Unraveling the immunologic vulnerabilities of diffuse hemispheric glioma, H3 G34-mutant.

2082 Background: Recent insights into the molecular basis of diffuse hemispheric glioma, H3 G34-mutant (DHG-G34), an incurable high-grade glioma, have opened new therapeutic possibilities. The oncohistone mutation in DHG-G34 impairs the epigenetic regulator SETD2, which normally orchestrates DNA-mismatch-repair and negatively regulates the transcriptional silencing activity of the Polycomb Repressive Complex 2 (PRC2). In DHG-G34, a mismatch-repair-deficiency phenotype is predicted, but an anti-tumor immune response is notably absent in clinical observations, which we posit is linked to the aberrant transcriptional repression by PRC2. Here, we investigate DHG-G34 immunosuppression mechanisms in a large real-world multi-omics cohort and human cell lines. Methods: Clinical, molecular, and immunologic characteristics of a DHG-G34 cohort, sequenced (DNA and RNA) at Caris Life Sciences (Phoenix, AZ), are contrasted against diffuse midline glioma (DMG) and pediatric low-grade glioma (LGG). A significant difference between genomic subgroups was defined as fold-change > 1.2. Statistical significance was determined using chi-square, Fishers-exact, and Mann Whitney U tests with corrections for multiple hypothesis testing (q < 0.05).Hazard ratio (HR) was calculated using the Cox proportional hazards model. KNS-42 DHG-G34 cell line was exposed in triplicate to 1 µM valemetostat (an EZH1/2 inhibitor) or DMSO in vitro. Flow cytometry and RNA-sequencing were utilized to determine differential immunologic characteristics. Results: 29 DHG-G34, 51 DMG, and 52 LGG were identified in Caris database. Median OS among DHG-G34 was 15.4 mo., similar to DMG (HR = 1.1, 95% CI: 0.70-1.73, p = 0.686). Most frequent alterations in DHG-G34 were TP53 (93.1%), ATRX (69.0%), PDGFRA (31.0%). Tumor mutational burden (TMB) was high, >=10mt/MB, in 10% of DHG-G34 and 0/51 in DMG and 0/52 LGG (p < 0.01, q > 0.05), but immune fractions including dendritic cells, T cells, macrophages, microglia, and antigen-presenting genes were decreased (fold-change: 0.19 – 0.83, q < 0.05). In KNS-42, PRC2 inhibition upregulated MHC class 1 (p < 0.001). Of 402 KNS-42 missense variants, transcription of 43 were upregulated following EHZ2 inhibition, and 4 were downregulated (q < 0.05). Among immunostimulatory genes, P2RX7, which engages T-cells and stimulates an inflammatory response, was downregulated in patient samples (q < 0.001) and upregulated in KNS-42 following EZH2 inhibition (q < 0.001). Conclusions: DHG-G34 has an increased propensity for high TMB but remains immunologically cold. Our pilot experiment suggests inhibiting PRC2’s unique role in DHG-G34 may sensitize DHG-G34 to an immune response by inducing MHC class 1 expression and transcriptional readthrough of mutated transcripts. Understanding whether treatment induces TMB in DHG-G34 may affect timing of an immunotherapeutic intervention.