The Role of the TBX5-driven Regulatory Network in the Manifestation and Sustainability of Atrial Fibrillation

Abstract

Atrial fibrillation (AF) is the most common cardiac arrhythmia and is associated with many well-known risk factors. Within the past decade, there has been a plethora of evidence to suggest that genetic factors also play an important role in the genesis of the arrhythmia, including the T-box transcription factor gene TBX5 and the paired-like homeodomain transcription factor 2 gene PITX2. Both transcription factors are primarily known for the roles they play in cardiac development during embryogenesis. The multi-tier transcriptional network driven by TBX5 and modulated by PITX2 links 7 AF loci. This regulates many gene expressions, which play a crucial role in AF susceptibility. The exact mechanism underlying AF due to the genetic impairment and mutations is not well understood, and an effective clinical approach needed to treat such patients remains to be established. Computational models provide a powerful framework for investigating the electrical properties of atrial myocytes. Through the utilization of human atrial myocyte models, the mechanism by which the TBX5 network modulate the ionic mechanisms in the atrial myocyte can be elucidated. Furthermore, the models can be utilized to simulate the effects of different antiarrhythmic treatment options to ascertain the conditions in which a specific treatment is effective or prone to adverse effects. Particularly, the representative common pool models which can predict the ionic remodelling accurately under the TBX5 network, i.e., the Courtemanche, Ramirez, Nattel (CRN) model, the Maleckar et al model and the Nygren et al model, and existing spatial models of the human atrial myocyte that encompasses both spatial-temporal and structural aspects of Ca2+ handling are further adapted and utilised in this thesis. In this thesis, I first developed a biophysics-based model of the human atrial myocyte that is capable of reproducing early and after/spontaneous depolarization (EADs and DADs/SDs) by updating the outdated Ca2+ handling of the CRN model with that from the more recent Grandi et al model. Under the assumption that changes in the gene expressions correlate proportionally to changes in the function of the channel, I implemented various combinations of the changes in the ionic and Ca2+ handling channels due to impaired TBX5. Simulations showed upregulated ICaL and downregulated IK1 to be the main cause of action potential duration (APD) prolongation, and impaired TBX5-induced EADs and DADs/SDs due to an elevation of intracellular Ca2+ ([Ca2+]i). However, the latter channel also caused unphysiological depolarization of the resting membrane potential (RMP), hence upregulated ICaL is the far more likely trigger of impaired TBX5 afterdepolarizations. The Maleckar et al common pool model was then used to investigate the mechanisms underlying the genesis of afterdepolarizations from an upregulation of ICaL and INaCa, and downregulation of the sarcoplasmic reticulum Ca2+ transport ATPase (SERCA) due to the impairment of TBX5. As it was the case in the Grandi-CRN hybrid common pool, upregulated ICaL was also the main trigger of DADs/SDs. Implementing the upregulation of INaCa alongside upregulated ICaL on the other hand suppressed these afterdepolarizations, as increasing INaCa caused an increase in the threshold [Ca2+]i for SDs to occur. Through bifurcation analyses of SERCA with respect to INaCa and ICaL and virtual intracellular Ca2+ injection experiments, I also showed that the suppression of afterdepolarizations was possible with the elevation of SERCA because within a certain parameter range, increasing SERCA also caused the elevation of the [Ca2+]i SD threshold. Lastly, I conducted a benchmarking study on the Koivumaki et al, Voigt et al, and Sutano et al spatial models. Through this study, I discovered that detailed descriptions of both the transverse and longitudinal aspects of Ca2+ handling, as well as accurate representations of the main internal structures of the atrial myocyte are necessary facets in generating frequent DADs and afterdepolarizations with a more physiological increase in the conductivity of the L-type Ca2+ channel ICaL (GICaL). With the implementation of tubules based on electron microscopy imaging data, the increase in GICaL needed to trigger both EADs and DADs decreases to a more physiological magnitude. Implementing a distribution of RyRs based on imaging data decreases the overall magnitude of the DADs, but increases the frequency in which they occur. However, none of the spatial models was able to capture the suppression of afterdepolarizations due to the normalization or elevation of SERCA. In summary, our modelling study suggests that either inhibiting upregulated ICaL or restoring SERCA function could suppress impaired TBX5-induced afterdepolarizations. Furthermore, it demonstrates that spatial models are necessary in accurately capturing the genesis of afterdepolarizations. Our computer models improve our knowledge beyond the known limitations of both clinical and animal model studies on the multi-tier transcriptional network and provide a powerful framework for testing alternative treatment approaches for patients with AF.