A study of familial breast cancer: Identifying additional breast cancer susceptibility loci

Breast cancer is a serious public health concern and despite intensive research, the etiology of breast cancer is poorly understood. Known risk factors explain only a small proportion of breast cancer, and yet familial aggregation, recent segregation analyses, and studies of low penetrance genes suggest that additional breast cancer susceptibility loci exist and perhaps act in a polygenic fashion. Here three strategies and their fulfillment, to aid in identification of additional breast cancer predisposition loci, are described. The first strategy, identification of a new breast cancer phenotype, was proposed to identify a homogenous subgroup of cases that could be clustered into a single gene or a minimal number of genes disorder. We show that the histological subtype lobular breast cancer, identified through a familial aggregation study, increases risk of breast cancer and specifically lobular breast cancer in lobular cancer families compared to risk associated with any type of breast cancer. Hence, study of lobular breast cancer families may be useful in locating additional breast cancer genes. Second, novel association methods that are adaptable to large pedigree resources, qualitative or quantitative data types, and the ability to perform a variety of statistical tests, correcting for correlations between related individuals, was proposed. This strategy was fulfilled with the development of the pedigree-based association tool, PedGenie. Lastly, strategies to improve the reproducibility of results and study design including use of tagging single nucleotide polymorphisms (tSNPs) to identify underlying haplotype and linkage disequilibrium structure, and control of potential population stratification were proposed. These concepts were incorporated into a study of five candidate genes involved in the DNA repair pathway. Linkage disequilibrium structure and identification of tSNPs for each gene are reported. For the gene XRCC4, we show that two 4-locus haplotypes were significantly associated with age at diagnosis of breast cancer (one earlier diagnosis and one later diagnosis) and that two 2-locus haplotypes were significantly associated with breast cancer risk. Each proposed strategy and its fulfillment in the Ph.D. dissertation aid in a small measure to the identification of additional breast cancer loci, which ultimately will benefit women worldwide.