P04.04.B BIOLOGICAL AND MOLECULAR CHARACTERIZATION OF NOVEL GLIOBLASTOMA TME SUBTYPES REVEALS DISTINCT THERAPEUTIC VULNERABILITIES

Abstract BACKGROUND New precision medicine therapies are urgently required for glioblastoma (GBM). Recently, we developed a novel GBM classification system, identifying three patient clusters uniquely characterized by tumour microenvironment (TME) composition: TMELow, TMEMedium, and TMEHigh1. Building on this work, our current study leverages transcriptomic profiling, integrative network analyses, and spatial multiplex proteomics to deeply characterize these subtypes. We aim to uncover novel molecular vulnerabilities and inform tailored therapeutic interventions in GBM. MATERIAL AND METHODS We performed comprehensive transcriptomic analyses across > 600 GBM samples from publicly available and in-house cohorts. All samples were classified into TME subtypes using MCP-counter for immune cell composition estimation, followed by unsupervised Partitioning Around Medoids (PAM) clustering, as previously described1. Differential expression analysis (DESeq2) and pathway analysis (PROGENy) delineated key biological differences among subtypes. To define subtype-specific master regulators (MRs), we utilized ‘Genome Enhancer’ network modelling. BulkSignalR analysis was applied to map ligand-receptor interactions (LRIs), identifying subtype-specific intercellular communication. Additionally, multiplex spatial proteomics (Cell Dive) is ongoing, planned to validate transcriptomic signatures and investigate the spatial interplay between key immune and tumour markers at the protein level. RESULTS Our analyses revealed that TMEHigh tumours were rich in tumour-associated macrophages (TAMs), with upregulated angiogenesis and TGF-β signalling. These inflamed tumours displayed immunosuppressive LRI networks (e.g., SPP1-ITGB1, TGFB1-TGFBR2). By contrast, TMELow tumours were immune-desert with neural and proliferative signatures, featuring neuronal interactions (e.g., NRXN-NLGN) and EGFR-driven signalling. TMEMedium tumours showed mixed features with moderate immune infiltration. Master regulator analysis highlighted distinct drivers: immunosuppressive factors (e.g., TGFB1, SPP1) dominated the TAM rich subtypes, whereas neural developmental regulators (e.g., NTRK3, PTPRZ1) defined the TAM negative cluster. CONCLUSION Our integrated transcriptomic and network analysis uncovers distinct molecular programs underpinning GBM TME subtypes and reveals candidate intercellular interactions and master regulators that may drive immune suppression or tumour progression. These findings lay the foundation for subtype-specific therapeutic targeting strategies. Ongoing spatial proteomic profiling will further refine our understanding of the GBM microenvironment and support the development of subtype-specific therapeutic interventions.