Research Grants Awarded
A Novel Strategy to Find New Genetic Risk Factors for Breast Cancer
In this study, we introduce a novel strategy identify genetic risk factors for breast cancer. The genes BRCA1 and BRCA2 are the best known examples, in that inherited mutations in these genes greatly increase risk. However, after a decade of research it is clear that these genes leave unexplained a large fraction of familial breast cancers, and it is strongly suspected that other risk genes must exist, although efforts to find them have been largely unsuccessful. Here we propose a new strategy to localize such genes, based on the founder effect, i.e. the situation where risk mutations in "unrelated" patients are in identical, having been inherited from a common ancestral carrier. This effect implies the pair actually share a much larger fragment of identical founder DNA surrounding the mutation, and we have developed a powerful method to directly search for such shared DNA fragments between individuals. This allows us to directly search pairs of unrelated, high-risk breast cancer patients for large shared DNA fragments, which then immediately become candidate regions for the risk factors. Such regions typically contain tens to hundreds of candidate genes. The net result is equivalent to a massive linkage study using pedigrees that connect the "unrelated" patients through distant ancestors, but here we only need to analyze the patients in hand, and we do not need to know the underlying pedigree at all, which enormously reduces the cost and complexity of such a linkage study. This "pedigree-free" approach relies on the latest advances in SNP genotyping technology, as well as new analytical methods invented by our group, to detect the shared DNA fragments. In order to enrich for the existence of the essential founder effects, we will conduct the search within a genetic isolate already shown to have strong founder effects for the BRCA genes, the Ashkenazi Jew population. Within this group, we will select for patients enriched for genetic risk factors for breast cancer, based on their family history and age of onset, but who are negative for the known BRCA mutations in this population, and thus more likely to implicate new risk genes. The necessary patient samples will be obtained from existing, well characterized breast cancer biospecimen repositories. In this pilot study, we hope to demonstrate that this new approach is a general, powerful and cost-effective way to broadly screen for new genetic risk loci, and thus candidate genes, which can then be subject to more focused analysis to search for the mutations and identify the specific risk genes. The ultimate value of this approach, if successful, would be to greatly accelerate the discovery of the remaining genetic risk factors for breast cancer, which should collectively be of comparable importance to the well known BRCA genes for understanding the overall genetic risks for breast cancer.
The present study introduces a new and powerful method to search for "breast cancer genes", that is, genes that greatly increase cancer risk if they are inherited in a mutated form. For high risk families, as well as the general population, it is important to identify these genetic risk factors in order to better assess personal risk levels and to make better decisions in the overall battle against breast cancer. The most well known breast cancer genes are BRCA1 and BRCA2, which were discovered a decade ago by searching for genetic mutations in high risk families. Since their discovery, they have been the subject of over 5000 scientific papers, and it has become clear that a small percentage of all women carry mutations in these genes, and that such women are at a greatly elevated risk for breast cancer. While the discovery of the BRCA genes was an enormous advance in breast cancer genetics, it turns out that they explain only a small portion of all breast cancers, perhaps 5% to 10%, and even in families with a strong history of breast cancer, these genes are the culprit in only half of such families. This suggests that there must be additional breast cancer genes to be discovered. Yet, despite extensive efforts to find such genes, little progress has been made. The ideal way to search for mutated genes in a cancer patient is to read all the letters of their DNA "text", and directly look for genetic "spelling errors". However, it currently costs millions of dollars to read all three billion letters in one person?s DNA, and it will be decades before the cost drops enough to make this approach possible. Our new approach makes use of the fact different patients often have inherited the exact same mutation, tracing back to a distant common ancestor, perhaps hundreds or thousands of years back. This implies they have inherited a large block of DNA text around the mutation as well, and therefore they have in common a large piece of their DNA text. To determine two large blocks of text are identical, we do not need to compare every letter---it is enough just to compare a tiny fraction of letters sampled from the text. In this way, by reading just a small fraction of all of each patient?s letters, we can use statistics to locate large, identical block of DNA text. In turn, this locates the shared mutation, down to a block of DNA containing only hundreds of genes. These genes are all candidates for the gene of interest, and they can be systematically searched for mutations in standard ways to ultimately identify the gene. Thus our procedure provides a highly efficient first pass of screening, to identify the genes that should be analyzed more carefully, by finding the blocks of DNA that are shared identically between different pairs of patients. The ultimate value of this work is to greatly accelerate finding the remaining genetic risk factors for breast cancer, beyond the BRCA genes, by providing the prime candidate genes for more detailed analysis.