Metabolic Profiling Defines Glioblastoma Subtypes with Distinct Prognoses and Therapeutic Vulnerabilities
A partir de trois séries de données portant sur des glioblastomes sans mutation de l'isocitrate déshydrogénase et à partir d'une analyse multiomique, cette étude identifie trois sous-types métaboliques présentant des pronostics distincts et des vulnérabilités thérapeutiques
Background : Glioblastoma (GBM) is a highly aggressive brain tumor with profound metabolic heterogeneity. However, a clinically actionable classification based on metabolic gene expression remains undefined.
Methods : We conducted a comprehensive multi-omics analysis of IDH-wildtype GBMs from three publicly available datasets. Prognostic metabolism-related genes were used to define transcriptional subtypes, which were validated in independent datasets and patient-derived cell (PDC) models. Functional assays and drug sensitivity studies were performed to explore therapeutic relevance.
Results : We identified three distinct metabolic subtypes: M1, enriched for synaptic signaling and amino acid metabolism, exhibited leading-edge anatomical features; M2, characterized by mitochondrial metabolism and cell cycle activity, was associated with favorable survival; and M3, marked by hypoxia, immune activation and suppression, and broad metabolic pathway engagement, correlated with poor prognosis. These subtypes were reproducible across cohorts and faithfully recapitulated in PDC models. Metabolomic profiling confirmed distinct subtype-specific metabolic signatures. Notably, M3 cells showed high sensitivity to inhibitors targeting glycosaminoglycan degradation, nicotinamide metabolism, and retinoic acid pathways in both in vitro and in vivo models.
Conclusion : Our study defines three biologically and clinically relevant metabolic subtypes of IDH-wildtype GBM. This classification reveals distinct metabolic programs and therapeutic vulnerabilities, providing a framework for precision metabolism-targeted strategies in glioblastoma.
Neuro-Oncology , article en libre accès, 2026