Computational investigation of drug-induced effects on human cardiac electro-mechanics

Abstract

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): NC3Rs Infrastructure for Impart Award (NC/P001076/1) Wellcome Trust Senior Research Fellowship in Basic Biomedical Sciences (214290/Z/18/Z) Background Human-based computer modelling and simulations have been widely used in cardiac electrophysiology, to provide a better understanding of the ionic mechanisms underlying risk of arrhythmias, and their modulation by drugs and diseased conditions. More recently, multiscale computer models of human cardiac electro-mechanics have been developed. These models provide the means for a comprehensive investigation of action potential, calcium transient, active force, and their variability in the population, and can predict drug-induced contractility changes in humans. Purpose This study aims to perform a computational investigation of variability in human cardiac contractility biomarkers and their modulation by well-known drugs. Methods We considered the most recent model of human ventricular electro-mechanics (Margara et al. 2020). We constructed a population of 300 cells by randomly varying the main ionic currents in the model, to represent the biological variability observed in human experimental data. We then simulated the effect of 10 reference compounds, 6 of which with known pro-arrhythmic risk. Simulations were performed at 1 Hz for multiple drug concentrations, using the Virtual Assay software. A set of action potential, calcium transient and active force biomarkers were computed, as well as the electro-mechanical window, and the occurrence of early after-depolarisations and after-contractions in the virtual population. Simulation results were compared against clinical risk of drug-induced arrhythmias and experimental recording from human ventricular myocytes from literature. Results Overall, biomarker variability in the virtual population increased following drug application compared to control conditions. All compounds had a negative inotropic effect in simulation, with a marked decrease of the active tension peak, e.g. -80% for nifedipine 8 nM. This is in agreement with human experimental data for all compounds except Dofetilide, for which no inotropic effect was observed in vitro. Compounds with known risk of arrhythmias provoked early after-depolarisations, which in turn caused after-contractions. Their occurrence in the population increased together with the drug concentration, e.g. 3.6% at 0.16 µM and 48% at 0.48 µM for droperidol. In addition, these compounds also displayed prolonged action potential and calcium transient, and a shortening of the electro-mechanical window, all known biomarkers of pro-arrhythmia. Conclusions We evaluated the effect and the cardiac safety of 10 reference compounds in a population of 300 human ventricular electro-mechanical models. Simulation results were in good agreement with experimental data in human ventricular cardiomyocytes, and they allowed to identify the compounds with a known pro-arrhythmic risk based on drug-induced early after-depolarisations and after-contractions. This methodology provides new insights into variability in human cardiac contractility and its modulation by drugs.