BIOM-17. DIFFERENCES IN THE IMMUNE MICROENVIRONMENT OF GLIOMAS HARBORING IDH2 VERSUS IDH1 MUTATIONS

Abstract INTRODUCTION IDH mutations are a defining feature of lower-grade glioma and secondary glioblastoma. Approximately 95% of glioma-associated IDH mutations are in codon 132 of IDH1, but a small proportion are in IDH2. IDH mutations produce the oncometabolite 2-hydroxyglutarate, which induces global DNA hypermethylation and is associated with an immunosuppressive tumor microenvironment. IDH1 is localized in the cytosol while IDH2 is found in the mitochondrial matrix, and mutations in these genes may have differing effects on the tumor microenvironment. METHODS Formalin-fixed, paraffin-embedded tissue from 633 IDH-mutant gliomas (615 IDH1-mutant, 18 IDH2-mutant) underwent whole-exome and whole-transcriptome sequencing at Caris Life Sciences (236 grade 2/3 astrocytoma, 158 grade 2/3 oligodendroglioma, 202 IDH-mutant glioblastoma, 37 glioma, NOS). QuantiSEQ was used to infer tumor-infiltrating immune cell populations from RNAseq data, and gene-set enrichment analyses (GSEA) were performed using Wikipathway. RESULTS IDH1-mutant gliomas had higher levels of pro-inflammatory M1 macrophages (P=0.04), modestly higher levels of monocytes (P=0.08), and lower levels of neutrophils (P=0.04) – typically considered immunosuppressive – compared with IDH2-mutant gliomas. No differences were observed in levels of B cells, dendritic cells, NK cells, or T cell subsets (Treg, CD4+, CD8+). IDH2-mutant gliomas were enriched for hallmark oligodendroglioma mutations (TERT promoter, CIC, FUBP1), while IDH1-mutant gliomas were enriched for hallmark astrocytoma mutations (ATRX, TP53). However, associations with tumor-infiltrating immune cells persisted after excluding 1p/19q co-deleted oligodendroglioma from analyses. GSEA revealed upregulation of the microglial TYROBP signaling pathway, the microglial phagocytic pathway, and of Type II Interferon signaling in IDH1-mutant gliomas versus IDH2-mutant gliomas. CONCLUSIONS Although IDH2 mutations are generally thought to function similarly to IDH1 mutations, we observe differences in tumor-infiltrating immune cells across groups. IDH2-mutant gliomas appeared to have a more immunosuppressive tumor microenvironment than their IDH1-mutant counterparts. Early-phase immunotherapy trials should consider covariate-adaptive randomization approaches to equally allocate IDH2-mutant gliomas across treatment arms.