Abstract
An accurate prediction of the intracranial infiltration tendency and drug response of individual glioblastoma (GBM) cells is essential for personalized prognosis and treatment for this disease. However, the clinical utility of mouse patient-derived orthotopic xenograft (PDOX) models remains limited given current technical constraints, including difficulty in generating sufficient sample numbers from small tissue samples and a long latency period for results. To overcome these issues, we established zebrafish GBM xenografts of diverse origin, which can tolerate intracranial engraftment and maintain their unique histological features. Subsequent single-cell RNA-sequencing (scRNA-seq) analysis confirmed significant transcriptional identity to that of invading GBM microtumors observed in the proportionally larger brains of model animals and humans. Endothelial scRNA-seq confirmed that the zebrafish blood-brain barrier is homologous to the mammalian blood-brain barrier. Finally, we established a rapid and efficient zebrafish PDOX (zPDOX) model, which can predict long-term outcomes of GBM patients within 20 days. The zPDOX model provides a novel avenue for precision medicine of GBM, especially for the evaluation of intracranial infiltration tendency and prediction of individual drug sensitivity.