Decoding atrial mechanics: human atrial slices as a model for atrial disease

Living myocardial slices (LMS) are ultrathin cardiac sections that preserve the native myocardial architecture in near-physiological conditions, traditionally sourced from animal-based ventricular biopsies. With recent advances, we have developed high-quality human atrial LMS (aLMS) for studying atrial pathology such as atrial fibrillation (AF). In cardiac disease, pathological remodelling disrupts excitation-contraction coupling, impairing atrial function and accelerating clinical decline. The biomechanical response to electrical cues, however, remains largely unexplored. Human-based aLMS offer an accurate model of atrial pathology, providing insights into atrial biomechanics and aiding targeted therapies.

To characterize biomechanical profiles of aLMS and expand the analysis across patient subsets of cardiac disease to uncover mechanisms underlying atrial dysfunction.

Atrial tissue samples were collected during cardiac surgery from donors and patients with AF, heart failure (HF), and echocardiographically confirmed atrial dilation. aLMS were created following our optimised protocol and cultured in biomimetic chambers with isotonic loading and electrical pacing to mimic in vivo conditions. Contraction metrics, including peak force (Fmax), time to peak (TTP), time to relaxation (TTR) and refractory periods (RP) were measured using MyoDish software. Linear mixed-effects models were used for statistical analysis.

Analysis of 250 aLMS from 48 patients revealed distinct biomechanical trends. Fmax was non-significantly lower in aLMS as compared to ventricular LMS (1336.9 vs 3076.1 µN, p=0.30) with a trend towards shorter TTP (86.8 vs 185.3 ms, p=0.09). Healthy aLMS exhibited a higher, though non-significant, Fmax than diseased aLMS (2465.2 vs 1319.1 µN, p=0.24). In AF samples, a non-significant shift towards higher Fmax (2319.7 vs 1328.0 µN, p=0.40), and an increased RP variability (173.6 vs 81.9 ms, p=0.02) was observed versus non-AF cases. HF aLMS displayed a significantly lower Fmax (1013.2 vs 2798.4 µN, p=0.04) and longer TTR (168.3 vs 125.4 ms, p=0.02) compared to non-HF patients. Dilated slices exhibited similar trends to HF, with a propensity for RP prolongation (251.7 vs. 206.7 ms, p=0.10), worsening with dilation severity.

Human aLMS offer a robust, patient-specific platform for exploring the biomechanical basis of atrial disease. Despite high biological variability, clear biomechanical trends point to meaningful changes that merit further study. Higher Fmax in AF slices may indicate compensatory adaptations, while reduced atrial contractility in HF suggests atrial remodelling, despite being a ventricular disorder. Future research will leverage optimized in vivo-like conditions and advanced techniques such as real-time calcium imaging, work loops, and electrical mapping to deepen understanding of atrial pathophysiology, refine disease modelling, and support targeted therapy development.