Abstract:Background High-grade serous ovarian cancer (HGSOC) is characterized by extensive intratumoral heterogeneity, which is associated with drug resistance. Our aim was to study the effects of drugs used in the treatment of HGSOC - carboplatin (C), paclitaxel (P) and olaparib (O) - on clonal dynamics in experimental models. Methods We tracked clonal evolution in PEO1, a BRCA2-mutant HGSOC cell line model, by transfecting an initial population of 1.714x106 cells with a barcode (BC) library (CloneTrackerXP). Barcoded populations of cells (18.73x106 cells/flask) were grown in HYPERflasks till confluence, and then replicates were exposed to C (3163 nM) for 5 weeks, P (6.32 nM) for 4 weeks, and O (28467 nM) for 5 weeks (as single agents), reducing the cell population to < 10% (controls kept till cell confluence in the absence of drugs). After that, drug pressure was removed, allowing cells to regrow fully. NGS was used to quantify BCs in the final cell populations and, to get information on the temporal dynamics, also in dead floating cells collected from the weekly culture media change. Heterogeneity of the cell populations was estimated using the Shannon diversity index (SDI). Results The SDI in the controls was 6.306 (95% Bootstrap CI 6.304-6.310) for C and 6.192 (95% Bootstrap CI 6.190-6.197) for P and O. After exposure to C, P and O, the SDI was reduced to 3.208, 2.520 and 4.753 (95% Bootstrap CI 3.209-3.211, 2.517-2.523, 4.750-4.759) respectively, using aggregated replicates (pairwise 2-sided Hutcheson t-test: p < 2.2e-16 pre-drug exposure vs C, P and O). The number of unique BCs in controls vs C, P and O were 1799 vs 545, 1463 vs 288 and 1463 vs 1270, respectively, demonstrating reduction of heterogeneity and change in frequency of clones following drug selection pressure. Cell populations exposed to C and O share some similar clone selection (approximately 60%). This is important as, although the degree of heterogeneity caused by individual drugs are different (C and P lead to less heterogeneous populations than O), there is some concordance between the remaining clones determined by DNA damaging/DNA damage response inhibitors compared to clone selection following exposure to antimicrotubule agents. The selection pattern of BCs among replicates of the different drugs was similar with conservation of BCs with the highest fitness, demonstrating that the initial conditions are determinants of selection under drug exposure. Conclusion Our work, for the first time, quantifies and shows that heterogeneity followed by O therapy is greater than by C and P. Furthermore, the emergent population of cells post-selection pressure was more similar in the C and O groups when compared to P-exposed cells, suggesting correlation in the mechanism of action of drugs and drug resistance/tolerance. Our results also indicate that evolutionary dynamics are largely deterministic and, therefore, can be predictable. Citation Format: Alvaro H. Ingles Russo Garces, Salvatore Milite, Javier Fernandez-Mateos, Bingjie Chen, Lisa Pickard, Adam Stewart, Sara Diaz Sanchez, Rachel Lau, Erica Oliveira, Susana Banerjee, Andrea Sottoriva, Udai Banerji. A study of evolutionary dynamics induced by carboplatin, olaparib and paclitaxel in high-grade serous ovarian cancer cell line models [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl) nr 3790.

Abstract 3790: A study of evolutionary dynamics induced by carboplatin, olaparib and paclitaxel in high-grade serous ovarian cancer cell line models