Abstract 1252: Evolution of transcriptomic and epigenomic intra-tumor heterogeneity in high-grade serous ovarian cancer with chemotherapy

Background: Intra-tumor heterogeneity (ITH) refers to the presence of diverse cell populations within a single tumor, each with distinct genetic, epigenetic, and phenotypic characteristics. This complexity presents significant challenges in treatment, as different subpopulations may respond variably to therapies, yet the non-genetic component of ITH and its contribution to treatment response remains poorly understood. We mapped transcriptomic and epigenomic ITH to explore evolution of non-genetic ITH in response to therapy in high-grade serous ovarian cancer (HGSOC). Methods: The SCANDARE study (NCT03017573) included patients with HGSOC that are treated with neo-adjuvant chemotherapy (NAC). Biopsies were performed at baseline, after NAC surgery and at recurrence for single-cell RNA sequencing (scRNA-seq, n=17 samples, 111,946 with 1,773 median gene coverage), single-nuclei RNA sequencing (snRNA-seq, n=35 samples, 151,719 nuclei with 2,843 median gene coverage), and single-cell epigenomics (snCUT&Tag, n=24 samples, 58,892 nuclei with 1,409 unique fragments median coverage). Our snCUT&Tag dataset is one of the first single-cell epigenomic map of cancer patients under treatment, focusing on the histone modification H3K4me1, which accumulates at primed and active enhancers and promoters. We integrated these clinically annotated single-cell/nuclei datasets with publicly available scRNA-seq data of HGSOC. Results: We constructed a consensus map of functional tumor states in HGSOC, based on scRNA-seq analysis of over 200,000 tumor cells. This analysis revealed 15 recurrent tumor cell phenotypes, ranging from stressed to cycling or inflammatory states. We identified the transcription factors and cell-cell communications specific to each tumor cell state. These tumor states are encoded at the epigenomic level, we could identify in each tumor epigenetic clones - clusters of cells with the same epigenome - each displaying a H3K4me1 landscape characteristic of one tumor state. These results show that tumor states correspond to tumor populations with distinct non-genetic features. Longitudinal sampling by snRNA-seq further showed that tumor state composition consistently evolves upon chemotherapy exposure in patients: tumors lose their cycling, hypoxic and interferon cells while gaining cells with partial mesenchymal and inflammatory characteristics. Conclusions: We propose a consensus map of transcriptomic and epigenomic tumor states in HGSOC. Our work is a proof of concept that we can monitor functional ITH with high resolution snRNA-seq from frozen biopsies of the standard of care. We show that these consensus tumor states consistently evolve upon treatment in all patients, pinpointing tumor cell states that are not successfully targeted by NAC, that would need to be targeted by adjuvant therapies. Citation Format: Yuna Landais, Juliette Bertorello, Marta Puerto, Baptiste Simon, Amélie Roehrig, Adeline Durand, Marceau Quatredeniers, Constance Lamy, Fabrice Lecuru, Christophe Le Tourneau, Celine Vallot. Evolution of transcriptomic and epigenomic intra-tumor heterogeneity in high-grade serous ovarian cancer with chemotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 1252.