Metabolic Pathway Alterations in Glioblastoma Stem Cells Under Hypoxia: Computational Analysis of Hypoxia-Activated Pathways

Abstract

Therapy resistance in glioblastoma stem cells (GSCs) often arises in hypoxic microenvironments, yet most computational studies analyze single pathways in isolation. Here, we introduce HypoPINN-lite, a graph-based framework that integrates differential expression, enrichment analysis, and pathway interaction modeling to capture multi-pathway crosstalk under hypoxia. Applying this approach to transcriptomic profiles from GSE77307, we identified a tightly coupled metabolism–inflammation module linking glycolysis (LDHA, GLUT1), fatty acid synthesis (FASN), oxidative phosphorylation, and COX-2 signaling (PTGS2). Unlike standard enrichment, which highlights pathways independently, our framework reveals that hypoxia drives these pathways to function as a single adaptive collapse network. This convergence suggests therapeutic leverage points: disrupting both metabolic and inflammatory edges may destabilize the module and prevent adaptation. Beyond glioblastoma, HypoPINN-lite is generalizable to other cancers and stress conditions, offering a systems-level strategy to identify multi-pathway vulnerabilities invisible to traditional analyses.