HLS-Cardiac Safety AI Trained Human Heart and Micro Heart Model

点击翻译按钮获取中文摘要

HLS17-12. The US FDA is considering to establish a new cardiac safety assessment approach defined by a new paradigm called, “Comprehensive in vitro Proarrhythmia Assay (CIPA)”. The CIPA will 1) assess drug effects on each cardiac ion channel type individually using a high- throughput assay ion channel assays, 2) compute net effect on repolarization and risks for torsade pointes (TdP) using a mathematical model, and 3) confirm the computational prediction by measuring the drug’s effects on action potentials in induced pluripotent stem cell (iPSC) derived human cardiac myocytes (CMs). This paradigm shift, if successful, could reduce the cost of cardiac safety analyses by replacing or lowering the requirement to perform an expensive ($2-4 million) thorough QT study during clinical trials. Protecting consumers from drug induced arrhythmia and sudden deaths is a paramount importance for the regulators and pharmaceutical companies as well as lowing the cost of drug development. Many cardiac safety scientists, however, are skeptical about CIPA’s approach since CMs derived from human iPSCs exhibit a poor excitation-contraction coupling due to their immaturity. In addition, a proposed CIPA mathematical model was developed to simulate electrophysiology of human adult CMs, so there is a mismatch between experimental system and computational tool. To address these concerns, we proposed three specific aims in tw0 phases by following Fast- Track SBIR processes. Phase I feasibility Aim 1 will measure drug-induced changes in AP and CaT using human adult heart slices isolated from human donors. Here we will confirm our successful handling and analyzing human adult heart slices, which will be based on a recently published protocol by our collaborator, Dr. Igor Efimov, at George Washington University. After this validation, we will move on to perform the following two studies: Aim 2. Compare drug-induced changes in AP and CaT in NuHearts generated from adult CMs and cardiac fibroblasts from same human donor hearts; Aim 3. Validate and improve the computational models and train an artificial intelligence to predict cardiac safety risks of unknow compounds. After our successful completion proposed projects, we can establish an unprecedented cardiac safety assessment platform that can predict safety issues using a well-trained AI without doing any experiments using human heart slices that are rarely accessible for most of the safety laboratories or biotech firms.