Abstract PR003: Integration of clonal composition and tumor heterogeneity reveals novel evolutionary states and intervention targets in ovarian cancer

Abstract

Abstract High-grade serous ovarian cancer (HGSC) is the most common form of epithelial ovarian cancer, typically diagnosed at advanced stage with five-year survival of only 38%. Herein, we used 149 treatment-naïve and 65 after-treatment samples subjected to whole-genome sequencing from 55 HGSC patients to reconstruct evolutionary histories, characterize clonal compositions and intra- and inter-tumor heterogeneity, in order to identify mechanisms that drive treatment failure. Using clonal information from the phylogenetic trees we quantified intra- and intertumor heterogeneity and identified three types of clonal compositions corresponding to three evolutionary states (“adaptive”, “maintaining” and “evolving”) of treatment-naïve HGSC tumors. The branch depths from phylogenetic trees were used to quantify “clonal age” for each state. The “adaptive” state showed the highest clonal age, followed by the “maintaining” state, and the “evolving” being the youngest, least evolved tumor state. The three states have significantly different survival association, “evolving” the longest and “maintaining” the shortest (p = 0.029). Pathway analysis of genes harboring branch mutations revealed that the three states are characterized by unique signaling cascades, with MAPK, PI3K/AKT, and NOTCH signaling significantly enriched at “evolving”, “maintaining”, and “adaptive” states, respectively. Furthermore, we suggest two evolutionary trajectories originating from “evolving” state towards “maintaining” via PI3K/AKT activation and extensive cytokine signaling or towards “adaptive” state via RAS/RAF/MAP cascade. Overall, altered PI3K/AKT signaling was found in half of the patients, followed by alterations in NOTCH signaling enriched in 20% of patients. To explore the effect of treatment on evolutionary states we investigated clonal compositions and enriched pathways in the after-treatment samples. Interestingly, 62% of the neoadjuvant chemotherapy treated (three cycles of platinum & paclitaxel) samples remained in the same evolutionary state as compared to the treatment-naïve samples from the same patients, indicating that the neoadjuvant therapy does not significantly alter tumor composition. Analysis of relapse samples revealed that signal transduction via PI3K/AKT and NOTCH was enriched after the full course of treatment (including platinum-taxane chemotherapy, surgery, possible maintenance therapy with PARP inhibitors or angiogenesis inhibitors), which testifies for their central role in treatment failure. Taken together, our results show that integrating clonal composition and heterogeneity at diagnosis allows allocation of HGSC tumors into three states with different prognosis. Aberrations in PI3K/AKT or NOTCH signaling, distinctive for higher evolved “maintaining” and “adaptive” states, are key pathways to the development of chemoresistant clones. As the herein identified pathways can be targeted by several clinically approved drugs, our results provide a means to identify effective, combinatorial personalized treatments for HGSC patients at the relapse setting. Citation Format: Alexandra Lahtinen, Kari Lavikka, Yilin Li, Sanaz Jamalzadeh, Anni Virtanen, Rainer Lehtonen, Olli Carpén, Sakari Hietanen, Kaisa Huhtinen, Antti Häkkinen, Johanna Hynninen, Jaana Oikkonen, Sampsa Hautaniemi. Integration of clonal composition and tumor heterogeneity reveals novel evolutionary states and intervention targets in ovarian cancer [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr PR003.