A shared vascular-hypoxic transcriptional program is associated with immune organization in glioblastoma

Abstract

Glioblastoma (GBM) contains spatially distinct vascular and hypoxia/necrosis-associated compartments, but the extent to which these classical anatomical regions share transcriptional programs linking vascular remodeling with immune-related processes remains incompletely defined. Here, Ivy Glioblastoma Atlas Project (IvyGAP) anatomical transcriptomic data were used as a discovery framework and integrated with independent bulk and single-cell transcriptomic resources to characterize the cellular, immune and clinical context of a shared vascular-hypoxic program. Analysis of regionally defined IvyGAP compartments identified 1,114 genes enriched in vascular-associated regions and 403 genes enriched in hypoxia/necrosis-associated regions, with 95 genes shared between the two comparisons. Functional enrichment implicated angiogenesis, extracellular matrix remodeling, inflammatory signaling, chemotaxis, complement/coagulation and cytokine-related processes. The shared signature was further resolved into angiogenesis and vascular remodeling, immune and inflammatory signaling, and shared immunovascular gene subsets. Single-cell mapping showed expression of these programs across endothelial, perivascular and myeloid populations, with immune/inflammatory components also detectable in vascular-associated cell populations. Vascular remodeling genes were positively associated with inflammatory genes, stromal and exclusion scores, myeloid-enriched immune populations and immunoregulatory markers. Representative genes from each functional subset showed preferential expression in mesenchymal GBM across TCGA and CGGA, while survival modeling using the predefined combined immunovascular program separated patients into groups with distinct overall survival across three GBM cohorts. Together, these findings identify an anatomically anchored vascular–hypoxic transcriptional program associated with vascular remodeling, inflammatory signaling, immune-context features and adverse clinical outcome, providing a cross-dataset framework for interpreting immunovascular relationships in GBM.