Integrative Computational Modeling of Cardiomyocyte Calcium Handling and Cardiac Arrhythmias: Current Status and Future Challenges

Figures

• 

  Figure 1. Alterations to calcium handling as a key determinant of cardiac arrhythmias. Subcellular calcium-handling abnormalities may initiate delayed afterdepolarizations (DADs) and triggered activity. At the cellular level, modifications of calcium-handling proteins can directly or indirectly affect cardiac ion-channel function, leading to substantial changes in AP properties. At higher spatial scales, calcium-handling abnormalities may promote structural remodeling, allowing the generation of reentrant waves. Together, those processes may create vulnerable substrates for cardiac arrhythmias. (CICR = calcium-induced calcium release; CV = conduction velocity; ERP = effective refractory period; K1 = inward-rectifier potassium channel; Kr = rapid delayed-rectifier potassium channel; Ks = slow delayed-rectifier potassium channel; LTCC = L-type calcium channel; NCX = sodium-calcium exchanger; NKA = sodium-potassium adenosine triphosphatase; PLN = phospholamban; RyR2 = ryanodine receptor type 2; SERCA2a = sarco/endoplasmic reticulum calcium adenosine triphosphatase-2a).

• 

  Figure 2. The proposed mechanisms involved in low intracellular sodium-induced calcium silencing. The reduction of LTCC function following rapid atrial pacing/persistent AF has been proposed as the cause of low intracellular sodium, which further induced the calcium extrusion via NCX and lowered the frequency of SR calcium leak and SCaEs, ultimately leading to calcium-signaling stabilization/silencing in the long term. (CICR = calcium-induced calcium release; ICa,L = L-type calcium current; LTCC = L-type calcium channel; NCX = sodium-calcium exchanger; NKA = sodium-potassium adenosine triphosphatase; PLM = phospholemman; PLN = phospholamban; RyR2 = ryanodine receptor type 2; SERCA2a = sarco/endoplasmic reticulum calcium adenosine triphosphatase-2a).

• 

  Figure 3. Schematic illustration of multiscale computational modeling. To integrate individual modeling scales into a holistic model, a subcellular model of cardiomyocyte calcium handling is coupled with other cardiac ion channels to form a cellular model. Subsequently, cell models propagate the electrical impulse through gap junctions to form a 1-dimensional strand or 2-dimensional tissue model. At the organ level, different types of tissues, representing different regions and layers of the heart (e.g., epicardium, mid-myocardium, endocardium, Purkinje fibers, etc.) are combined with fiber orientation and structural remodeling.

• 

  Figure 4. Interspecies difference of AP and calcium transient (CaT) in several computational models of ventricular cardiomyocytes. All simulations were carried out in Myokit [31]. (A,B) The AP and CaT were stimulated using the default setting of each model at a 1 Hz pacing frequency (basic cycle length [BCL] of 1000 ms) and quasi-steady-state was obtained following 1000 beats of pre-pacing. (C,D) The rate dependence of APD90 and CaT. Mouse model = Bondarenko et al. (2004; [48]), Rat model = Niederer et al. (2007; [49]), Guinea-Pig model = Noble et al. (1998; [50]), Rabbit model = Mahajan et al. (2008; [51]), Dog model = Decker et al. (2009; [21]) and Human model = O’Hara et al. (epicardial; 2011; [52]). By default, the Bondarenko and Niederer models were set at room temperature (22–25 °C), whereas the rest of the models were set at body temperature (±37 °C).

• 

  Figure 5. Comparison of APD and calcium transient (CaT) between several canine and human ventricular cardiomyocyte models. The AP and CaT were stimulated using the default setting of each model at a 1 Hz pacing frequency (basic cycle length [BCL] of 1000 ms; the left panels) and quasi-steady-state was obtained following 1000 beats of pre-pacing. (BPS model = [60]; Decker model = [21]; GPB model = [61]; Heijman model = [62]; ORd model = [52]; Passini model = [63]; ToR-ORd model = [64]; TTP model = [65]; Winslow model = [66]).

• 

  Figure 6. A schematic illustration summarizing multiscale integration of patient-specific data. The “bottom-up” approach starts from the ionic current/molecular level up to the organ level. At each scale, the acquired data can be utilized as an input for computational modeling. Ultimately, a personalized 3-dimensional computational model can be developed and used for patient-specific arrhythmia research. (MRI = Magnetic Resonance Imaging; mRNA = messenger Ribonucleic acid; STED = Stimulated Emission Depletion).