Abstract
As a potentially life threatening side effect, pharmaceutical compounds may trigger cardiac arrhythmias by impeding the heart’s electrical and mechanical function. For this reason, any new compound needs to be tested since 2005 for its proarrhythmic risk both during the preclinical and the clinical phase of the drug development process. While intensive monitoring of cardiac activity during clinical tests with human volunteers constitutes a major cost factor, preclinical in vitro tests with non cardiac cells and in vivo tests with animals are currently under serious debate because of their poor extrapolation to drug cardiotoxicity in humans. For about five years now, regulatory agencies, industry and academia are working on an overhaul of the cardiac drug safety paradigm that is built a) on human heart muscle cells, that can be abundantly bioengineered from donor stem cells without ethical concerns (human induced pluripotent stem cell derived cardiomyocytes, hiPSC-CMs), and b) on computational models of human cardiac electrophysiology both at the cellular and the organ level. The combined use of such human in vitro and human in silico models during the preclinical phase is expected to improve proarrhythmia test specificity (i.e. to lower the false-positive rate), to better inform about the need of thorough heart monitoring in the clinic, and to reduce or even replace animal experiments. This review article starts by concisely informing about the electrical activity of the human heart, about its possible impairment due to drug side effects, and about hiPSC-CM assays for cardiac drug safety testing. It then summarizes the mathematical description of human cardiac electrophysiology in terms of mechanistic ODE and PDE models, and illustrates how their numerical analysis may provide insight into the genesis of drug induced arrhythmias. Finally, this paper surveys proarrhythmic risk estimation methods, that involve the simulation of human heart muscle cells, and addresses opportunities and challenges for future interdisciplinary research.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Heart muscle cells have the ability to generate and conduct electrical signals referred to as action potentials. The concerted production and propagation of these action potentials throughout the heart forms the basis of its rhythmic contraction and relaxation. This rhythmic activity can be adversely affected by pharmaceutical compounds such that the blood pumping function of the heart is greatly diminished and a life-threatening situation occurs. In fact, drug induced cardiac arrhythmias belong to the most serious drug side effects, have led to the post-marketing withdrawal of several medicines and are a major reason for the discontinuation of candidate compounds during the drug development process. Consequently, the assessment of the proarrhythmic risk is an integral and mandatory part of both preclinical and clinical safety testing of new pharmaceutical compounds. In that regard, a major problem to the industry is that preclinical tests based on animal models show a significant false positive rate and hence may lead to the development stop of otherwise promising compounds based on rather non-specific markers of proarrhythmia.
Mathematical modelling of cardiac electrophysiology [235], [64] looks back on more than 50 years of history and has become a prime example of computational systems biology. While nonlinear ODE systems form the established tool for describing action potentials of single heart muscle cells, complex PDE systems of the reaction diffusion type model the propagation of the electrical excitation wave through interconnected heart muscle cells, cardiac tissue or the whole heart. Recently, the field gained additional momentum driven by international and multidisciplinary research initiatives [36], [98] that aim at overcoming the above mentioned non-specificity problem of preclinical drug cardiotoxicity testing. As one pillar of these efforts, in vitro data about drug effects on transmembrane ionic currents shall be integrated into a mechanistic action potential model of a native human heart muscle cell [138]. This model shall then serve as the basis for proarrhythmic risk prediction by means of machine learning techniques. Finally, these predictions shall be scrutinised by in vitro experiments with human stem cell derived heart muscle cells (hiPSC-CMs), that keep the genetic features of their human donor and provide - without ethical constraints - an abundant stock of human material for drug safety testing [34].
The overall goal is to better support the decision whether intensive (and hence expensive) electrocardiography (ECG) monitoring [262] is still necessary during late-stage clinical trials in humans. This decision might be further facilitated by means of in silico drug trials with 3D high-fidelity multiscale and multiphysics models of the human heart [234], [10], [207]. In return, the respective computational techniques also offer the opportunity to design and analyze experiments with 2D monolayers or 3D structures engineered from hiPSC-CMs as, e.g., macroscale human myocardia [240] or human minihearts in a jar [136].
Several reviews of computational methods for cardiac safety testing of drugs have been previously published [154], [270], [48], [75]. One distinctive feature of the article at hand is that it contrasts ODE and PDE modelling of drug induced cardiac arrhythmias with arrhythmic events that can also be experimentally observed in hiPSC-CMs. Furthermore, this paper covers the association of cardiac arrhythmias with bifurcations and pattern formation in dynamical systems, and surveys simulation based approaches for the classification of drugs according to their proarrhythmic liability. Section 2 gives a short introduction to cardiac electrophysiology and presents how its distortion by pharmaceutical compounds may lead to cardiac arrhythmias. Section 3 informs about the Comprehensive In Vitro Proarrhythmia Assay (CiPA) initiative, that is led by the US Food and Drug Administration (FDA) and works on defining a new safety paradigm, and illustrates how hiPSC-CMs are currently used in drug cardiotoxicity testing. Section 4 deals with the mathematical formulation of cardiac action potentials and their spatial propagation through cardiac tissue and organ. Furthermore, it addresses the rich dynamical repertoire of these models that corresponds with diverse types of arrhythmic behaviour both at the cellular and at the tissue level. Section 5 discusses methods that integrate mechanistic modelling of drug effects into statistical learning tools for the assessment of the proarrhythmic risk. Finally, Sect. 6 lists challenges related to the predictive ability of risk classifiers and to the credibility of cardiac electrophysiology models, in particular of hiPSC-CMs, and also touches on some topics that so far are hardly considered in simulation based drug cardiotoxicity testing. For the convenience of the reader, an appendix gives basic examples of cardiac AP models in full detail and features two prominent logistic regression models for proarrhythmic risk classification.
2 Drug Induced Cardiac Arrhythmias
2.1 Electrical Activity of the Heart
The human heart consists of billions of muscle cells (cardiomyocytes, CMs), that rhythmically contract in order to provide the mechanical force for pumping blood throughout the body [161]. The rhythmic contraction of the heart is regulated by a wave of electrical activation, which is based on the excitability of cardiac cells and their ability to generate and propagate action potentials. An action potential (AP) is a characteristic transient change in the transmembrane voltage that starts with a depolarization phase and ends with a repolarization phase, see Fig. 1 (left). Pacemaker cells located in the sinoatrial (SA) node regularily depolarize by themselves and generate spontaneous APs, a property that is referred to as automaticity. In contrast, atrial and ventricular CMs are quiescent, have a resting state and can generate an AP only if the transmembrane voltage is raised above a certain threshold value by an external stimulus. Still, the CMs are electrically connected via gap junctions, which are ion conducting channels formed by proteins. Hence, the currents created during an AP of a cell may flow to and stimulate an AP in the neighbouring cells. Once a cell has produced an AP, there is a refractory period during which the cell can generate no further AP. Together, excitability and refractoriness of cardiac tissue ensure that the wavefront of depolarization, initiated by the pacemaker cells, is propagated in a unidirectional manner from the atria through the atrioventricular (AV) node to the ventricles.
The propagation of APs throughout the heart causes changes in the potential on the body surface, which can be recorded in an electrocardiogram (ECG). Figure 1 (right, top) displays an ECG that represents a normal sinus rhythm of about 70 beats per minute for a resting person. The small P wave, see Fig. 1 (left), reflects the depolarization of the atria, and the larger QRS complex corresponds with the depolarization of the ventricles. The end of the T wave indicates their full repolarization, such that the QT interval is a measure of the ventricular AP duration (APD), see Fig. 2 (left).
The AP of a CM is orchestrated by voltage gated ion channels, which are membrane bound proteins that open and close in dependence on the transmembrane voltage [84]. These ion channels ensure a selective and chronological transmembrane flow of ions such as calcium, potassium and sodium into and out of the cell, see Fig. 2 (right). Of particular importance is the voltage gated \(\text{K}_{\text{V11.1}}\) potassium channel, that carries the rapidly activating delayed rectifier potassium current \(\text{I}_{\text{Kr}}\) and is encoded by the human ether a go-go related gene (hERG) [221]. Mutations of hERG may impair ventricular repolarisation and prolong the QT interval, a condition referred to as congenital long QT syndrome (LQTS) [44].
When the calcium channels are opened during an AP, calcium flows into the cell and triggers the quick release of calcium from the sarcoplasmic reticulum (SR) into the cytosol [15]. The rapid increase of cytosolic calcium leads to the contraction of the myofilaments, and the subsequent relaxation of the cell is enabled by a pump that drives calcium back into the SR. Oscillatory calcium dynamics not only are involved in the excitation-contraction coupling, but also are implicated in the automaticity of the pacemaker cells [277]. Accordingly, the pacemaking function is maintained by two coupled oscillatory mechanisms termed membrane clock, formed by the ionic currents across the membrane, and calcium clock. As part of the latter, the SR spontaneously releases calcium at local sites within the cell, which is in contrast to the global calcium release triggered by the AP.
2.2 Cardiac Arrhythmias
A cardiac arrhythmia is a pathophysiological deviation from the normal sinus rhythm and can be caused by perturbations in the excitation of the tissue or in the conduction of the excitation wave. Figure 1 (right, middle) displays an ECG that corresponds with a Torsade de Pointes (TdP) arrhythmia, which is a particular form of polymorphic ventricular tachycardia (VT) with repeated twisting of the QRS complexes. TdP may degenerate into ventricular fibrillation (VF), see Fig. 1 (right, bottom), which is a life threatening arrhythmia as it can lead to sudden cardiac death within minutes after onset. The basic electrophysiological mechanisms of cardiac arrhythmias [7], [248] can be roughly devided into a) focal activity, that arises in a localized tissue region due to abnormal automaticity or triggered activity of CMs, and b) reentry activity, where an action potential wave fails to terminate itself and reexcites a region of nonrefractory tissue.
An example of a premature trigger is given by early afterdepolarizations (EADs), which are abnormal depolarizations during the repolarization phase of the AP, see Fig. 3 (left). EADs can lead to ectopic beats outside the region of the sinoatrial node and may initiate polymorphic VT and TdP [265], [267]. EADs emerge in a condition referred to as reduced repolarization reserve [215], which is an ionic current imbalance due to reduction of outward and/or an increase of inward currents. Another example for abnormal activity is given by repolarization alternans, which are beat-to-beat oscillations in APD and/or calcium transients. Cellular alternans may manifest as T-wave alternans in the ECG and are involved in arrhythmogenesis including TdP formation [205].
Examples of reentry activity are formed by two-dimensional spiral and three-dimensional scroll waves, where action potentials travel in circular motion around an anatomical obstacle and reactivate their site of origin. Such spiral movement is also possible in absence of a structural obstacle, in which case the wave rotates around a functional obstacle formed by its own refractory tail, see Fig. 3 (right). Spiral waves may become self-sustained sources of high-frequency excitation and can drive atrial and ventricular tachycardia and fibrillation. In particular, spiral waves that rotate along a circular trajectory are linked to monomorphic VT, while spiral waves that meander through space are related with polymorphic VT and TdP, see Fig. 3 (right). As a result of conduction block, spiral waves may also break into several spiral waves that continuously terminate and redevelop in a chaotic manner, and such spiral wave breakups are considered as a possible basis for the transition from VT to VF [202]. Recently, it was demonstrated that also a single spiral wave can display chaotic tip trajectories that are consistent with fibrillation [143].
2.3 Drug Cardiotoxicity
Pharmaceutical compounds - not only cardiovascular drugs but also chemotherapeutics, antipsychotics, antibiotics and others - may interfere with the function of the heart and trigger life-threatening cardiac arrhythmias such as TdP [278], see Fig. 1 (right, middle). In the 1990s, a sequence of reports about the occurence of QT interval prolongation and TdP - presumably or demonstrably induced by drugs and in some cases with fatal outcome - led to the withdrawal of 14 drugs from the market [232]. By that time, it was already known that patients with congenital LQTS are predisposed to TdP [263]. As a number of drugs were found to block the hERG channel, see Fig. 4 (left), a mechanistic link [221] between the congenital (or inherited) LQTS and the acquired (or drug induced) LQTS was finally established. This further supported the association between drug induced LQTS and TdP [278].
According to the trigger-substrate model [105], APD and QT interval prolongation not only facilitate the occurence of EADs but also promote heterogeneity of repolarization across the tissue. A disrupted spatial dispersion of repolarization may form a substrate which maintains reentrant arrhythmias once initiated by EADs or other triggers. Based on this paradigm, illustrated in Fig. 4 (right), the International Conference on Harmonization (ICH) in 2005 released two mandatory guidelines for preclinical (ICH S7B) and clinical (ICH E14) cardiac safety testing of new pharmaceutical compounds, see [2] and [1]. ICH S7B first prescribes an in vitro assay in order to assess the potency of the test compound to block the hERG channel heterologously expressed in host cells. The compound is considered as a torsadogenic hERG blocker if its hERG \(\text{IC50}\) value is below the 30-fold of its effective free therapeutic plasma concentration (EFTPC) [211], see also Fig. 25 (right) in the appendix. \(\text{IC50}\) denotes the drug concentration at which the current through the channel is reduced by 50\(\%\). The second component of ICH S7B is an in vivo QT assay in animal models in order to assess whether the test compound induces QT prolongation. With respect to human trials, ICH E14 specifies a thorough QT study in order to assess a test compound’s impact on the heart rate corrected QT (QTc) interval of healthy volunteers during clinical phase I. This test rates the compound as possibly torsadogenic if the mean prolongation of QTc is above 10 ms. The ICH S7B and ICH E14 guidelines are still in use today and have proven to be successful - no drug that accordingly was approven to market has ever since been withdrawn for cardiac safety concerns.
However, since 2005 more than a billion dollar was spent on hundreds of intensive ECG monitorings in clinical phases IIb and III due to safety concerns raised during ICH S7B and ICH E14 testing [190]. Furthermore, it is meanwhile well established that, on the one hand, QT prolongation not necessarily leads to TdP [125], and on the other hand, hERG blockers not necessarily prolong QT as they may have compensating effects on other ion channels. Hence, hERG blockade and QT interval prolongation are no longer considered to be specific markers to assess a drug’s proarrhythmic liability. As a negative consequence, potentially benefitial compounds may be discontinued at an early stage, mostly based on their impact on hERG alone rather than due to actual proarrhythmic effects.
More complex preclinical assays include ventricular wedge preparations and isolated Langendorff-perfused hearts model from canine or rabbit [125], [105]. While they may grant more informative data about a drug’s proarrhythmic liability, they only offer low throughput recordings and do not bypass the error-prone animal-to-human extrapolation. Although also adult human primary cardiomyocytes, isolated from non-transplantable human donor hearts, form a preclinical model for proarrhythmia risk testing [167], [5], ethical concerns and limited availability do currently not render them suitable for high-throughput studies.
3 Cardiac Safety Testing with Human IPSC-Derived Cardiomyocytes
In contrast, human iPSC-CMs provide easily accessible human test material without ethical concerns and hence have become a promising tool for cardiac arrhythmia research [86]. In the context of preclinical cardiac safety testing of drugs, their potential is currently explored in multidisciplinary research projects run by, e.g., the Japanese IPS Cardiac Safety Assessment (JiCSA) consortium [98] and the Comprehensive in vitro Proarrhythmia Assay (CiPA) initiative [218], [36], [61], [70], see Fig. 5 (left). Within CiPA, hiPSC-CM technology shall be combined with computational modelling of human ventricular CMs [138], [139] see Sects. 4 and 5 for details, in order to better assess a drug’s torsadogenic liability before going into clinical trials. For that purpose, 28 reference compounds were classified based on expert opinion into low, medium and high proarrhythmic risk groups and divided into 12 training and 16 validation compounds, see Table 1. These CiPA compounds include the non-cardiovascular drugs terfenadine and cisapride, which are hERG blockers withdrawn from the market due to cardiac toxicity, but also the non-torsadogenic drug verapamil, whose hERG blocking effects are compensated by a simultaneous block of the calcium channel.
Human iPSC-CMs possess a heterogeneous AP phenotype characterized by spontaneous activity [99], see Fig. 5 (right) for an example of a ventricular-like AP from a single hiPSC-CM. For studying drug cardiotoxicity, multicellular hiPSC-CM settings are more suitable and range from 2D monolayers over 3D organoids to engineered tissue [135] as well as miniature heart chambers [136]. The associated assays [34] comprise the recording of electrical intra- and extracellular signals, calcium transients, beating movements and contractile forces. For instance, Fig. 6 (left) displays local extracellular APs (LEAPs) with drug induced EADs, while Fig. 6 (right) shows extracellular field potentials (FPs) and cell motion images from 3D human iPS cell-engineered cardiac tissue during drug induced reentrant acticity [102], [227], compare also with Fig. 3. Two recent studies [16] and [17] confirmed the utility of hiPSC-CMs for proarrhythmic risk prediction by showing that high and intermediate CiPA risk compounds consistently evoke EADs-like arrhythmic events across different electrophysiological readout platforms and different commercially available hiPSC-CM lines, see Fig. 7. Note that according to a scoring system suggested in [108], the occurence of EADs in hiPSC-CMs immediately attributes very high proarrhythmic risk to a compound, regardless of all other endpoints measured.
4 Modelling and Simulation of Cardiac Electrophysiology and Arrhythmias
4.1 Modelling of Cardiac APs
Computational models of single cell APs typically take the form of nonlinear systems of ordinary differential equations (ODEs). Based on viewing the cell membrane as a capacitor with capacitance \(C_{m}\), these equations describe the time course of the transmembrane voltage \(V\) according to
along with the dynamics of the voltage dependent ionic transmembrane currents \(I_{\text{ion}}\) into and out of the cell, compare with Fig. 2 (right). Furthermore, the equations may include a description of intracellular processes such as calcium handling by the SR. Over the past 50 years, more than 200 cardiac AP models have been developed [141] for animal and human species, including human SA node [55], human Purkinje fibre [231], [247], human atrial [42], [76], [229], [38], human ventricular [118], [251], [77], [177], [12], [243], [78], [244] CMs and human iPSC-CMs [185], [107], [104], [182]. The level of complexity ranges from phenomenological descriptions with a handful of state variables and model parameters to detailed mechanistic models based on experiments with dozens of state variables and hundreds of parameters, see the appendix for simple examples. Typically, the resulting ODE systems are stiff such that explicit methods with small step sizes or implicit methods are needed for their numerical integration.
Usually, the transmembrane currents flowing through the voltage gated ion channels are modelled [103] as
with the current-voltage relationship \(i_{\text{ion}}(V)\) of a single open channel and the total number \(N_{\text{ion}}\) of channels of the specific type under consideration. In most cases, the probability \(O_{\text{ion}}\) that the channel is open, is described by a monomial of voltage dependent gating variables, which are part of the state vector of the underlying cardiac AP model and take values between \([0,1]\), see the appendix for an example. An alternative is to derive \(O_{\text{ion}}\) from the Kolmogorov differential equation associated with a Markov process that describes voltage dependent random fluctuations between different open and closed configurations of the channel [62].
Viewing the channel as an ohmic resistor, the function \(i_{\text{ion}}(V)\) is commonly described by the linear relationship
Here, \(g_{\text{ion}}\) denotes the single channel conductance, such that the maximum conductance of the macroscopic transmembrane current \(I_{\text{ion}}\) is given by
At the Nernst potential
with universal gas constant \(R\), absolute temperature \(T\), valence \(z\) of the ion and Faraday constant \(F\), the transmembrane voltage is balanced by the concentration difference of the ion across the membrane, resulting in a zero net flow of the ion through the channel. An alternative model of the current-voltage relationship is the Goldman-Hodgkin-Katz (GHK) equation
where \(p_{\text{ion}}\) is the permeability of the membrane to the ion, normalized by the number of open channels. It can be derived by integrating the Nernst-Planck equation, which describes the ion flow in dependence of the concentration gradient and of the electric field through the membrane, under the assumption of a constant field. As in case of the ohmic model (3), the ionic current (5) is zero at the Nernst potential \(V = V_{\text{ion}}\).
Typically, APs of spontaneously active cells are modelled by autonomous ODE systems that feature corresponding stable limit cycles. Models of quiescent CMs rather possess a stable steady state in accordance with the resting membrane potential (RMP) of the cell. In this case, the simulation of an AP requires to provide a suprathreshold current stimulation \(I_{\text{stim}}\), see (1), that drives the system into an area of phase space from where it returns to the steady state only after a large detour. Figure 8 (left) shows APs computed with the Paci model [185] for a spontaneously beating hiPSC-CM, and Fig. 8 (right) illustrates the AP response of the O’Hara Rudy (ORd) model [177] of a human ventricular CM to a periodic stimulation with a pacing cycle length (CL) of 1000 ms. This pattern is referred to as a 1:1-rhythm as, after the vanishing of possible transients, each stimulus evokes one and the same AP.
Common readout parameters of cardiac AP simulations include the maximum upstroke velocity \(\text{(dV/dt)}_{\text{max}}\), the duration of the AP at \(x \%\) of repolarization (APDx), or the AP amplitude (APA), see Fig. 9 (left). Likewise, characteristics of the calcium dynamics may be of interest. For models of quiescent CMs, another important feature is the APD restitution APD = r(DI), which describes the nonlinear dependency of APD (typically APD90) on the previous resting interval referred to as the diastolic interval (DI), see Fig. 9 (right). As discussed in the following, steepness of the restitution curve r is implicated in arrhythmogenesis both at the cellular and the tissue level.
4.2 Nonlinear Dynamics of AP Models and Cellular Arrhythmias
Cardiac AP models may possess a rich repertoire of nonlinear dynamics [110] that includes proarrhythmic behaviour such as APD alternans and early afterdepolarizations (EADs), see Fig. 10. These abnormal activities are associated with bifurcations [100] in dynamical systems, which are qualitative changes in the dynamics upon variation of model parameters [121]. In case of APD alternans, their onset is linked with a period doubling bifurcation of the iterated map
with \(\phi (\text{APD}_{n}) = r(\text{CL}-\text{APD}_{n})\) and \(\text{CL} = \text{DI}_{n}+\text{APD}_{n}\). For large values of the parameter CL, the fixed point \(\text{ADP}_{*}\) of the map \(\phi \) lies on the flat part of the restitution curve r and is stable. This situation corresponds with the 1:1 rhythm shown in Fig. 8 (right). A reduction of CL moves the fixed point \(\text{APD}_{*}\) towards the steeper part of r until \(r'(\text{DI}_{*}) > 1\) (which corresponds with \(\phi '(\text{APD}_{*}) < -1\)) and the stability of \(\text{APD}_{*}\) is lost. Consequently, the period of the map \(\phi \) is doubled from 1 to 2 and one obtains \(\text{APD}_{n+2} = \text{APD}_{n}\) in correspondence with the 2:2 rhythm shown in Fig. 10 (left). As CL is further reduced, more complex m:n (m stimuli evoke n APs) or irregular patterns of APD alterations may occur [110].
Cardiac APs with EADs as illustrated in Fig. 10 (right) are an example of mixed mode oscillations (MMOs) [49], which are characterized by an alternation between oscillations of large and small amplitudes. MMOs may result from the multiple time scales at which the ODE state variables operate, and a separation of the variables into slow and fast ones can be used for their analysis. Using a parsimonious AP model with one slow channel gating variable, the occurence of EADs has been attributed to a supercritical Hopf bifurcation [245], [204] in the fast subsystem, where the slow variable then serves as the bifurcation parameter, see Fig. 11 (left). As the slow variable increases and passes through the bifurcation point, the trajectory of the full system may momentary coil around the emerging stable limit cycles, which consequently causes the appearance of EADs. In [114], a delayed subcritical Hopf bifurcation was identified as an alternative EADs generating mechanism, where the twisting of the trajectory only sets in after the slow variable has passed the region of the limit cycles that now are unstable and oriented towards the opposite direction. Using the same AP model but rather working with two slow gating variables than just one, a folded node (FN) singularity was detected in the slow subsystem [116]. Then, MMO theory implies that, in vicinity of the folded node, the trajectories of the full system are organized by twisted slow manifolds. This defines a further dynamical mechanism of EADs [116] and also explains their dependency on the pacing frequency in case of external stimulation [264]. In addition, FNs may jointly occur with saddle-focus (SF) equilibria of the full AP system, see Fig. 11 (right). In the SF mechanisms behind EADs [114], [116], the trajectory first approaches the equilibrium along its stable manifold and then spirals away on the unstable manifold associated with a pair of complex conjugate eigenvectors with positive real part. Viewing the transmembrane voltage as the only fast variable, sufficient parametric conditions for the genesis of EADs have recently been derived in [31], while isolas of periodic orbits have been held responsible for the different EADs patterns in [11].
In an alternative approach to EADs analysis that does not exploit multiple time scales, the dynamical mechanisms of EADs generation and termination in periodically paced models of the human ventricular AP were shown to depend on the bifurcation structure of the non paced model counterparts [120], [119]. Of note, AP models may also feature EADs alternans with an alternating number of small scale oscillations [223], [264] or alternating EADs amplitudes [183], as also observable in experiments, see Fig. 6 (left). Furthermore, AP models may possess multiple stable EADs dynamics [275], [249], [111] that differ in the number of small scale oscillations, where only the initial conditions used for the simulation decide which type of dynamics manifests itself. Likewise, multistable AP dynamics with different m:n patterns are possible [237], [236]. In particular, multistability serves as an explanation for transient arrhythmogenesis [274] and intermittent occurence of cardiac arrhythmias [275], [111].
While APD alternans and EADs are considered to be proarrhythmic and sometimes are referred to as cellular arrhythmias, the simulations shown in Fig. 10 correspond with periodic, hence rhythmic AP model behaviour. However, deterministic AP models may also feature chaotic dynamics, i.e., aperiodic long-term behaviour that is highly sensitive to the initial conditions. For instance, Fig. 12 (left) displays an AP trace with an irregular sequence of APDs, obtained with the Beeler Reuter AP model under periodic pacing at high rates [201]. Chaotic AP dynamics is also possible in the presence of EADs, as illustrated in Fig. 12 (right) for the periodically paced Luo-Rudy AP model [245]. In [245], [223] the appearance of chaotic EADs is linked to the homoclinic bifurcation in the fast subsystem, at which the stable limit cycles of the fast variables terminate, see Fig. 11 (left). In [115], it was demonstrated that this homoclinic bifurcation is not a necessary condition for chaotic EADs dynamics and that the latter may also result from a cascade of period doubling bifurcations of limit cycles in the full system. This period-doubling route to chaos was detected both in periodically paced models and in an unpaced model [115] of a spontaneously active cell. Furthermore, bifurcation analysis showed that chaotic EADs dynamics may coexist [115] in a stable manner with periodic AP behaviour without EADs, where only the initial conditions decide which type of dynamics occurs. Finally, irregular EADs dynamics are also associated with random ion channel fluctuations [223], that are modelled by Markov processes or stochastic ODEs [57]. Random dynamical systems [8] may show behaviour that is not present in corresponding deterministic systems, see [230] for a recent analysis of noise-induced EADs.
4.3 Modelling of AP Wave Propagation
The propagation of a wave of electrical excitation through some cardiac tissue preparation or the heart organ \(\Omega \) (or some part of it such as the ventricular chambers) is commonly described by the bidomain model [250], [235]
with the transmembrane potential \(V(x,t)\), the extracellular potential \(u_{e}(x,t)\) and a vector \(s(x,t)\) of cellular state variables. The nonlinear ODE system in (6) describes gating processes and electrochemical reactions in the cardiac cells, while the two partial differential equations (PDEs) describe the diffusion of the electrical signal. \(\sigma _{i}\) and \(\sigma _{e}\) denote the intra- and extracellular conductivities that are orthotropic due to the organization of the tissue into fibres and laminar sheets. Furthermore, cardiac tissue is heterogeneous such that model properties and components such as excitability, conductivity, ionic currents, channel conductances may be nonuniform and varying across \(\Omega \). Different action potential models (1) may be incorporated into (6) in order to model different cell types that occupy spatially distinct but coupled regions as, e.g., in case of the transmural heterogeneity of the ventricular wall. The bidomain model yields a macroscopic description of cardiac electrical activity, in which both the intracellular and the extracellular domain are continuous and occupy all of \(\Omega \), so both \(V\) and \(u_{e}\) are defined on all of \(\Omega \).
The numerical simulation of the bidomain model [235], [64], in particular on realistic three-dimensional heart geometries with detailed cellular models, requires advanced discretization and solution techniques in order to properly capture the spatial and temporal variations of the wavefront. Under the assumption that the intra- and extracellular conductivity tensors are aligned, i.e., \(\sigma _{e} = \lambda \sigma _{i}\), the bidomain model reduces to the monodomain model
which can be solved at less costs but still requires high performance computing in realistic applications, see for instance [25] for a coupled monodomain solver with optimal memory usage. Figure 13 shows the simulation of a normal AP wave propagation across the left and right ventricular chambers of a human heart model, using the monodomain equations with anisotropic conductivity and different AP models for human ventricular endocardial, mid-myocardial and epicardial cells as well as Purkinje fibre cells in order to account for regional specificity [40]. Numerical comparisons [65], [198], [19] between the mono- and the bidomain model indicate differences in wave propagation, while a theoretical result of [41] states that the error between the mono- and the bidomain solutions is bounded by the error between the respective conductivity operators.
A common postprocessing step both for the monodomain and the bidomain model is to calculate pseudo-ECG signals [69], [9], [91], [152] by a projection of the transmembrane voltage gradient according to
where \(x_{e}\) is the location of an electrode placed in some distance from \(\Omega \). More realistic simulations of ECG signals recorded at the body surface can be obtained by coupling the bidomain equation with a model of the human torso [235], [152] that views the latter as a passive volume conductor. Alternatively, ECG signals can be derived via the boundary element method [199], that involves integrals of \(u_{e}\) over the torso surface, or via the lead field method [197], that exploits the Green’s function of the bidomain diffusion operator.
4.4 Modelling of Cardiac Arrhythmias at the PDE Level
The mono- and bidomain equations are reaction-diffusion models of excitable cardiac tissue and can show a variety of spatiotemporal patterns, ranging from normally propagating waves over spiral waves to spatiotemporal chaos. Patterns associated with cardiac arrhythmias result from a collision of travelling waves with mutual annihilation due to refractoriness, and they even may arise in isotropic and homogeneous tissue, where all model properties are uniform and do not change across \(\Omega \).
APD Alternans and EADs in Cardiac Cables
A cardiac cable occupying \(\Omega = [0,L]\) is the simplest example of a spatially extended system, in which APD oscillations may vary throughout the domain. APD alternans are called concordant if the whole domain shows the same long-short pattern in APD. In contrast, discordant APD alternans are characterized by a spatial gradient of APD from short to long that, on the next beat, is inverted and gets long to short. Depending on the length \(L\) of the cable and the pacing period CL [58], discordant APD alternans may lead to a conduction block, where only every second beat propagates along the entire cable, see Fig. 14 (left).
If EADs are large enough, they may form premature triggers that propagate as premature ventricular complexes (PVC) into neighbouring areas. Figure 14 (right) illustrates PVC propagation through a heterogeneous cardiac cable, in which cells with repolarization abnormalities on the left are coupled to spontaneously active cells with normal repolarization on the right. For an example with a homogeneous cable, where multiple PVCs form due to partial regional synchronization of chaotic EADs and maintain uncoherent electrical activity even after termination of external stimulation, see [224].
Spiral Wave Break up in Cardiac Tissue
Figure 15 displays the break up of a spiral wave and the onset of chaotic spatiotemporal reentry activity in a 2D simulation of homogeneous cardiac tissue with \(\Omega = [0,L] \times [0,L]\). In this example [228], the spiral is first produced by partially blocking a plane wave that is traveling from left to right. The broken wave gains velocities in both longitudinal and transverse directions such that it curls and builds a spiral with multiple turns. As the spiral core meanders through space, a part of it dissociates and collides with the remaining part of the spiral. This leads to spontaneous break up and initiation of multiple smaller spirals. These spirals drift through space and interact in an irregular and self-sustained manner, that causes chaotic high frequency excitation of the tissue, until they finally disappear and the tissue comes to a rest.
Different mechanisms of spiral wave break up have been identified in simulation experiments [58], [228] with the monodomain model, that may be further promoted by tissue heterogeneities, anisotropy or three spatial dimensions. They include steepness of APD restitution, and dynamically caused dispersions in APD, refractoriness and conduction velocities, but also localized heterogeneities with impaired conduction or spatial gradients in material properties. The liability of APD alternans in 2D for spiral wave break up in both homogeneous and heterogeneous tissue has been addressed in several computational studies [203], see [251], [266] for examples with human anisotropic but homogeneous tissue and a focus on sodium channel function. The role of the latter in APD alternans-mediated chaotic reentry activities is also highlighted in [79] using an isotropic and homogeneous 2D tissue with a parsimonious rabbit AP model.
The particular role of EADs in the generation of cardiac arrhythmias has been investigated in a series of computational studies using the monodomain model in a variety of settings that include animal or human cell models, 2D or 3D geometries, isotropic or anisotropic conductivities \(\sigma _{i}\), and homogeneous or heterogeneous tissue, see [260] for a recent review. Typically, a part or all of \(\Omega \) is made prone to EADs by reducing the repolarization reserve of the underlying cell model, before the response of the affected tissue to an excitation wave is studied. The observed arrhythmic behaviour includes meandering of reentrant waves due to EADs formation at the spiral core [88], focal activity due to local synchronization of chaotic EADs [224], and spiral fibrillation with multiple small rotations due to wave breaks caused by EADs [259], [287], [260]. The occurence of EADs induced PVCs depends on a variety of factors that include the size of the EADs cell cluster, the coupling strength between the EADs cells but also the pacing cycle length CL [286]. In heterogeneous and anisotropic models of the human ventricular chambers, EADs may promote reentrant activity [52] and induce TdP arrhythmias via focal and/or reentrant activity [258] in dependence on the difference in repolarisation reserve between the heterogeneity and the surrounding tissue.
The majority of the computational electrophysiology studies of cardiac arrhythmias focuses on ventricular cells and chambers, as TdP and VF can lead to sudden cardiac death, see Fig. 1 (right). However, given the health burden of atrial fibrillation (AF) as the most common sustained type of arrhythmia [32], also the interest in modelling of atrial cells [83] and atrial chambers [37], [256] is continuously growing.
5 Simulation Based Cardiac Drug Safety Testing
Computational models of cardiac electrophysiology find applications [172] in cardiovascular disease research or medical device and treatment design, but they are also extensively used in the context of cardiac drug safety testing. The most common way of describing the impact of a drug on cardiac ion channels [21], [158] is the pore-block model. There, the maximum ion conductance (4) (or the maximum ion permeability) in the channel current formulation (2) is simply reduced by a factor
in dependence on the applied drug and its concentration \([D]\). Here, \(h_{\text{ion}}\) denotes the Hill coefficient, where \(h_{\text{ion}}=1\) assumes that one drug molecule is sufficient to block one ion channel. Typical non-integer values of \(h_{\text{ion}}\) lie between 0.4 and 1.6. For instance, the simulation of a compound that blocks both the hERG and the calcium channel requires knowledge of values of \(\text{IC50}_{\text{Kr}}\), \(h_{\text{Kr}}\) and \(\text{IC50}_{\text{CaL}}\) \(h_{\text{CaL}}\), to be derived from experiments or read out from data bases, in order to calculate the corresponding scaling factors for \(G_{\text{Kr}}\) and \(G_{\text{CaL}}\) of that drug. More elaborate descriptions of drug-channel interaction address the kinetics of drug binding to the channel [50], [158], [129], [137], [238] and are based on Markov models of the channel that involve drug bound states.
At the single cell level, one usually studies changes in AP morphology, such as APD90 prolongation or the occurence of EADs and repolarization failure, and in calcium transients, that arise in response to parameter modifications in the AP model (1) motivated by the presence of a drug, see [29], [175], [181] for examples with human ventricular and hiPSC cardiomyocyte models. In silico drug trials are also run on cardiac cell models adapted to mimic disease states such as LQTS and ischemia [149], [30] or on populations of cell models [22], [183], [160], [182] see Fig. 16 (left), that are obtained by random sampling of the parameter space, in order to take intra- and intersubject variability into account. Based on the fraction of models for which abnormalities occur under the influence of reference compounds with known TdP risk, risk scores can be derived from multiple simulations that possess higher accuracy than animal models in predicting proarrhythmicity [191]. Recently, blinded studies on AP model populations also suggested that the minimum set of ion channels necessary for reliable TdP risk predictions is given by Nav1.5(peak), Cav1.2 and hERG [283].
Single cell AP models (1) are also integrated into statistical learning algorithms [82] such as linear determinant analysis (LDA), logistic regression, support vector machine (SVM) or principal component analysis (PCA) in order to perform risk classification of drugs. One observation made is that risk classifiers based on indirect drug features derived from model simulations such as APD90 [153], APD50 and diastolic \(\text{Ca}^{2+}\) [122], [112] or EADs tendency [149], [189], [160], [94] may outperform those solely based on direct features [109], [200], [127], [16] obtained from experiments, see also the appendix. Within the CiPA initiative, several derived features have been tested by means of an optimized version [51] of the ORd human ventricular cell model, now incorporating \(I_{Kr}\)-drug binding dynamics [137], and the 12 training compounds listed in Table 1. The risk metric found to best stratisfy the drugs into the three TdP risk levels across various conditions is qNet, see Fig. 16 (right), which is defined as the charge passed by the net ionic current from the beginning to the end of the AP beat. Furthermore, qNet was shown to correlate with the system’s robustness against EADs, in the sense that low values of qNet make the cell more prone to EADs induction by an \(I_{\text{Kr}}\) reduction in addition to the drug block effect.
Likewise, PDE models of human cardiac electrophysiology find application in in-silico drug trials on 1D cardiac cables [246], 2D ventricular wedge preparations [113] and 3D models of the ventricular chambers [159], [280], [40], often also embedded into torso models [269], [281], [282], [178], [92] for computing impacts on ECG signals. Both the mono- [159], [269], [280], [40] and the bidomain [178], [282], [113] model are in use and may quickly pose high computational demands if anatomical geometries, realistic conductivity tensors, and detailed models of drug action and of heterogeneous cellular electrophysiology are applied. Typically, for known safe (e.g. verapamil) and known torsadogenic (e.g. dofetilide) drugs the underlying ion channel models first are manipulated based on corresponding pharmacological profiles. Then, excitation wave propagations and ECG signals are simulated under various drug doses, before concentration thresholds, at which arrhythmic behaviour possibly occurs, are reported, see Fig. 17 for an example. Furthermore, these simulation studies point to mechanisms that possibly are responsible for the different proarrhythmicity of the drugs, despite them having similar channel blocking potencies or impact on QT interval prolongation. They include accumulation and recovery times during frequency dependent sodium channel block [159], flattening of APD restitution [269], [280] or EADs translating into pronounced U waves in the ECG [40]. Given the computational burden of drug parameter space screening with 3D heart models, Gaussian process machine learning is suggested in [39] for building a surrogate model of the QT interval and in [219] for more efficiently constructing the classification boundary between safe and proarrhythmic domains.
Of particular interest for preclinical cardiac drug safety testing is modelling and simulation of monolayers of hiPSC-CMs, as used in high throughput in vitro recording tools such as MEA or VSD, compare with Fig. 7. These monolayers form a mixture of hiPSC-CMs with atrial-like, ventricular-like and nodal like phenotypes that may not be partitioned into clearly distinct regions and rather are spatially disorganized. In order to take a spectrum of AP morphologies throughout a well-mixed tissue into account, homogenisation is proposed in [20] to correspondingly extend the mono- and bidomain equations, while a probabilistic description of cell heterogeneity using correlation matrices is proposed in [241], [210]. There, the bidomain model furthermore is coupled to an electrode model in order to simulate field potential recordings from MEAs. These FP simulations are used to enrich experimental MEA measurements in order to improve the performance of SVM and LDA risk classifiers of drugs based on a composition of FP features such as FP duration or repolarization amplitude. An alternative and computationally much cheaper tool for the simulation of FPs from MEAs is discussed in [239], where the transmembrane voltage of the single cell ODE model (1) is postprocessed by means of SPICE based software through an electronic circuit that resembles the junction between the cell membrane and the MEA electrode, and that acts as a high pass filter with the FP signal as the output.
6 Challenges and Outlook
Computational cardiology has continuously grown over the past 50 years and has turned into a prominent showcase of systems biology [43]. One major area of application is cardiac drug safety testing [217], [68], [253], which recently gained novel momentum in course of the emergence of hiPSC-CM technology and the initiatives built upon it for an overhaul of the current safety paradigm [226], [61], [98]. While proarrhythmic risk classification of drugs based on mechanistic models of human cardiac electrophysiology [191], [262], [242], [219] gains more and more recognition also within pharmaceutical industries and regulatory agencies, still there are a couple of hurdles on the way to its full integration into critical decision making. Some of these challenges are listed in the following.
Validation, Verification and Uncertainty Quantification
One important aspect in the use of computational models for safety-critical applications is the establishment of their trustworthiness through validation, verification and uncertainty quantification [3]. Here, validation refers to the assessment of the degree to which the mathematical model properly reflects the reality of interest. For instance, a comparison of EADs [87] occuring in AP models with those observed in experiments revealed rather large discrepancies in inter-EAD intervals and attributed them to shortcomings of calcium current formulations. As another example, inconsistencies between AP models and experiments were found in [255] with respect to discordant alternans and linked with limitations in the description of voltage-calcium coupling. Often, cardiac AP models (1) are overparametrized and degenerate [53] such that completely different combinations of model parameters may lead to the same or to a nearly identical AP morphology [222], [96]. Furthermore, most AP models have been built by manually combining submodels based on experimental data, that were obtained from a range of cells, often even across species, and under a variety of environmental conditions. These problems can be addressed by collecting more informative and specific data sets, e.g., from single cardiomyocytes in a series of designed experiments [80] or from clinical trials [146], [112] in order to better constrain the model. This requires to differ between structurally and practically identifiable model parameters [46], [268] as well as to quantify their uncertainties [194], [132]. One approach is to use Bayesian methods in order to infer probability distributions of the AP model parameters instead of individual numbers [47], [95]. Motivated by biological variability [272], [169], an alternative is to build populations of models by statistical sampling of model parameters, where the population is either calibrated to ranges of data values [22] or to data distributions [126]. Uncertainties in both model inputs and outputs affect cardiac electrophysiology modelling across all scales [195], [155]. In particular at the ion channel level [35], computational strategies are under development that define experimental protocols for novel high-throughput recording systems [13], [130], in order to improve model credibility. For a mathematical model, that not only accounts for variability in ion current data but also in the underlying voltage-clamp experiment, see [131].
Verification deals with the correct implementation of the mathematical model and its accurate numerical solution [3]. In the context of computational cardiology, this in particular affects reaction diffusion models based on the mono- and the bidomain equations, see (8) and (6). Their spatial discretization by means of finite elements leads to a nonlinear system of algebraic-differential equations that needs to be integrated with respect to time either in a coupled or a decoupled manner [235], [64]. Decoupled methods, such as operator splitting [235] or semi-implicit time discretization [54], separate the diffusion term from the reaction term or the PDEs from the ODEs, treat them by different numerical schemes in order to increase numerical efficiency, and only have to deal with linear systems at each time step. Coupled methods yield more accurate results than decoupled ones and their time stepping is not subject to stability constraints [64]. However, coupled methods are much more expensive as they require to solve a nonlinear system at each time step. Consequently, they have been applied to PDE models only in 1D [279] or only in combination with phenomenological AP models [163], but so far are prohibitive in higher spatial dimension for physiological AP models with dozens of state variables. Recently, a coupled monodomain solver with optimal memory usage for 2D and 3D applications with mechanistic AP models has been introduced in [25] that exploits the sparsity of local matrices in the assembly of global FEM matrices. Currently, this methodology is extended to the solution of the bidomain equations with space-time adaptive discretization [26], see Fig. 18. While the progression of high performance computing (HPC) paths the way towards realistic simulations of human cardiac electrophysiology in real time at least with operator splitting methods for the monodomain model [171], [213], reduced order modelling of the mono- and bidomain equations [18], [186] and the reduction of physiological AP models [142] form further methodological strategies for bringing down complexity and computational costs. Another approach for the acceleration of mono- and bidomain solvers is the use of graphical processor units (GPU) [165], [273], [150], [97], which for benchmarks with 3D geometries enable speed ups by a factor of 10 to 200 in comparison with CPU implementations.
Simulation Based Proarrhythmic Risk Assessment
As outlined above, several methods for the computational assessment of a drug’s proarrhythmic risk have been suggested that involve mechanistic ODE or PDE models of cardiac electrophysiology. Given the computational burden and the higher degree of uncertainty of PDE models [242], [219], ODE models currently dominate the literature on simulation based cardiac drug safety testing [153], [122], [191]. In a proof of concept study [4], a single cell model was even shown to be equally accurate as a 3D ventricular chamber model in predicting abnormal electrical activity. In general, the performances of the proposed risk metrics and classification schemes are difficult to compare [271], also because nonuniform sets of drugs are considered and because individual compounds are, sometimes even inconsistently, placed into 2 up to 5 different risk categories (from low to high TdP risk) for the training of the respective classifiers. In the CiPA study [51], ternary classification based on proportional odds logistic regression and leave-one-out validation was chosen for the training compounds of Table 1. This determined qNet, see Fig. 16 (right), as the best performing metric out of 13 candidates derived from an optimized version of the ORd model. However, when exposed to uncertainties in the model parameters that describe drug binding and ionic current block, qNet may loose its ability to correctly separate drugs by their TdP risk [27], see Fig. 19 (left). In a study with a phenomenological AP model [116], qNet was shown to be a monotonically decreasing function of the normal distance from the region of EADs occurence in parameter space, see Fig. 19 (right), such that low risk drugs paradoxically are located closer to the critical domain than their high risk counterparts. Also, a global sensitivity analysis [188] of the CiPA ORd model revealed significant differences between the parameter sets that affect qNet and EADs generation, which further challenges the correlation of qNet with the robustness against EADs as postulated in [51]. Recently, a torsade mectric score [140], defined as an averaged qNet value, showed high accuracy in TdP risk prediction but still misclassified some of the validation compounds of Table 1. While further ODE simulation based risk markers, such as drug induced shortening of the electromechanical window [192] or concentration dependent repolarization reserve current [67], are currently under investigation, also a debate has been kicked off [189], [156], [157], [81] as to whether proarrhythmic risk classifiers built on features derived from cellular AP models actually provide increased prediction accuracies, in comparison with statistical learning approaches based on direct features such as ion channel blocking potencies [109], [156] or hiPSC-CM responses to drugs [16], [108], [193].
Modelling of hiPSC-CMs
A vital component of the CiPA initiative [262], [138] is to integrate drug effects on multiple ion channels into (an optimized version of) the ORd model of the human adult ventricular cardiomyocyte and to contrast model predictions with drug effects on human iPSC-CMs, see Fig. 5 (left). However, hiPSC-CMs differ from human adult CMs in morphology, electrophysiology and calcium handling, a situation often referred to as hiPSC-CM immaturity [214], [101], see Fig. 20. Furthermore, hiPSC-CMs possess a quite diverse spectrum of phenotypes, typically characterized by spontaneous beating, and only roughly categorizable into nodal-, atrial- and ventricular-like subtypes [99]. Hence, distinct computational models of hiPSC-CMs have been suggested as a tool to gain mechanistic insight into their immature phenotype [107], to interpret experimental findings in specific hiPSC-CM cell lines [133] and to translate from hiPSC-CMs to human adult CMs [73], [252], [93]. The series of papers [180], [181], [184], [185], [182] features corrective adaptations of a single cell ODE model (1) of a ventricular-like hiPSC-CM with membrane clock driven automaticity. In [107], the model of [181] is complemented by a PDE diffusion equation for the calcium release from the sacroplasmatic reticulum, which accounts for the calcium clock as the second mechanism behind the spontaneous activity of hiPSC-CMs. The dependency of the latter on the potassium current \(I_{K1}\) formulation is studied in [56], while [104] suggests a novel hiPSC-CM model that is based on a simplified description of voltage-dependent gating variable rate constants and incorporates experimental hiPSC-CM data from multiple laboratories. Still, all of the available hiPSC-CM models, in parts and due to lack of data, contain ionic current and calcium handling formulations from previously published models of human adult or even animal CM models. Furthermore, they so far only enable oversimplified mono- or bidomain simulations of experiments with hiPSC-CM monolayers [209], [20] or 3D engineered tissue, compare with Figs. 6 and 7, as statistical data about the phenotypic mix in such multi-cellular systems but also corresponding cell models, e.g. for nodal-like hiPSC-CMs, are missing. As a consequence, simulated drug effects on hiPSC-CMs may deviate from in vitro observations, see [117] for a study on the relation between beating frequency and duration of repolarization, such that more effort is needed to increase data availability and to refine hiPSC-CM models. With respect to an improved characterization of individual ionic currents, the dynamic clamp technique [14], [261] allows to inject computationally simulated currents into isolated hiPSC-CMs in real time and to test model-based hypotheses on ion channels in a closed loop setting. Further applications of dynamic clamp include the improvement of hiPSC-CM immaturity [74] and the investigation of drug induced EADs [179].
Cardiac Electro-Mechanic Coupling
In this paper, we surveyed computational methods in cardiac electrophysiology and their relevance for preclinical drug cardiotoxicity testing with hiPSC-CMs. In doing so, we followed the paradigm according to which cardiac arrhythmias result from disturbances of ion channel function and excitation wave propagation that cause desynchronized contractions of the heart. However, the mechanical forces produced by excitation can themselves impair ion channel function and calcium handling [106]. Mathematical models of excitation-contraction coupling (ECC) and mechano-electric feedback (MEF) exist both at the cellular and the tissue/organ level. In the first case, electrophysiological AP models are coupled to models of the myofilament protein dynamics [89], [170], [212] that describe tension and changes of sarcomere length (SL) in dependence on the intracellular calcium concentration. In turn, the calcium dynamics are affected, e.g., by calcium-troponin binding or stretch-activated ion channels. For human species, such cellular models are used to study the modulation of AP alternans [285] and EADs [284] formation by contractile mechanisms, but they also are integrated into electro-mechanic models of the human heart [123], [10], [124], see Fig. 21. In general, cardiac mechanics has a passive component, corresponding with the tissue response to external forces, and an active component, referring to the force generation for the muscle contraction. While the active stress approach [164], [173] considers the first-Piola-Kirchhoff tensor as the sum of the passive and the active component, the active strain approach [28], [6] is based on a factorization of the deformation gradient tensor into the product of a passive and an active part. The resulting equations of continuum mechanics [33], with the muscle deformation and the stress tensor as unknowns, are coupled to the mono- (8) or bidomain (6) equations and are used, e.g., to study the inotropic effects of ion channel mutations [168] or the impact of contraction on spiral wave activity in cardiac tissue [164], [28]. For a recent simulation study of simultaneous drug effects on contractility and electrophysiology of the 3D human left ventricular chamber, see [147]. As also multicellular hiPSC-CM modelling progresses, these computational tools will also be used in ECC and MEF studies of force-generating human myocardium models [240], [135], [136] that are bioengineered with stem cell derived CMs, see Fig. 22 for examples. In particular, such models of the human myocardium allow to screen compounds not only for side effects on electrophysiology but also on contractility [127], [220], [233], e.g., drug induced sarcomere shortening, and on structure [276], e.g., drug induced changes in morphology.
Further topics, that go beyond the CiPA paradigm, as illustrated in Fig. 5 (left), include drug-drug interactions [151], drug induced disruption of hERG trafficking [174], or hidden drug cardiotoxicity [59], where drugs only negatively affect the diseased heart. Moreover, the prediction of a drug’s proarrhythmic risk might be improved by taking clinical exposures and metabolites under stressed scenarios into account [134]. In tackling these problems, computational models of drug adverse effects on cardiac function can be of great value as they allow to run through thousands of complex scenarios, which is not possible in in vitro or in vivo experiments. Underlying mechanistic models of the human heart are continuously refined and extended, e.g., by consideration of blood flow [234], [10], [207], [206] or of physiology based pharmacokinetics [254]. However, with big data also interfusing the field of drug toxicity testing, these models face the competition with machine learning algorithms [148], [145], [72], [23] and related bioinformatics approaches [208], [144], [162], [128] such as quantitative structure activity relationship (QSAR) [66], [71]. Ultimately, all these in-silico disciplines need to build even stronger bonds with human-relevant testing platforms [187] in order to actually achieve replacement, reduction and refinement (3Rs) of animal testing [176].
References
ICH E14: Note for Guidance on the Clinical Evaluation of QT/ QTc Interval Prolongation and Proarrhythmic Potential for Nonantiarrhythmic Drugs; U.S. Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER), Center for Biologics Evaluation and Research (CBER) (2005). Reference CHMP/ICH/2/04
ICH S7B: Note for Guidance on the Nonclinical Evaluation of the Potential for Delayed Ventricular Repolarization (QT Interval Prolongation) by Human Pharmaceuticals; International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (2005). Reference CHMP/ICH/423/02
ASME V & V 40: Assessing Credibility of Computational Models Through Verification and Validation: Application to Medical Devices (2018). American Society of Mechanical Engineers
Abbasi, M., Small, B.G., Patel, N., Jamei, M., Polak, S.: Early assessment of proarrhythmic risk of drugs using the in vitro data and single-cell-based in silico models: proof of concept. Toxicol. Mech. Methods 27(2), 88–99 (2017). https://doi.org/10.1080/15376516.2016.1256460. PMID: 27813448
Abi-Gerges, N., Indersmitten, T., Truong, K., Nguyen, W., Ratchada, P., Nguyen, N., Page, G., Miller, P.E., Ghetti, A.: Multiparametric mechanistic profiling of inotropic drugs in adult human primary cardiomyocytes. Sci. Rep. 10(1), 7692 (2020). https://doi.org/10.1038/s41598-020-64657-2
Ambrosi, D., Arioli, G., Nobile, F., Quarteroni, A.: Electromechanical coupling in cardiac dynamics: the active strain approach. SIAM J. Appl. Math. 71(2), 605–621 (2011). https://doi.org/10.1137/100788379
Antzelevitch, C., Burashnikov, A.: Overview of basic mechanisms of cardiac arrhythmia. Cardiac Electrophysiol. Clinics 3(1), 23–45 (2011). https://doi.org/10.1016/j.ccep.2010.10.012. Basic Science for the Clinical Electrophysiologist
Arnold, L.: Random Dynamical Systems. Springer Monographs in Mathematics. Springer, Berlin (1998)
Baher, A., Qu, Z., Hayatdavoudi, A., Lamp, S.T., Yang, M.J., Xie, F., Turner, S., Garfinkel, A., Weiss, J.N.: Short-term cardiac memory and mother rotor fibrillation. Am. J. Physiol., Heart Circ. Physiol. 292(1), H180–H189 (2007). https://doi.org/10.1152/ajpheart.00944.2005. PMID: 16891403
Baillargeon, B., Rebelo, N., Fox, D.D., Taylor, R.L., Kuhl, E.: The Living Heart Project: a robust and integrative simulator for human heart function. Eur. J. Mech. A, Solids 48, 38–47 (2014). https://doi.org/10.1016/j.euromechsol.2014.04.001. Frontiers in Finite-Deformation Electromechanics
Barrio, R., Martínez, M.A., Pérez, L., Pueyo, E.: Bifurcations and slow-fast analysis in a cardiac cell model for investigation of early afterdepolarizations. Mathematics 8(6), 880 (2020). https://doi.org/10.3390/math8060880
Bartolucci, C., Paci, M., Hyttinen, J., Passini, E., Severi, S.: Evolution of the seminal o’hara rudy model to more accurately simulate the electrophysiology of human ventricular cardiomyocytes. In: 2019 Computing in Cardiology (CinC), pp. 1–4 (2019)
Beattie, K.A., Hill, A.P., Bardenet, R., Cui, Y., Vandenberg, J.I., Gavaghan, D.J., Boer, T.P., Mirams, G.R.: Sinusoidal voltage protocols for rapid characterisation of ion channel kinetics. J. Physiol. 596(10), 1813–1828 (2018). https://doi.org/10.1113/JP275733
Berecki, G., Verkerk, A.O., van Ginneken, A.C.G., Wilders, R.: Dynamic Clamp as a Tool to Study the Functional Effects of Individual Membrane Currents pp. 309–326. Springer, New York (2014)
Bers, D.M.: Excitation-Contraction Coupling and Cardiac Contractile Force. Developments in Cardiovascular Medicine. Springer, Berlin (2001). https://www.springer.com/de/book/9780792371571
Blinova, K., Dang, Q., Millard, D., Smith, G., Pierson, J., Guo, L., Brock, M., Lu, H.R., Kraushaar, U., Zeng, H., Shi, H., Zhang, X., Sawada, K., Osada, T., Kanda, Y., Sekino, Y., Pang, L., Feaster, T.K., Kettenhofen, R., Stockbridge, N., Strauss, D.G., Gintant, G.: International multisite study of human-induced pluripotent stem cell-derived cardiomyocytes for drug proarrhythmic potential assessment. Cell Rep. 24(13), 3582–3592 (2018). https://doi.org/10.1016/j.celrep.2018.08.079
Bot, C.T., Juhasz, K., Haeusermann, F., Polonchuk, L., Traebert, M., Stoelzle-Feix, S.: Cross - site comparison of excitation-contraction coupling using impedance and field potential recordings in hiPSC cardiomyocytes. J. Pharmacol. Toxicol. Methods 93, 46–58 (2018). https://doi.org/10.1016/j.vascn.2018.06.006. Special Issue: Fifteenth Annual Focused Issue on Methods in Safety Pharmacology
Boulakia, M., Schenone, E., Gerbeau, J.F.: Reduced-order modeling for cardiac electrophysiology. Application to parameter identification. Int. J. Numer. Methods Biomed. Eng. 28(6–7), 727–744 (2012). https://doi.org/10.1002/cnm.2465
Bourgault, Y., Pierre, C.: Comparing the bidomain and monodomain models in electro-cardiology through convergence analysis (2010). https://hal.archives-ouvertes.fr/hal-00545888. Working paper or preprint
Bowler, L.A., Gavaghan, D.J., Mirams, G.R., Whiteley, J.P.: Representation of multiple cellular phenotypes within tissue-level simulations of cardiac electrophysiology. Bull. Math. Biol. 81(1), 7–38 (2019). https://doi.org/10.1007/s11538-018-0516-1
Brennan, T., Fink, M., Rodriguez, B.: Multiscale modelling of drug-induced effects on cardiac electrophysiological activity. Eur. J. Pharm. Sci. 36(1), 62–77 (2009). https://doi.org/10.1016/j.ejps.2008.09.013. Biosimulations for Pharmaceutical Sciences
Britton, O.J., Bueno-Orovio, A., Van Ammel, K., Lu, H.R., Towart, R., Gallacher, D.J., Rodriguez, B.: Experimentally calibrated population of models predicts and explains intersubject variability in cardiac cellular electrophysiology. Proc. Natl. Acad. Sci. 110(23), E2098–E2105 (2013). https://doi.org/10.1073/pnas.1304382110
Cai, C., Guo, P., Zhou, Y., Zhou, J., Wang, Q., Zhang, F., Fang, J., Cheng, F.: Deep learning-based prediction of drug-induced cardiotoxicity. J. Chem. Inf. Model. 59(3), 1073–1084 (2019). https://doi.org/10.1021/acs.jcim.8b00769
Cannon, R.C., D’Alessandro, G.: The ion channel inverse problem: neuroinformatics meets biophysics. PLoS Comput. Biol. 2(8), e91 (2006). https://doi.org/10.1371/journal.pcbi.0020091
Chamakuri, N., Kügler, P.: A coupled monodomain solver with optimal memory usage for the simulation of cardiac wave propagation. Appl. Math. Comput. 378, 125,212 (2020). https://doi.org/10.1016/j.amc.2020.125212
Chamakuri, N., Kügler, P.: Parallel space-time adaptive numerical simulation of 3D cardiac electrophysiology with the bidomain model. J. Comput. Sci. (2020, under review)
Chang, K.C., Dutta, S., Mirams, G.R., Beattie, K.A., Sheng, J., Tran, P.N., Wu, M., Wu, W.W., Colatsky, T., Strauss, D.G., Li, Z.: Uncertainty quantification reveals the importance of data variability and experimental design considerations for in silico proarrhythmia risk assessment. Front. Physiol. 8, 917 (2017). https://doi.org/10.3389/fphys.2017.00917
Cherubini, C., Filippi, S., Nardinocchi, P., Teresi, L.: An electromechanical model of cardiac tissue: constitutive issues and electrophysiological effects. Prog. Biophys. Mol. Biol. 97(2), 562–573 (2008). https://doi.org/10.1016/j.pbiomolbio.2008.02.001. Life and Mechanosensitivity
Christophe, B.: Simulation of early after-depolarisation in non-failing human ventricular myocytes: can this help cardiac safety pharmacology? Pharmacol. Rep. 65, 1281–1293 (2013)
Christophe, B., Crumb, W.J.: Impact of disease state on arrhythmic event detection by action potential modelling in cardiac safety pharmacology. J. Pharmacol. Toxicol. Methods 96, 15–26 (2019). https://doi.org/10.1016/j.vascn.2018.12.004
Chu, Z., Yang, D., Huang, X.: Conditions for the genesis of early afterdepolarization in a model of a ventricular myocyte. Chaos 30(4), 043,105 (2020). https://doi.org/10.1063/1.5133086
Chugh, S.S., Havmoeller, R., Narayanan, K., Singh, D., Rienstra, M., Benjamin, E.J., Gillum, R.F., Kim, Y.H., McAnulty, J.H., Zheng, Z.J., Forouzanfar, M.H., Naghavi, M., Mensah, G.A., Ezzati, M., Murray, C.J.: Worldwide epidemiology of atrial fibrillation. Circulation 129(8), 837–847 (2014). https://doi.org/10.1161/CIRCULATIONAHA.113.005119
Ciarlet, P.G.: Mathematical Elasticity. Studies in Mathematics and Its Applications, vol. 20. Elsevier, Amsterdam (2009)
Clements, M., Roquemore, L.: Stem Cell-Derived Models in Toxicology. Methods in Pharmacology and Toxicology. Springer, New York (2018). https://books.google.de/books?id=24RFuQEACAAJ
Clerx, M., Beattie, K.A., Gavaghan, D.J., Mirams, G.R.: Four ways to fit an ion channel model. Biophys. J. (2019). https://doi.org/10.1016/j.bpj.2019.08.001
Colatsky, T., Fermini, B., Gintant, G., Pierson, J.B., Sager, P., Sekino, Y., Strauss, D.G., Stockbridge, N.: The comprehensive in vitro proarrhythmia assay (CiPA) initiative — update on progress. J. Pharmacol. Toxicol. Methods 81, 15–20 (2016). https://doi.org/10.1016/j.vascn.2016.06.002. Focused Issue on Safety Pharmacology
Colman, M.A., Ni, H., Liang, B., Schmitt, N., Zhang, H.: In silico assessment of genetic variation in kcna5 reveals multiple mechanisms of human atrial arrhythmogenesis. PLoS Comput. Biol. 13(6), 1–21 (2017). https://doi.org/10.1371/journal.pcbi.1005587
Colman, M.A., Saxena, P., Kettlewell, S., Workman, A.J.: Description of the human atrial action potential derived from a single, congruent data source: novel computational models for integrated experimental-numerical study of atrial arrhythmia mechanisms. Front. Physiol. 9, 1211 (2018). https://doi.org/10.3389/fphys.2018.01211
Costabal, F.S., Matsuno, K., Yao, J., Perdikaris, P., Kuhl, E.: Machine learning in drug development: characterizing the effect of 30 drugs on the QT interval using Gaussian process regression, sensitivity analysis, and uncertainty quantification. Comput. Methods Appl. Mech. Eng. 348, 313–333 (2019). https://doi.org/10.1016/j.cma.2019.01.033
Costabal, F.S., Yao, J., Kuhl, E.: Predicting the cardiac toxicity of drugs using a novel multiscale exposure–response simulator. Comput. Methods Biomech. Biomed. Eng. 21(3), 232–246 (2018). https://doi.org/10.1080/10255842.2018.1439479. PMID: 29493299
Coudière, Y., Bourgault, Y., Rioux, M.: Optimal monodomain approximations of the bidomain equations used in cardiac electrophysiology. Math. Models Methods Appl. Sci. 24(6), 1115–1140 (2014). https://doi.org/10.1142/S0218202513500784
Courtemanche, M., Ramirez, R.J., Nattel, S.: Ionic mechanisms underlying human atrial action potential properties: insights from a mathematical model. Am. J. Physiol., Heart Circ. Physiol. 275(1), H301–H321 (1998). https://doi.org/10.1152/ajpheart.1998.275.1.H301. PMID: 29586616
Crampin, E.J., Halstead, M., Hunter, P., Nielsen, P., Noble, D., Smith, N., Tawhai, M.: Computational physiology and the Physiome Project. Exp. Physiol. 89(1), 1–26 (2004). https://doi.org/10.1113/expphysiol.2003.026740
Curran, M.E., Splawski, I., Timothy, K.W., Vincen, G., Green, E.D., Keating, M.T.: A molecular basis for cardiac arrhythmia: Herg mutations cause long qt syndrome. Cell 80(5), 795–803 (1995). https://doi.org/10.1016/0092-8674(95)90358-5
Noble, D.: A modification of the Hodgkin—Huxley equations applicable to Purkinje fibre action and pacemaker potentials. J. Physiol. 160(2), 317–352 (1962). https://doi.org/10.1113/jphysiol.1962.sp006849
Daly, A.C., Gavaghan, D., Cooper, J., Tavener, S.: Inference-based assessment of parameter identifiability in nonlinear biological models. J. R. Soc. Interface 15(144), 20180, 318 (2018). https://doi.org/10.1098/rsif.2018.0318
Daly, A.C., Gavaghan, D.J., Holmes, C., Cooper, J.: Hodgkin Huxley revisited: reparametrization and identifiability analysis of the classic action potential model with approximate Bayesian methods. R. Soc. Open Sci. 2(12), 150499 (2015). https://doi.org/10.1098/rsos.150499
Davies, M.R., Wang, K., Mirams, G.R., Caruso, A., Noble, D., Walz, A., Lavé, T., Schuler, F., Singer, T., Polonchuk, L.: Recent developments in using mechanistic cardiac modelling for drug safety evaluation. Drug Discov. Today 21(6), 924–938 (2016). http://www.sciencedirect.com/science/article/pii/S1359644616300290
Desroches, M., Guckenheimer, J., Krauskopf, B., Kuehn, C., Osinga, H., Wechselberger, M.: Mixed-mode oscillations with multiple time scales. SIAM Rev. 54(2), 211–288 (2012). https://doi.org/10.1137/100791233
Di Veroli, G.Y., Davies, M.R., Zhang, H., Abi-Gerges, N., Boyett, M.R.: High-throughput screening of drug-binding dynamics to HERG improves early drug safety assessment. Am. J. Physiol., Heart Circ. Physiol. 304(1), H104–H117 (2013). https://doi.org/10.1152/ajpheart.00511.2012. PMID: 23103500
Dutta, S., Chang, K.C., Beattie, K.A., Sheng, J., Tran, P.N., Wu, W.W., Wu, M., Strauss, D.G., Colatsky, T., Li, Z.: Optimization of an in silico cardiac cell model for proarrhythmia risk assessment. Front. Physiol. 8, 616 (2017). https://doi.org/10.3389/fphys.2017.00616
Dutta, S., Mincholé, A., Zacur, E., Quinn, T.A., Taggart, P., Rodriguez, B.: Early afterdepolarizations promote transmural reentry in ischemic human ventricles with reduced repolarization reserve. Prog. Biophys. Mol. Biol. 120(1), 236–248 (2016). https://doi.org/10.1016/j.pbiomolbio.2016.01.008. Recent Developments in Biophysics & Molecular Biology of Heart Rhythm
Edelman, G.M., Gally, J.A.: Degeneracy and complexity in biological systems. Proc. Natl. Acad. Sci. 98(24), 13,763–13,768 (2001). https://doi.org/10.1073/pnas.231499798
Ethier, M., Bourgault, Y.: Semi-implicit time-discretization schemes for the bidomain model. SIAM J. Numer. Anal. 46(5), 2443–2468 (2008). https://doi.org/10.1137/070680503
Fabbri, A., Fantini, M., Wilders, R., Severi, S.: Computational analysis of the human sinus node action potential: model development and effects of mutations. J. Physiol. 595(7), 2365–2396 (2017). https://doi.org/10.1113/JP273259
Fabbri, A., Goversen, B., Vos, M.A., van Veen, T.A., de Boer, T.P.: Required GK1 to suppress automaticity of iPSC-CMs depends strongly on IK1 model structure. Biophys. J. (2019). https://doi.org/10.1016/j.bpj.2019.08.040
Fall, C., Marland, E., Wagner, J., Tyson, J.: Computational Cell Biology. Interdisciplinary Applied Mathematics. Springer, New York (2007). https://books.google.de/books?id=AdCTvbOzRywC
Fenton, F.H., Cherry, E.M., Hastings, H.M., Evans, S.J.: Multiple mechanisms of spiral wave breakup in a model of cardiac electrical activity. Chaos 12(3), 852–892 (2002). https://doi.org/10.1063/1.1504242
Ferdinandy, P., Baczkó, I., Bencsik, P., Giricz, Z., Görbe, A., Pacher, P., Varga, Z.V., Varró, A., Schulz, R.: Definition of hidden drug cardiotoxicity: paradigm change in cardiac safety testing and its clinical implications. Eur. Heart J. 40(22), 1771–1777 (2018). https://doi.org/10.1093/eurheartj/ehy365
Fermini, B., Fossa, A.A.: The impact of drug-induced QT interval prolongation on drug discovery and development. Nat. Rev. Drug Discov. 2, 439 EP (2003). https://doi.org/10.1038/nrd1108. Review Article
Fermini, B., Hancox, J.C., Abi-Gerges, N., Bridgland-Taylor, M., Chaudhary, K.W., Colatsky, T., Correll, K., Crumb, W., Damiano, B., Erdemli, G., Gintant, G., Imredy, J., Koerner, J., Kramer, J., Levesque, P., Li, Z., Lindqvist, A., Obejero-Paz, C.A., Rampe, D., Sawada, K., Strauss, D.G., Vandenberg, J.I.: A new perspective in the field of cardiac safety testing through the comprehensive in vitro proarrhythmia assay paradigm. J. Biomol. Screen. 21(1), 1–11 (2016). https://doi.org/10.1177/1087057115594589. PMID: 26170255
Fink, M., Noble, D.: Markov models for ion channels: versatility versus identifiability and speed. Philos. Trans. R. Soc., Math. Phys. Eng. Sci. 367(1896), 2161–2179 (2009). https://doi.org/10.1098/rsta.2008.0301
FitzHugh, R.: Impulses and physiological states in theoretical models of nerve membrane. Biophys. J. 1(6), 445–466 (1961). https://doi.org/10.1016/S0006-3495(61)86902-6
Franzone, P., Pavarino, L., Scacchi, S.: Mathematical Cardiac Electrophysiology. MS&A. Springer, Berlin (2014). https://books.google.de/books?id=H7shBQAAQBAJ
Franzone, P.C., Pavarino, L., Taccardi, B.: Simulating patterns of excitation, repolarization and action potential duration with cardiac bidomain and monodomain models. Math. Biosci. 197(1), 35–66 (2005). https://doi.org/10.1016/j.mbs.2005.04.003
Frid, A.A., Matthews, E.J.: Prediction of drug-related cardiac adverse effects in humans-b: use of QSAR programs for early detection of drug-induced cardiac toxicities. Regul. Toxicol. Pharmacol. 56(3), 276–289 (2010). https://doi.org/10.1016/j.yrtph.2009.11.005
Gaur, N., Ortega, F., Verkerk, A.O., Mengarelli, I., Krogh-Madsen, T., Christini, D.J., Coronel, R., Vigmond, E.J.: Validation of quantitative measure of repolarization reserve as a novel marker of drug induced proarrhythmia. J. Mol. Cell. Cardiol. (2020). https://doi.org/10.1016/j.yjmcc.2020.04.019
Ghosh, S., Matsuoka, Y., Asai, Y., Hsin, K.Y., Kitano, H.: Toward an integrated software platform for systems pharmacology. Biopharm. Drug Dispos. 34(9), 508–526 (2013). https://doi.org/10.1002/bdd.1875
Gima, K., Rudy, Y.: Ionic current basis of electrocardiographic waveforms. Circ. Res. 90(8), 889–896 (2002). https://doi.org/10.1161/01.RES.0000016960.61087.86
Gintant, G., Sager, P.T., Stockbridge, N.: Evolution of strategies to improve preclinical cardiac safety testing. Nat. Rev. Drug Discov. 15, 457 EP (2016). https://doi.org/10.1038/nrd.2015.34. Perspective
Gobbi, M., Beeg, M., Toropova, M.A., Toropov, A.A., Salmona, M.: Monte Carlo method for predicting of cardiac toxicity: hERG blocker compounds. Toxicol. Lett. 250–251, 42–46 (2016). https://doi.org/10.1016/j.toxlet.2016.04.010
Golgooni, Z., Mirsadeghi, S., Baghshah, M.S., Ataee, P., Baharvand, H., Pahlavan, S., Rabiee, H.R.: Deep learning-based proarrhythmia analysis using field potentials recorded from human pluripotent stem cells derived cardiomyocytes. IEEE J. Transl. Eng. Health Med. 7, 1900203 (2019)
Gong, J.Q.X., Sobie, E.A.: Population-based mechanistic modeling allows for quantitative predictions of drug responses across cell types. npj Syst. Biol. Appl. 4(1), 11 (2018). https://doi.org/10.1038/s41540-018-0047-2
Goversen, B., Becker, N., Stoelzle-Feix, S., Obergrussberger, A., Vos, M.A., van Veen, T.A.B., Fertig, N., de Boer, T.P.: A hybrid model for safety pharmacology on an automated patch clamp platform: using dynamic clamp to join ipsc-derived cardiomyocytes and simulations of ik1 ion channels in real-time. Front. Physiol. 8, 1094 (2018). https://doi.org/10.3389/fphys.2017.01094
Grandi, E., Morotti, S., Pueyo, E., Rodriguez, B.: Editorial: safety pharmacology – risk assessment qt interval prolongation and beyond. Front. Physiol. 9, 678 (2018). https://doi.org/10.3389/fphys.2018.00678
Grandi, E., Pandit, S.V., Voigt, N., Workman, A.J., Dobrev, D., Jalife, J., Bers, D.M.: Human atrial action potential and \(\text{Ca}^{2+}\) model. Circ. Res. 109(9), 1055–1066 (2011). https://doi.org/10.1161/CIRCRESAHA.111.253955
Grandi, E., Pasqualini, F.S., Bers, D.M.: A novel computational model of the human ventricular action potential and ca transient. J. Mol. Cell. Cardiol. 48(1), 112–121 (2010). https://doi.org/10.1016/j.yjmcc.2009.09.019. Special Issue: Ion Channels
Gray, R.A., Franz, M.R.: A model for human action potential dynamics in vivo. Am. J. Physiol., Heart Circ. Physiol. 318(3), H534–H546 (2020). https://doi.org/10.1152/ajpheart.00557.2019. PMID: 31951472
Gray, R.A., Pathmanathan, P.: A parsimonious model of the rabbit action potential elucidates the minimal physiological requirements for alternans and spiral wave breakup. PLoS Comput. Biol. 12(10), 1–21 (2016). https://doi.org/10.1371/journal.pcbi.1005087
Groenendaal, W., Ortega, F.A., Kherlopian, A.R., Zygmunt, A.C., Krogh-Madsen, T., Christini, D.J.: Cell-specific cardiac electrophysiology models. PLoS Comput. Biol. 11(4), 1–22 (2015). https://doi.org/10.1371/journal.pcbi.1004242
Han, S., Han, S., Kim, K.S., Lee, H.A., Yim, D.S.: Usefulness of Bnet, a simple linear metric in discerning Torsades de Pointes risks in 28 CiPA drugs. Front. Pharmacol. 10, 1419 (2019). https://doi.org/10.3389/fphar.2019.01419
Hastie, T., Tibshirani, R., Friedman, J.: The Elements of Statistical Learning, 2nd edn. Springer Series in Statistics. Springer, Berlin (2009)
Heijman, J., Erfanian Abdoust, P., Voigt, N., Nattel, S., Dobrev, D.: Computational models of atrial cellular electrophysiology and calcium handling, and their role in atrial fibrillation. J. Physiol. 594(3), 537–553 (2016). https://doi.org/10.1113/JP271404
Hille, B.: Ion Channels of Excitable Membranes, 3rd edn. Sinauer Associates Inc., Sunderland (2001). http://www.amazon.com/exec/obidos/redirect?tag=citeulike07-20&path=ASIN/0878933212
Hodgkin, A.L., Huxley, A.F.: A quantitative description of membrane current and its application to conduction and excitation in nerve. J. Physiol. 117(4), 500–544 (1952). https://doi.org/10.1113/jphysiol.1952.sp004764
Hoekstra, M., Mummery, C., Wilde, A., Bezzina, C., Verkerk, A.: Induced pluripotent stem cell derived cardiomyocytes as models for cardiac arrhythmias. Front. Physiol. 3, 346 (2012). https://doi.org/10.3389/fphys.2012.00346
Huang, X., Song, Z., Qu, Z.: Determinants of early afterdepolarization properties in ventricular myocyte models. bioRxiv (2018). https://doi.org/10.1101/373266
Huffaker, R.B., Weiss, J.N., Kogan, B.: Effects of early afterdepolarizations on reentry in cardiac tissue: a simulation study. Am. J. Physiol., Heart Circ. Physiol. 292(6), H3089–H3102 (2007). https://doi.org/10.1152/ajpheart.01309.2006. PMID: 17307992
Hunter, P., McCulloch, A., ter Keurs, H.: Modelling the mechanical properties of cardiac muscle. Prog. Biophys. Mol. Biol. 69(2), 289–331 (1998). https://doi.org/10.1016/S0079-6107(98)00013-3
Hunter, P.J., Crampin, E.J., Nielsen, P.M.F.: Bioinformatics, multiscale modeling and the IUPS Physiome Project. Brief. Bioinform. 9(4), 333–343 (2008). https://doi.org/10.1093/bib/bbn024
Hurtado, D.E., Kuhl, E.: Computational modelling of electrocardiograms: repolarisation and t-wave polarity in the human heart. Comput. Methods Biomech. Biomed. Eng. 17(9), 986–996 (2014). https://doi.org/10.1080/10255842.2012.729582. PMID: 23113842
Hwang, M., Han, S., Park, M.C., Leem, C.H., Shim, E.B., Yim, D.S.: Three-dimensional heart model-based screening of proarrhythmic potential by in silico simulation of action potential and electrocardiograms. Front. Physiol. 10, 1139 (2019). https://doi.org/10.3389/fphys.2019.01139
Jæger, K.H., Charwat, V., Charrez, B., Finsberg, H., Maleckar, M.M., Wall, S., Healy, K.E., Tveito, A.: Improved computational identification of drug response using optical measurements of human stem cell derived cardiomyocytes in microphysiological systems. BioRxiv (2019). https://doi.org/10.1101/787390
Jin, Q., Greenstein, J.L., Winslow, R.L.: Estimating the probability of cellular arrhythmias with simplified statistical models that account for experimentally observed uncertainty in underlying biophysical mechanisms. bioRxiv (2020). https://doi.org/10.1101/2020.04.09.034843
Johnstone, R.H., Chang, E.T., Bardenet, R., de Boer, T.P., Gavaghan, D.J., Pathmanathan, P., Clayton, R.H., Mirams, G.R.: Uncertainty and variability in models of the cardiac action potential: can we build trustworthy models? J. Mol. Cell. Cardiol. 96, 49–62 (2016). https://doi.org/10.1016/j.yjmcc.2015.11.018. Special Issue: Computational Modelling of the Heart
Jæger, K.H., Wall, S., Tveito, A.: Detecting undetectables: can conductances of action potential models be changed without appreciable change in the transmembrane potential? Chaos 29(7), 073,102 (2019). https://doi.org/10.1063/1.5087629
Kaboudian, A., Velasco-Perez, H.A., Iravanian, S., Shiferaw, Y., Cherry, E.M., Fenton, F.H.: A Comprehensive Comparison of GPU Implementations of Cardiac Electrophysiology Models, pp. 9–34. Springer, Cham (2019)
Kanda, Y., Yamazaki, D., Osada, T., Yoshinaga, T., Sawada, K.: Development of torsadogenic risk assessment using human induced pluripotent stem cell-derived cardiomyocytes: Japan iPS Cardiac Safety Assessment (JiCSA) update. J. Pharmacol. Sci. 138(4), 233–239 (2018). https://doi.org/10.1016/j.jphs.2018.10.010
Kane, C., Terracciano, C.M.N.: Concise review: criteria for chamber-specific categorization of human cardiac myocytes derived from pluripotent stem cells. Stem Cells 35(8), 1881–1897 (2017). https://doi.org/10.1002/stem.2649
Karagueuzian, H.S., Stepanyan, H., Mandel, W.J.: Bifurcation theory and cardiac arrhythmias. Am. J. Cardiovasc. Dis. 3(1), 16 (2013)
Karakikes, I., Ameen, M., Termglinchan, V., Wu, J.C.: Human induced pluripotent stem cell derived cardiomyocytes. Circ. Res. 117(1), 80–88 (2015). https://doi.org/10.1161/CIRCRESAHA.117.305365
Kawatou, M., Masumoto, H., Fukushima, H., Morinaga, G., Sakata, R., Ashihara, T., Yamashita, J.K.: Modelling Torsade de Pointes arrhythmias in vitro in 3D human iPS cell-engineered heart tissue. Nat. Commun. 8(1), 1078 (2017). https://doi.org/10.1038/s41467-017-01125-y
Keener, J., Sneyd, J.: Mathematical Physiology: I: Cellular Physiology. Interdisciplinary Applied Mathematics. Springer, Berlin (2009). https://books.google.de/books?id=VVlFAQAAIAAJ
Kernik, D.C., Morotti, S., Wu, H., Garg, P., Duff, H.J., Kurokawa, J., Jalife, J., Wu, J.C., Grandi, E., Clancy, C.E.: A computational model of induced pluripotent stem-cell derived cardiomyocytes incorporating experimental variability from multiple data sources. J. Physiol. 597(17), 4533–4564 (2019). https://doi.org/10.1113/JP277724
Killeen, M.J.: Cardiac Drug Safety. World Scientific, Singapore (2012). https://www.worldscientific.com/worldscibooks/10.1142/7839
Kohl, P., Sachs, F., Franz, M.R.: Cardiac Mechano-Electric Coupling and Arrhytmias, 2nd edn. Oxford University Press, London (2011)
Koivumäki, J.T., Naumenko, N., Tuomainen, T., Takalo, J., Oksanen, M., Puttonen, K.A., Lehtonen, Å., Kuusisto, J., Laakso, M., Koistinaho, J., Tavi, P.: Structural immaturity of human ipsc-derived cardiomyocytes: in silico investigation of effects on function and disease modeling. Front. Physiol. 9, 80 (2018). https://doi.org/10.3389/fphys.2018.00080
Kopljar, I., Lu, H.R., Van Ammel, K., Otava, M., Tekle, F., Teisman, A., Gallacher, D.J.: Development of a human ipsc cardiomyocyte-based scoring system for cardiac hazard identification in early drug safety de-risking. Stem Cell Rep. 11(6), 1365–1377 (2018). https://doi.org/10.1016/j.stemcr.2018.11.007
Kramer, J., Obejero-Paz, C.A., Myatt, G., Kuryshev, Y.A., Bruening-Wright, A., Verducci, J.S., Brown, A.M.: MICE models: superior to the herg model in predicting torsade de pointes. Sci. Rep. 3, 2100 EP (2013). https://doi.org/10.1038/srep02100. Article
Krogh-Madsen, T., Christini, D.J.: Nonlinear dynamics in cardiology. Annu. Rev. Biomed. Eng. 14(1), 179–203 (2012). https://doi.org/10.1146/annurev-bioeng-071811-150106. PMID: 22524390
Krogh-Madsen, T., Christini, D.J.: Slow \([{N}a^{+}]_{i}\) dynamics impacts arrhythmogenesis and spiral wave reentry in cardiac myocyte ionic model. Chaos 27(9), 093,907 (2017). https://doi.org/10.1063/1.4999475
Krogh-Madsen, T., Jacobson, A.F., Ortega, F.A., Christini, D.J.: Global optimization of ventricular myocyte model to multi-variable objective improves predictions of drug-induced Torsades de Pointes. Front. Physiol. 8, 1059 (2017). https://doi.org/10.3389/fphys.2017.01059
Kubo, T., Ashihara, T., Tsubouchi, T., Horie, M.: Significance of integrated in silico transmural ventricular wedge preparation models of human non-failing and failing hearts for safety evaluation of drug candidates. J. Pharmacol. Toxicol. Methods 83, 30–41 (2017). https://doi.org/10.1016/j.vascn.2016.08.007
Kügler, P.: Early afterdepolarizations with growing amplitudes via delayed subcritical Hopf bifurcations and unstable manifolds of saddle foci in cardiac action potential dynamics. PLoS ONE 11(3), 1–14 (2016). https://doi.org/10.1371/journal.pone.0151178
Kügler, P., Bulelzai, M., Erhardt, A.H.: Period doubling cascades of limit cycles in cardiac action potential models as precursors to chaotic early afterdepolarizations. BMC Syst. Biol. 11(1), 42 (2017). https://doi.org/10.1186/s12918-017-0422-4
Kügler, P., Erhardt, A.H., Bulelzai, M.A.K.: Early afterdepolarizations in cardiac action potentials as mixed mode oscillations due to a folded node singularity. PLoS ONE 13(12), 1–22 (2019). https://doi.org/10.1371/journal.pone.0209498
Kügler, P., Rast, G., Guth, B.D.: Comparison of in vitro and computational experiments on the relation of inter-beat interval and duration of repolarization in a specific type of human induced pluripotent stem cell-derived cardiomyocytes. PLoS ONE 14(9), 1–14 (2019). https://doi.org/10.1371/journal.pone.0221763
Kurata, Y., Hisatome, I., Matsuda, H., Shibamoto, T.: Dynamical mechanisms of pacemaker generation in IK1-downregulated human ventricular myocytes: insights from bifurcation analyses of a mathematical model. Biophys. J. 89(4), 2865–2887 (2005). https://doi.org/10.1529/biophysj.105.060830
Kurata, Y., Tsumoto, K., Hayashi, K., Hisatome, I., Kuda, Y., Tanida, M.: Multiple dynamical mechanisms of phase-2 early afterdepolarizations in a human ventricular myocyte model: Involvement of spontaneous SR Ca2+ release. bioRxiv (2020). https://doi.org/10.1101/613182
Kurata, Y., Tsumoto, K., Hayashi, K., Hisatome, I., Tanida, M., Kuda, Y., Shibamoto, T.: Dynamical mechanisms of phase-2 early afterdepolarizations in human ventricular myocytes: insights from bifurcation analyses of two mathematical models. Am. J. Physiol., Heart Circ. Physiol. 312(1), H106–H127 (2017). https://doi.org/10.1152/ajpheart.00115.2016. PMID: 27836893
Kuznetsov, Y.A.: Elements of Applied Bifurcation Theory, vol. 112. Springer, Berlin (2013)
Lancaster, M.C., Sobie, E.: Improved prediction of drug-induced torsades de pointes through simulations of dynamics and machine learning algorithms. Clin. Pharmacol. Ther. 100(4), 371–379 (2016). https://doi.org/10.1002/cpt.367
Land, S., Niederer, S.A., Smith, N.P.: Efficient computational methods for strongly coupled cardiac electromechanics. IEEE Trans. Biomed. Eng. 59(5), 1219–1228 (2012). https://doi.org/10.1109/TBME.2011.2112359
Land, S., Park-Holohan, S.J., Smith, N.P., dos Remedios, C.G., Kentish, J.C., Niederer, S.A.: A model of cardiac contraction based on novel measurements of tension development in human cardiomyocytes. J. Mol. Cell. Cardiol. 106, 68–83 (2017). https://doi.org/10.1016/j.yjmcc.2017.03.008
Lawrence, C.L., Pollard, C.E., Hammond, T.G., Valentin, J.P.: Nonclinical proarrhythmia models: predicting Torsades de Pointes. J. Pharmacol. Toxicol. Methods 52(1), 46–59 (2005). https://doi.org/10.1016/j.vascn.2005.04.011. Second Annual Focus on Safety Pharmacology
Lawson, B.A.J., Drovandi, C.C., Cusimano, N., Burrage, P., Rodriguez, B., Burrage, K.: Unlocking data sets by calibrating populations of models to data density: a study in atrial electrophysiology. Sci. Adv. 4(1), e1701676 (2018). https://doi.org/10.1126/sciadv.1701676
Lee, E.K., Tran, D.D., Keung, W., Chan, P., Wong, G., Chan, C.W., Costa, K.D., Li, R.A., Khine, M.: Machine learning of human pluripotent stem cell-derived engineered cardiac tissue contractility for automated drug classification. Stem Cell Rep. 9(5), 1560–1572 (2017). https://doi.org/10.1016/j.stemcr.2017.09.008
Lee, H.M., Yu, M.S., Kazmi, S.R., Oh, S.Y., Rhee, K.H., Bae, M.A., Lee, B.H., Shin, D.S., Oh, K.S., Ceong, H., Lee, D., Na, D.: Computational determination of hERG-related cardiotoxicity of drug candidates. BMC Bioinform. 20(10), 250 (2019). https://doi.org/10.1186/s12859-019-2814-5
Lee, W., Mann, S.A., Windley, M.J., Imtiaz, M.S., Vandenberg, J.I., Hill, A.P.: In silico assessment of kinetics and state dependent binding properties of drugs causing acquired LQTS. Prog. Biophys. Mol. Biol. 120(1), 89–99 (2016). https://doi.org/10.1016/j.pbiomolbio.2015.12.005. Recent Developments in Biophysics & Molecular Biology of Heart Rhythm
Lei, C.L., Clerx, M., Gavaghan, D.J., Polonchuk, L., Mirams, G.R., Wang, K.: Rapid characterization of hERG channel kinetics I: using an automated high-throughput system. Biophys. J. (2019). https://doi.org/10.1016/j.bpj.2019.07.029
Lei, C.L., Clerx, M., Whittaker, D.G., Gavaghan, D.J., de Boer, T.P., Mirams, G.R.: Accounting for variability in ion current recordings using a mathematical model of artefacts in voltage-clamp experiments. Philos. Trans. R. Soc., Math. Phys. Eng. Sci. 378(2173), 20190, 348 (2020). https://doi.org/10.1098/rsta.2019.0348
Lei, C.L., Ghosh, S., Whittaker, D.G., Aboelkassem, Y., Beattie, K.A., Cantwell, C.D., Delhaas, T., Houston, C., Novaes, G.M., Panfilov, A.V., Pathmanathan, P., Riabiz, M., dos Santos, R.W., Walmsley, J., Worden, K., Mirams, G.R., Wilkinson, R.D.: Considering discrepancy when calibrating a mechanistic electrophysiology model. Philos. Trans. R. Soc., Math. Phys. Eng. Sci. 378(2173), 20190, 349 (2020). https://doi.org/10.1098/rsta.2019.0349
Lei, C.L., Wang, K., Clerx, M., Johnstone, R.H., Hortigon-Vinagre, M.P., Zamora, V., Allan, A., Smith, G.L., Gavaghan, D.J., Mirams, G.R., Polonchuk, L.: Tailoring mathematical models to stem-cell derived cardiomyocyte lines can improve predictions of drug-induced changes to their electrophysiology. Front. Physiol. 8, 986 (2017). https://doi.org/10.3389/fphys.2017.00986
Leishman, D.J.: Improving prediction of torsadogenic risk in the cipa in silico model by appropriately accounting for clinical exposure. J. Pharmacol. Toxicol. Methods 101, 106,654 (2020). https://doi.org/10.1016/j.vascn.2019.106654
Lemoine, M.D., Krause, T., Koivumäki, J.T., Prondzynski, M., Schulze, M.L., Girdauskas, E., Willems, S., Hansen, A., Eschenhagen, T., Christ, T.: Human induced pluripotent stem cell derived engineered heart tissue as a sensitive test system for QT prolongation and arrhythmic triggers. Circ. Arrhythm. Electrophysiol. 11(7), e006,035 (2018). https://doi.org/10.1161/CIRCEP.117.006035
Li, R.A., Keung, W., Cashman, T.J., Backeris, P.C., Johnson, B.V., Bardot, E.S., Wong, A.O., Chan, P.K., Chan, C.W., Costa, K.D.: Bioengineering an electro-mechanically functional miniature ventricular heart chamber from human pluripotent stem cells. Biomaterials 163, 116–127 (2018). https://doi.org/10.1016/j.biomaterials.2018.02.024
Li, Z., Dutta, S., Sheng, J., Tran, P.N., Wu, W., Chang, K., Mdluli, T., Strauss, D.G., Colatsky, T.: Improving the in silico assessment of proarrhythmia risk by combining hERG (human ether-a-go-go-related gene) channel–drug binding kinetics and multichannel pharmacology. Circ. Arrhythm. Electrophysiol. 10, e004,628 (2017). https://doi.org/10.1161/CIRCEP.116.004628
Li, Z., Garnett, C., Strauss, D.G.: Quantitative systems pharmacology models for a new international cardiac safety regulatory paradigm: an overview of the comprehensive in vitro proarrhythmia assay in silico modeling approach. CPT: Pharmacometrics Syst. Pharmacol. 8(6), 371–379 (2019). https://doi.org/10.1002/psp4.12423
Li, Z., Mirams, G.R., Yoshinaga, T., Ridder, B.J., Han, X., Chen, J.E., Stockbridge, N.L., Wisialowski, T.A., Damiano, B., Severi, S., Morissette, P., Kowey, P.R., Holbrook, M., Smith, G., Rasmusson, R.L., Liu, M., Song, Z., Qu, Z., Leishman, D.J., Steidl-Nichols, J., Rodriguez, B., Bueno-Orovio, A., Zhou, X., Passini, E., Edwards, A.G., Morotti, S., Ni, H., Grandi, E., Clancy, C.E., Vandenberg, J., Hill, A., Nakamura, M., Singer, T., Polonchuk, L., Greiter-Wilke, A., Wang, K., Nave, S., Fullerton, A., Sobie, E.A., Paci, M., Musuamba Tshinanu, F., Strauss, D.G.: General principles for the validation of proarrhythmia risk prediction models: an extension of the cipa in silico strategy. Clin. Pharmacol. Ther. 107(1), 102–111 (2020). https://doi.org/10.1002/cpt.1647
Li, Z., Ridder, B.J., Han, X., Wu, W.W., Sheng, J., Tran, P.N., Wu, M., Randolph, A., Johnstone, R.H., Mirams, G.R., Kuryshev, Y., Kramer, J., Wu, C., Crumb, W.J. Jr., Strauss, D.G.: Assessment of an in silico mechanistic model for proarrhythmia risk prediction under the cipa initiative. Clin. Pharmacol. Ther. 105(2), 466–475 (2019). https://doi.org/10.1002/cpt.1184
Lloyd, C.M., Lawson, J.R., Hunter, P.J., Nielsen, P.F.: The cellml model repository. Bioinformatics 24(18), 2122–2123 (2008). https://doi.org/10.1093/bioinformatics/btn390
Lombardo, D.M., Rappel, W.J.: Systematic reduction of a detailed atrial myocyte model. Chaos 27(9), 093,914 (2017). https://doi.org/10.1063/1.4999611
Lombardo, D.M., Rappel, W.J.: Chaotic tip trajectories of a single spiral wave in the presence of heterogeneities. Phys. Rev. E 99, 062,409 (2019). https://doi.org/10.1103/PhysRevE.99.062409
Luechtefeld, T., Hartung, T.: Computational approaches to chemical hazard assessment. ALTEX 34(4), 459–478 (2017). https://doi.org/10.14573/altex.1710141
Luechtefeld, T., Marsh, D., Rowlands, C., Hartung, T.: Machine learning of toxicological big data enables Read-Across Structure Activity Relationships (RASAR) outperforming animal test reproducibility. Toxicol. Sci. 165(1), 198–212 (2018). https://doi.org/10.1093/toxsci/kfy152
Mann, S.A., Imtiaz, M., Winbo, A., Rydberg, A., Perry, M.D., Couderc, J.P., Polonsky, B., McNitt, S., Zareba, W., Hill, A.P., Vandenberg, J.I.: Convergence of models of human ventricular myocyte electrophysiology after global optimization to recapitulate clinical long QT phenotypes. J. Mol. Cell. Cardiol. 100, 25–34 (2016). https://doi.org/10.1016/j.yjmcc.2016.09.011
Margara, F., Whang, Z.J.W., Levrero-Florencio, F., Santiago, A., Vázquez, M., Bueno-Orovio, A., Rodriguez, B.: in-silico human electro-mechanical ventricular modelling and simulation for drug-induced pro-arrhythmia and inotropic risk assessment. Prog. Biophys. Mol. Biol. (2020, under review)
Mayr, A., Klambauer, G., Unterthiner, T., Hochreiter, S.: Deeptox: toxicity prediction using deep learning. Front. Environ. Sci. 3, 80 (2016). https://doi.org/10.3389/fenvs.2015.00080
McMillan, B., Gavaghan, D.J., Mirams, G.R.: Early afterdepolarisation tendency as a simulated pro-arrhythmic risk indicator. Toxicol. Res. 6, 912–921 (2017). https://doi.org/10.1039/C7TX00141J
Mena, A., Ferrero, J.M., Matas, J.F.R.: Gpu accelerated solver for nonlinear reaction–diffusion systems. Application to the electrophysiology problem. Comput. Phys. Commun. 196, 280–289 (2015). https://doi.org/10.1016/j.cpc.2015.06.018
Millard, D.C., Strock, C.J., Carlson, C.B., Aoyama, N., Juhasz, K., Goetze, T.A., Stoelzle-Feix, S., Becker, N., Fertig, N., January, C.T., Anson, B.D., Ross, J.D.: Identification of drug–drug interactions in vitro: a case study evaluating the effects of sofosbuvir and amiodarone on hiPSC-derived cardiomyocytes. Toxicol. Sci. 154(1), 174–182 (2016). https://doi.org/10.1093/toxsci/kfw153
Mincholé, A., Zacur, E., Ariga, R., Grau, V., Rodriguez, B.: MRI-based computational torso/biventricular multiscale models to investigate the impact of anatomical variability on the ECG QRS complex. Front. Physiol. 10, 1103 (2019). https://doi.org/10.3389/fphys.2019.01103
Mirams, G.R., Cui, Y., Sher, A., Fink, M., Cooper, J., Heath, B.M., McMahon, N.C., Gavaghan, D.J., Noble, D.: Simulation of multiple ion channel block provides improved early prediction of compounds’ clinical torsadogenic risk. Cardiovasc. Res. 91(1), 53–61 (2011). https://doi.org/10.1093/cvr/cvr044
Mirams, G.R., Davies, M.R., Cui, Y., Kohl, P., Noble, D.: Application of cardiac electrophysiology simulations to pro-arrhythmic safety testing. Br. J. Pharmacol. 167(5), 932–945 (2012). https://doi.org/10.1111/j.1476-5381.2012.02020.x
Mirams, G.R., Niederer, S.A., Clayton, R.H.: The fickle heart: uncertainty quantification in cardiac and cardiovascular modelling and simulation. Philos. Trans. R. Soc., Math. Phys. Eng. Sci. 378(2173), 20200, 119 (2020). https://doi.org/10.1098/rsta.2020.0119
Mistry, H.B.: Complex versus simple models: ion-channel cardiac toxicity prediction. PeerJ 6, e4352 (2018). https://doi.org/10.7717/peerj.4352
Mistry, H.B.: Comprehensive in vitro proarrhythmic assay complexity bias. Clin. Pharmacol. Ther. 105(6), 1323–1324 (2019). https://doi.org/10.1002/cpt.1400
Moreno, J.D., Lewis, T.J., Clancy, C.E.: Parameterization for in-silico modeling of ion channel interactions with drugs. PLoS ONE 11(3), 1–22 (2016). https://doi.org/10.1371/journal.pone.0150761
Moreno, J.D., Zhu, Z.I., Yang, P.C., Bankston, J.R., Jeng, M.T., Kang, C., Wang, L., Bayer, J.D., Christini, D.J., Trayanova, N.A., Ripplinger, C.M., Kass, R.S., Clancy, C.E.: A computational model to predict the effects of class i anti-arrhythmic drugs on ventricular rhythms. Sci. Transl. Med. 3(98), 98ra83–98ra83 (2011). https://doi.org/10.1126/scitranslmed.3002588
Morotti, S., Grandi, E.: Logistic regression analysis of populations of electrophysiological models to assess proarrythmic risk. MethodsX 4, 25–34 (2017). https://doi.org/10.1016/j.mex.2016.12.002
Mulroney, S., Myers, A.: Netter’s Essential Physiology. Student Consult. Elsevier, Amsterdam (2015). https://www.elsevier.com/books/netters-essential-physiology/mulroney/978-0-323-35819-4
Munawar, S., Windley, M.J., Tse, E.G., Todd, M.H., Hill, A.P., Vandenberg, J.I., Jabeen, I.: Experimentally validated pharmacoinformatics approach to predict herg inhibition potential of new chemical entities. Front. Pharmacol. 9, 1035 (2018). https://doi.org/10.3389/fphar.2018.01035
Murillo, M., Cai, X.C.: A fully implicit parallel algorithm for simulating the non-linear electrical activity of the heart. Numer. Linear Algebra Appl. 11(2–3), 261–277 (2004). https://doi.org/10.1002/nla.381
Nash, M.P., Panfilov, A.V.: Electromechanical model of excitable tissue to study reentrant cardiac arrhythmias. Prog. Biophys. Mol. Biol. 85(2), 501–522 (2004). https://doi.org/10.1016/j.pbiomolbio.2004.01.016. Modelling Cellular and Tissue Function
Neic, A., Liebmann, M., Hoetzl, E., Mitchell, L., Vigmond, E.J., Haase, G., Plank, G.: Accelerating cardiac bidomain simulations using graphics processing units. IEEE Trans. Biomed. Eng. 59(8), 2281–2290 (2012). https://doi.org/10.1109/TBME.2012.2202661
Nekouzadeh, A., Rudy, Y.: Conformational changes of an ion-channel during gating and emerging electrophysiologic properties: application of a computational approach to cardiac kv7.1. Prog. Biophys. Mol. Biol. 120(1), 18–27 (2016). https://doi.org/10.1016/j.pbiomolbio.2015.12.014. Recent Developments in Biophysics & Molecular Biology of Heart Rhythm
Nguyen, N., Nguyen, W., Nguyenton, B., Ratchada, P., Page, G., Miller, P.E., Ghetti, A., Abi-Gerges, N.: Adult human primary cardiomyocyte-based model for the simultaneous prediction of drug-induced inotropic and pro-arrhythmia risk. Front. Physiol. 8, 1073 (2017). https://doi.org/10.3389/fphys.2017.01073
Ni, H., Adeniran, I., Zhang, H.: In-silico investigations of the functional impact of kcna5 mutations on atrial mechanical dynamics. J. Mol. Cell. Cardiol. 111, 86–95 (2017). https://doi.org/10.1016/j.yjmcc.2017.08.005
Ni, H., Morotti, S., Grandi, E.: A heart for diversity: simulating variability in cardiac arrhythmia research. Front. Physiol. 9, 958 (2018). https://doi.org/10.3389/fphys.2018.00958
Niederer, S., Hunter, P., Smith, N.: A quantitative analysis of cardiac myocyte relaxation: a simulation study. Biophys. J. 90(5), 1697–1722 (2006). https://doi.org/10.1529/biophysj.105.069534
Niederer, S., Mitchell, L., Smith, N., Plank, G.: Simulating human cardiac electrophysiology on clinical time-scales. Front. Physiol. 2, 14 (2011). https://doi.org/10.3389/fphys.2011.00014
Niederer, S.A., Lumens, J., Trayanova, N.A.: Computational models in cardiology. Nat. Rev. Cardiol. 16(2), 100–111 (2019). https://doi.org/10.1038/s41569-018-0104-y
Niederer, S.A., Smith, N.P.: An improved numerical method for strong coupling of excitation and contraction models in the heart. Prog. Biophys. Mol. Biol. 96(1), 90–111 (2008). https://doi.org/10.1016/j.pbiomolbio.2007.08.001. Cardiovascular Physiome
Nogawa, H., Kawai, T.: herg trafficking inhibition in drug-induced lethal cardiac arrhythmia. Eur. J. Pharmacol. 741, 336–339 (2014). https://doi.org/10.1016/j.ejphar.2014.06.044
Obejero-Paz, C.A., Bruening-Wright, A., Kramer, J., Hawryluk, P., Tatalovic, M., Dittrich, H.C., Brown, A.M.: Quantitative profiling of the effects of vanoxerine on human cardiac ion channels and its application to cardiac risk. Sci. Rep. 5, 17,623 EP (2015). https://doi.org/10.1038/srep17623
O’Connor, M.D.: The 3R principle: advancing clinical application of human pluripotent stem cells. Stem Cell Res. Therapy 4(2), 21 (2013). https://doi.org/10.1186/scrt169
O’Hara, T., Virág, L., Varró, A., Rudy, Y.: Simulation of the undiseased human cardiac ventricular action potential: model formulation and experimental validation. PLoS Comput. Biol. 7(5), 1–29 (2011). https://doi.org/10.1371/journal.pcbi.1002061
Okada, J.i., Yoshinaga, T., Kurokawa, J., Washio, T., Furukawa, T., Sawada, K., Sugiura, S., Hisada, T.: Screening system for drug-induced arrhythmogenic risk combining a patch clamp and heart simulator. Sci. Adv. 1(4), e1400142 (2015). https://doi.org/10.1126/sciadv.1400142
Ortega, F.A., Grandi, E., Krogh-Madsen, T., Christini, D.J.: Applications of dynamic clamp to cardiac arrhythmia research: role in drug target discovery and safety pharmacology testing. Front. Physiol. 8, 1099 (2018). https://doi.org/10.3389/fphys.2017.01099
Paci, M., Hyttinen, J., Aalto-Setälä, K., Severi, S.: Computational models of ventricular- and atrial-like human induced pluripotent stem cell derived cardiomyocytes. Ann. Biomed. Eng. 41(11), 2334–2348 (2013). https://doi.org/10.1007/s10439-013-0833-3
Paci, M., Hyttinen, J., Rodriguez, B., Severi, S.: Human induced pluripotent stem cell-derived versus adult cardiomyocytes: an in silico electrophysiological study on effects of ionic current block. Br. J. Pharmacol. 172(21), 5147–5160 (2015). https://doi.org/10.1111/bph.13282
Paci, M., Passini, E., Klimas, A., Severi, S., Hyttinen, J., Rodriguez, B., Entcheva, E.: All-optical electrophysiology refines populations of in silico human iPS-CMs for drug evaluation. Biophys. J. 118(10), 2596–2611 (2020). https://doi.org/10.1016/j.bpj.2020.03.018
Paci, M., Passini, E., Severi, S., Hyttinen, J., Rodriguez, B.: A population of in silico models to face the variability of human induced pluripotent stem cell-derived cardiomyocytes: the herg block case study. In: 2016 Computing in Cardiology Conference (CinC), pp. 1189–1192 (2016)
Paci, M., Passini, E., Severi, S., Hyttinen, J., Rodriguez, B.: Phenotypic variability in LQT3 human induced pluripotent stem cell-derived cardiomyocytes and their response to antiarrhythmic pharmacologic therapy: an in silico approach. Heart Rhythm 14(11), 1704–1712 (2017). https://doi.org/10.1016/j.hrthm.2017.07.026
Paci, M., Pölönen, R.P., Cori, D., Penttinen, K., Aalto-Setälä, K., Severi, S., Hyttinen, J.: Automatic optimization of an in silico model of human ipsc derived cardiomyocytes recapitulating calcium handling abnormalities. Front. Physiol. 9, 709 (2018). https://doi.org/10.3389/fphys.2018.00709
Pagani, S., Manzoni, A., Quarteroni, A.: Numerical approximation of parametrized problems in cardiac electrophysiology by a local reduced basis method. Comput. Methods Appl. Mech. Eng. 340, 530–558 (2018). https://doi.org/10.1016/j.cma.2018.06.003
Pang, L., Sager, P., Yang, X., Shi, H., Sannajust, F., Brock, M., Wu, J.C., Abi-Gerges, N., Lyn-Cook, B., Berridge, B.R., Stockbridge, N.: Workshop report. Circ. Res. 125(9), 855–867 (2019). https://doi.org/10.1161/CIRCRESAHA.119.315378
Parikh, J., Di Achielle, P., Kozloski, J., Gurev, V.: Intrinsic structure of model-derived metrics for in silico proarrhytmic risk assessment identified by global sensitivity analysis. bioRxiv (2019). https://doi.org/10.1101/543926
Parikh, J., Gurev, V., Rice, J.J.: Novel two-step classifier for torsades de pointes risk stratification from direct features. Front. Pharmacol. 8, 816 (2017). https://doi.org/10.3389/fphar.2017.00816
Park, E., Willard, J., Bi, D., Fiszman, M., Kozeli, D., Koerner, J.: The impact of drug-related qt prolongation on fda regulatory decisions. Int. J. Cardiol. 168(5), 4975–4976 (2013). https://doi.org/10.1016/j.ijcard.2013.07.136
Passini, E., Britton, O.J., Lu, H.R., Rohrbacher, J., Hermans, A.N., Gallacher, D.J., Greig, R.J.H., Bueno-Orovio, A., Rodriguez, B.: Human in silico drug trials demonstrate higher accuracy than animal models in predicting clinical pro-arrhythmic cardiotoxicity. Front. Physiol. 8, 668 (2017). https://doi.org/10.3389/fphys.2017.00668
Passini, E., Trovato, C., Morissette, P., Sannajust, F., Bueno-Orovio, A., Rodriguez, B.: Drug-induced shortening of the electromechanical window is an effective biomarker for in silico prediction of clinical risk of arrhythmias. Br. J. Pharmacol. 176(19), 3819–3833 (2019). https://doi.org/10.1111/bph.14786
Patel, D., Stohlman, J., Dang, Q., Strauss, D.G., Blinova, K.: Assessment of proarrhythmic potential of drugs in optogenetically paced induced pluripotent stem cell-derived cardiomyocytes. Toxicol. Sci. 170(1), 167–179 (2019). https://doi.org/10.1093/toxsci/kfz076
Pathmanathan, P., Cordeiro, J.M., Gray, R.A.: Comprehensive uncertainty quantification and sensitivity analysis for cardiac action potential models. Front. Physiol. 10, 721 (2019). https://doi.org/10.3389/fphys.2019.00721
Pathmanathan, P., Gray, R.A.: Validation and trustworthiness of multiscale models of cardiac electrophysiology. Front. Physiol. 9, 106 (2018). https://doi.org/10.3389/fphys.2018.00106
Pollard, C., Abi Gerges, N., Bridgland-Taylor, M., Easter, A., Hammond, T., Valentin, J.P.: An introduction to qt interval prolongation and non-clinical approaches to assessing and reducing risk. Br. J. Pharmacol. 159(1), 12–21 (2010). https://doi.org/10.1111/j.1476-5381.2009.00207.x
Potse, M.: Scalable and accurate ecg simulation for reaction-diffusion models of the human heart. Front. Physiol. 9, 370 (2018). https://doi.org/10.3389/fphys.2018.00370
Potse, M., Dube, B., Richer, J., Vinet, A., Gulrajani, R.M.: A comparison of monodomain and bidomain reaction-diffusion models for action potential propagation in the human heart. IEEE Trans. Biomed. Eng. 53(12), 2425–2435 (2006). https://doi.org/10.1109/TBME.2006.880875
Potse, M., Dubé, B., Vinet, A.: Cardiac anisotropy in boundary-element models for the electrocardiogram. Med. Biol. Eng. Comput. 47(7), 719–729 (2009). https://doi.org/10.1007/s11517-009-0472-x
Qu, Y., Vargas, H.M.: Proarrhythmia risk assessment in human induced pluripotent stem cell-derived cardiomyocytes using the maestro MEA platform. Toxicol. Sci. 147(1), 286–295 (2015). https://doi.org/10.1093/toxsci/kfv128
Qu, Z.: Chaos in the genesis and maintenance of cardiac arrhythmias. Prog. Biophys. Mol. Biol. 105(3), 247–257 (2011). https://doi.org/10.1016/j.pbiomolbio.2010.11.001. Muscle Excitation-Contraction Coupling: Elements and Integration
Qu, Z., Garfinkel, A.: Chap. 36 - nonlinear dynamics of excitation and propagation in cardiac muscle. In: Zipes, D.P., Jalife, J. (eds.) Cardiac Electrophysiology, 4th edn., pp. 327–335. Saunders, Philadelphia (2004). https://doi.org/10.1016/B0-7216-0323-8/50039-7
Qu, Z., Garfinkel, A., Chen, P.S., Weiss, J.N.: Mechanisms of discordant alternans and induction of reentry in simulated cardiac tissue. Circulation 102(14), 1664–1670 (2000). https://doi.org/10.1161/01.CIR.102.14.1664
Qu, Z., Xie, L.H., Olcese, R., Karagueuzian, H.S., Chen, P.S., Garfinkel, A., Weiss, J.N.: Early afterdepolarizations in cardiac myocytes: beyond reduced repolarization reserve. Cardiovasc. Res. 99(1), 6–15 (2013). https://doi.org/10.1093/cvr/cvt104
Qu, Z., Xie, Y., Garfinkel, A., Weiss, J.: T-wave alternans and arrhythmogenesis in cardiac diseases. Front. Physiol. 1, 154 (2010). https://doi.org/10.3389/fphys.2010.00154
Quarteroni, A., Dede’, L., Manzoni, A., Vergara, C.: Mathematical Modelling of the Human Cardiovascular System: Data, Numerical Approximation, Clinical Applications. Cambridge Monographs on Applied and Computational Mathematics. Cambridge University Press, Cambridge (2019). https://doi.org/10.1017/9781108616096
Quarteroni, A., Lassila, T., Rossi, S., Ruiz-Baier, R.: Integrated heart—coupling multiscale and multiphysics models for the simulation of the cardiac function. Comput. Methods Appl. Mech. Eng. 314, 345–407 (2017). https://doi.org/10.1016/j.cma.2016.05.031. Special Issue on Biological Systems, Dedicated to William S. Klug
Raies, A.B., Bajic, V.B.: In silico toxicology: computational methods for the prediction of chemical toxicity. Wiley Interdiscip. Rev. Comput. Mol. Sci. 6(2), 147–172 (2016). https://doi.org/10.1002/wcms.1240
Raphel, F., Boulakia, M., Zemzemi, N., Coudière, Y., Guillon, J., Zitoun, P., Gerbeau, J.: Identification of ion currents components generating field potential recorded in MEA from hiPSC-CM. IEEE Trans. Biomed. Eng. 65(6), 1311–1319 (2018). https://doi.org/10.1109/TBME.2017.2748798
Raphel, F., De Korte, T., Lombardi, D., Braam, S., Gerbeau, J.F.: A greedy classifier optimisation strategy to assess ion channel blocking activity and pro-arrhythmia in hiPSC-cardiomyocytes (2019). https://hal.inria.fr/hal-02276945. Working paper or preprint
Redfern, W., Carlsson, L., Davis, A., Lynch, W., MacKenzie, I., Palethorpe, S., Siegl, P., Strang, I., Sullivan, A., Wallis, R., Camm, A., Hammond, T.: Relationships between preclinical cardiac electrophysiology, clinical qt interval prolongation and torsade de pointes for a broad range of drugs: evidence for a provisional safety margin in drug development. Cardiovasc. Res. 58(1), 32–45 (2003). https://doi.org/10.1016/S0008-6363(02)00846-5
Rice, J.J., Wang, F., Bers, D.M., de Tombe, P.P.: Approximate model of cooperative activation and crossbridge cycling in cardiac muscle using ordinary differential equations. Biophys. J. 95(5), 2368–2390 (2008). https://doi.org/10.1529/biophysj.107.119487
Richards, D.F., Glosli, J.N., Draeger, E.W., Mirin, A.A., Chan, B., luc Fattebert, J., Krauss, W.D., Oppelstrup, T., Butler, C.J., Gunnels, J.A., Gurev, V., Kim, C., Magerlein, J., Reumann, M., Wen, H.F., Rice, J.J.: Towards real-time simulation of cardiac electrophysiology in a human heart at high resolution. Comput. Methods Biomech. Biomed. Eng. 16(7), 802–805 (2013). https://doi.org/10.1080/10255842.2013.795556. PMID: 23734785
Robertson, C., Tran, D.D., George, S.C.: Concise review: maturation phases of human pluripotent stem cell-derived cardiomyocytes. Stem Cells 31(5), 829–837 (2013). https://doi.org/10.1002/stem.1331
Roden, D.M.: Repolarization reserve. Circulation 118(10), 981–982 (2008). https://doi.org/10.1161/CIRCULATIONAHA.108.798918
Roden, D.M., Viswanathan, P.C.: Genetics of acquired long qt syndrome. J. Clin. Invest. 115(8), 2025–2032 (2005). https://doi.org/10.1172/JCI25539
Rodriguez, B., Burrage, K., Gavaghan, D., Grau, V., Kohl, P., Noble, D.: The systems biology approach to drug development: application to toxicity assessment of cardiac drugs. Clin. Pharmacol. Ther. 88(1), 130–134 (2010). https://doi.org/10.1038/clpt.2010.95
Sager, P.T., Gintant, G., Turner, J.R., Pettit, S., Stockbridge, N.: Rechanneling the cardiac proarrhythmia safety paradigm: a meeting report from the cardiac safety research consortium. Am. Heart J. 167(3), 292–300 (2014). https://doi.org/10.1016/j.ahj.2013.11.004
Sahli Costabal, F., Seo, K., Ashley, E., Kuhl, E.: Classifying drugs by their arrhythmogenic risk using machine learning. BioRxiv (2019). https://doi.org/10.1101/545863
Sala, L., van Meer, B.J., Tertoolen, L.G., Bakkers, J., Bellin, M., Davis, R.P., Denning, C., Dieben, M.A., Eschenhagen, T., Giacomelli, E., Grandela, C., Hansen, A., Holman, E.R., Jongbloed, M.R., Kamel, S.M., Koopman, C.D., Lachaud, Q., Mannhardt, I., Mol, M.P., Mosqueira, D., Orlova, V.V., Passier, R., Ribeiro, M.C., Saleem, U., Smith, G.L., Burton, F.L., Mummery, C.L.: Musclemotion. Circ. Res. 122(3), e5–e16 (2018). https://doi.org/10.1161/CIRCRESAHA.117.312067
Sanguinetti, M.C., Jiang, C., Curran, M.E., Keating, M.T.: A mechanistic link between an inherited and an acquird cardiac arrthytmia: HERG encodes the IKr potassium channel. Cell 81(2), 299–307 (1995). https://doi.org/10.1016/0092-8674(95)90340-2
Sarkar, A.X., Sobie, E.A.: Regression analysis for constraining free parameters in electrophysiological models of cardiac cells. PLoS Comput. Biol. 6(9), 1–11 (2010). https://doi.org/10.1371/journal.pcbi.1000914
Sato, D., Xie, L.H., Nguyen, T.P., Weiss, J.N., Qu, Z.: Irregularly appearing early afterdepolarizations in cardiac myocytes: random fluctuations or dynamical chaos? Biophys. J. 99(3), 765–773 (2010). https://doi.org/10.1016/j.bpj.2010.05.019
Sato, D., Xie, L.H., Sovari, A.A., Tran, D.X., Morita, N., Xie, F., Karagueuzian, H., Garfinkel, A., Weiss, J.N., Qu, Z.: Synchronization of chaotic early afterdepolarizations in the genesis of cardiac arrhythmias. Proc. Natl. Acad. Sci. 106(9), 2983–2988 (2009). https://doi.org/10.1073/pnas.0809148106
Saxena, P., Hortigon-Vinagre, M.P., Beyl, S., Baburin, I., Andranovits, S., Iqbal, S.M., Costa, A., IJzerman, A.P., Kügler, P., Timin, E., Smith, G.L., Hering, S.: Correlation between human ether-a-go-go-related gene channel inhibition and action potential prolongation. Br. J. Pharmacol. 174(18), 3081–3093 (2017). https://doi.org/10.1111/bph.13942
Servick, K.: A painstaking overhaul for cardiac safety testing. Science 353(6303), 976–977 (2016). https://doi.org/10.1126/science.353.6303.976
Shaheen, N., Shiti, A., Huber, I., Shinnawi, R., Arbel, G., Gepstein, A., Setter, N., Goldfracht, I., Gruber, A., Chorna, S.V., Gepstein, L.: Human induced pluripotent stem cell-derived cardiac cell sheets expressing genetically encoded voltage indicator for pharmacological and arrhythmia studies. Stem Cell Rep. 10(6), 1879–1894 (2018). https://doi.org/10.1016/j.stemcr.2018.04.006
Sinha, S., Sridhar, S.: Patterns in Excitable Media - Genesis, Dynamics and Contrl. CRC Press/Taylor & Francis, Boca Raton/London (2015)
Skibsbye, L., Jespersen, T., Christ, T., Maleckar, M.M., van den Brink, J., Tavi, P., Koivumäki, J.T.: Refractoriness in human atria: time and voltage dependence of sodium channel availability. J. Mol. Cell. Cardiol. 101, 26–34 (2016). https://doi.org/10.1016/j.yjmcc.2016.10.009
Slepukhina, E., Ryashko, L., Kügler, P.: Noise-induced early afterdepolarizations in a three-dimensional cardiac action potential model. Chaos Solitons Fractals 131, 109515 (2019). https://doi.org/10.1016/j.chaos.2019.109515
Stewart, P., Aslanidi, O.V., Noble, D., Noble, P.J., Boyett, M.R., Zhang, H.: Mathematical models of the electrical action potential of Purkinje fibre cells. Philos. Trans. R. Soc., Math. Phys. Eng. Sci. 367(1896), 2225–2255 (2009). https://doi.org/10.1098/rsta.2008.0283
Stockbridge, N., Morganroth, J., Shah, R.R., Garnett, C.: Dealing with global safety issues. Drug Safety 36(3), 167–182 (2013). https://doi.org/10.1007/s40264-013-0016-z
Stoelzle-Feix, S., Juhasz, K., Ulrich, T., Linder, P., George, M., Gossmann, M.: Higher throughput in vitro cardiac contractility recordings under physiological mechanical conditions (2019). https://www.nanion.de/en/products/flexcyte-white-paper-download.html. Nanion Technologies Whitepaper
Sugiura, S., Washio, T., Hatano, A., Okada, J., Watanabe, H., Hisada, T.: Multi-scale simulations of cardiac electrophysiology and mechanics using the university of Tokyo heart simulator. Prog. Biophys. Mol. Biol. 110(2), 380–389 (2012). https://doi.org/10.1016/j.pbiomolbio.2012.07.001. SI: Beating Heart
Sundnes, J., Lines, G., Cai, X., Nielsen, B., Mardal, K., Tveito, A.: Computing the Electrical Activity in the Heart. Monographs in Computational Science and Engineering. Springer, Berlin (2007). https://books.google.de/books?id=MCfh_FnHpgIC
Surovyatkina, E.: Bifurcations and Multistability in Periodically Stimulated Cardiac Cells pp. 253–284. Springer, Vienna (2012)
Surovyatkina, E., Noble, D., Gavaghan, D., Sher, A.: Multistability property in cardiac ionic models of mammalian and human ventricular cells. Prog. Biophys. Mol. Biol. 103(1), 131–141 (2010). https://doi.org/10.1016/j.pbiomolbio.2010.01.004. Ion channels: Growing evidence for the role in disease
Sutanto, H., Laudy, L., Clerx, M., Dobrev, D., Crijns, H.J., Heijman, J.: Maastricht antiarrhythmic drug evaluator (manta): a computational tool for better understanding of antiarrhythmic drugs. Pharmacol. Res. 148, 104,444 (2019). https://doi.org/10.1016/j.phrs.2019.104444
Tertoolen, L., Braam, S., van Meer, B., Passier, R., Mummery, C.: Interpretation of field potentials measured on a multi electrode array in pharmacological toxicity screening on primary and human pluripotent stem cell-derived cardiomyocytes. Biochem. Biophys. Res. Commun. 497(4), 1135–1141 (2018). https://doi.org/10.1016/j.bbrc.2017.01.151
Tiburcy, M., Hudson, J.E., Balfanz, P., Schlick, S., Meyer, T., Liao, M.L.C., Levent, E., Raad, F., Zeidler, S., Wingender, E., Riegler, J., Wang, M., Gold, J.D., Kehat, I., Wettwer, E., Ravens, U., Dierickx, P., van Laake, L.W., Goumans, M.J., Khadjeh, S., Toischer, K., Hasenfuss, G., Couture, L.A., Unger, A., Linke, W.A., Araki, T., Neel, B., Keller, G., Gepstein, L., Wu, J.C., Zimmermann, W.H.: Defined engineered human myocardium with advanced maturation for applications in heart failure modeling and repair. Circulation 135(19), 1832–1847 (2017). https://doi.org/10.1161/CIRCULATIONAHA.116.024145
Tixier, E., Raphel, F., Lombardi, D., Gerbeau, J.F.: Composite biomarkers derived from micro-electrode array measurements and computer simulations improve the classification of drug-induced channel block. Front. Physiol. 8, 1096 (2018). https://doi.org/10.3389/fphys.2017.01096
Tixier, E., Raphel, F., Lombardi, D., Gerbeau, J.F.: Composite biomarkers derived from micro-electrode array measurements and computer simulations improve the classification of drug-induced channel block. Front. Physiol. 8, 1096 (2018). https://doi.org/10.3389/fphys.2017.01096
Tomek, J., Bueno-Orovio, A., Passini, E., Zhou, X., Minchole, A., Britton, O., Bartolucci, C., Severi, S., Shrier, A., Virag, L., Varro, A., Rodriguez, B.: Development, calibration, and validation of a novel human ventricular myocyte model in health, disease, and drug block. eLife 8, e48,890 (2019). https://doi.org/10.7554/eLife.48890
Tomek, J., Bueno-Orovio, A., Rodriguez, B.: ToR-ORd-dynCl: an update of the ToR-ORd model of human ventricular cardiomyocyte with dynamic intracellular chloride. bioRxiv (2020). https://doi.org/10.1101/2020.06.01.127043
Tran, D.X., Sato, D., Yochelis, A., Weiss, J.N., Garfinkel, A., Qu, Z.: Bifurcation and chaos in a model of cardiac early afterdepolarizations. Phys. Rev. Lett. 102, 258,103 (2009). https://doi.org/10.1103/PhysRevLett.102.258103
Trenor, B., Gomis-Tena, J., Cardona, K., Romero, L., Rajamani, S., Belardinelli, L., Giles, W.R., Saiz, J.: In silico assessment of drug safety in human heart applied to late sodium current blockers. Channels 7(4), 249–262 (2013). https://doi.org/10.4161/chan.24905. PMID: 23696033
Trovato, C., Passini, E., Nagy, N., Varró, A., Abi-Gerges, N., Severi, S., Rodriguez, B.: Human Purkinje in silico model enables mechanistic investigations into automaticity and pro-arrhythmic abnormalities. J. Mol. Cell. Cardiol. 142, 24–38 (2020). https://doi.org/10.1016/j.yjmcc.2020.04.001
Tse, G.: Mechanisms of cardiac arrhythmias. J. Arrhythmia 32(2), 75–81 (2016). https://doi.org/10.1016/j.joa.2015.11.003
Tsumoto, K., Kurata, Y., Furutani, K., Kurachi, Y.: Hysteretic dynamics of multi-stable early afterdepolarisations with repolarisation reserve attenuation: a potential dynamical mechanism for cardiac arrhythmias. Sci. Rep. 7(1), 10,771 (2017). https://doi.org/10.1038/s41598-017-11355-1
Tung, L.: a bi-domain model for describing ischemic myocardial d-C potentials. Ph.D. thesis, Massachusetts Institute of Technology, Cambridge, Mass, USA (1978)
ten Tusscher, K.H.W.J., Panfilov, A.V.: Alternans and spiral breakup in a human ventricular tissue model. Am. J. Physiol., Heart Circ. Physiol. 291(3), H1088–H1100 (2006). https://doi.org/10.1152/ajpheart.00109.2006. PMID: 16565318
Tveito, A., Jæger, K.H., Huebsch, N., Charrez, B., Edwards, A.G., Wall, S., Healy, K.E.: Inversion and computational maturation of drug response using human stem cell derived cardiomyocytes in microphysiological systems. Sci. Rep. 8(1), 17,626 (2018). https://doi.org/10.1038/s41598-018-35858-7
Tylutki, Z., Polak, S., Wiśniowska, B.: Top-down, bottom-up and middle-out strategies for drug cardiac safety assessment via modeling and simulations. Curr. Pharmacol. Rep. 2(4), 171–177 (2016). https://doi.org/10.1007/s40495-016-0060-3
Tylutki, Z., Szlek, J., Polak, S.: CardiacPBPK: a tool for the prediction and visualization of time-concentration profiles of drugs in heart tissue. Comput. Biol. Med. 115, 103,484 (2019). https://doi.org/10.1016/j.compbiomed.2019.103484
Uzelac, I., Ji, Y.C., Hornung, D., Schröder-Scheteling, J., Luther, S., Gray, R.A., Cherry, E.M., Fenton, F.H.: Simultaneous quantification of spatially discordant alternans in voltage and intracellular calcium in langendorff-perfused rabbit hearts and inconsistencies with models of cardiac action potentials and ca transients. Front. Physiol. 8, 819 (2017). https://doi.org/10.3389/fphys.2017.00819
Vagos, M., van Herck, I.G.M., Sundnes, J., Arevalo, H.J., Edwards, A.G., Koivumäki, J.T.: Computational modeling of electrophysiology and pharmacotherapy of atrial fibrillation: recent advances and future challenges. Front. Physiol. 9, 1221 (2018). https://doi.org/10.3389/fphys.2018.01221
Vandenberg, J.I., Perry, M.D., Perrin, M.J., Mann, S.A., Ke, Y., Hill, A.P.: herg k+ channels: structure, function, and clinical significance. Physiol. Rev. 92(3), 1393–1478 (2012). https://doi.org/10.1152/physrev.00036.2011. PMID: 22988594
Vandersickel, N., de Boer, T.P., Vos, M.A., Panfilov, A.V.: Perpetuation of Torsade de Pointes in heterogeneous hearts: competing foci or re-entry? J. Physiol. 594(23), 6865–6878 (2016). https://doi.org/10.1113/JP271728
Vandersickel, N., Kazbanov, I.V., Nuitermans, A., Weise, L.D., Pandit, R., Panfilov, A.V.: A study of early afterdepolarizations in a model for human ventricular tissue. PLoS ONE 9(1), 1–19 (2014). https://doi.org/10.1371/journal.pone.0084595
Vandersickel, N., Van Nieuwenhuyse, E., Seemann, G., Panfilov, A.V.: Spatial patterns of excitation at tissue and whole organ level due to early afterdepolarizations. Front. Physiol. 8, 404 (2017). https://doi.org/10.3389/fphys.2017.00404
Verkerk, A.O., Veerman, C.C., Zegers, J.G., Mengarelli, I., Bezzina, C.R., Wilders, R.: Patch-clamp recording from human induced pluripotent stem cell-derived cardiomyocytes: improving action potential characteristics through dynamic clamp. Int. J. Mol. Sci. 18(9), 1873 (2017). https://doi.org/10.3390/ijms18091873
Vicente, J., Zusterzeel, R., Johannesen, L., Mason, J., Sager, P., Patel, V., Matta, M.K., Li, Z., Liu, J., Garnett, C., Stockbridge, N., Zineh, I., Strauss, D.G.: Mechanistic model-informed proarrhythmic risk assessment of drugs: review of the “CiPA” initiative and design of a prospective clinical validation study. Clin. Pharmacol. Ther. 103(1), 54–66 (2018). https://doi.org/10.1002/cpt.896
Viskin, S.: Long QT syndromes and Torsade de Pointes. Lancet 354(9190), 1625–1633 (1999). https://doi.org/10.1016/S0140-6736(99)02107-8
Vo, T., Bertram, R.: Why pacing frequency affects the production of early afterdepolarizations in cardiomyocytes: an explanation revealed by slow-fast analysis of a minimal model. Phys. Rev. E 99, 052,205 (2019). https://doi.org/10.1103/PhysRevE.99.052205
Volders, P.G., Vos, M.A., Szabo, B., Sipido, K.R., de Groot, S., Gorgels, A.P., Wellens, H.J., Lazzara, R.: Progress in the understanding of cardiac early afterdepolarizations and Torsades de Pointes: time to revise current concepts. Cardiovasc. Res. 46(3), 376–392 (2000). https://doi.org/10.1016/S0008-6363(00)00022-5
Wang, W., Zhang, S., Ni, H., Garratt, C.J., Boyett, M.R., Hancox, J.C., Zhang, H.: Mechanistic insight into spontaneous transition from cellular alternans to arrhythmia—a simulation study. PLoS Comput. Biol. 14(11), 1–27 (2018). https://doi.org/10.1371/journal.pcbi.1006594
Weiss, J.N., Garfinkel, A., Karagueuzian, H.S., Chen, P.S., Qu, Z.: Early afterdepolarizations and cardiac arrhythmias. Heart Rhythm 7(12), 1891–1899 (2010). https://doi.org/10.1016/j.hrthm.2010.09.017
Whittaker, D.G., Clerx, M., Lei, C.L., Christini, D.J., Mirams, G.R.: Calibration of ionic and cellular cardiac electrophysiology models. WIREs Syst. Biol. Med. 12(4), e1482 (2020). https://doi.org/10.1002/wsbm.1482
Wilhelms, M., Rombach, C., Scholz, E.P., Dössel, O., Seemann, G.: Impact of amiodarone and cisapride on simulated human ventricular electrophysiology and electrocardiograms. Europace 14(suppl5), v90–v96 (2012). https://doi.org/10.1093/europace/eus281
Wiśniowska, B., Mendyk, A., Fijorek, K., Polak, S.: Computer-based prediction of the drug proarrhythmic effect: problems, issues, known and suspected challenges. Europace 16(5), 724–735 (2014). https://doi.org/10.1093/europace/euu009
Wiśniowska, B., Polak, S.: Am I or am I not proarrhythmic? Comparison of various classifications of drug TdP propensity. Drug Discov. Today 22(1), 10–16 (2017). https://doi.org/10.1016/j.drudis.2016.09.027
Wiśniowska, B., Tylutki, Z., Polak, S.: Humans vary, so cardiac models should account for that too! Front. Physiol. 8, 700 (2017). https://doi.org/10.3389/fphys.2017.00700
Xia, Y., Wang, K., Zhang, H.: Parallel optimization of 3D cardiac electrophysiological model using GPU. Comput. Math. Methods Med. 2015, 862735 (2015). https://doi.org/10.1155/2015/862735
Xie, Y., Izu, L.T., Bers, D.M., Sato, D.: Arrhythmogenic transient dynamics in cardiac myocytes. Biophys. J. 106(6), 1391–1397 (2014). https://doi.org/10.1016/j.bpj.2013.12.050
Xie, Y., Liao, Z., Grandi, E., Shiferaw, Y., Bers, D.M.: Slow \([{N}a]_{i}\) changes and positive feedback between membrane potential and \([{C}a]_{i}\) underlie intermittent early afterdepolarizations and arrhythmias. Circ. Arrhythm. Electrophysiol. 8(6), 1472–1480 (2015). https://doi.org/10.1161/CIRCEP.115.003085
Yang, X., Papoian, T.: Moving beyond the comprehensive in vitro proarrhythmia assay: use of human-induced pluripotent stem cell-derived cardiomyocytes to assess contractile effects associated with drug-induced structural cardiotoxicity. J. Appl. Toxicol. 38(9), 1166–1176 (2018). https://doi.org/10.1002/jat.3611
Yaniv, Y., Lakatta, E.G., Maltsev, V.A.: From two competing oscillators to one coupled-clock pacemaker cell system. Front. Physiol. 6, 28 (2015). https://doi.org/10.3389/fphys.2015.00028
Yap, Y.G., Camm, A.J.: Drug induced QT prolongation and Torsades de Pointes. Heart 89(11), 1363–1372 (2003). https://doi.org/10.1136/heart.89.11.1363
Ying, W., Rose, D.J., Henriquez, C.S.: Efficient fully implicit time integration methods for modeling cardiac dynamics. IEEE Trans. Biomed. Eng. 55(12), 2701–2711 (2008). https://doi.org/10.1109/TBME.2008.925673
Yuan, Y., Bai, X., Luo, C., Wang, K., Zhang, H.: The virtual heart as a platform for screening drug cardiotoxicity. Br. J. Pharmacol. 172(23), 5531–5547 (2015). https://doi.org/10.1111/bph.12996
Zemzemi, N., Bernabeu, M.O., Saiz, J., Cooper, J., Pathmanathan, P., Mirams, G.R., Pitt-Francis, J., Rodriguez, B.: Computational assessment of drug-induced effects on the electrocardiogram: from ion channel to body surface potentials. Br. J. Pharmacol. 168(3), 718–733 (2013). https://doi.org/10.1111/j.1476-5381.2012.02200.x
Zemzemi, N., Rodriguez, B.: Effects of L-type calcium channel and human ether-a-go-go related gene blockers on the electrical activity of the human heart: a simulation study. Europace 17(2), 326–333 (2015). https://doi.org/10.1093/europace/euu122
Zhou, X., Qu, Y., Passini, E., Bueno-Orovio, A., Liu, Y., Vargas, H.M., Rodriguez, B.: Blinded in silico drug trial reveals the minimum set of ion channels for torsades de pointes risk assessment. Front. Pharmacol. 10, 1643 (2020). https://doi.org/10.3389/fphar.2019.01643
Zile, M.A., Trayanova, N.A.: Myofilament protein dynamics modulate ead formation in human hypertrophic cardiomyopathy. Prog. Biophys. Mol. Biol. 130, 418–428 (2017). https://doi.org/10.1016/j.pbiomolbio.2017.06.015. Cardiac Mechanics and Electrics: it takes two to tango
Zile, M.A., Trayanova, N.A.: Increased thin filament activation enhances alternans in human chronic atrial fibrillation. Am. J. Physiol., Heart Circ. Physiol. 315(5), H1453–H1462 (2018). https://doi.org/10.1152/ajpheart.00658.2017. PMID: 30141984
Zimik, S., Nayak, A.R., Pandit, R.: A computational study of the factors influencing the PVC-triggering ability of a cluster of early afterdepolarization-capable myocytes. PLoS ONE 10(12), 1–24 (2015). https://doi.org/10.1371/journal.pone.0144979
Zimik, S., Vandersickel, N., Nayak, A.R., Panfilov, A.V., Pandit, R.: A comparative study of early afterdepolarization-mediated fibrillation in two mathematical models for human ventricular cells. PLoS ONE 10(6), 1–20 (2015). https://doi.org/10.1371/journal.pone.0130632
Acknowledgements
Open Access funding provided by Projekt DEAL. I would like to acknowledge research membership at The Living Heart Project run by Dassault Systèmes and to thank the Safety Pharmacology Society for repeatedly making me part of the program committee for its annual meeting.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
Foundations of Cardiac AP Modelling
A prominent example of a phenomenological description of excitability is the FitzHugh-Nagumo model [63]. If adapted and reparametrized for cardiac AP simulation, the model reads as
where \(V\) represents the transmembrane voltage and \(W\) is referred to as the recovery variable. Furthermore, \(v_{\text{rest}}\), \(v_{\text{th}}\) and \(v_{\text{peak}}\) denote the resting potential, the threshold potential and the peak potential with \(v_{\text{amp}} = v_{\text{peak}} - v_{\text{th}}\), see Table 2 for parameter values.
Mechanistic AP models are based on viewing the membrane as an electric circuit with capacitive and conductive branches in parallel, see Fig. 23 for a schematic representation of the Noble model [45], which marks the beginning of cardiac AP modelling. The model describes the Purkinje fibre action and pacemaker potentials and reads as
with the ohmic current formulations
and \(V_{Na}=40\) mV, \(V_{K} = -100\) mV, \(V_{L} = -60\) mV, \(C_{m} = 12\) \(\mu \text{F}/\text{cm}^{2}\), \(g_{Na} = 400\) \(\text{mS}/\text{cm}^{2}\), \(g_{K_{2}} = 1.2\) \(\text{mS}/\text{cm}^{2}\), \(g_{L} = 0.075\) \(\text{mS}/\text{cm}^{2}\). Here, \(m\) is the sodium activation variable, \(h\) is the sodium inactivation variable and \(n\) is the potassium activation variable, and the voltage dependent rate functions are given by
in units of \(1/\text{ms}\). One interpretation [103] of these equations is that, e.g., the sodium channel consists of three gates of type \(m\) and one gate of type \(h\), each of which can be either closed or open. Under the assumption that the gates operate independently, the fraction of open sodium channels is given by \(m^{3}h\). Then, \(\alpha _{m}(V)\) and \(\alpha _{h}(V)\) denote the transition rates from the closed to the open state of the respective gates, while \(\beta _{m}(V)\) and \(\beta _{h}(V)\) are associated with the transitions from open to closed. The equations for the gating variables \(m\), \(h\), \(n\), which can be interpreted as open propabilities of subunits that constitute the channel, are often rewritten in the form
with the steady state function \(x_{\infty }(V)\) and the time constant function \(\tau _{x}(V)\) given by
for \(x \in \left \{ n,m,h \right \} \).
The Noble model is a modification of the Hodgkin-Huxley (HH) equations [85] for the description of excitation in neural cells. Today, the HH formalism is present in nearly all of the modern mechanistic AP models and used for the description of most of the cardiac ionic currents, either in combination with the linear current-voltage relation (3) or the GHK equation (5). The HH description can be generalized to Markov models [62] that represent the channel by a discrete and finite set of open and closed states, which allows to also take relationships between gating processes into account. A Markov model is not a unique description of the ion channel function and different model structures that fit the same experimental data can have different interpretations with respect to underlying mechanisms [24]. Furthermore, Markov states are abstract objects that do not necessarily represent structural conformations of the ion channel protein [166]. Still, such phenomenological Markov models are the current method of choice for describing kinetic drug-channel interaction [50], [158], [129], [137] after a model extension by drug bound states.
In addition to detailed descriptions of a number of transmembrane currents, state of the art cardiac AP models (1) also feature ion pumps and exchangers. Furthermore, calcium fluxes into and out of the sacroplasmatic reticulum (SR) as well as calcium buffering are taken into account, which makes, e.g., the SR calcium concentration \([Ca]_{SR}\) and the myoplasm calcium concentration \([Ca]_{i}\) part of the ODE state vector in (1). Figure 24 illustrates the ORd model of the human ventricular CM and the Paci model of ventricular-like hiPSC-CMs and indicates the complexity of cardiac AP modelling.
Examples of Risk Classification Based on Experimental Features
The ICH S7B guidelines [2] for preclinical cardiac safety testing consider a compound to be torsadogenic if its hERG IC50 value, see Fig. 25 (right), and its effective free therapeutic plasma concentration (EFTPC) satisfy
Figure 25 (right) illustrates that this criterion is too conservative and not specific, as not only the CiPA high TdP risk training compounds would accordingly be rated as torsadogenic, but also most of those with only medium or even low TdP risk. For instance, verapamil is a potent hERG blocker but still considered safe as it is in therapeutic use without any signs of torsadogenicity. One explanation for the high false-positive rate of the hERG assay is that negative hERG blocking effects on cardiac repolarization may be compensated by additional blockage of other channels such as the sodium NaV1.5 or the calcium CaV1.2 channel. In [109], several logistic regression models with \(\text{IC50}_{\text{hERG}}\), \(\text{EFTPC}\), \(\text{IC50}_{\text{CaV1.2}}\) and \(\text{IC50}_{\text{NaV1.5}}\) as direct features were trained based on data for 32 torsadogenic (TdP = 1) and 23 non-torsadogenic (TdP = 0) drugs. The multichannel effect (MICE) model, that was found to most significantly reduce false positives and false negatives in comparison with the hERG test, is given by
with \(\beta _{0} = -0.75\) and \(\beta _{1} = 3.04\). Note that it only involves the ratio of \(\text{IC50}_{\text{hERG}}\) and \(\text{IC50}_{\text{CaV1.2}}\). However, Fig. 26 (left) indicates that the model still lacks specificity in particular if applied to compounds not used for model development. In [225], several derivatives of the torsadogenic hERG blocker dofetilide were tested on hiPSC-CMs. While Dofe54 induced EADs, Dofe42 and Dofe44 had hardly any impact on AP morphology even at high dosage. Still, the MICE model would rate all of them as torsadogenic.
Within the CiPA initiative, logistic regression models for proarrhythmic risk prediction have been tested [16], with up to seven direct features describing electrophysiological responses of spontaneously beating or optogenetically paced [193] hiPSC-CMs to the 28 drugs of Table 1. One model finally chosen reads as
with \(\beta _{0} = -0.1311\), \(\beta _{1} = 0.6583\) (in case of EADs of type A, see Fig. 7 (left)), \(\beta _{2} = 0.00687\) and \(\beta _{3} = 0.0232\), where \(P\) is the probability of a drug to be of high or intermediate TdP risk. Feature \(\text{F}_{1}\) is categorical and indicates whether the compound induced arrhythmia-like events at any concentration across all replicates, feature \(\text{F}_{2}\) is the maximum compound induced repolarization change observed at any concentration, and feature \(\text{F}_{3}\) denotes the drug induced repolarization change at the clinical maximum concentration, with both \(\text{F}_{2}\) and \(\text{F}_{3}\) averaged across all replicates. A threshold of low TdP risk versus high or intermediate TdP risk was set at 0.8, see Fig. 26 (right), such that all of the drugs in the low-risk category of Table 1 correctly fell below it. However, also one drug from the high TdP risk category and four drugs from the intermediate TdP risk category fell below the threshold.
Examples of misclassification, as those shown in Fig. 25 and in Fig. 26, motivate the use of mechanistic models of cardiac electrophysiology in order to build more accurate risk classifiers that (also) involve features derived from model simulations, see Sect. 5.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Kügler, P. Modelling and Simulation for Preclinical Cardiac Safety Assessment of Drugs with Human iPSC-Derived Cardiomyocytes. Jahresber. Dtsch. Math. Ver. 122, 209–257 (2020). https://doi.org/10.1365/s13291-020-00218-w
Published:
Issue Date:
DOI: https://doi.org/10.1365/s13291-020-00218-w