Abstract AP18: FUNCTIONAL GENOMIC SCREEN FOR ATTACHMENT AND INVASION IN METASTATIC OVARIAN CANCER

PURPOSE: To perform a biologically relevant functional genomic screen for genes involved in attachment and early invasion incorporating tumor microenvironment cells cultured from human omentum. METHODS: Primary fibroblasts cultured from the human omentum were plated as one component to recapitulate the tumor microenvironment. An arrayed small interfering RNA (siRNA) panel targeting 719 genes encoding kinases was used. Each well contained 2 oligos per gene. Three ovarian cancer cells with known metastatic capabilities (OVCAR8, ES2, and A2780) were selected and stably labeled with GFP. Following exposure to siRNA, ovarian cancer cells are overlaid upon a mixed submesothelial matrix (NOFs, collagen I, and fibronectin). After 72 hours of attachment/early invasion, a wash step was performed, and the functional readout of the intensity of fluorescence was performed. Cell death due to the siRNA was evaluated by staining with resazurin. Data was analyzed by expressing the effect of increased or decreased attachment using the median average deviation (MAD). RESULTS: MAD values were generated for each gene. A baseline MAD value of 0 was used and positive MAD values denoted genes that influenced an increase in adhesion when knocked down whereas negative values represented genes that had decreased adhesion when knocked down. Negative controls for adhesion/early invasion were used and found to have a MAD value of 0.52 for mock transfection reagent control and 2.45 for siNEG control. Positive controls for adhesion/early invasion had a MAD value of -6.13 for siB1- integrin and -8.91 for a known inhibitor of attachment – canthardin. The transfection efficiency control, siDEATH, had a MAD value of -8.8. Of the 719 genes encoding kinases, we found 17 genes with MAD values above +3 and 19 genes with MAD values below -3. Of these hits, JAK1 and Wee1 were identified. JAK1 has been found to contribute to ovarian cancer invasion, migration, and metastasis in preclinical models. Wee1 is involved in cell cycle progression, and a Wee1 inhibitor is currently in a Phase II trial in relapsed ovarian cancer. Both targets confirm that biologically-relevant genes can be identified through this screen by incorporating stromal cells cultured from human omentum. Additionally, across 3 cell lines, we identified kinases that overlapped in at least 2 cell lines with a MAD lower range of -1.85. There were 94 kinases identified: 9 with a role in adhesion, 7 in immune response, 17 in proliferation, 6 in apoptosis, 16 in metabolism, 33 in cell cycle, 2 in migration, 3 in ribosomal, 1 in autophagy, and 23 genes with overlapping functions. CONCLUSIONS: By incorporating the tumor microenvironment into a functional genomic screen, biologically relevant genes for ovarian cancer cell attachment and early invasion can be identified and will be validated as potential novel therapeutic targets. Citation Format: Katherine Fuh, Anne Gibson, Daniel Wilke, Maxene Ilagan, Andrea Hagemann, Carolyn McCourt, Premal Thaker, David Mutch, Matthew Powell, and Gregory Longmore. FUNCTIONAL GENOMIC SCREEN FOR ATTACHMENT AND INVASION IN METASTATIC OVARIAN CANCER [abstract]. In: Proceedings of the 11th Biennial Ovarian Cancer Research Symposium; Sep 12-13, 2016; Seattle, WA. Philadelphia (PA): AACR; Clin Cancer Res 2017;23(11 Suppl):Abstract nr AP18.