Despite considerable research, the relationship between cigarette smoking and breast cancer remains controversial. The Carolina Breast Cancer Study (CBCS) and the Carcinoma In Situ (CIS) Study are population-based, case-control studies of invasive breast cancer and carcinoma in situ in North Carolina. Half of the subjects are white and half are African American. In the CBCS, we have recently found that cigarette smoking influences both the prevalence and spectrum of p53 mutations in breast tumors. Most notably, current smokers were twice as likely to harbor p53 mutations (OR=2.11; 95% C1=1.17-3.78), over 3 times as likely to harbor transversions (OR=3.37; 95% C1=1.03-11.0) and over 10 times as likely to harbor G:C--)T:A transversions (OR=10.53; 95% CI= 1.77-62.5) compared to never smokers. Case-control analyses support our case-case studies, and indicate that current smoking (OR=7.30; 95% C1=1.38-38.5) and long (>20 year) duration of smoking (OR=8.65; 95% C1=1.60-46.8) are significantly associated with increased risk of breast cancers carrying G:C->T:A transversions. These and other results clearly demonstrate that the genotoxic effect of smoking is manifested as a p53 mutational fingerprint in the breast tumors of smokers, and implicate smoking in breast cancer development. In the proposed study, we will evaluate 800 invasive and 200 in situ breast tumors from the CBCS and CIS studies for LOH at a series of chromosomal loci that are frequently deleted in both lung cancer and early forms of breast cancer. Furthermore, several of these loci were previously found to undergo preferential deletion in the lung tumors of smokers; these loci include FHIT (3p14.2), 3p21.3, 9p21, 8p21-23 and p53 (17p13). Additional loci that undergo LOH in breast cancer, but which have no specific relationship with smoking will also be evaluated, including 17q21 (BRCA1), 13q12 (BRCA2), and 16q22.1 (E-cadherin). Genetic variation in genes involved in carcinogen metabolism or DNA repair may influence susceptibility to tobacco smoke carcinogens, and thus the formation of somatic alterations in tumors. We will first compare the prevalence of LOH, both overall and at individual loci, in smokers and nonsmokers, and the relationship between LOH and the occurrence of p53 mutations in breast tumors. As a more exploratory aim, we will then determine whether common germline variants of the GST and NAT metabolizing enzymes, or of DNA repair enzymes (XRCC1, APE, HOGG1, XPA, XPD, XPF, and others), modify the prevalence of LOH or p53 mutations associated with smoking. The characterization of subsets of breast cancer defined by somatic genetic alterations such as LOH or p53 mutation, together with analysis of germline polymorphisms and tobacco smoke exposure should provide a comprehensive approach for evaluating the effects of smoking in breast cancer. If breast cancer can be attributed, in part, to cigarette smoking, then smoking prevention or cessation should reduce exposure and thus decrease the incidence of breast cancer.
