Individual Contributions of Cardiac Ion Channels on Atrial Repolarization and Reentrant Waves: A Multiscale In-Silico Study

The excitation, contraction, and relaxation of an atrial cardiomyocyte are maintained by the activation and inactivation of numerous cardiac ion channels. Their collaborative efforts cause time-dependent changes of membrane potential, generating an action potential (AP), which is a surrogate marker of atrial arrhythmias. Recently, computational models of atrial electrophysiology emerged as a modality to investigate arrhythmia mechanisms and to predict the outcome of antiarrhythmic therapies. However, the individual contribution of atrial ion channels on atrial action potential and reentrant arrhythmia is not yet fully understood. Thus, in this multiscale in-silico study, perturbations of individual atrial ionic currents (I(Na), I(to), I(CaL), I(Kur), I(Kr), I(Ks), I(K1), I(NCX) and I(NaK)) in two in-silico models of human atrial cardiomyocyte (i.e., Courtemanche-1998 and Grandi-2011) were performed at both cellular and tissue levels. The results show that the inhibition of I(CaL) and I(NCX) resulted in AP shortening, while the inhibition of I(Kur), I(Kr), I(Ks), I(K1) and I(NaK) prolonged AP duration (APD). Particularly, in-silico perturbations (inhibition and upregulation) of I(Kr) and I(Ks) only minorly affected atrial repolarization in the Grandi model. In contrast, in the Courtemanche model, the inhibition of I(Kr) and I(Ks) significantly prolonged APD and vice versa. Additionally, a 50% reduction of I(to) density abbreviated APD in the Courtemanche model, while the same perturbation prolonged APD in the Grandi model. Similarly, a strong model dependence was also observed at tissue scale, with an observable I(K1)-mediated reentry stabilizing effect in the Courtemanche model but not in the Grandi atrial model. Moreover, the Grandi model was highly sensitive to a change on intracellular Ca(2+) concentration, promoting a repolarization failure in I(CaL) upregulation above 150% and facilitating reentrant spiral waves stabilization by I(CaL) inhibition. Finally, by incorporating the previously published atrial fibrillation (AF)-associated ionic remodeling in the Courtemanche atrial model, in-silico modeling revealed the antiarrhythmic effect of I(Kr) inhibition in both acute and chronic settings. Overall, our multiscale computational study highlights the strong model-dependent effects of ionic perturbations which could affect the model’s accuracy, interpretability, and prediction. This observation also suggests the need for a careful selection of in-silico models of atrial electrophysiology to achieve specific research aims.