In Silico Populations Optimized on Optogenetic Recordings Predict Drug Effects in Human Induced Pluripotent Stem Cell-Derived Cardiomyocytes

Abstract

All-optical high-throughput systems allow simultaneous high resolution action potential (AP) and Ca2+ transient (CaTr) measurements from cardiomyocytes within multicellular context, offering means to speed up in vitro drug tests. Here, we aim to develop experimentally-constrained in silico models of human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) and hiPSC-CM populations to predict drug effects in humans, by leveraging functional data obtained by all-optical means. Using multi-objective genetic algorithms (MoGAs), we constructed three control populations of in silico hiPSC-CMs, constrained with experimental data of APs and CaTrs recorded at room temperature and non-paced conditions from three different plates containing hiPSC-CM syncytia. We then simulated the effect of increasing doses of Diltiazem (130 models), Cisapride (200 models) and Astemizole (200 models) in the three populations, respectively. Comparing model predictions with the experimental drug administration (not used for the optimization/calibration of the populations) revealed good agreement with experiments: e.g. Diltiazem shortened APs while Astemizole and Cisapride prolonged APs.