Integrated multi-omics reveals ferro-aging as a signature of glioblastoma malignancy and therapeutic potential of valproic acid

Glioblastoma (GBM) is the most lethal primary brain tumor with poor prognosis and inevitable relapse. Ferro-aging, a newly defined chronic iron-overload and ACSL4-driven lipid peroxidation process, has not been explored in GBM. We integrated bulk transcriptomic (TCGA-GBM, CGGA), single-cell (GSE182109), and spatial transcriptomic (GSE253080) data to systematically characterize ferro-aging heterogeneity, pseudotemporal trajectory, cell communication, spatial distribution, and immune microenvironment remodeling in GBM. A prognostic model was constructed using multi-omics intersection, LASSO, and Cox regression, followed by drug sensitivity and molecular docking analyses. High-ferro-aging GBM cells exhibited enrichment of EMT, hypoxia, angiogenesis, and TNFα/NF-κB signaling, adopted an invasive pseudotemporal trajectory. Spatial transcriptomics validated malignant pathway enrichment in high-ferro-aging regions. A four-gene signature (PLAUR, STEAP3, VGF, NNAT) stratified patients into high- and low-risk groups with significantly different survival. The high-risk group featured EGFR-dominant mutations, elevated CAF infiltration, T-cell exclusion, and reduced predicted immunotherapy response. Molecular docking identified valproic acid as a potential agent targeting STEAP3 and PLAUR (binding energy ≤ − 5.0 kcal/mol). This first integrative multi-omics analysis reveals ferro-aging as a signature of GBM malignancy and immune evasion, provides a robust prognostic model, and proposes VPA as a potential ferro-aging-targeted therapeutic strategy.