Ependymoma is a rare tumor arising from brain or spinal ependymal cells. Due to limited treatment options besides surgery and radiotherapy, the management of ependymoma remains challenging. This study aims to establish a novel ex vivo model and to find effective systemic drugs for the treatment of ependymoma.

Tumor Organoids (TOs) were generated by the reaggregation of single cell suspensions from freshly resected ependymoma tissue. Live/dead staining and ATP-based CellTiterGlo3D were used to assess the viability. To perform the high-throughput drug screening, TOs were generated from a spinal NF2-mutated ependymoma. The robotic liquid handler Hamilton Microlab STAR ® was used to screen 179 FDA-approved antineoplastic drugs (NCI AOD XI) in the range of 10 nmol/l to 30 µmol/l. A multi-step approach was employed to generate dose-response curves for all drugs.

TOs reaggregated within 2-3 days to dense mini-tumors resulting in a reduced diameter by 50% independent of the seeded cell number. TOs primarily comprised viable cells, whereas dead cells were mainly located on the outside. Half-maximum inhibitory concentrations (IC50) were obtained for 34% (n=61/179) of the drugs, the remaining drugs showed no efficacy up to 30 µmol/l. To assess if potential therapeutic concentrations were reached, IC50 values were compared to the peak serum drug concentrations (Cmax). By using this approach, 11.7% of the drugs (n=21/179) can be classified as sensitive (Cmax/IC50 ratio > 1). Of those, the drug classes HDAC inhibitors (Belinostat and Romidepsin), anthracyclines (Doxorubicin and Epirubicin), and the proteasome inhibitor Carfilzomib demonstrated the highest Cmax/IC50 ratios, suggesting that therapeutic drug levels can be achieved.

Taken together, we generated standardized ependymoma TOs and identified HDAC, anthracyclines, and proteasome inhibitors as therapeutic candidates for ependymoma. Our findings may help guide the clinical decision-making for the systemic treatment of ependymoma in the future.