A model-guided pipeline for drug cardiotoxicity screening with human stem-cell derived cardiomyocytes

New therapeutic compounds go through a preclinical drug cardiotoxicity screening process that is overly conservative and provides limited mechanistic insight, leading to the misclassification of potentially beneficial drugs as proarrhythmic. There is a need to develop a screening paradigm that maintains this high sensitivity, while ensuring non-cardiotoxic compounds pass this phase of the drug approval process. In this study, we develop an in vitro-in silico pipeline using human induced stem-cell derived cardiomyocytes (iPSC-CMs) to address this problem. The pipeline includes a model-guided optimization that produces a voltage-clamp (VC) protocol to determine drug block of seven cardiac ion channels. Such VC data, along with action potential (AP) recordings, were acquired from iPSC-CMs before and after treatment with a control solution or a low-, intermediate-, or high-risk drug. We identified significant AP prolongation (a proarrhythmia indicator) in two high-risk drugs and, from the VC data, determined strong ion channel blocks that led to the AP changes. The VC data also uncovered an undocumented funny current (I f ) block by quinine, which we confirmed with experiments using a HEK-293 expression line. We present a new approach to cardiotoxicity screening that simultaneously evaluates proarrhythmia risk (e.g. AP prolongation) and mechanism (e.g. channel block) from iPSC-CMs.