Mechano-electrical feedback in transgenic rabbit models of long QT syndrome Type 2 and short QT syndrome Type 1

Abstract

Aims

Electromechanical coupling and mechano-electrical feedback (MEF) are crucial for cardiac function, but their pro-arrhythmic roles in short and long QT syndromes (SQT1 and LQT2) are not fully understood. We aimed to evaluate MEF-induced electrical changes, their arrhythmic impact, and the involvement of stretch-activated channels (SACs) in transgenic rabbit models of SQT1 and LQT2.

Methods and results

Patch-clamp and fluorescence imaging were used to analyse action potential duration (APD), Ca²⁺ transients, and contractility in ventricular cardiomyocytes (VCMs) from LQT2, SQT1 and wild-type (WT) rabbits. LQT2 cells showed prolonged APD and Ca²⁺ transients, increased early afterdepolarizations, Ca²⁺ oscillations, and impaired mechanics compared to WT and SQT1. The cellular electromechanical window (Ca²⁺-transient duration minus APD) was more negative in LQT2 and more positive in SQT1 than in WT. QTc prolonged with preload/afterload increase and decreased with preload reduction across all genotypes, but MEF-induced QTc changes and dispersion were most pronounced in LQT2.

Ex vivo Langendorff experiments showed that increased right ventricular (RV) pressure prolonged APD and QTc in WT hearts. This was attenuated by the SAC blocker GSMTx4, suggesting a role for SACs in MEF.

In silico models of human VCMs including SACs confirmed higher vulnerability to stretch/MEF-induced arrhythmias, including re-entry, in SQT1 and LQT2.

Conclusion

Mechano-electrical feedback-induced electrical changes, partly mediated by SACs, occur in WT, SQT1, and LQT2, but MEF effects are strongest in LQT2. Mechano-electrical feedback induces pro-arrhythmic effects in silico more prominently in LQT2 and SQT1 than in WT, highlighting the potential pro-arrhythmic role of MEF in a vulnerable electrophysiological substrate.

Graphical Abstract

Graphical Abstract.

What’s new?

• Mechano-electrical feedback (MEF) induces electrical changes in long QT (LQT2) and short QT syndrome (SQT1).

• Increased preload and afterload prolong QTc and reduced preload shortens QTc across all genotypes, but MEF-induced QTc changes and dispersion are most pronounced in LQT2.

• Increased RV pressure in Langendorff-perfused hearts prolongs action potential duration and QTc. This is attenuated by the stretch-activated channel (SAC)-blocker GSMTx4, suggesting a role for SACs in MEF.

• In silico models reveal increased vulnerability to stretch-/MEF-induced arrhythmias in SQT1 and LQT2.

Introduction

Electromechanical coupling and mechano-electrical feedback (MEF) play essential roles in maintaining normal cardiac function.¹ The importance of this bidirectional interaction between electrical and mechanical functions, termed ‘electromechanical reciprocity’,² is well recognized in both physiological and pathological states.

Understanding this electromechanical reciprocity in health, and its perturbation in disease, is crucial for uncovering pathophysiological mechanisms underlying various cardiac diseases, including the QT syndromes. In the congenital channelopathies long QT (LQTS) and short QT syndrome (SQTS), historically classified as ‘purely electrical’ diseases, accumulating clinical and experimental evidence demonstrates that these diseases are, in fact, ‘electromechanical’ disorders.^2,3,4,5 In particular, patients with LQTS or SQTS not only exhibit pathologically prolonged or shortened cardiac repolarization and increased regional repolarization heterogeneity, predisposing them to ventricular arrhythmias, but also mechano-electrical coupling-induced secondary global and regional alterations in mechanical cardiac function:^2,5,6,7,8 In LQTS Type 2 (LQT2), pathogenic loss-of-function variants in the KCNH2 gene, leading to an impaired rapid delayed-rectifier potassium current (I_Kr), result secondarily in subclinically altered global and regional left ventricular (LV) function—including reduced strain, strain rate, and diastolic dysfunction, a prolonged contraction duration,^6,7,9,10,11 and a pronounced mismatch between durations of mechanical systole and electrical QT, the so-called electromechanical window (EMW).^6,12 Similarly, in SQTS Type 1 (SQT1), pathogenic gain-of-function variants in KCNH2, leading to enhanced I_Kr, cause an abnormally short QT interval, accompanied by mechanical dysfunction manifesting as reduced systolic function with decreased LV ejection fraction and altered myocardial strain, accelerated diastolic relaxation along with increased mechanical heterogeneity,^5,13 and a pathologically positive EMW and its cellular surrogate both in clinical and in silico.^5,14

These electromechanical changes correlate with the arrhythmogenic risk, supporting a potential pro-arrhythmic effect of mechanical alterations.^10,11 Furthermore, acute regional disparities in myocardial contraction and relaxation due to aftercontractions may additionally contribute to the formation of electrical heterogeneities via MEF, exacerbating the risk of re-entrant ventricular arrhythmias in LQTS.^15,16 Moreover, mechano-induced ventricular fibrillation in SQTS due to direct mechanical impact of the catheter on the heart¹⁷ highlight the potential role of MEF in arrhythmogenesis. A direct assessment of the extent of mechano-induced electrical changes, their underlying mechanisms, and their impact on arrhythmogenesis in LQT2 and SQT1, however, is lacking. Given the complexity of electromechanical reciprocity, investigations into MEF and its impact on disease phenotype and arrhythmia formation in LQTS and SQTS must be conducted using appropriate animal models. Rabbits exhibit excellent similarity to human cardiac electrophysiology, including action potential (AP) morphology, underlying cardiac ion channels/currents,^18,19 and regional systolic and diastolic mechanical properties.²⁰ Additionally, rabbits and humans share similar electromechanical coupling and MEF mechanisms.²¹ Our transgenic rabbit model of LQT2 was generated by cardiac-specific overexpression of the human dominant-negative loss-of-function pathogenic variant KCNH2-G628S,²² while our SQT1 rabbit was generated by cardiac-specific overexpression of the human dominant gain-of-function pathogenic variant KCNH2-N588K.¹³ They both closely mimic the corresponding human disease phenotypes, displaying prolonged or shortened QT and AP duration (APD), respectively, as well as altered mechanical function, and an increased propensity for ventricular arrhythmias. Thus, the LQT2 and SQT1 rabbit models provide a unique opportunity to study the impact of MEF on prolonged or shortened QT/APD caused by different variants in the very same gene, excluding other influences.

Here, we used these transgenic rabbit models to investigate the electrical consequences of mechanical changes induced by acutely altering preload or afterload and associated cardiac stretch in LQT2 and SQT1. Given that stretch-activated channels (SACs) are potential mediators of MEF-induced electrical changes, including the modulation of APD, initiation of afterdepolarizations, and promotion of arrhythmic activity at the cellular and tissue levels,^1,23 we further evaluated the effect of their inhibition on stretch-induced QT/APD changes on the whole heart level.

Finally, these experimental data served as input for in silico models of LQTS and SQTS ventricular cardiomyocytes (VCMs) to investigate and expand the mechanistic understanding of SAC-related MEF-induced electrical alterations and their impact on arrhythmogenesis at the cellular and tissue levels.

Methods

A more detailed methods section can be found in the online supplement.

Animal studies

Adult transgenic LQT2, SQT1, and wild-type (WT) New Zealand white rabbits^13,22 of both sexes were utilized according to European Union (EU) regulations and Swiss animal welfare laws (licences BE115-19, BE107-22, and BE55-21).

In vivo ECG recordings

Twelve-lead ECGs were obtained under ketamine/xylazine anaesthesia.²⁴ QTc and QTc dispersion were assessed at baseline and during preload/afterload changes (i.v. 0.9% NaCl bolus, abdominal compression, and aortic balloon inflation).

Cardiomyocyte isolation

Following anaesthesia with ketamine/xylazine and euthanasia with pentobarbital, hearts were excised and perfused on a Langendorff system with collagenase digestion to obtain VCMs.²⁵

Action potentials

Action potentials were recorded in isolated LV cardiomyocytes (VCM) (from LQT2, SQT1, and WT) at 1–2 Hz at 37°C in current clamp mode. Action potential duration (APD₉₀), amplitude (APA), and resting membrane potential (RMP) were analysed.

Ca²⁺ transients and contractility

Fura-2AM–loaded (LQT2, SQT1, and WT) VCMs were field stimulated at 0.5–1 Hz. Ca²⁺ transient duration (Ca²⁺T₉₀) and fractional shortening were quantified at 37°C.²⁶

Cellular electromechanical window

Average APD₉₀ and average Ca²⁺T₉₀ from VCMs stimulated at 1 Hz at 37°C were used to estimate the cellular EMW (average Ca²⁺T₉₀–average APD₉₀).^27,28

Monophasic action potentials ex vivo

Langendorff-perfused WT hearts were used to assess preload-dependent APD/QT changes. A right ventricular balloon catheter modified preload (6–75 mmHg). Experiments were repeated after applying 30 µM GSMTx4, a blocker of SACs.^29,30

RNA extraction and RT-qPCR

Total RNA was isolated from isolated (LQT2, SQT1, and WT) LVCMs and analysed by RT-qPCR (MIC cycler, SensiFAST SYBR). Expression of Piezo1, TRPM4, and KCNK2/TREK-1 was calculated as relative expression to reference gene beta-actin (ACTB) and normalized to WT.

Computational modelling

We employed our previously developed in silico model of human VCM electrophysiology incorporating MEF through three different SAC types: non-selective (SAC_ns), potassium-selective (SAC_K), and calcium-selective (SAC_Ca).^31,32 The model simulates stretch-induced SAC effects by modelling a step input with adjustable amplitude, timing, and duration.

SQT1 and LQT2 phenotypes were reproduced by scaling I_Kr by factors of 1.4 and 0.63, yielding APD₉₀ values of 262.1 ms and 374.1 ms, respectively (control = 308.5 ms). Simulations included sustained (500 beats) and transient (20 beats) sub-threshold stimuli (5% stretch) to reflect experimental conditions.

At the tissue level, 2D simulations (500 × 500 grid, conductance = 0.85 mS) were performed in Myokit³³ to assess local and re-entrant activity following circular SAC activation under different stretch timings and magnitudes across Control, SQT1, and LQT2 genotypes.³⁴

Statistical analyses

Data analysis was performed using GraphPad Prism. Results are presented as mean ± SD. Group differences were evaluated using t-tests or one-way analysis of variance (ANOVA) with appropriate post hoc tests, or Mann–Whitney tests for non-parametric data. Potential clustering of VCMs was investigated using hierarchical statistical methods from Sikkel et al.³⁵ Statistical significance was set at P < 0.05.

Results

We first characterized baseline electrical and mechanical features and their electromechanical interplay—e.g. the EMW—in the different genotypes.

Baseline differences in QT characteristics between genotypes in vivo

Using 12-lead surface ECG, differences in QT duration were assessed in transgenic SQT1 and LQT2 rabbits compared to WT controls (Figure 1A and B). Consistent with previous work,^13,22 heart rate-corrected QT duration (QTc) was significantly prolonged in LQT2 and shortened in SQT1 compared to WT (LQT2, 246.1 ± 15.4 ms; WT, 174.5 ± 15.8 ms; SQT1, 159.2 ± 12.8 ms; Figure 1A and B). In addition, QT dispersion—a marker for regional repolarization heterogeneity—was increased only in LQT2 (QT_max − QT_min: LQT2 19.9 ± 6.5 ms vs. WT 14.1 ± 6.2 ms; P < 0.05; Figure 1C), indicating a pronounced arrhythmogenic substrate predisposing the LQT2 hearts to spontaneous VTs.²²

Figure 1.

Genotype differences in ECG and AP characteristics. (A) Representative ECG traces. SQT1, WT, and LQT2 traces at baseline. (B) Baseline genotype differences in QTc: QTc in transgenic SQT1 (N = 12, blue circles) and LQT2 (N = 10, red triangles) rabbits compared to WT controls (N = 12, grey squares). Prolonged QTc in LQT2 and shortened QTc in SQT1 rabbits (*P < 0.05 for SQT1 vs. WT, ***P < 0.001 for LQT2 vs. WT and SQT1). (C) Baseline differences in QT dispersion. Increased QT dispersion in LQT2 (N = 10) compared to WT (N = 12) (*P < 0.05). Male and female rabbits are represented by darker and lighter colour shades, respectively. (D) Representative AP recordings in isolated VCMs from WT (grey), SQT1 (blue), and LQT2 (red) rabbits stimulated at 1 Hz. (E) EADs in LQT2. Left: representative LQT2 action potential traces showing EADs. Right: comparison of EAD occurrence among the different genotypes in individual cardiomyocytes. Frequent EADs were observed only in LQT2 cardiomyocytes (***P < 0.01 vs. WT and SQT1). (F) AP characteristics at 1 Hz. LQT2 cardiomyocytes show prolonged APD compared to WT and SQT1 (P < 0.001), while SQT1 cardiomyocytes have significantly shorter APD compared to WT and LQT2 (P < 0.001). WT (n = 50/N = 5), SQT1 (n = 60/N = 7), and LQT2 (n = 30/N = 7). n indicates the number of cells; N indicates the number of hearts included in the experiments.

Genotype differences in cellular action potential characteristics

Whole-cell patch-clamp recordings showed that LQT2 VCMs exhibited a significantly prolonged APD₉₀ compared to WT and SQT1 at 1 Hz (LQT2, 475.5 ± 80.8 ms; WT, 332.1 ± 95.0 ms; SQT1, 246.1 ± 15.4 ms; P < 0.0001; Figure 1D and F) and displayed more frequent early afterdepolarizations (EADs) (P < 0.01 vs. WT and SQT1; Figure 1E). These EADs, observed exclusively in LQT2 VCMs, indicate that pro–arrhythmic-triggered activity is enhanced in LQT2. In contrast, SQT1 VCMs showed a significantly shorter APD₉₀ than WT VCMs (P < 0.0001; Figure 1D and F). Similar differences were observed at a stimulation frequency of 2 Hz, with a longer LQT2 APD₉₀ and shorter APD₉₀ in SQT1 compared to WT (P < 0.01 for both) (see Supplementary material online, Figure  S1). No differences were observed between the genotypes in the remaining AP parameters (APA, dV/dtmax, and RMP) (Figure 1F; Supplementary material online, Figure  S1).

Genotype differences in cellular Ca²⁺-transient characteristics, contractile function, and electromechanical reciprocity

Ca²⁺T duration and sarcomere shortening were evaluated in isolated rabbit VCMs (at 1 Hz LQT2, 484 ± 144 ms; WT, 446 ± 0.092 ms; SQT1, 474 ± 96 ms and 0.5 Hz LQT2, 765 ± 365 ms; WT, 526 ± 224 ms; SQT1, 550 ± 234 ms; Figure 2). While no genotype differences in Ca²⁺T duration were observed at 1 Hz (Figure 2A and D), 15 out of 65 LQT2 VCMs (∼23%) exhibited Ca²⁺ oscillations (Figure 2A and B) previously identified as pro-arrhythmic, preceding the initiation of EADs.³⁶ This led to Ca²⁺T persisting beyond a single pacing interval (1000 ms). As such, the measured Ca²⁺T₉₀ values in the remaining cells shown in Figure 2D underestimate the true Ca²⁺T₉₀ duration at 1 Hz. To assess and visualize the presumed average Ca²⁺T₉₀ at 1 Hz if these excessively long Ca²⁺T₉₀ would also be considered (and set to 1000 ms thus still underestimating the real, but not assessable, duration), an additional analysis including these 15 cells was performed (see Supplementary material online, Figure  S2A), indicating that the Ca²⁺T₉₀ at 1 Hz is significantly longer in LQT2 than in WT and SQT1. To better assess potential differences in Ca²⁺T₉₀, similar experiments were also performed at a lower pacing rate of 0.5 Hz. At this frequency, LQT2 VCMs showed significantly prolonged Ca²⁺T₉₀ compared to WT and SQT1 (Figure 2C and D). Additionally, LQT2 VCMs demonstrated slower Ca²⁺T upstroke velocity and smaller Ca²⁺T amplitude as compared to WT (see Supplementary material online, Figure  S2A).

Figure 2.

Genotype differences in calcium transients and contractility in cardiomyocytes. (A) Representative calcium transients in isolated cardiomyocytes from each genotype stimulated at 1 Hz. (B) Calcium oscillation occurrence in isolated cardiomyocytes. LQT2 (N = 6) cardiomyocytes show more frequent pro-arrhythmic calcium oscillations as compared to WT (N = 8) (P < 0.01) and SQT1 (N = 6) (P < 0.01). (C) Representative calcium transients in isolated cardiomyocytes from each genotype stimulated at 0.5 Hz. (D) Genotype differences in calcium transient characteristics. LQT2 (n = 44/N = 6, red) cardiomyocytes display prolonged calcium transients as compared to WT (n = 60/N = 8, grey) and SQT1 (n = 45/N = 6, blue) at 0.5 Hz (P < 0.001 and P < 0.01, respectively). (E) Representative overlay of calcium transients and action potential traces at 1 Hz for the assessment of cellular EMW in SQT1, WT, and LQT2. (F) Cellular EMW. LQT2 (n = 44/N = 6, for calcium transient; n = 30/N = 7, for APD) cells exhibit a negative EMW (calcium transient duration minus APD), while SQT1 (n = 45/N = 6, for calcium transient; n = 60/N = 7, for APD) cells have a more positive EMW than WT (n = 60/N = 8, for calcium transient; n = 50/N = 5, for APD). *P < 0.05; **P < 0.01; ***P < 0.001. n indicates the number of cells; N indicates the number of hearts included in the experiments.

Moreover, LQT2 VCMs showed a reduced cellular contraction velocity and peak sarcomere shortening, indicating mechanical impairment compared to WT and SQT1 (see Supplementary material online, Figure  S2B). These experiments were conducted both with and without the fluorescent dye Fura-2, a calcium chelator, to rule out potential effects of the experimental conditions during Ca²⁺ imaging. The results were consistent at both 0.5 and 1 Hz (see Supplementary material online, Figure  S2C).

Genotype differences in cellular electromechanical reciprocity

Using Ca²⁺T₉₀ and APD₉₀ data at 1 Hz, we calculated the difference between average Ca²⁺T₉₀ and APD₉₀ as a cellular surrogate of the EMW.^2,12 Consistent with in vivo patient data, LQT2 VCMs exhibited a negative EMW, while SQT1 VCMs had a more positive EMW than WT VCMs (Figure 2E and F), demonstrating an impaired electromechanical interplay.

Genotype differences in mechano-electrical feedback-induced QT changes in vivo

Mechano-electrical feedback-induced changes in QTc were assessed during increased preload, decreased preload, and increased afterload.

An increased preload (Figure 3A and B), induced by an acute i.v. bolus injection of isotonic NaCl (0.9%), prolonged QTc in all groups. This transient QTc increase was observed ∼20 s after bolus injection, with the QTc returning to baseline within 40 s (see Supplementary material online, Figure  S3C). Notably, increased preload led to a rise in QT dispersion only in LQT2 (from 19.9 ± 6.5 to 30.3 ± 7.2 ms, P < 0.001), with significantly greater change in QT dispersion than in WT and SQT1 (see Supplementary material online, Figure  S3A). The extent of MEF-induced changes in QTc (ΔQTc) differed among genotypes: QTc changes were more pronounced in LQT2 (46.38 ± 11.34 ms) than in WT rabbits (28.37 ± 8.09 ms, P < 0.001), while QTc in SQT1 rabbits (22.83 ± 8.44 ms) did not differ from WT (P = 0.32). The ΔQTc correlated with the baseline QT, meaning that a longer baseline QT was associated with a greater MEF-induced increase in QTc after increased preload, irrespective of the underlying genotype (Figure 3B).

Figure 3.

Genotype differences and MEC-induced changes in QTc. (A) Representative ECG traces in SQT1, WT, and LQT2 at baseline and after bolus. QT intervals are indicated in red. (B) Effect of increased preload on QTc. Left: QTc prolongation induced by acute i.v. injection of NaCl bolus, observed in all genotypes (P < 0.01 for LQT2 [red], N = 10, P < 0.0001 for WT [grey], N = 12, and SQT1 [blue], N = 12) and all individual animals. Middle: comparison of mechano-induced QTc changes (ΔQTc) demonstrates significant differences between LQT2 and WT (P < 0.01) and SQT1 (P < 0.01). Right: linear regression of baseline QTc and observed QTc changes (ΔQTc) based on the combined data from all genotypes. WT (grey squares), SQT1 (blue circles), LQT2 (red triangles). (C) Effect of decreased preload on QTc. Left: acute reduction in QTc following inferior vena cava compression, observed in all genotypes (P < 0.001 for LQT2, N = 10 and WT, N = 11 P < 0.01 for SQT1, N = 10) and all individual animals. Middle: comparison of mechano-induced QTc changes (ΔQTc) demonstrates significant differences between LQT2 and WT (P < 0.01) and SQT1 (P < 0.05). Right: linear regression of baseline QTc and observed QTc changes (ΔQTc). (D) Effect of increased afterload on QTc. Left: QTc prolongation after acute aortic occlusion, observed in all genotypes (P < 0.01 for LQT2, N = 4; P < 0.001 for WT, N = 5 and for SQT1, N = 7) and all individual animals. Middle: comparison of mechano-induced QTc changes (Delta-QTc) demonstrates significant differences between LQT2 and WT (P < 0.01) and SQT1 (P < 0.05). Right: linear regression of baseline QTc and observed QTc changes (ΔQTc). *P < 0.05; **P < 0.01; ***P < 0.001.

A decreased preload, induced by abdominal compression, acutely shortened QTc in all groups (Figure 3C). This transient QTc decrease occurred ∼10 s after compression (see Supplementary material online, Figure  S3B). As seen with the increased preload, the decreased preload-induced ΔQTc was more pronounced in LQT2 than in WT and SQT1 (P < 0.001; Figure 3C).

An increased afterload, induced by acute proximal aortic occlusion using a Swan–Ganz catheter, acutely prolonged QTc in all groups (Figure 1D). The transient QTc prolongation occurred ∼20 s after aortic occlusion (see Supplementary material online, Figure  S3B). ΔQTc was more pronounced in LQT2 compared to WT and SQT1 (P < 0.001; Figure 3D). And here again, the extent of all MEF-induced changes was dependent on the underlying QTc duration as shown by linear regression analysis (R² = 0.438, P < 0.001 in increased preload, R² = 0.345, P < 0.001 in decreased preload, and R² = 0.319, P < 0.001 in increased afterload; Figure 3B–D).

Mechano-electrical feedback-induced QT and action potential duration changes at the whole-heart level

To investigate potential mechanisms underlying MEF-induced electrical changes, we conducted ex vivo Langendorff perfusion experiments in WT rabbit hearts, in which the end diastolic pressure in the right ventricle (RV) was sequentially increased to simulate an increase in preload. We first investigated the effects of different intraventricular pressure levels on QT and monophasic APD (MAP). As previously demonstrated, physiological intraventricular preload pressure changes from 6 to 15 mmHg had no effect on the QT/APD in Langendorff-perfused hearts ex vivo,³⁵ likely due to differences in the sensitivity to MEF in explanted hearts compared to the in vivo situation. We thus tested higher preload pressures of 30, 50, and 75 mmHg. No significant changes in APD₉₀ were observed at 30 mmHg, and only moderate differences were seen at 50 mmHg, which prompted the decision to test the effects at 75 mmHg, where the effects were most prominent (see Supplementary material online, Figure  S4). These changes were evident in RV-MAP upon the application of increased RV preload, while no changes occurred in the LV MAPs, supportive of a direct, local, stretch-mediated effect.

By increasing RV pressure to 75 mmHg, RV APD₉₀ prolonged significantly compared to baseline at 6 mmHg (6 mmHg, 119.8 ± 22.4 ms vs. 75 mmHg, 131 ± 24.2 ms, P < 0.05; Figure 4A and B). Similarly, QT interval prolonged upon increasing RV preload to 75 mmHg (6 mmHg 181.2 ± 15.3 ms vs. 75 mmHg 190.4 ± 15.8 ms, P < 0.05; Figure 4C and D). Application of GSMTx4, a pharmacological blocker of SACs,³⁶ attenuated the MEF effect of increased preload on APD₉₀ and QT, leading to a less pronounced preload-induced RV APD₉₀ prolongation (delta-APD₉₀ KH/baseline, 11.4 ± 5.5 ms vs. GSMTx4, 8.0 ± 5.5 ms, P < 0.05; Figure 4B) and a less pronounced QT prolongation (delta-QT, KH/baseline, 9.2 ± 5.1 ms vs. GSMTx4, 5.4 ± 3.4 ms, P < 0.05; Figure 4D). Importantly, this GSMTx4-induced decrease in delta-APD₉₀ was observed in all individual hearts, indicating a consistent attenuation of the MEF effect. These experiments suggest that SACs are at least partially responsible for the MEF-induced changes in APD/QT.

Figure 4.

MEC-induced changes in APD and QT interval in ex vivo whole heart. (A) Representative MAP recordings at baseline (6 mmHg, light grey) and during increased preload (75 mmHg, dark grey). (B) Effect of increased preload on APD₉₀. From left to right: (1) increased APD₉₀ in WT rabbit hearts (N = 5) at 75 mmHg compared to baseline (P < 0.05) and (2) slightly increased APD₉₀ in WT rabbit hearts (N = 5) at 75 mmHg under GSMTx4 perfusion compared to baseline (P < 0.05). The effect is reduced by the pharmacological agent GSMTx4 as (3) APD₉₀ changes (Delta-APD₉₀) are smaller with GSMTx4, P < 0.05, as well as the (4) %-changes at baseline and with GSMTx4 (P < 0.05). (C) Representative ECG recordings at baseline (6 mmHg, light grey) and during increased preload (75 mmHg, dark grey) showing QT prolongation. (D) Effect of increased preload on QT interval. From left to right: (1) increased QT in WT rabbit hearts (N = 5) at 75 mmHg compared to baseline (P < 0.05) and (2) slightly increased QT in WT rabbit hearts (N = 5) at 75 mmHg under GSMTx4 perfusion compared to baseline (P < 0.05). The effect is reduced by the pharmacological agent GSMTx4 as (3) QT changes (Delta-QT) are smaller with GSMTx4, P < 0.05, and (4): %-changes at baseline and with GSMTx4 (P < 0.05). *P < 0.05.

Expression of key mechano-sensitive ion channels

We quantified mRNA levels of three key mechano-sensitive/stretch-activated channel genes: Piezo1 (a mechanically gated calcium-selective cation channel, SAC_Ca), TRPM4 (a Ca²⁺-activated, stretch-activated non-selective cation channel, SAC_ns), and KCNK2/TREK-1 (a stretch-activated K⁺ channel, SAC_K). All transcripts were detected in isolated VCMs from WT (N = 3), SQT1 (N = 3), and LQT2 (N = 3) rabbits without significant genotype-dependent differences in expression levels (see Supplementary material online, Figure  S5).

In silico characterization of stretch-activated channel-mediated changes in repolarization and arrhythmogenesis

Continuous and transient sub-threshold stretch application prolonged the APD₉₀ in all model variants as observed experimentally (see Supplementary material online, Figures  S6 and S7). LQT2 exhibited the greatest difference in APD₉₀ compared to the non-stretched APDs (LQT2 APD₉₀ increased by 29.5%, compared to +2.6% in SQT1 and +10.9% in WT Control).

Next, we investigated the potential pro-arrhythmic consequences of SAC activation in response to acute (10 ms) supra-threshold stretch (30%) by performing a systematic analysis of the effects of stretch-related SAC activation on electrical parameters and triggered activity in all model variants.

Stretch applied slightly before APD₉₀ (e.g. at APD₉₀ – 20 ms) depolarized the membrane potential in each model, but was unable to elicit a new AP (Figure 5A and B, third column). When stretch was applied after the APD₉₀ instead (e.g. at APD₉₀ + 100 ms in Phase 4), it produced delayed afterdepolarization (DAD)-induced triggered APs in all three cases (Figure 5A, yellow colour; Figure 5B, fourth column). However, when applied in a vulnerable window during Phase 3 of the AP (e.g. APD₉₀ – 80 ms, and APD₉₀ – 100 ms), it caused distinct effects in the three model variants, producing mild depolarizations for the SQT1 model and minor to moderate depolarizations for the Control model while triggering EAD-like events in the LQT2 model (Figure 5A, blue colour for SQT1 and Control and lighter blue colour for LQT2; Figure 5B, panels a–c, first and second columns).

Figure 5.

Stretch applied at the cellular level provokes distinct genotype-specific effects. (A) Net amplitudes of afterdepolarizations triggered by 30% stretch applied for 10 ms at various times during the time course of the AP for SQT1, Control, and LQT2 model variants. Stretch applied during Phase 1 to Phase 3 could not induce afterdepolarization in any of the model variants, while it triggers APs (afterdepolarization amplitude > 80 mV, yellow colour) when delivered at the end of the APD₉₀. Specifically in the LQT2 model, stretch applied 70–100 ms before the end of the APD₉₀ triggers EADs (first region of light blue colour). (B) Cellular APs in the SQT1 (blue), Control (grey), and LQT2 (red) model variants. Stretch (step signal of 10 ms duration) is applied 100, 80, and 20 ms prior to each model’s APD₉₀, as well as 100 ms after the end of each APD₉₀. Early afterdepolarizations are triggered, e.g. at APD₉₀ - 80 ms, and APD₉₀ – 100 ms, in LQT2; new APs are generated at APD₉₀ + 100 ms in all the three models; afterdepolarizations occur for the remaining cases.

At the tissue level, stretch-induced SAC activation was applied far from tissue borders (in the centre). Depending on the timing and size of the area where the stretch effects were applied, we observed four distinct outcomes in all model variants (shown for the LQT2 variant in Figure 6A, panels a–d). When the effects of stretch were applied in a circular area in a refractory region, they only induced a mild depolarization of the membrane, which prevented propagation of the stretch-induced wave (Figure 6A, panel a). When the SACs were activated in a circular area partly located in an excitable tissue, their activation could instead result in a full AP and therefore induce propagation of a stretch-induced wave forming two symmetric lobes of excitation eventually colliding, disrupting propagation (Figure 6A, panel b). When they occurred at a time such that there was (i) sufficient excitability and (ii) a sufficiently long path length, SAC activation initiated a sustained figure-of-eight re-entry (Figure 6A, panel c). Furthermore, in some LQT2 simulations, the stretch-induced electrical activity persisted long enough to block the subsequent paced wave (Figure 6A, panel d).

Figure 6.

Local stretch can yield four distinct outcomes in 2D simulations. (A) Snapshots of electrical activity in simulated LQT2 tissue. Stretch applied at various time points showing four different outcomes in the LQT2 model variant, depending on the timing and relative size of the stretch stimulus. (a) Stretch applied in refractory tissue (stretched/tissue area = 0.2, stretch applied 16 ms after the APD₉₀) does not depolarize the tissue and provokes only mild effects on AP. (b) Stretch applied in an excitable region (same relative size, but later in time) can induce depolarization and propagation of a stretch-induced wave (stretched/tissue area = 0.2, stretch applied 46 ms after the APD₉₀). (c) At the same moment in time but with a larger stretched area, stretch can provide a longer line of block, creating a sufficiently long path length for sustained figure-of-eight re-entry initiation (stretched/tissue area = 0.3, stretch applied 71 ms after the APD₉₀). (d) Stretch can block paced propagation when persisting long enough (same relative size, later in time) (stretched/tissue area = 0.3, stretch applied 131 ms after the APD₉₀). (B) Multiparameter analysis with stretch applied in the centre of the tissue for the three model variants SQT1, Control, and LQT2. Stretch applied in refractory tissue, i.e. stretch timing < APD₉₀ line (dotted grey line), does not propagate (black crosses). With later stretch stimuli, stretch can trigger sustained figure-of-eight re-entry (yellow dots). In LQT2, late stretch stimuli block the following propagation (green squares). In the remaining instances, stretch-induced depolarizations propagate in the tissue but do not lead to re-entry (blue circles). (C) Arrhythmogenicity bar chart: re-entry occurrences in yellow—SQT1 (76 occurrences among 180 simulations, 42%), Control (56 occurrences, 31%), and LQT2 (39 occurrences, 22%). Paced wave conduction block occurrences in green: LQT2 (38 occurrences, 21%).

Finally, we conducted a multiparameter analysis, varying the time of application and the relative size (stretched area/tissue) of the stretch stimulus applied in the centre in all the model variants to quantify the occurrence of pro-arrhythmic events (Figure 6B). Stretch could more readily initiate figure-of-eight re-entries in the SQT1 model (76 occurrences among 180 simulations, 42%), compared to Control (56 occurrences, 31%).

Furthermore, stretch-induced SAC activation could block the following paced activity exclusively in the LQT2 model (38 occurrences, 21%; Figure 6C), providing an additional substrate for subsequent arrhythmogenesis. Finally, stretch-induced SAC activation could provoke re-entries with smaller stretched area/tissue ratios in both QT syndromes (SQT1, LQT2: min stretch/tissue area = 0.2) compared to Control (0.3) (Figure 6C), highlighting the increased vulnerability of both LQT2 and SQT1 to stretch-related arrhythmogenesis.

Discussion

Using a combined experimental and computational approach, the present study elucidates the interplay between electrical and mechanical cardiac properties, known as electromechanical reciprocity² in SQTS and LQTS compared to healthy WT conditions. Building on our previous work and that by others on electromechanical coupling-induced subclinical mechanical dysfunction in LQTS and SQTS,^{5–7,9–13,37} our novel findings on (i) the cellular EMW in the different channelopathies, (ii) MEF-induced changes in electrical properties and arrhythmogenesis, and (iii) the role of SACs in this mechano-electrical interaction provide important insights into the potentially pro-arrhythmic mechano-electrical interplay in the ‘electrical diseases’ LQTS and SQTS.

Cellular electromechanical window in long QT syndrome type 2 and short QT syndrome type 1

We set out to investigate the cellular surrogate of the in vivo EMW¹² to complement our previous in vivo and whole-heart data on electromechanical disturbances in LQTS and SQTS.^{9–11,13,37,38}

At the cellular level, whole-cell patch-clamp recordings at body temperature revealed significantly prolonged APD₉₀ in LQT2 VCMs compared to WT, accompanied by frequent EADs, evidencing a high propensity to develop cellular triggered activity in LQT2, while SQT1 VCMs showed a significantly shorter APD as shown in previous studies^13,22 without EADs. Similarly, LQT2 VCMs demonstrated prolonged Ca²⁺T₉₀ at lower pacing rates, slower Ca²⁺T upstroke velocities, smaller Ca²⁺T amplitude, more frequent pro-arrhythmic Ca²⁺ oscillations, and impaired contractile function with decreased contraction velocity. These observations are consistent with previous publications on mechanical abnormalities in LQT2 at the in vivo level using MRI and strain analysis via echocardiography.^6,9

Using these data, we calculated the cellular EMW, which we defined as Ca²⁺T₉₀ minus APD₉₀ similar to Passini et al. and Nimani et al.^27,28 Other studies have emphasized the importance of EMW to assess the arrhythmogenic potential, as it reflects the timing between electrical repolarization and mechanical relaxation.¹² In our study, LQT2 VCMs exhibited a more negative cellular EMW than WT, indicative of an increased risk for arrhythmogenesis, while SQT1 cardiomyocytes displayed a more positive cellular EMW compared to WT, which has also been identified as a pro-arrhythmic feature.¹⁴ While the EMW negativity can be driven by the prolonged cardiac repolarization (QT/APD), e.g. as observed during pharmacological I_Kr blockade,³⁹transient accentuations of EMW negativity can similarly be driven by shortening of mechanical systole despite persistent QT prolongation.⁴⁰ These transient accentuations were observed to precede ventricular tachyarrhythmias in patients with inherited LQTS and drug-induced QT prolongation. Importantly, the EMW negativity in LQTS patients is observed despite a general prolongation of the contraction duration in LQTS,^7,10 indicating an overall altered electromechanical interaction. These cellular EMW findings, consistent with clinical data, together with our previous MRI-based assessment of the altered mechanical function^9,13 further underscore that LQT2 and SQT1 rabbit models recapitulate key aspects of electrical–mechanical dysfunction and reciprocity observed in human LQTS/SQTS patients on all levels.

Mechano-electrical feedback-induced electrical changes in long QT syndrome type 2 and short QT syndrome type 1

We then further focussed on investigating MEF-induced electrical changes in LQTS and SQTS, to provide more insights into the thus far underexplored MEF loop of electromechanical reciprocity. In this study, we observed pronounced MEF-induced changes in QT duration due to acute changes in pre- and afterload. While an increased preload due to a bolus application induces myocardial stretch and SAC activation, this intervention may simultaneously activate atrial volume receptors and autonomic reflex pathways such as the Baroreflex. However, we have demonstrated in a previous study in rabbits under complete autonomic blockade³⁸ that increased preload-driven QTc changes persist despite the absence of autonomic feedback, indicating that the observed QT prolongation predominantly reflects intrinsic myocardial MEF responses rather than systemic reflex mechanisms. In all genotypes, acute increases in preload and consecutive mechanical changes led to QTc prolongation, with the effect being most pronounced in LQT2 rabbits and least pronounced in SQT1. This observation aligns with our prior study, in which we observed a more pronounced MEF-induced QT prolongation in acquired, drug-induced LQT2 rabbits than in WT.³² Similarly, an increased afterload induced QT prolongation in all groups, while a reduced preload shortened QT in all groups, which was again in both settings most pronounced in LQT2. The more pronounced QT changes in LQT2, which is characterized by a longer baseline QTc than WT and SQT1, may be attributed to a dependency of QT alterations on the initial QT/APD (i.e. the longer the initial QT/APD, the greater the effect). Indeed, such APD dependency of APD change has already been observed for certain drugs lengthening or shortening the APD.^41–44 In line with this, we observed a linear relation between baseline QT and the extent of QT changes. In addition, LQT2 rabbits exhibited a significant increase in QT dispersion following increased preload, unlike WT and SQT1 rabbits. These data support the observation that QT/APD prolongation per se may amplify regional APD heterogeneities, which may further accentuate the pro-arrhythmic risk. Thus, the more pronounced stretch-induced QT/APD prolongation and heterogeneity in LQT2 may promote arrhythmias more readily compared to WT as observed in general in the context of increased QT variability⁴⁵—and as we also observed in our in silico simulations discussed below. In a broader clinical context, the preload-dependent QTc changes we observed in the RV are consistent with reports in pulmonary hypertension linking RV pressure overload to QTc prolongation and adverse outcomes. While pulmonary hypertension is a different disease entity than the QT syndromes, these parallels support a mechanistic role for RV stretch and MEF in modulating cardiac repolarization.^46,47

The role of stretch-activated channels in mediating mechano-electrical feedback-induced electrical changes

Determining the role of SACs in modulating cardiac electrophysiology is crucial to better understand the mechanisms underlying MEF and their contribution to arrhythmogenesis. Mechanical stimuli such as stretch or pressure can influence transmembrane ion fluxes through SACs, leading to changes in APD and, ultimately, arrhythmic events, depending on the timing of the stretch event in relation to the phase of the AP.^48–53

GSMTx4 is a blocker of non-selective cation SACs.⁴⁷ It acts as a membrane tension buffer at the lipid interface: under mechanical stress, it inserts more deeply into the outer leaflet of the membrane, partially relieving lateral membrane tension and thus reducing the effective mechanical stimulus sensed by the channel gate.⁵⁴ It is typically used to investigate whether an observed phenomenon is mediated by SACs.⁵³ We therefore investigated whether GSMTx4 could attenuate the high preload–induced APD increase in Langendorff-perfused rabbits heart ex vivo.

We first optimized our Langendorff perfusion setup to reliably produce high preload–induced changes in QT and APD. Previously, we did not observe any electrical changes when increasing the preload to physiological levels of 15 or 30 mmHg.³³ However, motivated by the recent work of other groups, which applied up to double the normal volume in the LV of Langendorff-perfused rat hearts to elicit stretch-activated Purkinje-based ectopic triggers or investigate stretch-induced APD changes,^55,56 we opted to investigate similarly high preloads in the rabbit heart. Although these higher preload values (50 and 75 mmHg) exceed physiological levels, we reasoned that such elevations may be necessary to compensate for the ex situ experimental context. In the isolated Langendorff-perfused heart, the absence of autonomic innervation and thoracic constraints likely reduce the MEF sensitivity compared to the in vivo situation. Therefore, higher-than-physiological preload may be required to achieve a comparable degree of myocardial stretch and cellular response.

Using MAP measurements, primarily focusing on the RV (to mimic the in vivo situation of the i.v. bolus injection), we observed a significant increase in RV APD and QT duration at higher preload pressures of 50 and 75 mmHg, but not at the previously tested 15 and 30 mmHg. Therefore, we decided to continue our investigations on MEF-induced APD changes using an RV preload of 75 mmHg. This intervention consistently increased RV APD₉₀ and QT by around 10 ms. This stretch-induced prolongation of cardiac repolarization was attenuated in each individual heart to ∼5 ms by GSMTx-4, indicating that SACs are at least partially responsible for the MEF-induced changes in APD/QT. qPCR analysis of mRNA levels of three candidate SAC channel genes representing non-selective SAC_ns, potassium-selective SAC_K, and calcium-selective SACs, e.g. Piezo1, TRPM4, and KCNK2/TREK-1, confirmed that all these SACs are expressed in rabbit cardiomyocytes, yet no significant genotype-specific differences in expression levels were detected. Although confirmation in larger cohorts is needed, these findings support the concept that alterations in MEF in LQTS and SQTS may not primarily arise from differences in the expression of certain SACs. Rather, they are likely driven by differences in APD that affect the consequences of SAC activation under mechanical load. This interpretation aligns well with our computational results, which revealed that stretch-induced SAC activation strongly depends on the underlying electrophysiological substrate. Given the non-selectivity of GSMTx4, the residual, but small, MEF effect observed upon GSMTx4 perfusion likely stems from (i) incomplete blockade of SAC channels not affected by GSMTx4, (ii) MEF-induced activation of mechano-sensitive voltage-gated channels,²³ (iii) and stretch-dependent alterations in Ca²⁺ handling and myofilament buffering.⁵⁷

The role of stretch-activated channels in mediating mechano-electrical feedback-induced arrhythmogenesis: ex vivo and in silico synergy

Finally, with our study, we provide insights into the role of MEF-induced arrhythmogenesis in the pathophysiology of LQT2 and SQT1. Previous research has shown that mechanical stimuli, such as localized pressure, stretch, or changes in pre- or afterload, can trigger arrhythmias via the activation of SACs, thus provoking electrical disturbances and triggered activity.^56,58–60 Notably, blocking SACs using GSMTx4 has been shown to mitigate this MEF and reduce the risk of arrhythmogenesis.⁵⁹ However, these experiments are extremely challenging, particularly when involving SQTS and LQTS rabbit models, precluding a systematic characterization of all stretch-related parameters that may influence arrhythmogenesis, also in view of the 3R principles (replacement, refinement, and reduction) in animal experiments.

The computational model employed here is a valuable tool to corroborate and expand the experimental findings. Although our simulations were not designed to identify specific SACs involved in the modulation of cardiac electrophysiology, they demonstrated that SAC activation reproduced the experimentally observed phenotypes in WT, SQT1, and LQT2 and extended the findings by revealing distinct pro-arrhythmic responses to acute SAC activation. For instance, sustained and transient stretch simulations led to increased APD₉₀. This reflects the experimental findings of APD/QT prolongation upon sustained pressure applied in RV or the in vivo bolus injection to increase preload in WT, which cause myocardial stretch and SAC activation. Building on this, we extended the simulations to the SQT1 and LQT2 model variants and observed a more pronounced APD₉₀ prolongation in LQT2 compared to WT and SQT1 similar to the in vivo experiments, further validating the usefulness of these models to investigate stretch-induced arrhythmogenesis in the QT syndrome settings.

At the cellular level, all three model variants (LQT2, SQT1, and Control) exhibited stretch-triggered ectopic APs when supra-threshold, acute stretch was applied during phase 4 of the AP. However, stretch-induced EADs due to a prolonged plateau phase of the AP occurred only in the LQT2 model. Ectopic beats or EADs can disrupt normal pacing, potentially triggering arrhythmias and re-entries even in healthy myocardium. If they occur in specific regions, and at the onset of the T-wave in the ECG, the risk of ventricular fibrillation and sudden death increases.⁵³ The prolonged QT/APD in the LQT2 model further increases the duration of the vulnerable window, during which such mechanical stimuli can trigger EADs and consecutively induce ventricular tachycardia / fibrillation, indicating that LQTS patients may be particularly susceptible to MEF-induced arrhythmogenesis. Different effects on AP, such as minor depolarizations or APD prolongation, would arise for different stretch conditions (e.g. amplitude, timing, or duration changes), as highlighted by the ex vivo findings of Figure 4, showing the dependence on preload magnitude, and as previously characterized in detail in our computational model.³³

We thus went on to also test stretch-induced arrhythmogenesis (including re-entry generation) in all three model variants at the tissue level. We could show that stretch-induced re-entrant arrhythmias depend on stretch stimulus characteristics (amplitude, time of application, and duration) and spatial parameters (location, relative size of stretched area, and amplitude gradient), as well as on a genotypic-specific baseline substrate. Among the different combinations of stretch stimuli and spatial parameters, SQT1 and LQT2 were more arrhythmogenic than the Control model. Specifically, re-entry was sustained for smaller stretched areas in both QT syndrome conditions and for a longer time window in the SQT1 compared to Control facilitated by the shorter APD and, hence, a shorter path length. This increased vulnerability to stretch-induced re-entry formation may contribute to the clinically observed enhanced susceptibility to mechano-induced arrhythmias in SQTS during electrode placements.⁶¹

In LQT2, the prolonged plateau phase increased susceptibility to stretch-triggered EADs, which created larger obstacles in the 2D tissue due to the formation of refractory areas, thus extending the re-entry path, thereby facilitating re-entry despite the longer APD.²² Also, the stretch-induced AP, due to its prolonged duration compared to the other model variants, could persist long enough to prevent the occurrence of the following paced beat.

In summary, both SQT1 and LQT2 more readily exhibited stretch-induced arrhythmias than Control, indicating that MEF may indeed contribute to the pro-arrhythmic phenotype in LQTS and SQTS.

Limitations and outlook

Some limitations of our in silico approach have to be considered: stretch is indirectly simulated through the application of well-defined electrical modifications triggered by the activation of SACs, instead of altering tissue dimensions directly, thus ignoring potential changes in myofilament-related Ca²⁺ buffering. While the model was adapted to reflect APD characteristics in rabbits, the underlying ion currents and Ca²⁺ handling remain human-like since it is based on an extensively validated human ventricular electrophysiology model²⁴ incorporating three types of SACs.³⁰ The conduction velocity employed in our study was unphysiologically slow for human hearts without overt structural remodelling. Conduction velocity was chosen to enable predefined analyses of re-entry in a highly simplified 2D tissue of 5 × 5 cm with unconnected borders. In a fully connected 3D tissue, the re-entry path length could be longer, potentially allowing re-entry formation with faster conduction velocities while preserving the other parameter values currently represented in the model. Similarly, our simulations are based on APDs calibrated to single-cell data, which are longer than in vivo MAPs, thus contributing to the slow conduction velocities required to simulate re-entry in our 2D tissue. Furthermore, our 2D model assumes isotropic conduction in transverse and longitudinal direction of propagation. The complex 3D organization of cardiac tissue and spatial heterogeneity in electrical and mechanical properties in vivo may alter or amplify the conditions under which MEF induces arrhythmias.⁶³ Therefore, by neglecting anisotropy, re-entry occurrence might be underestimated. Nevertheless, genotype differences in re-entry development are likely preserved. As such, our simulation results should be interpreted qualitatively rather than quantitatively as they were primarily intended as a proof of concept to highlight the potential differences in MEF-related pro-arrhythmia between the three models.

While the increase in preload and afterload was performed using standardized procedures, abdominal compression was applied manually. Despite standardized position of the hands and comparable pressure application in all animals, minor variations in the mechanical stimulus may have contributed to some variability. However, as the procedure consistently produced a reproducible QTc shortening of similar magnitude in all animals, we assume that the mechanical stimulus was comparable across experiments.

Although the SAC-blocking ex vivo experiments on MEF-induced APD/QT changes were only performed in WT hearts, using a non-specific SAC blocker, this study represents a first step in our understanding of whether and how SACs mediate MEF-induced electrical and pro-arrhythmic changes in the QT syndromes.⁶⁴

Moving forwards, we aim to extend our investigation to LQT2 and SQT1 rabbits to elucidate potential genotype-specific differences in the impact of different SAC subtypes in mediating MEF-induced electrical changes. Rather than performing similar ex vivo whole-heart experiments as in the WT with the above-discussed limitations and the need for high numbers of animals or investigations of MEF-induced electrical changes in isolated cardiomyocytes, we are planning to employ cardiac slice preparations for these more in-depth future comparative investigations, which preserve the native tissue architecture with different cell types and local mechanical environment⁶² to more precisely investigate region-specific and genotype-specific SAC-mediated electrophysiological responses.

Additionally, we aim to elucidate which SACs are the most important in mediating these MEF effects. As our in silico models have three different types of SACs incorporated, e.g. SAC_ns, SAC_K, and Piezo1, this will enable us to further dissect the individual contribution of each of these channels to MEF-induced arrhythmogenesis in LQT2 and SQT1. This will help us to identify the most promising candidates for further experimental testing using specific blockers and activators.

Conclusions

This study highlights the complex electromechanical interplay determining the arrhythmogenic potential of LQT2 and SQT1 rabbit models. The sensitivity of LQT2 rabbits to preload and afterload alterations, the role of SACs in modulating repolarization, the increased susceptibility to stretch-induced arrhythmias in our in silico model, and the utility of EMW in predicting arrhythmic risk underscore the importance of considering both mechanical and electrical factors in understanding cardiac arrhythmias. The impaired calcium handling observed in LQT2 further emphasizes the link between electrical dysfunction and mechanical abnormalities.

Moving forwards, targeting SACs may offer promising therapeutic strategies for managing arrhythmias in patients with LQT2 and SQT1. By integrating these mechanistic insights, we can advance our understanding of arrhythmogenesis and develop more targeted interventions for patients at risk of life-threatening cardiac events.

Supplementary material

Supplementary material is available at Europace online.

Funding

This work was funded by the Swiss National Science Foundation (SNSF) grant 310030_197595 to K.E.O., Bern Center of Precision Medicine Lighthouse Project grant to K.E.O., and German Research Foundation (DFG) project #394630089 to K.E.O.

Data availability

Data are made available upon reasonable request.