BIOM-86. MAPPING REGION-SPECIFIC TRANSCRIPTOMIC PROGRAMS IN GLIOBLASTOMA FROM HISTOLOGICALLY ANNOTATED RNA-SEQ DATA

Abstract Glioblastoma (GBM) exhibits spatial and molecular heterogeneity. Understanding region-specific transcriptional programs within distinct tumor microenvironments is crucial for identifying niche-targeted therapeutic strategies. This study uses the Ivy Glioblastoma Atlas Project (Ivy GAP) to characterize transcriptional variation across histologically annotated tumor regions. We analyzed annotated bulk RNA-seq data from Ivy GAP, covering five tumor regions: cellular tumor (CT, n=111), infiltrating tumor (IT, n=24), leading edge (LE, n=19), microvascular proliferation (MVP, n=50), and perinecrotic zone (PNZ, n=65). Differential gene expression (DGE) analysis was performed using CT as reference. Enrichment analysis was conducted using Gene Ontology and KEGG databases. Prognostic significance of differentially expressed genes (DEG) was assessed using CGGA-GBM dataset for survival analysis. Finally, we performed a pseudo-spatial analysis by plotting normalized Fragments Per Kilobase of transcript per Million mapped fragments (FPKM) expression levels of DEG genes across tumor regions to infer expression dynamics across distinct regions. Each region exhibited distinct transcriptional profiles. IT and LE were enriched for synaptic transmission, neuronal signaling, calcium signaling and excitable membrane potential genes; MVP for angiogenesis and extracellular matrix remodeling; and PNZ for hypoxia-induced metabolic adaptation. Several genes, including CACNG3 (20.5 vs 13.6 m, p=0.03), HCN1 (19.8 vs 13.9 m, p=0.02), SULT4A1(19.2 vs 13.9 m, p=0.03), and SYCE1 (21.7 vs 13.6 m, p=0.0006), were strongly associated with survival in CGGA-GBM dataset. Pseudo-spatial mapping allowed visualization of region-specific gene expression gradients. This integrative analysis reveals spatially distinct molecular landscapes in GBM and identifies prognostic gene markers linked to specific tumoral microenvironments. Our findings define region-specific transcriptional programs in GBM, offering mechanistic insights into tumor microenvironment heterogeneity and nominating candidate genes for future functional validation.