Introduction: Patients with heart failure with preserved ejection fraction (HFpEF) frequently develop atrial fibrillation (AF). Impaired diastolic function of the left ventricle (LV), leading to elevated filling pressures and subsequent left atrial remodeling, creates a potent substrate for AF initiation and persistence. LV diastolic dysfunction can therefore precede and even predict AF. Conversely, impaired atrial contraction during AF disrupts LV filling and hemodynamics. This study investigated how AF may influence LV diastolic function in humans.
Methods and Results: Fifty patients with either rate-controlled AF (n=23) or sinus rhythm (SR, n=27) with preserved LV function (LV ejection fraction >50%) undergoing aortic valve replacement for aortic stenosis were included, and LV myocardium from septal resections was analyzed. Myofilament stiffness in LV cardiomyocytes was significantly higher in AF patients than in SR patients, a change reversible with protein kinase G (PKG), Sodium-glucose co-transporter 2 inhibitors (SGLT2i) like empagliflozin (EMPA), and soluble guanylate cyclase (sGC) stimulator treatment. AF patients also showed reduced Ca2+-activated maximum tension in LV cardiomyocytes, indicating impaired cardiomyocyte function, and lower phosphorylation of myofilament proteins, including diminished PKG-dependent phosphorylation. Additionally, nitric oxide (NO) levels were impaired in the LV of AF patients, associated with reduced sGC activity, lower cyclic guanosine monophosphate (cGMP) levels, and impaired PKG activity. In line with these findings, isolated and pressurized coronary arterioles from AF patients exhibited defective bradykinin-induced endothelium-dependent vasodilation.
Culturing healthy human ventricular cardiomyocytes and endothelial cells with serum from AF patients induced secretion of pro-inflammatory factors, suggesting a serum-driven inflammatory response. Co-culturing ventricular cardiomyocytes with endothelial cells further amplified this inflammatory response, highlighting a potential mechanism linking systemic inflammation and cardiac dysfunction in HFpEF.
For translational analysis, we used in vitro AF-simulation with human ventricular iPSC-cardiomyocytes (hiPSC-CM) derived from healthy donors exposed to chronic arrhythmic pacing (60 bpm, 40% beat-to-beat variability vs. 60 bpm in controls) for 7 days. AF-simulated hiPSC-CM showed reduced NO levels, sGC and cGMP levels, and impaired PKG activity, along with increased oxidative stress markers, such as 3-nitrotyrosine, and elevated NADPH oxidase 2 (NOX2) and 4 (NOX4) expression in the LV of AF patients.
Conclusion: This study provides mechanistic evidence linking AF to human LV diastolic dysfunction. While diastolic dysfunction predisposes to AF, the interaction between AF and LV dysfunction appears unidirectional and detrimental, exacerbating diastolic performance and increasing HFpEF risk.
