Several years ago, we used resources of the Australian Ovarian Cancer Study (AOCS) to systematically analyze the cancer genomes of tumors from women with ovarian cancer that were poorly responsive to platinum/taxol-based chemotherapy. In particular, we focused on women with high grade serous ovarian cancer (HGSC), the most common type of epithelial ovarian cancer. Our work implicated a region on chromosome 19 (19q12), which included the CCNE1 gene, as an important determinant of poor treatment response. CCNE1 makes a protein, cyclinE1, that controls cell division but which can also damage the cell's genome when present in excess. We subsequently demonstrated that HGSC with an increased number of copies (amplification) of the CCNE1 gene was especially dependent on the cyclinE1 protein. This was an important finding as it indicated a specific vulnerability of HGSC with this genetic change that might be exploited therapeutically. About 15%-20% of HGSC appears to be driven by an increase of CCNE1 -- a similar frequency to Her2 amplification in breast cancer, which is treated with Herceptin and related new drugs. Therefore, CCNE1 could be as important a therapeutic target for women with HGSC as Her2 is in breast cancer. Of particular note, we find that tumors with CCNE1 amplification are unlikely to respond to PARP inhibitors as they lack key molecular properties that make other HGSC sensitive to this new class of drug. Hence, women with CCNE1-amplified cancers represent an important unmet need.

There are a number of drugs currently in clinical development in other solid cancers and leukemias that may be effective in CCNE1-amplified HGSC. However, there are important unresolved questions about their use: Which drug, among the several possible choices, is likely to be the most effective? Should these drugs be used as single agents or in combination with existing therapy? If used in combination, how should they be scheduled to achieve maximum effect? Would drugs that narrowly target the cyclinE1 protein have greater activity and reduced side effects? What resistance mechanisms are likely to emerge when we target cyclinE1 and can we circumvent them? Our ongoing work strongly suggests that we will need drugs that are more closely targeted to cyclinE1 than those currently available. Our collaborators in the UK are world experts in the development of this class of inhibitor, and we anticipate they will generate novel agents that must also be tested preclinically. To address these questions, we therefore urgently need a mouse model of Ccne1-driven HGSC. We also foresee a requirement for an animal model with an intact immune system to evaluate new immune checkpoint inhibitors in conjunction with targeted agents to cylinE1.

This application describes our aim to develop a mouse model that mimics the human disease, thereby providing a preclinical platform for testing novel cyclinE1 inhibitors. We will construct transgenic mice where the Ccne1 gene can be precisely activated when and where we want it. Using well-established technology, we will engineer mice to express the cyclinE1 protein at high levels in fallopian tube secretory epithelial cells, the progenitors of HGSC. We are very experienced in the generation of such animals and their analysis. We will cross in two other genetic alterations that are common to human HGSC with CCNE1 amplification, a mutation in Tp53 and loss of Pten. Importantly, TP53 is always defective in human HGSC and its loss appears to be a requirement for cells to tolerate the presence of excess cyclinE1. Based on our laboratory studies and work of other researchers, there is a good chance that these three genetic alterations -- Ccne1 overexpression, Tp53 mutation, and Pten loss -- will be enough to generate tumors. It is also possible, however, that additional mutations will be needed for the mice to develop HGSC that mimic the human disease. In the second part of the grant, we will use data from genetic screens in fallopian tube cells and information from thousands of cancer genomes to find genes that may cooperate with CCNE1 and that could be added to the mice.

The study is likely to provide a very high return on the invested effort, as the mice will provide an enduring resource -- once the model is established, it can be used in many settings, providing a powerful ongoing platform for development of CCNE1 inhibitors. Without such a model, the development of CCNE1 inhibitors will be hampered on an ongoing basis. The study brings together collaborators from around the world with a very strong track record and commitment to developing novel treatments for HGSC with amplified CCNE1. They have complementary skills in cancer genomics and patient cohort studies (Bowtell), mouse models and pathology (Drapkin), genetic cell line screens (Rottapel), bioinformatics (Beroukhim), and drug development (Newell, Venkitaraman). Successful generation of this model will be of great value for development of new therapies to attack ovarian cancer and other solid cancers with amplification of CCNE1, including breast, stomach, and lung cancer.