Aims: This study aimed to investigate how drug-hERG (IKr) binding kinetics modulate action potential (AP) prolongation in endocardial and midmyocardial cells, and how these localized effects translate into a realistic three-dimensional (3D) biventricular model to better evaluate proarrhythmic risk compared to traditional single-cell (0D) approaches.
Methods: Dynamic Markov chain IKr block models for three drugs with different blocking mechanisms (Astemizole, Diltiazem, and Ranolazine) were adapted to reproduce experimental data at 37ºC. These models were integrated into an optimized O'Hara-Rudy (ORd-Dutta) ventricular action potential model. Simulations were performed at a physiological pacing frequency of 1 Hz using IC50 concentrations. The 0D simulations evaluated AP prolongation and the occurrence of early afterdepolarizations (EADs). The 3D biventricular model, which incorporated realistic geometry and transmural heterogeneity (endocardial, M-cells, and epicardial layers), was used to evaluate the spatial propagation of electrical signals and the generation of arrhythmias.
Results: In 0D simulations, all three drugs prolonged the AP. However, only Astemizole induced EADs, which occurred specifically in midmyocardial cells. In 3D tissue simulations, Diltiazem and Ranolazine prolonged the AP but did not trigger any arrhythmias. Conversely, Astemizole generated EADs in midmyocardial nodes within the septal region. The 3D environment allowed the observation of these local alterations propagating towards neighboring regions, successfully initiating anomalous activations and transmural reentry phenomena.
Conclusion: Integrating dynamic Markov IKr models into 3D ventricular simulations reveals proarrhythmic vulnerabilities—such as midmyocardial EADs scaling into full tissue reentries—highlighting the critical importance of incorporating both channel-drug kinetics and three-dimensional geometry for accurate in silico cardiac safety evaluations.