UC San Diego

Chemotherapies are systemic treatments designed specifically to target various cancer types. However, there are still unknowns about the effectiveness of different chemotherapy drugs in individual cancer patients. Here, we study a mechanism of chemotherapy resistance related to autophagy, a cellular process involving degradation of cellular components through a lysosome-dependent mechanism. We design our experiment by using DrugCell, a visible neural network created by the Ideker lab, which models the hierarchical organization of human cancer cells and predicts the responses to 216 anti-cancer drugs. More specifically, we leverage DrugCell’s visible neural network data to identify top-candidate autophagy-modulating drugs: Vitamin A derivatives, ROS-p38 MAPK-1 Activator, AMPK activators, BL-2 Activators/Inhibitors, CDK Inhibitors, HDAC Inhibitors, microtubule depolymerization, PIK Inhibitors, PLK1 Inhibitors, and Replicative Stress Inducers. Since replicative stress drugs are commonly used to treat head and neck cancers, we characterize this drug group by studying the dose-dependent sensitivities to various head and neck cancer cell-lines. We also test the replicative stress response through western blot analyses checking for the phosphorylation status of ATR and Chk1. To identify the genetic components which dictate the autophagy-drug interactions, we conduct a pilot CRISPR chemogenetic screen and establish the experimental parameters for conducting a screen using autophagy-modulating drugs.