Various approaches to identify breast cancer genes.

Breast cancer is a major malignant disease affecting females worldwide. It is the most common invasive cancer among Australian women and is one of the two leading causes of cancer-related death for Australian females. After age and gender, the most significant risk factor for breast cancer is a family history of the disease, suggesting the presence of inherited predisposing germline alterations. There are a number of known breast cancer susceptibility genes that confer a moderate to high risk of disease, most of which are associated with very rare cancer predisposing syndromes. Mutations in the hereditary breast cancer susceptibility genes, BRCA1 and BRCA2, are highly penetrant and germline mutations in these genes explain about 30% of the multiple-case breast cancer families. Even in combination with other known breast cancer susceptibility genes of high-, moderate- and low- penetrance, about 65% of the familial aggregation of breast cancer remains unexplained. The overall objective of this study was to use various strategies to identify new breast cancer susceptibility genes and gain further insight into the molecular basis of breast cancer. The first approach was to apply a genome-wide screen for truncating mutations using a technique called gene identification by nonsense-mediated mRNA decay inhibition (GINI) on the lymphoblastoid cell lines (LCLs) of individuals from high-risk multiple-case non-BRCA1/2 (“BRCAx”) breast cancer families. Two different methods of nonsense-mediated mRNA decay (NMD) inhibition were attempted: pharmacological inhibition using caffeine and RNA interference using short interfering RNAs targeting critical components of the NMD pathway. RNA interference proved to be an inefficient method of inhibiting NMD and did not stabilise target genes in the positive control cell lines tested, particularly in LCLs. Pharmacological inhibition of NMD using caffeine was instead applied to the LCLs of members from three selected “BRCAx” families. Gene expression microarray analysis of a total of 144 samples identified several candidate genes per family. However, direct sequencing of the exonic and flanking splice-site regions of each gene did not reveal a nonsense mutation. The second approach to identify additional high-risk genes underlying breast cancer susceptibility was a candidate gene approach that involved mutation screening of two biologically plausible breast cancer genes, RAD51B (RAD51L1) and RAD51C (RAD51L2), in 188 breast cancer families and 190 controls. High resolution melt (HRM) analysis, followed by direct sequencing, identified variants of both RAD51B and RAD51C. However, no nonsense mutations or potentially pathogenic variants were found in either gene. Three novel truncating mutations were identified by collaborators who had screened RAD51C in an additional 1,200 familial breast and ovarian cancer families plus 267 sporadic ovarian cancers. This suggests that highly penetrant mutations in RAD51C, and possibly RAD51B, are likely to exist and influence breast and/or ovarian cancer risk but are extremely rare. In another attempt to identify new breast cancer genes that confer a high risk for the disease, applications were made to tissue banks to receive fresh metastatic tumour tissue from breast cancer patients with no known mutations in either BRCA1 or BRCA2 with the intention of establishing breast cancer cell lines from them, and subjecting them to GINI analysis. As a result of the applications to the tissue banks, fresh metastatic tumour tissue from sporadic breast cancer patients was also received. For one particular patient, there were ample resources to perform molecular characterisation to investigate the changes from germline material to primary tumour to metastasis. Thus, a third objective was added to the study that involved identifying genes involved in breast cancer metastasis to bone. Molecular profiling of the primary tumour was compared to that from the cell line that I established from this bone metastasis (BCBM1). Integration of copy number and gene expression data was used to distinguish putative driver from passenger genes. A number of genes that have previously been implicated in breast cancer metastasis to bone were identified, both within regions of homozygous loss and with no detectable gene expression, as well as within regions of high-level copy number gain accompanied by transcript overexpression. Novel candidate metastasis genes were found also within these regions that warrant further investigation. The results from this thesis provide insight into the challenges that cancer geneticists face with finding the missing heritability for breast cancer. However, the newly established BCBM1 cell line is a good in vitro model of studying breast cancer metastasis to bone and further characterisation of this new cell line is currently being performed in vivo prior to publication and making it publicly available