Atrial fibrillation (AF) is often found in patients with heart failure (HF). Growing clinical evidence indicate that the arrhythmic component of AF alone could contribute to ventricular dysfunction. However, the pathophysiological mechanisms of a clinically relevant normofrequent ventricular arrhythmia as it occurs in AF on the human ventricle are unknown.

To investigate the effects of normofrequent AF on the human ventricle we investigated ventricular myocardium from patients with sinus rhythm (SR) or normofrequent AF in the absence of HF (compensated hypertrophy, mean EF>50%, matched clinical characteristics, derived from morrow resections during AVR). In histological analysis we detected no difference between SR (n=9 patients) vs. AF (n=6) regarding the amount and distribution of fibrosis. As Ca-handling is a major determinant of contractile function, we isolated human ventricular cardiomyocytes and studied cellular Ca-handling (Fura-2 AM). Interestingly, systolic Ca-transient amplitude was significantly reduced, and we found a significantly prolonged Ca-elimination time in AF patients (n=82 cells/ 8 patients) compared to SR (n=115/11). Patch-clamp experiments revealed a prolonged action potential duration in AF (n=16/6 vs 29/6).

For the standardized evaluation of the mechanisms of a persistent normofrequent arrhythmia, we simulated AF in vitro by using arrhythmic (60 bpm, 40% beat-to-beat variability) or rhythmic (60 bpm) field stimulation. For studying the cellular effects in a model suitable for chronic pacing (up to 7 days), we utilized human iPSC cardiomyocytes (iPSC-CM) from healthy donors (n=6). After 7 days, irregularly paced iPSC-CM (n=69 vs. 75 cells/6 differentiations each) exhibited a significantly reduced systolic Ca-transient amplitude, a prolonged Ca-elimination time as well as a reduced SR Ca-load and a trend towards a lower SERCA2a activity compared to control. Confocal line-scans of arrhythmic paced cells (Fluo-4 AM, n=68 vs. 63 cells/7 differentiations) showed an increased diastolic SR Ca-release, which might also explain the reduced SR Ca-content. Moreover, in irregularly paced iPSC-CM we found significantly increased levels of cytosolic Na (n=80 vs. 69 cells/7 differentiations each) and in patch-clamp experiments a significantly prolonged action potential duration (n=11 vs. 14 cells/3 differentiations). Therefore, distinct electrophysiological remodelling may underlie the ventricular contractile dysfunction upon arrhythmia in AF patients. We further elucidated the underlying mechanisms by demonstrating that markers of oxidative stress (H2O2, LPO) are increased in human ventricular myocardium from patients with AF (n=6 patients each), while the reduced state of Glutathione (GSH) is decreased, which was associated with an enhanced NOX2/-4 activity in patients with AF (n=7) compared to SR (n=6). Consecutively, Ca²⁺/calmodulin-dependent protein kinase II δ (CaMKII) was found to be more oxidized (CaMKII-Met281/282) in the ventricle of AF patients (n=7 each) leading to an increased CaMKII activity, which adversely regulated EC-coupling protein phosphorylation including RyR2 hyperphosphorylation.

This study demonstrates that normofrequent arrhythmia/AF impairs human ventricular EC-coupling via increased oxidative stress and CaMKII oxidation. Thus, this study provides the first translational mechanistic characterization and the potential negative impact of isolated AF in the absence of tachycardia on the human ventricle.