DDDR-24. INTEGRATED ANALYSIS OF SINGLE CELL CHROMATIN ACCESSIBILITY AND RNA EXPRESSION IDENTIFIED COMMON VULNERABILITY DESPITE GLIOBLASTOMA HETEROGENEITY

Abstract

Abstract INTRODUCTION In 2021, the World Health Organization (WHO) reclassified glioblastoma, the most common form of adult brain cancer, into isocitrate dehydrogenase (IDH) wild-type glioblastomas and grade IV IDH mutant (G4 IDHm) astrocytomas. For both tumor types, intra-tumoral heterogeneity is a key contributor to therapeutic failure. METHODS we applied integrated genome-wide chromatin accessibility (snATACseq) and transcription (snRNAseq) profiles to clinical specimens derived IDHwt glioblastomas and G4 IDHm) astrocytomas, with goal of therapeutic target discovery. RESULTS The integrated analysis achieved resolution of intra-tumoral heterogeneity not previously possible, providing a molecular landscape of extensive regional and cellular variability. snATACseq delineated focal amplification down to an ~40 KB resolution. The snRNA analysis elucidated distinct cell types and cell states (neural progenitor/oligodendrocyte cell-like or astrocyte/mesenchymal cell-like) that were superimposable onto the snATACseq landscape. Paired-seq (parallel snATACseq and snRNAseq using the same clinical sample) provided high resolution delineation of extrachromosomal circular DNA (ecDNA), harboring oncogenes including CCND1 and EGFR. Importantly, the copy number of ecDNA genes correlated closely with the level of RNA expression. Integrated analysis across all specimens profiled suggests that IDHm grade 4 astrocytoma and IDHwt glioblastoma cells shared a common chromatin structure defined by open regions enriched for Nuclear Factor 1 transcription factors (NFIA and NFIB). Silencing of NF1A or NF1B suppressed in vitro and in vivo growth of patient-derived IDHwt glioblastomas and G4 IDHm astrocytoma models that mimic distinct glioblastoma cell states. CONCLUSION Our findings suggest despite distinct genotypes and cell states, glablastoma/G4 astrocytoma cells share dependency on core transcriptional programs, yielding an attractive platform for addressing therapeutic challenges associated with intra-tumoral heterogeneity.