Electromechanical feedback with reduced cellular connectivity alters electrical activity in an infarct injured left ventricle: a finite element model study

Abstract

Myocardial infarction (MI) significantly alters the structure and function of the heart. As abnormal strain may drive heart failure and the generation of arrhythmias, we used computational methods to simulate a left ventricle with an MI over the course of a heartbeat to investigate strains and their potential implications to electrophysiology. We created a fully coupled finite element model of myocardial electromechanics consisting of a cellular physiological model, a bidomain electrical diffusion solver, and a nonlinear mechanics solver. A geometric mesh built from magnetic resonance imaging (MRI) measurements of an ovine left ventricle suffering from a surgically induced anteroapical infarct was used in the model, cycled through the cardiac loop of inflation, isovolumic contraction, ejection, and isovolumic relaxation. Stretch-activated currents were added as a mechanism of mechanoelectric feedback. Elevated fiber and cross fiber strains were observed in the area immediately adjacent to the aneurysm throughout the cardiac cycle, with a more dramatic increase in cross fiber strain than fiber strain. Stretch-activated channels decreased action potential (AP) dispersion in the remote myocardium while increasing it in the border zone. Decreases in electrical connectivity dramatically increased the changes in AP dispersion. The role of cross fiber strain in MI-injured hearts should be investigated more closely, since results indicate that these are more highly elevated than fiber strain in the border of the infarct. Decreases in connectivity may play an important role in the development of altered electrophysiology in the high-stretch regions of the heart.

Fig. 1.

Images obtained by magnetic resonance imaging (MRI) used to build epicardial and endocardial surfaces (A) and delineate regions of infarct, border zone, and remote tissue (B). Fiber fields (C) assigned from previously obtained magnetic resonance-diffusion tensor imaging.

Fig. 2.

Simulation results of 5 consecutive beats of the left ventricle (LV) showing representative changes to action potential (A), calcium transients (B), LV pressure (C) and pressure-volume loops (D). Steady state reached after ∼3 cycles.

Fig. 3.

Comparison of MRI and simulation. Circumferential strain through systole for the remote (A) and border (B) zone regions. Solid lines indicate averages across the midwall strain in the elements or MRI results of the region while error bars are SDs.

Fig. 4.

Depiction of calculated (A) and experimentally determined (B) end-systolic circumferential strain of the LV. C–H compare the individual experimental and computational strain components along the circumferential direction in the highlighted cross section. E_CR, shear circumferential radial strain component; E_CL, shear circumferential longitudinal strain component; E_LR, shear radial longitudinal shear component.

Fig. 5.

Fiber (A–C) and cross fiber (D–F) strains in the different material regions of the simulated LV. Solid lines show average midwall strains of the elements of the region while the gray areas depict areas within 1 SD.

Fig. 6.

Effect of strain cycling, stretch-activated currents (SAC), and cellular connectitity on action potentials of a representative remote region (A, solid line) and high-fiber stretch region (A, broken line). No difference is seen in the remote region between regular (B) and low-border-zone (C) connectivity, whereas the duration of the action potentials in the high-stretch fibers increases between the regular connectivity (D) and the low connectivity (E).

Fig. 7.

Mean 90% action potential duration (APD₉₀) in the border zone (A) and remote (B) regions as a function of SAC linear response slope and border zone electrical connectivity. Dispersion of the action potential duration as quantified by its SD for the border zone (C) and remote (D) regions. Arrows depict direction of trend with increasing SAC magnitude.