PATH-37. SPATIAL GENE EXPRESSION PROGRAMS AND PROTEIN SIGNALING MECHANISMS DRIVE MENINGIOMA EVOLUTION AND THERAPEUTIC VULNERABILITY

Abstract

Abstract Intratumor heterogeneity drives cancer evolution and resistance to therapy, but unbiased approaches to elucidate mechanisms driving intratumor heterogeneity have been lacking. Here, we integrate spatial gene expression programs and protein signaling mechanisms across 16 meningioma samples to define how intratumor heterogeneity drives molecular classification, temporal evolution, or spatial evolution of the most common primary intracranial tumor. Spatial transcriptomic analysis was performed on 38,718 regions and spatial profiling of 72 proteins comprising proliferation, stress, microenvironment, immune, or signaling modules was performed on 82 regions. DNA methylation, copy number variant, targeted gene expression profiling, targeted DNA sequencing, histologic, or immunohistochemical analyses (Ki67, H3K27me3, p16) were performed on all meningiomas to study intratumor heterogeneity in the context of pre-existing classification schemes. Primary meningioma cells, CRISPR interference, pharmacology, and 3D co-culture models were used for mechanistic and functional validation. Spatial analyses revealed significant intratumor or intertumor heterogeneity irrespective of meningioma histologic subtype or grade, DNA methylation group (Merlin-intact, Immune-enriched, Hypermitotic), copy number variant (1p loss, 1q gain, 6p gain, 9p loss, 14q loss, 22q loss), gene expression risk score, or driver mutation (NF2, CDKN2A/B, TERT promoter, BAP1, SMARCB1, ARID1A). Spatial analyses on paired histologically or molecularly distinct regions from individual meningiomas (n=4), or on paired primary and recurrent meningioma samples (n=9), revealed conserved proliferation, immune, or signaling mechanisms underlying meningioma evolution. Mechanistic and functional studies validated therapeutic vulnerabilities across spatial clusters to combinations of FDA-approved small molecules inhibiting the cell cycle (abemaciclib), epigenetic regulators (vorinostat), the DNA damage response (nariparib), MAPK signaling (erlotinib, selumetinib), or PI3K-AKT signaling (copanlisib). In summary, these data define spatial gene expression programs and protein signaling mechanisms driving meningioma evolution and therapeutic vulnerability, shedding light on past clinical failures and elucidating novel combinations of concurrent or sequential molecular therapies to treat meningiomas that are resistant to standard interventions.