Invasive breast cancer (IBC) is a heterogeneous disease with different molecular and histological features as well as a variety of biological behaviors and clinical courses. For personalized cancer medicine to be successful, the critical underlying genetic changes have to be identified and developed both as biomarkers and as treatment targets. Gene fusions, which are the result of chromosome translocations that fuse together pieces of two different genes or a gene and an inappropriate regulatory unit, have an important role in the initial steps of tumorigenesis. For example, the Philadelphia translocation is a specific chromosomal abnormality found in certain leukemias, resulting from a translocation between chromosome 9 and 22. This forms a BCR-ABL1 fusion gene, and drugs such as imatinib that target this product have been very effective in many of these leukemia patients. Due to their critical role in causing some cancers and their presence exclusively in the tumor tissues, fusion genes and their products provide highly specific drug targets with only minimal side effects.

While a large collection of chromosome abnormalities have been documented in breast tumors, identifying gene fusions that are important players in breast tumor development and progression ("driving gene fusions") is in its infancy. In order to identify these chromosome aberrations that can serve as new therapeutic targets, we developed a novel bioinformatics analysis called "fusion copy number signature analysis." This is based on the observation that driving gene fusions and their original (wild-type) genes are often amplified in fusion-positive tumors and tumors without the fusions, respectively. This suggests that gene fusions and amplifications may synergize to activate important oncogenes; if directly druggable, these aberrations will provide substantial opportunities for new therapies. Taking advantage of large sets of gene copy number data for breast tumors publicly available, this approach will be highly cost-effective.

Our application of this strategy has revealed 19 gene abnormalities as candidate drug targets in IBC. Three out of the four candidates examined thus far are supported by our preliminary experimental and clinical data. All three validated genes encode druggable kinases not previously reported in IBC. Chromosome aberrations at these gene regions affect approximately 21% of invasive breast cancer patients, suggesting that this population of patients may be potentially manageable by new drug therapies. In this proposal, we intend to explore the clinical relevance and the biological role of these three oncogene targets and to press on to nominate and test additional candidate targets.

We expect that this study will establish novel therapeutic targets that may benefit a substantial portion of IBC patient populations, as has happened recently with the report of a recurrent EML4-ALK fusion detected in ~6% of lung cancer patients. Patients with this alteration demonstrated a striking response to the ALK inhibitor crizotinib. In a remarkably short period of time, ALK targeted therapies are now in advanced clinical development for this subset of lung cancer. As in the clinical trials with ALK inhibitors, molecular assays detecting these fusions will be applied to identify which patients may benefit from the new targeted therapies. Further, if chromosome abnormalities at these loci endow resistance or sensitivity to other specific therapies, assays detecting these aberrations may have predictive value for personalized medicine. For example, our preliminary data suggest that MAP3K3, one of our three lead genomic candidates, conveys chemotherapy resistance. Therefore, genomic assays detecting MAP3K3 aberrations might be useful in the clinic for appropriate use of chemotherapies.

Together, our new analysis approach will yield novel insights into recurring genetic abnormalities leading to breast cancer and establish robust targets for effective and personalized drug therapies. We expect that some of our newly identified potential targets will get into clinical development in a short time, first to evaluate their prognostic and predictive value (within 1-2 years), and then, depending on availability of appropriate drugs, even for therapeutic trials (within 2-4 years).