Fibrotic Remodeling during Persistent Atrial Fibrillation: In Silico Investigation of the Role of Calcium for Human Atrial Myofibroblast Electrophysiology

Figure 1. Three different cellular models to explore the effects of myofibroblast Ca2+ handling. (a) Baseline myofibroblast model without Ca2+. (b) Myofibroblast model including the L-type Ca2+ current. (c) Myofibroblast model including the L-type Ca2+ current and an intracellular Ca2+ handling system.

Figure 2. Scheme of the genetic algorithm. Available in vitro data were used as an input to fit the model parameters. The population is comprised of individuals (A1 to Am) with different parameter values representing the “genes” (P1 to Pn, n=11). After the transmembrane potential was simulated for all individuals, the fitness function (F) was evaluated. If the error was less than 0.1, the algorithm terminated. Otherwise, it continued by selecting the fittest individuals and allowing crossover and mutation of the genes to generate the next generation population. The process continued until the convergence condition was met.

Figure 3. Myofibroblast electrophysiology including the ICaL current and the intracellular Ca2+ handling system. (a) In silico patch-clamp experiment for ICaL (top) and comparison (bottom) with in vitro data (dots) from [11]. (b) Transmembrane potential of three different non-paced models: baseline model (blue), myofibroblast with ICaL (red), and myofibroblast with ICaL and intracellular Ca2+ handling (yellow).

Figure 4. Fibroblast coupling effect on the myocyte action potential. (a) Uncoupled AF-remodeled myocyte action potential and four different numbers of myofibroblasts coupled to a single myocyte. (b) Action potential restitution curves for AF-remodeled myocytes and four different numbers of myofibroblasts coupled to a single myocyte.

Figure 5. Three different fibrosis patterns with clustering of the myofibroblast infiltration in the central part of the tissue for a density of 20%. From left to right, patterns that resemble interstitial to patchy fibrosis.

Figure 6. Snapshots of the transmembrane potential at different time instances for three different fibrosis patterns with 20% myofibroblast density (black areas in the left column represent myofibroblasts). White arrows indicate the reentry trajectory, and the white star shows the exit point of the reentry. Reentrant activity was initiated by a cross-field stimulus. The top row shows the reentry around the fibrotic region of high entropy that resembles interstitial fibrosis. The middle row shows the reentrant activity with an exit point due to the separation of the clusters with a mean entropy value that resembles disperse patchy fibrosis. The bottom row shows reentrant activity around an area of low entropy that resembles a pattern of patchy fibrosis.

Figure 7. Tissue vulnerable window for all myofibroblast infiltration patterns. Vulnerability to arrhythmia increased for all patterns. Outliers are marked with a plus (+) on the boxplot.