EPND-01. THE ACTIVE REGULATORY LANDSCAPE OF EPENDYMAL TUMORS

Abstract

AbstractGenomic sequencing has driven precision-based oncology therapy; however, genetic drivers remain unknown or non-targetable for many malignancies, demanding alternative approaches to identify therapeutic leads. Ependymomas comprise histologically similar tumour entities driven by distinct molecular mechanisms, such as fusion oncoproteins, genome-wide chromosomal instability, or disruption of DNA methylation patterns. Despite these differences, ependymomas commonly resist chemotherapy and lack available targeted agents for clinical trial development. In the case of genomically balanced ependymomas, and those driven by fusion oncoproteins, we hypothesized that the chromatin landscapes could uncover oncogenes that would inform actionable targets for therapy and reveal specific transcriptional circuitries to identify the molecular origins of the disease. To this end, we mapped active chromatin landscapes in 42 primary ependymomas in discovery and validation cohorts using H3K27 acetylation ChIP-seq. Enhancer regions revealed novel oncogenes, molecular targets, and pathways, which when subjected to small molecule inhibitor or shRNA treatment, increased survival and halted proliferation in mouse and neurosphere patient-derived models of ependymomas. Reconstruction of enhancer networks permitted the identification of core transcription factors (TFs) that establish ependymoma cell state, and lineage-specifying TFs that dictate molecular subgroup identity. Lineage-associated TFs point to distinct spatio-temporal origins of ependymoma subgroups such as FOXJ1 TF activity and expression observed preferentially in subsets of hindbrain ependymomas. To translate our results into a potential clinical paradigm, we demonstrate cancer dependencies on super enhancer associated genes and lineage TFs, and establish the utility of chromatin landscape analysis to predict novel targets for cancer therapy.