BIOM-47. DECIPHERING GLIOBLASTOMA HETEROGENEITY: INTEGRATING MOLECULAR SIGNATURES AND GENOMIC LANDSCAPES TO PREDICT DRUG SENSITIVITY AND ADVANCE PRECISION ONCOLOGY

Abstract Glioblastoma (GBM) is characterized by significant heterogeneity and a complex tumor microenvironment, frequently leading to treatment failure and underscoring the urgent need for innovative therapeutic strategies. A cornerstone of our work involves utilizing diverse GBM patient-derived xenograft (PDX) models to rigorously evaluate drug efficacy and uncover robust biomarkers of response. Comprehensive drug screening was performed on these models, evaluating agents such as temozolomide, arsenic trioxide (ATO), TRC102, pevonedistat, TAS4464, candesartan cilexetil, selinexor, GB13, DSP-0390, and the combination of ATO with a MNK1 inhibitor. Transcriptome analyses revealed distinct pathway activation profiles associated with drug sensitivity and resistance. Generally, sensitive PDX lines exhibited activation of DNA repair and translation pathways, whereas resistant lines showed upregulation of interferon signaling, antigen presentation, and extracellular matrix pathways. Notably, sensitivity to GB13 correlated with activated Golgi-to-ER retrograde transport, and DSP-0390 sensitivity was linked to metabolic pathway activation, suggesting these pathways as potential predictive biomarkers. Furthermore, genomic sequencing of five PDX models with varying drug sensitivities confirmed characteristic GBM hallmarks: recurrent chromosome 7 gain and chromosome 10 loss. Despite shared aberrations, individual samples displayed unique focal gains and losses, indicating intertumoral heterogeneity. The broadly resistant GBM10 model exhibited elevated copy number variations (CNVs), including gains on chromosomes 1 and 7, and losses on chromosomes 10, 13, 14, and 22. Six patient tumor samples have also been sequenced. Future research will focus on analyzing CNV changes and transcriptomic differential expression within patient data and validating drug efficacy on patient-derived tumor cultures. This comprehensive strategy aims to translate preclinical discoveries into clinically relevant biomarkers, ultimately paving the way for personalized and more effective GBM treatments.