P01.45.A FUNCTIONAL DRUG SCREENING REVEALS THE LIMITATIONS OF GENOTYPE-BASED TARGETED THERAPY IN GLIOBLASTOMA

Abstract BACKGROUND Glioblastoma (GBM) remains one of the most aggressive brain cancers, with standard treatment unchanged since 2005. Although targeted therapies have shown promise in other cancers, multiple clinical trials in GBM have failed to improve overall survival. This limited success may be explained by poor drug penetration across the blood-brain barrier, the highly invasive growth pattern of GBM, and its inter- and intratumoral heterogeneity. Together, these factors may influence the effectiveness of traditional genotype-based targeted therapy approaches in GBM. In this study, we evaluated a genotype-driven approach by screening patient-derived glioma stem-like cells (GSCs) with targeted therapies chosen based on frequently observed gene amplifications and deletions, herinafter referred to as copy number variants (CNVs). MATERIAL AND METHODS CNV profiling was performed on GBM tissue samples and GSCs using whole genome sequencing, followed by CNV calling with the tool CNVkit. GSCs were only included if CNVs identified in the original tissue were retained in the corresponding GSCs, ensuring that selected models accurately reflect the genomic profile targeted for treatment. Twelve GSCs were treated with a panel of twelve drugs targeting CDKN2A, CDK4, EGFR, KIT, PIK3CA and NF1, which were selected based on six frequently occurring CNVs in the TCGA dataset. Drug responses were assessed to evaluate efficacy and correlation with the CNVs. RESULTS Drug screening revealed no consistent pattern in therapeutic response between samples with specific CNVs and those without, suggesting that these genomic alterations alone did not predict sensitivity to the corresponding targeted therapies. Notably, GSCs with an NF1 loss showed a trend toward increased sensitivity to MEK inhibitors compared to NF1 wildtype cultures — a difference primarily driven by response to one of the two MEK inhibitors tested. These findings emphasize the limited predictive value of common CNVs for targeted monotherapy and support the use of personalized screening approaches to identify more effective treatment options for GBM. CONCLUSION Our study shows that selecting targeted therapies based on frequently altered CNVs in GBM does not result in consistent or effective treatment responses, underscoring the biological complexity of the disease. While GSCs with an NF1 loss revealed a trend toward increased sensitivity to MEK inhibitors, this effect was primarily observed with one of the MEK targeted compounds, highlighting that even within genetically defined subgroups, drug response can be dependent on the specific drug applied. These results emphasize the limitations of using CNVs alone to guide personalized therapy and suggest that alternatives like functional drug screening approaches warrants further exploration.