Modelling of atrial fibrillation at physiologically relevant scales enabled by massive expansion of native human atrial cardiomyocytes

Funding Acknowledgements Type of funding sources: Public hospital(s). Main funding source(s): LUMC Background Current in vitro models of atrial fibrillation have limited translational potential due to a lack of relevant human physiology or the inability to reach the high activation frequencies present in human atrial fibrillation. Absence of relevant models is the result of a general deficit of readily available and standardized sources of well-differentiated human atrial cardiomyocytes. Therefore, we aimed to immortalize native human atrial cardiomyocytes to produce natural and standardized lines of these cells. Methods Human fetal atrial cardiomyocytes were transduced with a lentiviral vector directing myocyte-specific and doxycycline-inducible expression of simian virus 40 large T antigen. Addition of doxycycline to the culture medium pushed cardiomyocytes towards a highly proliferative phenotype (proliferation up to 10^12 cells). These cells were labelled hiAMs (human immortalised Atrial Myocytes). After differentiation upon doxycycline removal, hiAM cells were characterized using various molecular, biological and electrophysiological assays. Results Following cardiomyogenic differentiation, hiAMs no longer expressed the proliferation marker Ki67, revealed striated α-actinin and troponin T staining patterns and displayed synchronous contractions. Optical voltage mapping of hiAM monolayers revealed excitable cells showing homogeneous spreading of action potentials at 22.5 ± 3.1 cm/s with a mean APD80 of 139 ± 22 ms. Addition of flecainide (10 µM) to hiAM monolayers decreased the conduction velocity by 35% and increased the APD80 by 107%. Dofetilide (10 nM) addition had no effect on the conduction velocity, but did increase the APD80 by 81%. Due to their scalability, monolayers of hiAMs as big as 10 cm2 showing homogenous action potential propagation could easily be created. Following high-frequency electrical pacing, rotors could be induced with an average activation frequency of 7.5 ± 0.9 Hz. Infusion of flecainide during arrhythmic activity resulted in termination of the rotor in 18 of 24 attempts (75%), whereas addition of 0.1% DMSO (vehicle control) did not result in termination in any of the attempts. Dofetilide infusion did not result in termination. However, it did lower the average activation frequency to 2.1 ± 0.7 Hz. Conclusion We have generated first-of-a-kind lines of human atrial cardiomyocytes, allowing massive cell expansion under proliferation conditions and robust formation of cross-striated, contractile and excitable cardiomyocytes after differentiation. These characteristics allow, for the first time, the modelling, at a large-scale, of human atrial arrhythmias with frequencies similar to human atrial fibrillation. With the generation of hiAMs, a user-friendly, clinically-relevant and much-anticipated human atrial research model has been produced. Abstract Figure. hiAM AF Model