Patient-derived Pediatric Glioblastoma Models Provide Key Insights into IDH1-driven Drug Resistance

Background

Pediatric glioblastoma (GBM) is a rare, aggressive brain cancer with molecular characteristics distinct from adult GBM. IDH1 mutations, which occur more frequently in pediatric cases, are associated with unique clinical outcomes and resistance mechanisms. Despite its severity, GBM remains incurable, with limited treatment options. Pediatric populations rarely participate in clinical trials, making clinically relevant in vitro models critical for preclinical research and therapeutic development. The Human Cancer Models Initiative (HCMI) has developed patient-derived brain tumor models, including organoids and spheroids, annotated with comprehensive clinical and molecular data. These models offer a platform to study tumor biology and drug response in IDH1-mutant GBM.

Methods

Patient-derived glioblastoma models from the HCMI biobank—representing primary and recurrent tumors—were genomically profiled for key pediatric GBM mutations (IDH1, ATRX, TP53, KRAS, RELA). Histopathology confirmed IDH1 status and supported molecular classification. Genomic data were compared to patient records and The Cancer Genome Atlas (TCGA) to validate model fidelity. A subset of models was exposed to a panel of four compounds, including standard and experimental drugs. Drug sensitivity was assessed via 12-point dose curves, with IC50 values calculated. Cytotoxicity was measured using live/dead staining and ATP-based viability assays. This integrated approach links genomic alterations to drug resistance, supporting the development of targeted therapies for high-risk pediatric GBM.

Results

Patient-derived GBM organoid models were sequenced and confirmed to carry key mutations, including IDH1 p.R132H, ATRX p.C1590Y/p.R1739, TP53 p.R273C, KRAS Q22R/G13R, and RELA rearrangements. Drug screening across a 10-compound panel revealed variable responses based on genotype and dosage. Taxanes and platinum agents showed broad cytotoxicity, while PARP and KRAS inhibitors exhibited limited effects except at high concentrations. Viability was assessed via fluorescent staining and ATP-based assays, demonstrating the utility of these models for genotype-informed therapeutic screening in pediatric GBM.

Conclusion

HCMI-derived pediatric GBM models recapitulate IDH1-driven genomic features and reveal mutation-specific drug responses. These 3-D platforms support high-throughput screening and integrated genomic-pharmacologic profiling to inform targeted therapy development for high-risk pediatric GBM. Importantly, these preclinical models bridge a critical gap by providing insights from patient populations that cannot participate in clinical trials, reinforcing their role in precision oncology.