Enhancing Genome Analysis with Whole Proteome Profiling for Improved Therapeutic Decision Making in Diffuse Midline Glioma

Diffuse midline glioma (DMG) is a devastating brain tumor with a median overall survival (OS) of 11 months. Although more patients are now participating in precision medicine clinical trials, DMGs often lack clearly druggable genomic targets, limiting effective treatment options. To address this gap, we performed integrative proteogenomic analyses on 210 tumor and control samples, including DMGs and pediatric high-grade gliomas (pHGG). Whole-genome sequencing (90X) of these 210 samples (112 H3.3K27M DMGs, 19 H3.1K27M DMGs, 7 H3-WT DMGs, 31 pHGGs, and 41 normal controls) identified recurrent driver alterations in genes associated with DNA damage response (71%), PI3K/mTOR (24%), RTK (18%), and MAPK (12%) pathways. Among pHGGs, 15 tumors lacked any known DMG driver mutations, while 16 harbored two or more canonical DMG drivers, highlighting biological overlap between these tumors. Despite comprehensive profiling, FDA-approved therapies could only be matched to 38% of samples, with just 17% of variants deemed targetable, underscoring the need for alternative strategies. Hence, we performed complementary proteomic profiling to uncover targetable signaling vulnerabilities not evident at the genomic level. Proteomics of H3.3K27M DMG cell lines (UON-DMG-05 and SU-DIPG-XIII), one of which harbors a known targetable mutation (UON-DMG-05, PDGFRA:D842V), identified elevated activity of ATM and PI3K-Akt signaling, suggesting convergent, actionable dependencies. In vivo validation confirmed the therapeutic potential of targeting these pathways. Monotherapy with avapritinib (30 mg/kg/day; PDGFRAi) or paxalisib (5 mg/kg/day; PI3Ki) extended OS to 57 and 59 days, respectively, compared to 51 days in untreated controls (P