MDB-24. A DATA-DRIVEN APPROACH FOR THE IDENTIFICATION OF NEW SYNERGISTIC DRUG COMBINATIONS AGAINST NEURAL CANCERS

High-risk medulloblastomas are tumours with a poor survival prognosis (≤ 50%), which are currently treated with radiation therapy and DNA-damaging chemotherapy. The severe side effects of these treatments, including developmental problems and secondary tumours, underscore the need for new treatments with reduced toxicity. An attractive target for such therapy is the Myc-family proteins since high-risk medulloblastomas amplify either N-Myc (SHH-MB, G4-MB) or c-Myc (G3-MB). While the proteins from the Myc family themselves are often referred to as pharmacologically intractable, their levels and activity can be modulated by indirect targeting, e.g. inhibition of proteins located upstream or downstream of c-Myc/N-Myc. The diverse cellular processes and pathways that c-Myc and N-Myc are involved in, however, suggest that several Myc pathways must be interrupted simultaneously to achieve a lasting effect. To identify such multimodal interventions against Myc, we performed a principal component analysis of RNA sequencing data from patient-derived medulloblastoma cells (MB002, G3, MYC-amplified) and DAOY cells (SHH-group, no MYC amplification) that were treated with 120 small molecules. In the subsequent analysis, we focused on a selection of ~ 40 genes that represent common upstream regulators and downstream targets of either c-Myc or N-Myc. Applying a principal component analysis to these particular markers, we expect that the drugs that affect different pathways associated with c-Myc activity and are located at the opposite ends of the PC1 and PC2 axis should have a synergistic effect. Additionally, we expected that the drugs located closer to each other, which affect closely related pathways, would either have no effect or would be antagonistic. Our PCA distribution method can be used to select synergistic combinations in a manner that is easily extended to new compounds. This approach may allow fast screening of compounds for estimation of the treatment effect.