associ- (2-4), it is unclear whether the same SNP is associated ated with the risk of breast cancer, the rs3803662 (a C-to-T with risk of breast cancer in populations of African transition) was the most strongly correlated with dis- origin. In a subgroup of African American women ease; each copy of the T allele of the rs3803662 SNP (422 cases and 447 controls) from the Multiethnic was associated with a 20% increase in the risk of breast Cohort study, the T-rs3803662 allele was associated cancer. Subsequent studies, also in European ancestry with a lower risk of breast cancer, the opposite direc- tion from the results in the other ethnic groups (2). In Authors' Affiliations: 1Slone Epidemiology Center at Boston University, addition, a recent analysis of African American women Departments of 2Epidemiology and 3Biostatistics, Boston University (810 cases and 1,784 controls) from the Southern Com- School of Public Health, Boston, Massachusetts; and 4Georgetown munity Cohort Study and the Nashville Breast Health University Medical Center, Washington, DC Study, found no significant association between Note: Supplementary data for this article are available at Cancer Epidemi- TOX3 ology Biomarkers and Prevention Online (http://cebp.aacrjournals.org/). rs3803662 and seven other SNPs in the gene and risk of breast cancer (5). Corresponding Author: Edward A. Ruiz-Narvaez, ScD, Slone Epidemi- ology Center at Boston. University, 1010 Commonwealth Avenue, Bos- The reasons for the differences in results between Af- ton MA 02215. Phone: 617-734-6006, Fax: 617-738-5119. E-mail: rican American women and women of European or [email protected] Asian ancestry remain to be determined. The high allele doi: 10.1158/1055-9965.EPI-09-1250 frequency of the T-rs3803662 allele in African Ameri- ©2010 American Association for Cancer Research. cans (∼50%) as well as the considerable sample size 1320 Cancer Epidemiol Biomarkers Prev; 19(5) May 2010 Downloaded from cebp.aacrjournals.org on September 25, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst April 20, 2010; DOI: 10.1158/1055-9965.EPI-09-1250 TOX3/LOC643714 Locus and Breast Cancer Figure 1. LD structure of the TOX3/LOC643714 region in HapMap CEU and YRI populations. LD structure was assessed using LD units (LDU) estimated with the SNPBrowser software (25). LD units are proportional to the decay of LD along the physical distance. Regions of constant (“flat”) LD units represent LD blocks. Note that the big block (from 51.09 to 51.18 Mb) that covers parts of the TOX3 and LOC643714 genes in the HapMap CEU population and it is split into smaller blocks and gaps in the HapMap YRI population. of the Southern Community Cohort Study/Nashville Design and Methods Breast Health Study (810 cases and 1,784 controls) makes low statistical power an unlikely explanation. Study population. We conducted a nested case-control Difference in the linkage disequilibrium (LD) structure study within the ongoing Black Women's Health Study in the TOX3/LOC643714 region between populations (BWHS) that has been described elsewhere (6). Briefly, of European and African origin is a more likely reason the study began in 1995 when women 21 to 69 years of (Fig. 1). In the HapMap CEU population, the rs3803662 age from across the United States completed a 14-page SNP resides in a LD block that is more than 80 kb and postal health questionnaire. The initial cohort comprises covers part of the TOX3 and LOC643714 loci. In the 59,000 women who self-identified as “black” and had a HapMap YRI population, this big LD block was split valid address. Follow-up questionnaires are sent every into smaller blocks and included some gaps of low 2 years. Average follow-up of the baseline cohort through LD. In particular, the rs3803662 SNP is located inside the completed 2-year cycles to date is greater than 80%. a small 4 kb LD block in HapMap Yoruba samples. It We used medical records and cancer registry data to is therefore possible that the causal variant(s) tagged by confirm self-reported cases of breast cancer, as well as rs3803662 SNP in populations of European ancestry will to gain information on tumor characteristics such as es- be tagged by different SNPs in populations of African trogen and progesterone receptor (PR) status. We have ancestry. If we assume the same causal polymorphisms obtained records or registry data for 1,151 breast cancer in Europeans and Africans, the smaller LD blocks ob- cases reported on the BWHS questionnaires, of which served in the latter allows a more accurate mapping 99.4% were confirmed. Self-reported cases that were not of the causal variants. confirmed have been excluded. We conducted a nested case-control study to narrow DNA samples were obtained from BWHS participants the position of the causal variant for breast cancer in by the mouthwash-swish method (7), with all samples the TOX3/LOC643714 region. To this end, we performed stored in freezers at −80°C. Approximately 50% of parti- fine-scale mapping of the entire TOX3/LOC643714 locus cipants (27,800 women) provided a sample. Women who including genotyping of the index rs3803662 SNP. provided samples were slightly older than women who www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 19(5) May 2010 1321 Downloaded from cebp.aacrjournals.org on September 25, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst April 20, 2010; DOI: 10.1158/1055-9965.EPI-09-1250 Ruiz-Narváez et al. did not, but the two groups were similar with regard to was obtained among the blinded duplicates. All SNPs educational level, geographic region of residence, body with a calling rate of <90% or a deviation from Hardy- mass index, and family history of breast cancer. Weinberg equilibrium in the control sample of P < 0.001 The present study includes all cases of breast cancer were excluded. We also excluded samples with calling who provided a DNA sample and were diagnosed rates of <80%. We first genotyped 41 tagging SNPs cho- through the end of the 2007 follow-up cycle. Controls sen for the TOX3 gene among 769 cases and 833 controls. were selected from among BWHS participants with After the addition of breast cancer cases and controls DNA samples who were free of breast cancer at the identified after the first genotyping plates were made, end of the 2007 follow-up period. Controls were we genotyped 27 tagging SNPs along the LOC643714 lo- matched to cases approximately 1:1 on year of birth cus in 865 cases and 1,073 controls. We successfully gen- (±1 year), and geographic region of residence (North- otyped 35 of the TOX3 variants, 25 of the LOC643714 east, South, Midwest, and West). The study protocol variants, and 29 of the AIMs. Mean call rate in the final was approved by the Institutional Review Board of data set for both SNPs and samples was 98.4%. Boston University. Data analysis. We used PLINK (16) software version Selection of tag SNPs and ancestral informative 1.06 to calculate summary statistics for the genotype data. markers. We downloaded SNPs covering the entire We tested for association with breast cancer using the TOX3/LOC643714 locus from the HapMap Yoruba Cochran-Armitage trend test of an additive genetic model (YRI) database (8).