Skip to main content
eScholarship
Open Access Publications from the University of California

UC San Diego

UC San Diego Electronic Theses and Dissertations bannerUC San Diego

A Neural Network Model of Cancer Identifies Chemotherapies Synergistic with Autophagy Inhibition

Abstract

In cancer, autophagy has been shown to promote resistance to chemotherapy by inducing pro-survival mechanisms. This observation has led to basic and clinical studies combining chloroquine, an autophagy inhibitor, with standard chemotherapies. However, the range of particular chemotherapies likely to benefit from this combination remains unclear. Here we apply DrugCell, a deep learning model of the molecular pathways that govern cancer therapeutic response, to systematically prioritize drugs that promote cell survival through the activation of autophagy. Using this approach, we analyze the response of tumor cell lines to 684 drugs, identifying 21 for which resistance or sensitivity is modulated by genetic mutations in autophagy pathways. We systematically screen these prioritized drugs against a fluorescent readout of autophagic flux, confirming that 14 stimulate this pathway and 1 leads to pathway suppression, for a hit rate of 15/22 or 71%. In contrast, only 1 in 6 of a control set of drugs is found to affect autophagy activity. Finally, we test eight of the autophagy activators for synergy with chloroquine treatment, finding that four are highly synergistic, producing a BLISS score greater than 10. These results suggest that, for these drugs (JQ1, tosedostat, leptomycin B, vorinostat), chloroquine inhibition of autophagic flux interferes with autophagy as a survival response.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View