Human atrium-scaled in vitro modelling of atrial fibrillation: size-dependent arrhythmia dynamics and pharmacological responses

Background Atrial size strongly influences atrial fibrillation (AF) development, maintenance, and therapeutic outcome, including pharmacological cardioversion. Yet, preclinical research relies heavily on models considerably smaller than the human atrium, thereby not supporting the spacious arrhythmia dynamics seen in human AF. Consequently, there is a need for human AF models that enable mechanistic and therapeutic studies at a patient-relevant scale, thereby allowing for the evaluation of size-dependent arrhythmia dynamics and pharmacological responses. Purpose To develop full human atrium-sized models of AF at a patient-relevant scale for mechanistic and translational research. Methods Layers of conditionally immortalized human atrial myocytes (hiAM)1, were created in three sizes: 9.5 cm², 55 cm², and 152 cm², corresponding, respectively, to the surface areas of a standard in vitro model, and young adult and enlarged human atria. Optical voltage mapping visualized electrical activity during electrical pacing and burst pacing-induced re-entry. Arrhythmia complexity was quantified by the number of concurring circuits, which cores were tracked to assess their trajectories. Subsequent pharmacological testing included flecainide, ibutilide, or both. Results Upon 1 Hz electrical pacing all monolayers showed homogeneous wave propagation and similar wavelength between sizes. Tachypacing-induced re-entrant circuits showed a clear size-dependent increase in arrhythmia complexity and circuit movement, while dominant frequency declined with increasing size (Figure 1).Specifically, the human atrium-sized models supported multiple interacting circuits. Subsequent therapeutic testing revealed size-dependent drug effects. Vehicle infusion (0.1% DMSO) had no effect on dominant frequency and arrhythmia complexity. Ibutilide (20 and 60 nM) decreased dominant frequency, yet failed to increase arrhythmia termination. Yet, flecainide (3 µM), besides reducing dominant frequency, also reduced arrhythmia complexity across all models. However, termination efficacy decreased size-dependently, showing arrhythmia termination in the enlarged model only at a high dose (10 µM). In contrast, combined treatment with 3 µM flecainide and 20 nM ibutilide showed synergistic effects on re-entry termination. Notably, this combination effectively terminated arrhythmias in the enlarged model, an outcome not observed after single agent administration. Mechanistically, the larger models supported dynamic wave behaviors incompatible with the small model, including prolonged path lengths, reduced wavefront curvatures and increased temporal excitable gaps. Conclusion Our human atrium-sized in vitro modelling of AF enables mechanistic and translational research at a patient-relevant scale, demonstrating how size can critically influence arrhythmia dynamics and pharmacological responses.