COMPUTATIONAL ANALYSIS OF THE NOVEL LQT3 MUTATIONS G1481V AND Q1491H IN MYOCARDIAL AND PURKINJE CELLS

Abstract

Long QT Syndrome type 3 (LQT3) arises from mutations in SCNA5 gene encoding the Nav1.5 sodium channel. We studied two novel mutations (G1481V and Q1491H) found in infants who suffered sudden cardiac death. Using computational modeling, we incorporated experimental data into a Na⁺ Markovian model and simulated effects in rabbit cardiac cells. Both mutations caused tissue-specific effects, with Purkinje cells showing more severe action potential prolongation than myocardial cells. At slow heart rates, these mutations triggered early afterdepolarizations (EADs) and delayed afterdepolarizations (DADs) exclusively in Purkinje cells, with Q1491H displaying greater arrhythmogenic potential which was absent in myocardial cells. Enhanced late sodium current underlies these effects, especially in Purkinje cells with fewer repolarization reserves. Our findings suggest targeted therapies: late Na⁺ current blockers or Purkinje fiber ablation for patients with these mutations.

1. INTRODUCTION

In the heart (Figure 1, Panel A), ion channels and transporters traverse cells’ membrane, facilitating the transport of ions in and out of a cell. The ions influx and efflux across the membrane occur under and create an electrochemical gradient which forms the basis of an action potential (AP) (Figure 1, Panel B) that propagates through the heart and triggers its contraction. The time course of the various APs in the different compartments of the heart aggregate to generate an electrocardiogram (ECG). A derangement of an AP within a heart compartment can cause arrhythmias with faster or slower than normal heartbeat rate, or irregular spatiotemporal patterns of AP propagation.

Figure 1:

The human heart (A) and the action potentials in different heart regions combine to produce an ECG signal (B). The figure was reproduced from Bartos et al. [6].

Long QT Syndrome (LQTS) is an electrophysiological disorder, congenital [1] or drug-induced [2], manifesting on an ECG as the prolongation of the QT interval. This repolarization abnormality potentiates the risk of developing lethal arrhythmias and sudden cardiac death. There are three common subtypes of LQTS: LQT1, LQT2, and LQT3.

Among the three subtypes of LQTS, LQT3 has the lowest prevalence [3]. LQT1 and LQT2 result from the complete or partial loss of the slowly delayed and rapidly delayed rectifier potassium current, respectively. In contrast to LQT1 and LQT2, which result from the loss of K⁺ current, LQT3 emanates from a mutation in the SCNA5 gene (the gene encoding the Nav1.5 Na⁺ channel), leading to Na⁺ current gain. The SCNA5 gene mutation allows Na⁺ current to conduct at depolarized potentials [4]. LQT3 has been reported as one of the most lethal forms of LQTS [5] yet least responsive to beta-blockers (common therapeutic intervention for LQTS patients); therefore an understanding of LQT3-mediated arrhythmias is essential for proper treatment of patients.

Purkinje cells, the primary electrical conduction cells of the heart, form an intricated network of fibers that mediate the AP signals to the surrounding myocardium via Purkinje-myocardial junctions [7]. The myocardium is a heterogeneous mixture of three distinct cell types: epicardial (Epi), midmyocardial (mid-Myo) and endocardial (Endo), which form the inner, middle and outer layers of the ventricular wall, respectively [8]. Purkinje cells have a few repolarization reserves [2] compared to myocardial cells; hence, they most likely will present a severe phenotypic malfunction in the event of prolonged depolarization.

Two congenital novel mutations in the SCNA5 gene leading to LQT3 [9], namely G1481V and Q1491H, which destabilize voltage-gated Na⁺ inactivation, were identified in two infants that suffered sudden infant death. The G1481V mutation caused a substitution of glycine-to-valine 4 amino upstream from the isoleucine, phenylamine, and methionine motif. The Q1491H mutation caused a glutamine-to-histidine substitution 4 amino acids downstream from the isoleucine, phenylamine, and methionine motif.

Computational cardiac electrophysiology has been very useful in gaining mechanistic insights into arrhythmia initiation and maintenance since its emergence. Here, we use a computational modeling approach to study the different phenotypes expressed by both Purkinje and myocardial cells in the presence of G1481V and Q1491H mutations. Electrophysiological data from whole current recordings from WT(wild type) and mutant channels (G1481V and Q1491H) was fitted to a Markovian model of Na⁺ channel from Zhang et al. [10]. The optimized kinetic model of Na⁺ was integrated into well-validated rabbit Purkinje and myocardial ionic models that were developed by Aslanidi and co-authors [8]. Single-cell simulations were performed at different pacing frequencies in openCARP [11], an opensource finite element software for in-silico cardiac experiments.

Our simulation results revealed that both G1481V and Q1491H mutant channels manifest in a tissue-specific manner. Mutant channels (G1481V and Q1491H) prolonged the AP in myocardial and Purkinje cells. However, AP duration was markedly prolonged in Purkinje cells compared to myocardial cells. A significant late Na⁺ current was found to underlie the severe AP prolongation observed in the Purkinje cell. At slow heart rates, mutant channels developed arrhythmogenic EAD (abnormal depolarizations occurring during the plateau or repolarization phase of AP) and DAD (abnormal depolarizations occurring during the resting phase of AP) in Purkinje but was absent in myocardial cells.

2. METHODS

Experimental data for this study was obtained from Plumereau et al. [9]. Briefly, whole-cell patch clamp recordings using Human Embryonic Kidney (HEK) 293 cells expressing WT and mutant (G1481V and Q1491H) SCNA5 channels were peformed. A total of eight voltage protocols were used to biophysically characterize WT and mutant channels. The experimental data was then used to parameterize a Markovian kinetic model of Na⁺ [10] shown in Figure 2.

Figure 2:

Kinetic Model of Na⁺ channel by Zhang et al. [10].

Detailed description of the model can be found here [10]. Succinctly stated, the kinetic model herein is comprised of three closed states (C₁ – C₃) states and one open (O) state. Additionally, there are four inactivated (I₁₁ - 1₁₄) and four slow inactivated (I₂₁– I₂₄) states. The channel transition from I₁₁ - I₁₄ to I₂₁ – I₂₄ is very slow, leading to slow inactivation.

Before selecting the kinetic model in this study, a plethora of Na⁺ ionic models were explored from literature. This comprised of the Hodgkin-Huxley Na⁺ channel formulation in the rabbit Purkinje model by Aslanidi et al. [8], a Markovian Na⁺ model which was devloped by Clancy and co-authors [4], a Markovian human cardiac Na⁺ model [12], that addressed the very fast deactivation and removed the steep voltage dependence of inactivation and recovery in [4] and a 12-state Markovian Na⁺ channel model from Zhang et al. [10]. All the models were subjected to the protocol that was used to obtain the experimental recovery curve. This was to ascertain which of the models could recapitulate the experimental recovery current traces and account for the biexponential recovery process that was observed in the G1481V and Q1491H mutant channels after prolonged depolarization. None but the Zhang et al.[10] model could satisfy the two conditions.

In fitting the experimental data to the kinetic model of Na⁺, the cost function was defined as the weighted sum of the root mean squared error of a simulated curve and its experimental counterpart. Our optimization targets were the current voltage (IV) relationship, steady state activation, steady state inactivation, recovery from inactivation curves and the ratio of late Na⁺ current to peak Na⁺ current. The systems of differential equations governing the model during the optimization routine were solved using ode23t, an ordinary differential equation (ODE) solver for moderately stiff ODEs. Optimization of the objective function combined Nelder-Mead (NM) downhill simplex [13] and Particle Swarm Optimization (PSO) [14], [15] algorithms. The starting values for the NM algorithm were extracted from Zhang et al. [10].The initial assumption was that, only utilizing NM, a local search algorithm, could effectively minimize cost, however the resulting parameters after 150 iterations could not generalize well for the mutant channels. Even doubling the number of iterations barely improved the cost. To circumvent this, the optimized parameters from the NM algorithm, were used as a starting point for the PSO, a global optimization algorithm. A swarm size of 100 which was twice the size of the parameters to be optimized was used.

After the optimization routine, the optimized kinetic model of Na⁺ was integrated into well-validated rabbit Purkinje and myocardial ionic models and simulated at different pacing frequencies in openCARP to study the evolution of AP with time. The optimized WT and mutant Na⁺ channels (G1481V and Q1491H) failed to repolarize when incorporated in the myocardial and Purkinje ionic models. Previous researchers encountered a similar challenge when they substituted Hodgkin-Huxley equations to a Markovian model for describing a Na⁺ channel behavior[16], [17] in cardiac ionic models. We resolved this problem by scaling the maximal Na⁺ conductance by 40% so that AP was stable for both the WT and mutant channels. The scaling of maximal Na⁺ conductance reduced current peaks, albeit with no alteration to the channel kinetics.

3. RESULTS

3.1. Optimization Outcomes

Figure 3 compares the experimental data with the simulated protocols. The bold trace depicts the simulated counterpart for an experimental curve. As evidenced by Figure 3, the kinetic model of Na⁺ generalized well for both the WT and mutant channels albeit better for the WT. The Na⁺ model herein could account for the biexponential recovery process which could not be recapitulated by the Hodgkin Huxley Na⁺ model by Aslanidi et al.[8] or the Markovian model of Na⁺ that was used to study Na⁺ mutation which resulted in both Brugada syndrome and LQT3 [4].

Figure 3:

Comparison of kinetic model characteristics between experimental and simulated data. A. Current Voltage (I-V) Curve B. Steady state activation curve. C. Steady state inactivation curve. D. Recovery from inactivation curve. WT-sim – Simulated WT, WT-Exp – Experimental WT data, Q1491H-sim – Simulated Q1491H, Q1491H-Exp – Experimental Q1491H data, G1481V-Exp – Experimental G1481V.

3.2. Electrophysiological Consequences of Mutation

Identical mutant and WT optimized models were integrated into myocardial and Purkinje ionic models and paced at different frequencies, which captured fast, regular, and slow heartbeats. At each pacing frequency, the models were paced for 100 beats until the models attained a steady state. Effects of G1481V and Q1491H mutations on the rabbit Purkinje and myocardial ionic models at 2Hz pacing are shown in Figure 4. A significant peak current was observed in the G1481V mutation and lowest in Q1491H compared to the WT. This observation was consistent with their corresponding I-V curve (Figure 3, Panel A). Moreover, action potential duration (APD) at 90% repolarization was prolonged for the mutant channels compared to the WT in all cell types. However, percentage change in APD was markedly pronounced in Purkinje cells compared to myocardial myocytes.

Figure 4:

Outcome of LQT3 mutations in rabbit Purkinje and myocardial ionic model at 2Hz. Action potentials simulated at 2 Hz pacing frequency in wild type (WT) and LQT3 mutant cardiac cells. Simulations compare responses across different cardiac cell types: epicardial, endocardial, midmyocardial, and Purkinje cells. A) WT, B) Q1491H mutation, and C) G1481V mutation. Action potentials were recorded from epicardial (Epi), endocardial (Endo), midmyocardial (mid-Myo), and Purkinje (Purk) cells.

Panels A, B, and C in Figure 5 represent the 99^th and 100^th beat at 1 Hz pacing in WT, Q1491H, and G1481V, respectively. Similar to the 2Hz pacing, mutant channels (G1481V and Q1491H) showed prolonged action potential in all cell types (Figure 5). Peak current was highest and lowest in G1481V and Q1491H, respectively, juxtaposed to the WT channels. Furthermore, the APD was more prolonged in all cell types. Specifically, in Purkinje cells, at 1Hz, APD was prolonged by 138% and 128% (Table 1), respectively, in G1481V and Q1491H. Additionally, 1 Hz pacing did not only lead to markedly prolonged APD in Purkinje in the presence of the mutant channels but arrhythmogenic EAD and DAD (Figure 5, Panels B and C).

Figure 5:

Outcome of LQT3 mutations in rabbit Purkinje and myocardial ionic model at 1 Hz. Action potentials simulated at 1 Hz pacing frequency in wild type (WT) and LQT3 mutant cardiac cells. Simulations compare responses across different cardiac cell types: epicardial, endocardial, midmyocardial, and Purkinje cells. A) WT, B) Q1491H mutation, and C) G1481V mutation. Action potentials were recorded from epicardial (Epi), endocardial (Endo), midmyocardial (mid-Myo), and Purkinje (Purk) cells.

Table 1:

Average of the last ten action potential duration at 90% repolarization during 1 Hz pacing.

Action Potential Duration at 90% Repolarization During 1 Hz Pacing
Cell Type	WT	Q1491H	Q1491H prolongation from WT (%)	G1481V	G1481V prolongation from WT (%)
Epi	187	206	10	210	12
Mid-Myo	214	237	11	237	11
Endo	199	221	11	224	13
Purk	238	567	138	543	128

Notably, Q1491H exhibited a beat-to-beat alteration in AP which was absent from the G1481V mutation. After every eight beats in the Q1491H mutation AP, the ninth beat resulted in an AP, which was plagued with an EAD (Figure 6).

Figure 6:

Beat to beat alteration in action potential at a basic cycle length of 1000 ms.

3.3. Ionic Mechanism Underlying Pronounced AP Prolongation, EAD and DAD Formation in Mutant Channels in the Purkinje

EADs and DADs arise from different cellular mechanisms. EADs can occur through either reactivation of L-type calcium channels ($I_{\mathit{\text{CaL}}}$) [18] or enhanced late sodium current ($I_{\mathit{\text{NaL}}}$) [4]. DADs emerge during the diastolic phase when elevated intracellular Ca²⁺ activates the Na-Ca exchanger, generating a depolarizing inward current[19]. Our simulations at 1 Hz reveal distinct behaviors between WT and mutant channels (G1481V and Q1491H). In Purkinje cells, both mutations dramatically prolonged APD due to an enhanced $I_{\mathit{\text{NaL}}}$. This effect was notably more pronounced in mutant channels compared to WT (Figure 7, Panel B). In contrast, myocardial cells showed minimal differences in $I_{\mathit{\text{NaL}}}$ between mutant and WT channels at the same pacing frequency. The mechanism of EAD formation (Figure 7, marked by violet arrow) relates to Na⁺ channel recovery. A significant late Na⁺ current is generated at slow heart rate leading to prolonged APD. The sufficiently long diastolic interval allows $I_{\mathit{\text{CaL}}}$ to recover during the action potential plateau generating EADs. DAD formation (Figure 7, marked by cyan arrow) occurred during the diastolic phase of AP. The elevated intracellular calcium activates the Na-Ca exchanger current. As this Na-Ca exchanger removes Ca²⁺ by bringing in Na⁺, the net current is an inward current that depolarizes the Purkinje cell membrane.

Figure 7:

Electrophysiological basis for pronounced AP prolongation, EADs and DADs in the Purkinje cell at 1 Hz. A. Evolution of transmembrane potential with time. B. Na⁺ current in WT and mutant channels. C. Intracellular calcium current. Larger late Na⁺ in mutant channels and reactivation of $I_{\mathit{\text{CaL}}}$, respectively underlie pronounced AP prolongation and EAD formation in the Purkinje cell. DADs are caused by the elevated levels of intracellular calcium. Violet arrows point to an EAD. Cyan arrow points to a DAD. $I_{\mathit{\text{NaF}}}$ indicates the fast component of Na⁺ current. $I_{\mathit{\text{NaL}}}$ indicates the slow component of Na⁺ current.

4. DISCUSSION

Several studies have implicated the Purkinje cells, the primary conduction cells of the heart, as the likely source of ectopy in LQTS. In this study, we utilized a computational modeling approach to gain insight into the arrhythmogenic potentials of two novel mutations (G1481V and Q1491H) in the SCNA5 gene identified in two infants, leading to their sudden death.

Specifically, both G1481V and Q1491H mutant channels caused a prolongation in APD in Purkinje and myocardial cells; however, the Purkinje cells exhibited a severe phenotype. A large late Na⁺ current facilitated the significantly pronounced APD in Purkinje by the mutant channels.

During slow heart rates, $I_{\mathit{\text{CaL}}}$ recover to depolarize the membrane generating EADs in Purkinje cells. In addition, subthreshold DADs could be observed. Finally, Q1491H could readily generate EADs (Figure 6) at the slow heart rate between the two mutations and hence may be deleterious compared to G1481V.

This investigation illuminates the role of Na⁺ in regulating the cardiac rhythm. The kinetic model of Na⁺ presented here showed that mutant channels exhibit a slow recovery from inactivation (Figure 3, Panel D) together with a positive shift in the steady state inactivation curve (Figure 3, Panel C) compared to WT. This behavior culminated in a prolonged APD in myocardial cells, explaining the QT prolongation observed in the probands, as reported by Plumereau et al [9].

A computational modeling approach was utilized by Iyer and Co-authors [20] to study the difference in response to different LQT3 syndrome mutations in Purkinje and myocardial cells. The authors concluded that mutations leading to LQT3 depend not only on the mutation and biophysical alteration but also on the tissue expression in a rate-dependent manner. Consistent with our results, G1481V and Q1491H mutant channels expressed different phenotypic behavior in Purkinje and myocardial cells rate-dependently. This may explain why AP was only prolonged in Purkinje and myocardial cells at 2 Hz, but AP prolongation together with abnormal AP depolarizations were observed at 1 Hz in Purkinje cells but absent in myocardial cells. Our simulations showed a larger late Na⁺ current in Purkinje than myocardial cells in the presence of G1481V and Q1491H mutations.

In the event of the loss of a repolarization current, other repolarization currents compensate for the loss of the current. This ability of ionic currents is referred to as the concept of repolarization reserve [21]. Repolarization currents reduce the duration of APD. Few repolarization reserves have been reported in Purkinje [21], [22] compared to the myocardial cells; therefore, myocardial cells have adequate repolarization reserve to counteract the effects of the prominent late Na⁺ current, hence abolishing EAD and DAD that was observed in Purkinje. In patients with G1481V or Q1491H mutations, treatment strategies could focus on two approaches. First, administering class 1b antiarrhythmic medications like ranolazine and tetrodotoxin, which specifically target the late Na⁺ current, may prove beneficial. Second, given the severe phenotype observed in Purkinje cells, ablation of these specialized conducting fibers might remedy ventricular arrhythmias in affected patients.

A population-based approach at the cellular level [16], [23] in which ionic currents were varied within a wide range was used to explain the incomplete penetrance of patients with LQT3 from genetic mutation. This approach explored individual cells with different ionic properties, albeit showing AP and calcium transient characteristics within the physiological range of WT channel models. Differing from their study, we only integrated our optimized Na⁺ kinetic model into the well-validated rabbit Purkinje and myocardial, which ignored possible heterogeneities in ionic currents within a population. Since midmyocardial cells have longer APD compared to both epicardial and endocardial cells, within a population of midmyocardial cells, depending on the balance between inward and outward currents, some may generate EADs or DADs in the presence of the mutant channels despite not being observed in the midmyocardial ionic model used in this study. This possibility, however, seems less likely than in Purkinje cells which were found to be the most likely cell type to generate EADs or DADs due to the reported less repolarization reserves, and also predicted to be subjected to less electrotonic load compared to myocardial cells when not close to the Purkinje-myocardium junctions [24], [25].

Biographies

ANTHONY OWUSU-MENSAH recently graduated from Old Dominion University in December 2024, USA, where he obtained a Ph.D. in Biomedical Engineering. His research interests include using computational models and simulations of cardiac electrophysiology to study arrhythmia initiation and maintenance. His email address is aowus003@odu.edu.

OMER BERENFELD is a Professor of Internal Medicine - Cardiology, Biomedical Engineering and Applied Physics at the University of Michigan. He is also a senior member of the Center for Arrhythmia Research at the University of Michigan. He received his Ph.D. in Medical Physics from Tel Aviv University, Israel. His research interests are in the general field of cardiac electrophysiology, and more specifically in the theoretical, experimental and clinical understanding of mechanisms of impulse propagation in the heart and arrhythmias. His email address is oberen@med.umich.edu.

QUENTIN PLUMEREAU is a Ph.D. student in Clinical and Biomedical Sciences at Laval University where he is supervised by Professor Mohammed Chahine. His research lies in studying ion channels to elucidate our understanding of channelopathies resulting in cardiac diseases. His email address is quentin.plumereau.1@ulaval.ca.

MOHAMED CHAHINE is a currently a Professor at Laval University at the Department of Medicine in Québec City, Canada and holds the Canada Research Chair in Neuromuscular Disorders and Channelopathies. The focus of his research has been the study of ion channels in order to acquire a comprehensive understanding of their involvement in several human diseases. Dr. Chahine team were the first to clone and characterize the human cardiac Na⁺ channel, now known as Nav1.5. The first long QT mutation in Canada was reported by his laboratory. His email address is Mohamed.Chahine@phc.ulaval.ca.

MICHEL AUDETTE is an Associate Professor in the Department of Electrical and Computer Engineering at Old Dominion University. He obtained a Ph.D. in Biomedical Engineering from McGill university, Canada. His research interests are in medical/surgical simulations, surgery planning, and medical device facilitation. His email address is maudette@odu.edu.