Modeling substrates for atrial fibrillation: from acute stretch to chronic remodeling

Background Atrial Fibrillation (AF) is linked to dilatative remodeling, which promotes the substrate that sustains arrhythmia. Current preclinical models to study AF substrates are largely animal-based or rely on engineered tissues that tend to lack biomimicry. Atrial living myocardial slices (aLMS) provide a human-based translational platform to study atrial remodeling, from acute stretch to chronic adaptation, and to interrogate AF-related triggers under controlled conditions. Aim To assess the effects of acute and chronic atrial dilatation (AD) on refractoriness and biomechanics using a biomimetic aLMS platform, and to model AF-related triggers through chronic stretch and tachypacing in vitro. Methods 57 atrial tissue samples were obtained from surgical or explanted hearts with and without preoperative AD and AF, yielding 299 aLMS cultured under biomimetic conditions (1 mN preload, 1 Hz pacing). The effect of progressive stretch on the refractory period (RP) was assessed in 27 aLMS without prior AD or AF (10–40% increase). Contraction metrics, including peak force (Fmax), contraction duration (CD), contraction and relaxation velocity, and RP, were quantified in aLMS from non-dilated and dilated atria, and from patients with and without AF. To model chronic stress relevant to AF, 16 aLMS from non-dilated, non-AF patients were subjected to increased preload (30% stretch) or tachypacing (180 bpm) for 7 days. Results In vitro, incremental stretch of aLMS (10–40%) progressively prolonged RP (221.7 → 372.5 ms, p=0.04). Compared with patients without AF or AD, aLMS from patients with severe AD showed mechanical impairment, with reduced Fmax (2668.7 vs 613.9 µN, p=0.04), slower contraction and relaxation velocities (35623.8 vs 7670.1 µN/s and 26489.8 vs 6696.1 µN/s, p=0.04), and longer RP (253.3 vs 285.4 ms, p=0.03). Mild-to-moderate AD showed a similar pattern, albeit less pronounced. aLMS from patients with AF largely preserved force and velocities but demonstrated greater RP variability (IQR 169.3 vs 75.3 ms, p=0.02). When AF coexisted with AD, force and velocities declined, while both CD and RP prolonged. Chronic stretch in aLMS reproduced a dilatative phenotype with lower Fmax (-659.9 µN) and longer RP (+20.4 ms), while tachypacing produced trends toward lower Fmax (-475.3 µN) and shorter CD (-24.2 ms). Conclusion aLMS provide mechanistic insight into atrial remodeling from acute load to chronic adaptation, offering a human-based platform to study AF substrate formation. Acute AD prolongs refractoriness, while chronic AD impairs contractile performance and further extends RP. Patient phenotypes can be recapitulated in vitro, enabling controlled investigation of AF-related triggers. Future work will integrate work loop measurements, real-time calcium imaging, and high-density epicardial mapping in patients with chronic dilatation and simulated acute atrial volume overload to elucidate mechanisms linking AD to AF, and inform future therapies.