Alia Henedi, Jalal Cherkaoui, Stelian Camara Dit Pinto, Steven M Levine, Mohammed Cherkaoui, Nicolas R Gallo, Kenza E Benzeroual
Our work provides a foundation for the creation of digital twin models that incorporate patient-specific electrophysiology, offering a scalable platform for in-silico cardiac safety. By linking genetic polymorphisms to context-dependent functional outcomes, this approach supports early-stage candidate prioritization, offers a scalable platform for genotype-specific risk stratification, and advances the implementation of precision cardiotoxicity screening in drug development and clinical safety assessments.
INTRODUCTION: Drug-induced cardiotoxicity remains a leading cause of drug development failures and market withdrawals. Despite advances in preclinical testing, current in-silico models often overlook genetic variability, limiting their ability to capture patient-specific responses. We present a computational modeling framework that integrates known genetic variants in cardiac ion channels to predict individual differences in drug-induced arrhythmogenic risk.
METHODS: We simulated concentration-dependent effects of amiodarone across combinations of hERG and Nav1.5 alleles in five human cardiac cell types: endocardial, epicardial, midwall, Purkinje, and atrial cells. APD90 values were evaluated across all cell types, whereas qNet was calculated in ventricular cells only to assess genotype and cell type dependent differences in electrophysiological response and torsadogenic risk.
RESULTS: Our results highlight how genetic variations and cell type context influence electrophysiological response to amiodarone, uncovering high-risk profiles otherwise masked in population-averaged models. These simulations reveal key insights with translational and regulatory relevance: genetic background meaningfully alters drug response; midmyocardial cells are disproportionately vulnerable; the same mutation can produce different effects across cell types; Purkinje cells may serve as silent proarrhythmic substrates; and celltype- specific differences in APD90 and qNet may provide additional insight beyond APD prolongation alone, and consequently improve prediction of torsades de pointes risk.
CONCLUSION: Our work provides a foundation for the creation of digital twin models that incorporate patient-specific electrophysiology, offering a scalable platform for in-silico cardiac safety. By linking genetic polymorphisms to context-dependent functional outcomes, this approach supports early-stage candidate prioritization, offers a scalable platform for genotype-specific risk stratification, and advances the implementation of precision cardiotoxicity screening in drug development and clinical safety assessments.