Image-based morphological profiling of autophagy phenotypes in Zika virus infected cells.

Autophagy is a dynamic intracellular process that is essential in maintaining cellular homeostasis. Its potential as a therapeutic target is exemplified by its dysregulation in many disease states such as Zika virus (ZIKV) infection. ZIKV poses a significant burden to human health and hijacks autophagy to disrupt development. Here, we develop an experimental and computational pipeline to dissect ZIKV hijacking of autophagy in live cells. We build on our previously developed high-throughput scalable image-based profiling approach applied to small molecule perturbation of autophagy. In this study, we expand and modify the image-based profiling pipeline to observe autophagy and virus infection side-by-side using ZIKV encoding a fluorescent reporter. We observe cell and autophagy morphology changes after ZIKV infection and found distinct feature profiles in both infected and uninfected cells treated with ZIKV. We compare ZIKV-infected morphologies to those of cells treated with autophagy-regulating drugs and find ZIKV-induced changes similar to treatment with a combination of autophagy inducer and inhibitor. Using the pipeline, we also train and build a deep learning classifier to identify infected cells using a cellular autophagy reporter without the help of a fluorescent reporter virus. This shows the versatility of image-based profiling in infection-based systems.