Drug combinations provide effective treatments for diverse diseases, but also represent a major cause of adverse reactions. Currently there is no systematic understanding of how the complex cellular perturbations induced by different drugs influence each other. Here, we introduce a mathematical framework for classifying any interaction between perturbations with high-dimensional effects into 12 interaction types. We apply our framework to a large-scale imaging screen of cell morphology changes induced by diverse drugs and their combination, resulting in a perturbome network of 242 drugs and 1832 interactions. Our analysis of the chemical and biological features of the drugs reveals distinct molecular fingerprints for each interaction type. We find a direct link between drug similarities on the cell morphology level and the distance of their respective protein targets within the cellular interactome of molecular interactions. The interactome distance is also predictive for different types of drug interactions.

The undruggable nature of oncogenic Myc transcription factors poses a therapeutic challenge in neuroblastoma, a pediatric cancer in which MYCN amplification is strongly associated with unfavorable outcome. Here, we show that CYC065 (fadraciclib), a clinical inhibitor of CDK9 and CDK2, selectively targeted MYCN -amplified neuroblastoma via multiple mechanisms. CDK9 — a component of the transcription elongation complex P-TEFb — bound to the MYCN -amplicon superenhancer, and its inhibition resulted in selective loss of nascent MYCN transcription. MYCN loss led to growth arrest, sensitizing cells for apoptosis following CDK2 inhibition. In MYCN -amplified neuroblastoma, MYCN invaded active enhancers, driving a transcriptionally encoded adrenergic gene expression program that was selectively reversed by CYC065. MYCN overexpression in mesenchymal neuroblastoma was sufficient to induce adrenergic identity and sensitize cells to CYC065. CYC065, used together with temozolomide, a reference therapy for relapsed neuroblastoma, caused long-term suppression of neuroblastoma growth in vivo, highlighting the clinical potential of CDK9/2 inhibition in the treatment of MYCN -amplified neuroblastoma.