EPCO-12. DOMINANT MALIGNANT LINEAGE LEVERAGE RESTRICTED EPIGENOMIC PROGRAMS TO DRIVE EPENDYMOMA DEVELOPMENT

Abstract Pediatric cancers are often driven by unique alterations suggesting mechanistic ties to underlying developmental programs. Ependymoma (EPN) is an example of a brain cancer subtype driven by gene fusions involving ZFTA-RELA (ZR) that are exclusive to this disease. This leads to our hypothesis that specific chromatin states in developmental lineage programs are at risk of oncogenic transformation, and that these chromatin signatures persist during tumor development and govern tumor cell heterogeneity. To address this question, we constructed an integrated single-nucleus Multiome (snRNA-seq and snATAC-seq) atlas of the developing mouse forebrain and compared it with snMultiome analysis of ZR-driven mouse and human EPN. We identified specific developmental lineage programs present in glial progenitor cells enriched with the Plagl family transcription factors; the precise DNA binding sites of ZFTA fusion oncoproteins. In both mouse and human EPN, ZR expression maintains accessibility of developmental epigenomic programs leading to persistent cell proliferation and tumorigenicity. Cross-species analysis of mouse and human EPN revealed significant cell type heterogeneity mirroring incomplete neurogenic and gliogenic differentiation, with a small percentage of cycling intermediate progenitor-like cells that established a putative tumor cell hierarchy. These findings were expanded by applying in vivo lineage barcode tracing methods in mouse EPN, which revealed dominant cell lineages that aggressively expanded and recapitulated the entire tumor cell diversity. Our findings unravel the intersection between developmental and oncogenic epigenomic states critical for ZR driven brain cancer. These findings shape our understanding of the earliest stages of epigenetic programs in brain cancer, and how cancer drivers such as ZR intersect with developmental programs to establish tumor cell heterogeneity.