Deep Learning Predicts Patterns of Cardiotoxicity in a High-Content Screen Using Induced Pluripotent Stem Cell–Derived Cardiomyocytes

Abstract

Drug-induced cardiotoxicity and hepatotoxicity are major causes of drug attrition. To decrease late-stage drug attrition, pharmaceutical and biotechnology industries need to establish biologically relevant models that use phenotypic screening to predict drug-induced toxicity. In this study, we sought to rapidly detect patterns of cardiotoxicity using high-content image analysis with deep learning and induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs). We screened a library of 1280 bioactive compounds and identified those predicted to have cardiotoxic liabilities using a single-parameter score based on deep learning. Compounds with major predicted cardiotoxicity included DNA intercalators, ion channel blockers, epidermal growth factor receptor, cyclin-dependent kinase, and multi-kinase inhibitors. We also screened a diverse library of molecules with unknown targets and identified chemical frameworks with predicted cardiotoxic liabilities. By using this screening approach during target discovery and lead optimization, we can de-risk early-stage drug discovery. We show that the broad applicability of combining deep learning with iPSC technology is an effective way to interrogate cellular phenotypes and identify drugs that protect against diseased phenotypes and deleterious mutations.

Functional analysis of a common BAG3 allele associated with protection from heart failure

“…This adds a layer of information to the putative interaction list by quantifying the contribution of each one of the factors to the myofibrillar structure, allowing us to prioritize targets by identifying which ones are involved in BAG3-associated maintenance of sarcomeric integrity. A similar approach has been used to study small molecule perturbations on iPS-CMs and other cell types 41,42 . The use of iPS-CMs as a model for cardiomyopathies is not without limitations: iPS-CMs are known to lack some important features of mature myocytes and are not exposed to mechanical stress from pressure load, which is a predominant factor in the challenging of the myocardial protein quality control machinery 28,[43][44][45] .…”

“…Nine images per well were used in all conditions. Images were used to train a supervised machine learning image classification model using PhenoLearn framework (www.phenolearn.com), as described elsewhere 41,42 . This model was used to generate a classification score for each image from the screening plates indicating the similarity to the BAG3 siRNA (score closer to 0) or Scramble siRNA (score closer to 1) training sets.…”

Integrating nonlinear analysis and machine learning for human induced pluripotent stem cell‐based drug cardiotoxicity testing

“…Specifically, decision trees and quadratic discriminant analysis achieved 92% accuracy, somewhat outperforming SVM and kNN at 91% accuracy (Teles et al., 2021). In addition to the functional analyses based upon contractile motion or calcium transients, deep learning models trained with microscopic images of hiPSC‐CMs have also been proven to be effective at detecting adverse effects of cardiotoxic drugs (Grafton et al., 2021; Maddah et al., 2020; Orita et al., 2019).…”