Ovarian carcinosarcoma (OCS) is the most lethal gynaecological cancer1,2. Evidence-based treatment is limited to first-line platinum-based chemotherapy and most patients develop lethal, relapsed disease within one year of treatment2,3. Therefore, identifying new therapeutic strategies is vital.
We generated a genetically-engineered mouse model (GEMM) of OCS, driven by over-expression of Lin28b and inactivation of p53 in fallopian tube (FT) secretory epithelial cells. Tumours expressed the FT marker PAX8, as well as HMGA2, which is downstream of LIN28B and is over-expressed in 60% of human OCS tumours, indicating this model is representative of the human disease. We developed a cell line from the GEMM and used this cell line to screen a drug library of over 4000 compounds and carry out a CRISPR screen. We repeated these drug and CRISPR screens in the presence of cisplatin, a current standard treatment for OCS, and eribulin, a drug that targets epithelial-to-mesenchymal transition (EMT) which we have recently proven to be effective in the treatment of OCS4.
We developed novel OCS cell lines and organoid models to validate hits from the drug and CRISPR screens. We also used these models to test responses to drugs expected to target N-MYC and EMT, which we have previously shown are important features of these tumours4. High-throughput organoid drug screening has enabled us to identify novel cisplatin- and eribulin-based combination therapies that synergistically kill OCS organoids. We have chosen a number of combination therapies to test in our OCS patient-derived xenograft (PDX) models in vivo. Through analysis of these PDX models post-treatment we hope to extend our understanding of these combinatorial responses, and ultimately improve treatment options for women with this aggressive cancer.