Abstract BACKGROUND Hypoxia is a hallmark of the glioblastoma (GB) and a critical driver of tumor progression, cellular plasticity, and resistance to standard therapy. While its role in metabolic reprogramming is established, the mechanisms by which hypoxia shapes tumor evolution at the genomic and epigenetic levels remain poorly defined. MATERIAL AND METHODS To investigate the impact of hypoxia on glioblastoma evolution, we employed a multimodal approach that integrated multi-omic spatial transcriptomics and whole genome DNA methylation profiling, to explore spatial-hypoxia depended genomic alterations. We developed a hypoxia ex-vivo culturing system using human GB tissue slices analyzed by RNA-seq whole genome sequencing and whole genome methylation profiling. Further, an immunofluorescence imaging based machine learning method was established to evaluate cell cycle arrest and lineage transitions within an intact tissue architecture. RESULTS Spatial analysis of patient-derived samples revealed that hypoxia-enriched niches are closely associated with chromosomal instability, particularly marked by the published CX2 signature—defined by short, oscillatory single-copy alterations indicative of homologous recombination defects. In ex vivo tumor slices, hypoxia induced a distinct transcriptional program characterized by activation of the unfolded protein response, metabolic reprogramming, and G0-G1 cell cycle arrest, as confirmed by immunofluorescence image analysis and consistent with prior spatial transcriptomics data. Simultaneously, hypoxia promoted epigenetic remodeling at the promoters of key neurodevelopmental genes, including WNT8B and HMGB4, pointing toward a priming of cells into an early neuronal progenitor-like state and increased chromatin accessibility of genes regulating the G0-G1 to G1-S transition. Notably, whole-genome profiling of hypoxia-treated slices identified copy number gains converging in MAPK signaling genes of BRAF/RAF fusions—alterations absent in conventional glioblastoma cell lines. Collectively, these findings highlight a tissue-specific mechanism by which hypoxia drives oncogenic signaling and neuronal lineage plasticity. CONCLUSION Our results define a hypoxia-driven evolutionary program in GB, characterized by concomitant BRAF/MAPK pathway activation and neuronal epigenetic reprogramming. These adaptations promote the emergence of slow-cycling, therapy-resistant subclones poised for recurrence. Targeting the hypoxia-MAPK-lineage axis may present a promising therapeutic strategy to overcome resistance and delay GB relapse.
Villa et al. (Wed,) studied this question.