| dc.description.abstract | Since the introduction of paclitaxel in 1992, the overall survival for women diagnosed with ovarian cancer has remained relatively stagnant. This unsatisfactory outcome is in part attributed to the presence of subpopulations of tumor cells with stem cell-like properties, which are resistance to front-line therapies. Efforts to identify unappreciated drugs to combat these resistant cell populations typically have used traditional two-dimensional (2D) cell culture models which do not always reflect the tumor heterogeneity seen in clinical samples including reduced cell proliferation, hypoxia, and cancer stem-like cells. The lack of congruency between in vitro screening models and the patient’s outcome has limited the number of successful anti-cancer drugs that demonstrate clinical benefit. In order to identify drugs that are more likely to benefit patients, a drug screening model that better represents clinical disease was required. Towards this goal, I developed and characterized a three-dimensional (3D) cell culture model in which ovarian cancer cells were grown as multicellular tumor spheroids (MCTS) and directly compared drug activity in 2D and 3D models. Compared to a panel of ovarian cancer cell lines grown in 2D, MCTS formation induced many different phenotypes associated with clinical drug resistance including reduced cell proliferation, cellular hypoxia, and stem-like gene expression and function. Cells grown as MCTS better mimicked the response to chemotherapy seen in the clinic as compared to 2D. Specifically, paclitaxel treatment of MCTS revealed significant increases in expression of stem cell-related genes, e.g., CD133, CD44, OCT4, SOX2, and NANOG. Correspondingly, significant reduction in cell proliferation with accompanying resistance to paclitaxel was observed in MCTS compared to 2D cultures, providing an in vitro model that reflects certain aspects of clinical progression and drug resistance. Interestingly, the drug resistance and stem-like properties observed in MCTS were maintained in cells grown under 2D conditions when these cells were derived from MCTS, indicating a potential long-term change in phenotype. These changes provided an in vitro model that better reflects clinical disease that can be exploited for drug screening. To exploit these differences, the MCTS model was used to compare the activity of 304 repurposed drugs, most that are FDA approved for indication other than cancer, in four established ovarian cancer cell lines grown in 2D and 3D. I identified a number of compounds with selective activity in MCTS that can target subpopulations of resistant cells with stem-cell like properties. In a primary drug screen between 2D and 3D cultures designed to directly compare response to 10 µM of the 304 clinically approved drugs revealed fifteen compounds with preferential activity against cells grown as MCTS as measured by CellTiter-Glo. Drug hits were classified to have 3D specific activity based on >75% viability reduction in MCTS but 75% viability reduction in MCTS but <50% viability reduction in 2D cultures. Secondary screening using a range 0 to 25 µM drug identified licofelone and glafenine to have the most robust and selective activity in MCTS across a genetically diverse panel of ovarian cancer cell lines. Both are anti-inflammatory drugs that had not previously been considered for the treatment of cancer. Since an increasing number of drugs are removed from the market because of unacceptable toxicity, subsequent experiments focused on licofelone. Licofelone has demonstrated a better safety profile as compared to glafenine, i.e., reported increased risk of anaphylaxis and kidney failure. Importantly, treatment of MCTS using IC50 concentration of licofelone significantly reduced expression of stem cell markers by up to 90%, as measured using qRT-PCR and immunofluorescent (IF) immunohistochemistry in cells grown as MCTS. Consequently, Ki67 positive cells were increased from vehicle treatments (30%) in response to licofelone (57%) treatment in MCTS. Combination treatment in MCTS showed significant synergy (Combination Index <1) for sequential additions of licofelone followed by paclitaxel, or vice-versa. Likewise, licofelone addition reversed stemness (gene expression, colony forming and spheroid forming assays) in ovarian cancer cells that were induced by paclitaxel treatment in MCTS. By performing drug screens using a MCTS model, I identified two previously unappreciated candidate anti-cancer drugs that had preferential activity on ovarian cancer cells with stem cell-like properties. When used in combination with paclitaxel, licofelone showed significant synergy and reversed the drug resistant phenotypes of MCTS. | |
