Synchronous Systolic Subcellular Ca²⁺-Elevations Underlie Ventricular Arrhythmia in Drug-Induced Long QT Type 2

Abstract

Background

Repolarization-delay is a common clinical problem which can promote ventricular arrhythmias. In myocytes, abnormal sarcoplasmic reticulum Ca²⁺-release is proposed as the mechanism that causes early afterdepolarizations, the cellular equivalent of ectopic-activity in drug-induced long QT syndrome. A crucial missing link is how such a stochastic process can overcome the source-sink mismatch to depolarize sufficient ventricular tissue to initiate arrhythmias.

Methods and Results

Optical maps of action potentials (APs) and Ca²⁺-transients (CaT) from Langendorff rabbit hearts were measured at low (150×150 μm²/pixel) and high (1.5×1.5 μm²/pixel) resolution before and during arrhythmias. Drug-induced long QT type 2, elicited with dofetilide inhibition, produced spontaneous Ca²⁺-elevations during diastole and systole, before the onset of arrhythmias. Diastolic Ca²⁺⁻ waves appeared randomly, propagated within individual myocytes, were out-of-phase with adjacent myocytes and often died-out. Systolic secondary Ca²⁺⁻ elevations were synchronous within individual myocytes, appeared 188±30ms after the AP-upstroke, occurred during high cytosolic-Ca²⁺ (40–60% of peak-CaT), appeared first in small islands (0.5×0.5 mm²) that enlarged and spread throughout the epicardium. Synchronous systolic Ca²⁺-elevations preceded voltage-depolarizations (9.2±5ms, n=5) and produced pronounced Spatial Heterogeneities of CaT-durations and AP-durations. Early afterdepolarizations originating from sites with the steepest gradients of membrane-potential propagated and initiated arrhythmias. Interestingly, more complex subcellular Ca²⁺-dynamics (multiple chaotic Ca²⁺-waves) occurred during arrhythmias. K201, a ryanodine receptor stabilizer, eliminated Ca²⁺-elevations and arrhythmias.

Conclusions

The results indicate that systolic and diastolic Ca²⁺-elevations emanate from sarcoplasmic reticulum Ca²⁺-release and systolic Ca²⁺-elevations are synchronous because of high cytosolic and luminal-sarcoplasmic reticulum Ca²⁺ which overcomes source-sink mismatch to trigger arrhythmias in intact hearts.

Introduction

Delayed ventricular repolarization is frequently encountered in clinical practice. It is manifested by prolongation of QT interval on surface ECG and a propensity to potentially lethal polymorphic ventricular tachycardia.^{1, 2} Occasionally, the condition is congenital, caused by mutation(s) in one of the genes encoding cardiac ion channel subunits or auxiliary proteins.^3–7 More commonly, it is acquired,^8–10 caused by exposure to certain medications, electrolyte abnormalities or coronary ischemia. The mechanism linking action potential (AP) prolongation to ventricular arrhythmias remains incompletely understood.

At the single-cell level, early afterdepolarizations are observed as a result of AP prolongation.¹¹ It is believed that triggered activity caused by afterdepolarizations, along with increased dispersion of AP duration (APD), result in functional reentry and ventricular arrhythmias.¹² The precise sequence of events leading to early afterdepolarizations remains disputed. It was suggested that early afterdepolarizations are caused by the “window-current” of L-type Ca²⁺-channels during phase 3 of the AP, and the spontaneous reactivation of I_Ca,L.^{13, 14} However, subsequent compelling evidence showed that early afterdepolarizations are caused by Ca²⁺-release from overloaded sarcoplasmic reticulum,¹⁵ resulting in an elevation of intracellular Ca²⁺ (Ca_i) and subsequent augmentation of the Ca_i-dependent depolarizing current carried by the forward mode of the sodium-calcium exchanger.^{16, 17} Another related issue is whether early afterdepolarizations caused by sarcoplasmic reticulum Ca²⁺-overload and spontaneous Ca²⁺-release can overcome the source-sink mismatch to successfully propagate and initiate an arrhythmia in intact hearts. Synchronization of afterdepolarizations was experimentally demonstrated by a local injection of a high concentration of norepinephrine which elicited Ca²⁺-overload and a focal arrhythmia.¹⁸ Similar mechanisms cause delayed afterdepolarizations in digoxin toxicity or catecholaminergic polymorphic ventricular tachycardia.^{19, 20}

We have recently reported that development of secondary systolic Ca²⁺-elevations (SSCEs) precedes the development of ventricular arrhythmias by several minutes in a rabbit Langendorff heart with drug-induced repolarization-delay.²¹ In this model, SSCEs preceded membrane depolarizations at sites of ectopic foci. The temporal relationship between Ca_i and V_m and pharmacological manipulation supported sarcoplasmic reticulum Ca²⁺-release as the primary mechanism of early afterdepolarizations. However, the limited spatial resolution did not allow the detection of subcellular Ca²⁺ phenomena such as Ca²⁺-waves, which have been described in excitable²² and non-excitable²³ cells. Extensive studies have established that cytosolic Ca²⁺-elevation can elicit voltage depolarization by reverse Ca_i to voltage coupling and may contribute to arrhythmogenesis under specific circumstances. A critical challenge remains to explain how a stochastic process such as sarcoplasmic reticulum Ca²⁺-release can depolarize the cellular membrane sufficiently to produce a propagating early afterdepolarization and thereby initiate an arrhythmia. Here, we investigate Ca²⁺-dynamics at subcellular and the whole-heart resolution in perfused hearts (rather than dissociated cells) to gain new insights on the role of Ca²⁺-handling in eliciting afterdepolarizations and ventricular arrhythmias in rabbit model of drug-induced long QT type-2 (LQT2).

Methods

Langendorff preparation

Adult New Zealand White rabbits (females, 60 to 120 days old) were anesthetized with pentobarbital sodium (75 mg/kg, intravenously) and anticoagulated with heparin (200 U/kg intravenously). The heart was rapidly isolated, cannulated and perfused with Tyrode’s solution containing (in mM) 130 NaCl, 24 NaHCO₃, 1.0 MgCl₂, 4 KCl, 1.2 NaH₂PO₄, 50 dextrose, and 1.25 CaCl₂ (at pH 7.2–7.4), gassed with 95% O₂ plus 5% CO₂. Temperature was continuously monitored and regulated by a feedback system at 37 ± 2°C. Epicardial bipolar pseudo-ECG was continuously monitored.²⁴ The atrioventricular node was ablated by electrocautery to allow full control of heart rate by pacing.

This investigation conformed to the current Guide for Care and Use of Laboratory Animals, published by the National Institutes of Health, and was approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh.

Optical mapping

Dual optical mapping setups for macroscopic and microscopic measurements of CaTs and V_m in Langendorff perfused whole hearts were previously described.²⁵ Briefly, the isolated heart was perfused with blebbistatin (5–10 μM; Sigma, St Louis, MO) for 5–10 min to minimize motion artifact, immobilized in a chamber and stained with a voltage-sensitive dye (PGH1 or RH237: 200 μl of 1 mg/ml in dimethyl sulfoxide) and loaded with a Ca²⁺ indicator (Rhod-2 AM: 200 μl of 1 mg/ml in dimethyl sulfoxide). The anterior surface of the heart was illuminated with a 520 ± 30-nm excitation beam, and the fluorescence emitted by Rhod-2 and PGH1 was separated by a dichroic mirror (660 nm) and was focused on two complementary metal-oxide-semi-conductor cameras.²⁵ For macroscopic imaging, each camera viewed 1.5 × 1.5 cm² from the anterior surface of the heart using a Nikon camera lens (50 mm, 1:1.2). For subcellular imaging, the heart was perfused horizontally in a chamber on the stage of an upright microscope (Olympus BX61W1) using a ×40 water immersion objective (Olympus, ×40/LUMPLFL).²⁶ The sampling frequency was 500 Hz for macroscopic and 200–500 Hz for microscopic imaging.

Study protocol

The tolerance for motion artifact is much lower for high magnification compared to low-magnification imaging. To address this issue, we used a custom-designed plastic Lucite which has a heart-shaped cavity in the bottom, carved out of sylgard. Here, the heart is perfused and imaged in a horizontal position and is further stabilized by a plastic plate suspended over the heart by 4 metal bars fastened to the rim of the chamber. The pressure on the heart by the plate is adjustable by 4 spring-loaded screws. The epicardium is imaged from the top with an upright microscope where the 40× water-immersion objective accesses the heart through an oval opening on the compression plate allowing for the short working distance between the objective and epicardial surface. For these measurements, the heart was perfused continuously with a high concentration of blebbistatin (15 μM).

Simultaneous optical measurement of CaT and AP was performed at baseline and during drug-induced LQT2 conditions. Isolated whole hearts were paced with an epicardial unipolar electrode placed on the lateral wall of the right ventricle, approximately halfway between the apex and base. Hearts were paced at a cycle length of 0.5 sec (120 beats/minute), and subsequently slowed to 1.2 seconds (50 beats/minute; profound bradycardia for rabbit hearts). After baseline recordings, LQT2 was induced by perfusion of Tyrode’s solution containing dofetilide (500 nM, Pfizer, New York, NY), a selective I_Kr blocker, and by lowering K⁺ and Mg²⁺ concentrations by 50%.^{21, 26} In some experiments, 3-(4-benzylcyclohexyl)-1-(7-methoxy-2,3-dihydrobenzo[f][1,4]thiazepin-4(5H)-yl)propan-1-one (K201, 1 μM) was added to the perfusate before or after the induction of long QT type 2. K201 was a kind gift from Dr. Jonathan J. Abramson and Dr. Robert Strongin from Portland State University, Portland OR.

Data analysis

The amplitude of optically measured CaT and AP in each pixel was normalized between 0 and 1, and activation time at each site was calculated from maximum first derivative of the fluorescent signal [(dF/dt)_max] of the local AP or CaT upstroke. APD and CaT duration (CaTD) at each site were calculated from the interval between (dF/dt)_max and the recovery of V_m and CaT traces to 20% of baseline (APD₈₀ or CaTD₈₀), respectively. Automatic measurement of APD₈₀ and CaTD₈₀ from all pixels (100 × 100 pixels) was used to calculate mean APD₈₀ and CaTD₈₀. The dispersion of APD₈₀ was defined as the standard deviation of APD₈₀ measured from all pixels. The amplitude of the V_m gradient vector was defined as

$$|\mathrm{grad}(V\mathrm{m})|=\sqrt{{\left(\frac{\partial V\mathrm{m}}{\partial x}\right)}^2+{\left(\frac{\partial V\mathrm{m}}{\partial x}\right)}^2}$$

The gradients were calculated from discrete data (100 × 100 pixels) and spatial step of 3 pixel size was used to approximate the partial derivatives in the formula above. Standard deviation of V_m and CaT amplitude was considered a measure of spatial heterogeneity (SH) of the respective signals. It was calculated from all 100 × 100 pixels at each time and then averaged over repolarization time-interval (from 100 ms after action potential upstroke to the longest APD₈₀). SH values are reported as a natural logarithm of the time-averaged standard deviation. Phase angles of CaT and AP at each pixel were calculated based on modified phase analysis and used to identify the rise-time of CaT and AP. Briefly, after local maxima and local minima were defined at a given pixel, individual segments between neighboring local minima and maxima were normalized between 0 and π (rising phase) or –π and 0 (falling phase). Phase maps were used to visualize the spatiotemporal patterns of CaT/AP oscillation. Zero phase angles were used to identify the rise-times (take-off times) of Ca_i and V_m elevation during AP plateau. Correlation coefficient (r) between amplitudes of normalized CaT/AP at a given time or APD/CaTD was calculated by using MATLAB software. In microscopic imaging, changes in CaT and AP were analyzed by converting our two-dimensional data to stacked line scans along the individual cardiomyocyte as previously described. Wilcoxon signed-rank test (XLSTAT, Addinsoft) was used for statistical analysis and box-whisker diagrams displaying data range, interquartile range and median were used to visualize the distribution of the data.

Cell Isolation

Ventricular myocytes were isolated from female New Zealand white rabbits (10–12 weeks) with a standard protocol. Briefly, rabbits were euthanized as described above, the heart was excised and Langendorff-perfused with a Ca²⁺ containing Tyrode’s solution (containing (in mM): 140 NaCl, 5.4 KCl, 1 MgCl₂, 1.8 CaCl₂, 5 HEPES, 5.5 Glucose, at pH 7.4) for 5 min, then with a Ca²⁺-free Tyrode’s solution supplemented with 0.02% BSA for 10 min, after which collagenase type II (Worthington, 0.6 mg/ml) was added for a 15 min digestion at 37°C, followed by Ca²⁺-free Tyrode for 5 min. Hearts were removed and placed in a high K⁺, Kraft-Bruhe (KB) solution containing (in mM): 110 potassium glutamate, 25 KCl, 10 KH₂PO₄, 2 MgSO₄, 20 taurine, 5 creatine, 0.5 EGTA, 5 HEPES, 20 glucose, pH 7.4 with KOH. The sub-epicardium at the base of the left ventricle was excised, minced and myocytes were obtained by filtering through 200 μm nylon mesh. Cells were allowed to settle to form a pellet and washed with KB solution, then re-suspended in an incubation medium: M199 (free of phenol red) with 5% FBS (fetal bovine serum) and 100 μg/ml primocin (InvivoGen).^27–29

Electrophysiology

I_Ca,L was measured with whole-cell configuration of the patch-clamp technique. The pipettes were filled with (in mM): 130 CsCl, 20 tetraethylammonium chloride, 5 EGTA, 5 HEPES, 5 Mg-ATP, 0.1 Tris-GTP, 0.1 cAMP, 10 Creatine phosphate (pH 7.2 with CsOH). The external solution contained (in mM): 140 NaCl, 5.4 CsCl, 2.5 CaCl₂, 0.5 MgCl₂, 11 glucose, and 5.5 HEPES (pH 7.4 with NaOH). Currents were measured with an Axopatch 1D amplifier. I_Ca,L was measured from a holding potential of −80 mV and peak I_Ca-L was recorded during 100 ms voltage-clamp steps from −30 to +60 mV in 10 mV increments following a 40 ms pre-pulse at −30 mV which served to inactivate I_Na. Capacitance measurements were obtained from membrane test parameters. The current density (pA/pF) for each cell was normalized in relation to its capacitance.^{27, 28}

Results

Seven low-resolution experiments and a separate set of eight high-resolution experiments were analyzed. Ventricular tachycardia, either sustained or in the form of recurrent brief runs, eventually developed in all hearts under LQT2 conditions.

Diastolic and systolic Ca²⁺-dynamics before the onset of LQT2-related arrhythmias

During 2 Hz pacing, subcellular Ca²⁺-transients (CaTs) were smooth, monophasic and spatially homogenous (Fig. 1A, Movie A). Under LQT2 conditions, abnormal Ca_i elevations appeared in all hearts in systole and diastole during paced beats (Fig. 1B–C, Movies B–C), before the development of ectopy (Fig. 1D, Movie D). At the subcellular level, diastolic Ca²⁺-elevations corresponded to Ca²⁺-waves, propagating typically along the long axis of myocytes with a velocity of ~100–500 μm/s (n=8 hearts), and were not synchronous among neighboring cells. In a given cell, a single Ca²⁺-wave was typically observed per diastolic interval; occasionally 2–3 waves occurred consecutively and less frequently, more than one wave occurred simultaneously.

Figure 1.

Optical imaging of cytoplasmic Ca²⁺ at high resolution. Each of the square images displays color-coded intensity of Ca²⁺ signal from an epicardial region of ventricular myocardium 150×150 μm² in size. A: During baseline paced rhythm at 2 Hz, Ca²⁺ concentration is spatially homogenous and CaT tracings from two distinct pixels (a and b) are nearly superimposable. B: In paced rhythm under LQT2 conditions, a secondary systolic Ca²⁺ elevation occurs. The Ca²⁺ concentration remains homogenous within the field of view. C: Another example of slow pacing in LQT2 illustrates the contrast between the propagating diastolic Ca²⁺ wave (left and middle panels) and the synchronous systolic Ca²⁺ elevation (right). D: Multiple systolic peaks of CaT in LQT2 during arrhythmia are also spatially homogenous. See also Movies A–D.

In contrast to diastolic Ca²⁺-waves, SSCEs were cell-synchronous within myocytes when observed at subcellular-resolution and had similar amplitudes in neighboring myocytes. Most significant, SSCEs occurred at about the same time in myocytes in the field-of-view or 188±30 ms after the AP upstroke (n=5 hearts). Systolic and diastolic Ca_i events are more easily visualized in pseudo line-scan format (Fig. 2).

Figure 2.

High-resolution Ca²⁺ imaging displayed as pseudo line-scans in LQT2. A and B: The scan lines are identified by the red lines in the subcellular images in left most panels (field-of-view 150×150 μm). The diastolic Ca_i elevations (arrows) correspond to propagating waves, as evidenced by the oblique orientation of the corresponding pseudo line-scan feature. The systolic Ca_i oscillations are cell-synchronous, with vertical pattern on pseudo line-scan. C: An example of simultaneous subcellular Ca_i and V_m imaging. The pseudo line-scan again shows cell-synchronous systolic Ca²⁺ oscillations during the long CaTs. The simultaneous Ca_i and V_m tracings show that the corresponding V_m oscillations are less pronounced than CaT oscillations.

Spatial heterogeneity of CaT and APD in LQT2

In LQT2, SSCE amplitudes varied in different regions of epicardium when viewed at low-resolution (15×15 mm²) (Fig. 3A). Spatial variations of CaTs during paced beats (before ventricular arrhythmias) led to an increase in SH of CaT during phase 2 and 3 of APs compared to control (−4.56±0.67 vs. −3.28±0.45; p<0.05). SH of APDs also increased in LQT2 (−4.40±0.67 vs. −3.55±0.41; p<0.05), although to a significantly smaller degree (n=7, p<0.05) (Fig. 3B&E). SH of CaT and AP were low and similar at baseline, increased in LQT2 and the increase was higher for CaTD than for APD (AP: 0.85±0.71; CaT: 1.28±0.76; p<0.05) (Fig. 3E).

Figure 3.

Relationship between Ca²⁺ and V_m dynamics in LQT2 at low resolution. A: In paced rhythm under LQT2 conditions, the Ca_i and V_m maps (15×15 mm²) are similar, but the V_m map is substantially “smoother”. The images are taken 490 ms after pacing during regular paced rhythm. B: Superimposed CaT and AP from the 3 pixels (a–c) labeled in A. C: The scatterplot of Ca_i and V_m from different pixels at 490 ms shows that the signals are highly correlated. D: During AP phase 2 and 3, V_m and Ca_i signals remain well correlated except at the onset of SSCEs precedes the V_m rise during the AP plateau. See Movie E for another example. E: During baseline slow pacing (CTRL), Spatial Heterogeneity (SH) of both AP and CaT is low and similar. During LQT2 (paced rhythm), SH increases for both signals, but CaT heterogeneity is higher and its increase from baseline is more pronounced than for AP. * indicates p<0.05.

In LQT2, areas with elevated CaT correlated well with areas of depolarized V_m (membrane potential) at the time of SSCE (r=0.95±0.03, n=5; Fig. 3B, C–D; Movie E). The durations of AP (APD₈₀) and CaT (CaTD₈₀) were tightly correlated (r=0.96±0.01 at baseline, r=0.99±0.01 in LQT2; Fig. 1S). Areas of SSCE were irregular and did not correspond to obvious anatomical features. With appropriate contrast setting, initial Ca_i rise could be resolved in small “islands” of tissue (<1-mm) (Fig. 4D), which spread and fused during later plateau phase. SSCEs occurred before ectopy and their appearance was constant on beat-to-beat basis over a few seconds.

Figure 4.

Ca²⁺ dynamics in LQT2 at low resolution during paced rhythm. A: Field-of-view (15×15 mm) is selected on the anterior surface of the ventricles. B: Spatial heterogeneity (SH) of Ca²⁺ signal in LQT2 is evidenced by a different course of CaT in nearby pixels (a–c). The vertical lines (1–3) indicate the timing of the snapshots D1–D3. C: Superposition of AP and CaT from the same pixel show that the onset of the SSCE precedes the rise in membrane potential; the time delay, TD=14 ms in this example. D: SSCEs start to rise in several small (~ 0.5–1 mm) distinct areas (D1) that form isolated “islands” of Ca²⁺ elevation that spread and enlarge (panels D 2–3). The fluorescence intensity scale differs in these panels and in the corresponding Movie E to enhance the contrast of spatial heterogeneities of CaT on a millimeter-scale.

Propagated early afterdepolarizations arise in regions of steep V_m gradients

When the earliest ectopic depolarization occurred in the field-of-view, it typically corresponded to an early afterdepolarizations. Interestingly, the sites of earliest ectopy occurred in regions of steep spatial V_m gradients. At the origin of ectopic beats, V_m gradient was at 94.9±3.2 percentile (median 95.7, interquartile range 95.0–96.7; n=6 hearts) immediately before the ectopic beat (Fig. 2S). Mean gradient of CaT calculated from all pixels was steeper than mean gradient of V_m (0.28±0.05 vs. 0.20±0.04; p<0.05, n=6) immediately before the onset of ectopic activity. At the origins of early afterdepolarizations, SSCEs upstrokes preceded V_m upstroke by median 12.2 ms (interquartile range 2.8–14.6 ms).

Effect of RyR2 stabilization on SSCEs, SH of CaT, and ventricular arrhythmia

Addition of K201 (1 μΜ) to the perfusate either before or after switching to LQT2 blocked SSCEs, markedly reduced spatial heterogeneities of CaT and APs, and abolished ectopy (n=4/4) (Fig. 5A,B,C). Addition of K201 to dofetilide-induced LQT2 suppressed Torsade de Pointes (Fig. 5C) and in additional subcellular imaging experiments, K201 blocked SSCEs and diastolic Ca²⁺-waves, as well as the development of ventricular arrhythmia (n=4/5 hearts). In LQT2, K201 (1 μM) also caused a marked shortening of APDs (from 538.5±53 to 294.5±12, p<0.05) and CaTD (from 559±6 to 307.7±16.7, p<0.05). Besides its ability to stabilize RyR2, K201 was reported to have off-target effects and to inhibit I_Ca,L, I_Kr and SERCA pumps in a concentration dependent manner.³⁰ Voltage-clamp experiments showed that K201 at 1μM does not inhibit the L-type Ca²⁺ current (Fig. 5 D–E, n=5), whereas at 10 μM K201 suppressed I_Ca,L by > 20% (not shown) in agreement with previous reports.³⁰ To safeguard against possible off-target effects, additional pilot experiments were repeated with 0.5 μM K201 and showed that the lower concentration was equally effective at suppressing early afterdepolarization and Torsade de Pointes, n=3/3 (Figure 3S). Hence, the suppression of SSCEs by 1μM K201 was primarily attributed to inhibition of SR Ca²⁺ release since at that concentration K201 did not alter I_Ca,L, possible suppression of I_Kr would have prolonged rather than reduced action potential durations and inhibition of SERCA would reduce SR Ca²⁺ content and like RyR2 stabilization would suppress SSCEs.

Figure 5.

Effect of K201 on I_Ca,L. K201 suppresses systolic Ca²⁺ oscillations (A), dispersion of APD and CaTD (B) and ventricular arrhythmia (C). The data shown in panels A–C are from a low-resolution experiment. D&E: Rabbit ventricular myocytes were isolated from the base of the epicardium to measure I_Ca,L before and after 1μM K201. Current-to-voltage plots show that K201 did not alter I_Ca,L.

Ectopic beats due to delayed afterdepolarizations in LQT2

Diastolic and systolic Ca²⁺-elevations were observed in low-resolution experiments, and occurred in the same regions of the ventricles during paced beats (Fig. 4S). During ventricular arrhythmias, propagated delayed and early afterdepolarizations were observed in most experiments (Fig. 5S, traces). At the origin of delayed afterdepolarizations (black arrow-heads), Ca_i rose before V_m (Fig. 5S, green arrow-heads). This is supported by the phase-zero maps of Ca²⁺ and V_m signals, where phase zero of the signal corresponds to the onset of the signal upstroke. The timing is consistently earlier for Ca²⁺ than for V_m (n=5).

Patterns of subcellular Ca²⁺ handling during ventricular arrhythmias

In addition to stochastic diastolic waves and cell-synchronous SSCEs, more complex patterns of Ca²⁺-handling abnormalities were observed in ventricular arrhythmias by high-resolution imaging:

1. Multiple CaT spikes. During runs of non-sustained ventricular tachycardia, the first CaT rise did not fully recover to baseline, but was followed by a series of CaT spikes with sharp upstrokes. Similar to systolic Ca²⁺-elevations seen during paced rhythm, they were spatially synchronous in myocytes within the field-of-view (Fig. 1D, Movie D).

2. Ca²⁺ “ripples” superimposed on the plateau of long CaTs. We use the term “Ca²⁺ ripples” to denote the presence of multiple brief, small-amplitude Ca²⁺ elevations during prolonged elevated “plateau-like” CaTs, occurring in cell-asynchronous manner (as seen in Movie F). During episodes of non-sustained ventricular tachycardia, rectangular shaped CaTs lasting several seconds were often observed, with small amplitude Ca²⁺ ripples superimposed on the CaT (Fig. 6S, Movie F). Ca²⁺ ripples were not synchronized among adjacent myocytes and appeared to propagate within individual cells. Ripples can only be observed with high-resolution imaging, since these subcellular, regional signals will cancel themselves out if a pixel contains more than 1 cell. The phenomenon is mechanistically important as it represent an instability of sarcoplasmic reticulum Ca²⁺ uptake and release from multiple Ca²⁺ release units, since it is very difficult to imagine that I_Ca,L reactivation could drive ripples which occur out-of-phase in different regions of the same cell.

3. “Calcium fibrillation”. Chaotic Ca²⁺ dynamics was occasionally observed in a single cell or in all cells in the field-of-view during ventricular arrhythmias. These high frequency subcellular Ca_i oscillations (Fig. 6, Movie G) appeared as a persistent generation and annihilation of Ca²⁺ waves reminiscent of reentrant arrhythmias, though the scale and mechanism is obviously different.

Figure 6.

Random Subcellular Ca²⁺ oscillations during ventricular arrhythmia. Disorganized subcellular Ca²⁺ dynamics is occasionally observed under LQT2 conditions, with multiple Ca²⁺ wavefronts which collide, break and annihilate each other (bottom, 1–5). Ca²⁺ signal tracings (top right) from 3 pixels indicated in high-resolution image (top left, 150×150 μm² image size) illustrate the lack of spatial synchrony. Stable Ca²⁺ baseline is not reached in any cell in the field-of-view, and the signals from different regions of the same cell are poorly correlated. This may reflect an extreme degree of cellular Ca²⁺ overload. See also Movie G.

Discussion

The mechanism of arrhythmogenesis in LQT2 is disputed. One hypothesis posits that the decrease in repolarization rate allows the L-type Ca²⁺ channel to occupy a state where neither the activation nor the inactivation gates are completely closed, enhancing the I_CaL “window current”, which overwhelms the repolarizing currents and gives rise to early afterdepolarizations.¹⁴ In support of this mechanism, Marban et al.,³¹ reported that the chelation of cytosolic free Ca²⁺ with BAPTA and the depletion of sarcoplasmic reticulum Ca²⁺ with ryanodine failed to prevent early afterdepolarizations, but they were prevented by an L-type Ca²⁺ channel blocker. These data obtained in papillary muscles bathed in a cesium containing Tyrode’s solution as a model of LQT2 suggested that early afterdepolarizations are elicited by the I_CaL window current and are independent of cytosolic Ca²⁺ handling (assuming BAPTA chelates Ca_i in all myocytes) and sarcoplasmic reticulum Ca²⁺ overload.

DeFerrari et al.,³² loaded isolated cardiac myocytes with a Ca²⁺ indicator dye to investigate the mechanisms that induce delayed and early afterdepolarizations during β-adrenergic stimulation. Despite low temporal resolution, cytosolic Ca²⁺ waves were found to correspond to delayed afterdepolarizations and secondary Ca²⁺ elevations to early afterdepolarizations. The latter appeared to be synchronous within the cell rather than a random Ca²⁺ release from sarcoplasmic reticulum Ca²⁺-release units, consistent with a process driven by a voltage-dependent sarcolemmal channel such as I_CaL. Similar systolic secondary CaT elevations were recently described in mouse ventricular myocytes with increased sarcoplasmic reticulum load; in contrast to diastolic waves, they could not be easily suppressed with verapamil.³³

In contrast, Volders et al.,¹⁶ argued that both delayed and early afterdepolarizations are caused by spontaneous intracellular Ca²⁺ release in isolated canine myocytes exposed to bradycardia and β-adrenergic stimulation. This interpretation was based on the timing of myocyte contractions and Ca²⁺ release that preceded the onset of early afterdepolarizations, and on dissociation of Ca²⁺ release from early afterdepolarizations by sodium-calcium exchange blockade. A similar finding was obtained in isolated ventricular rabbit myocytes, where we elicited early afterdepolarization by I_Kr blockade which were suppressed by sodium-calcium exchange inhibition,²⁹ implicating SR Ca²⁺ release as the underlying mechanism. Choi and Salama²⁴ proposed that sarcoplasmic reticulum Ca²⁺ release elicited early afterdepolarizations induced by I_Kr blockade in isolated rabbit hearts because the upstroke of CaTs preceded the upstroke of early afterdepolarizations at the sites of early afterdepolarization foci. More recently, we reported that in LQT2, abnormal Ca²⁺ handling and secondary CaT elevations precede development of early afterdepolarizations and voltage-instabilities during the AP plateau by minutes, and that multiple Ca²⁺ re-elevations occurred during regular paced rhythm even before the onset of polymorphic ventricular tachycardia. In contrast to previous reports, depletion of sarcoplasmic reticulum Ca²⁺ with ryanodine and thapsigargin prevented secondary CaT oscillations, early afterdepolarizations and polymorphic ventricular tachycardia, in perfused rabbit hearts.²¹ In acute bradycardia without I_Kr block, the repolarization delay prolonged CaT durations resulting in SSCEs of sufficient magnitude to elicit early afterdepolarizations.²⁶

Finally, Horvath et al.,³⁴ observed SSCEs in isolated ventricular myocytes treated with ATX-II (LQT3 model). Both SSCEs and early afterdepolarizations were suppressed by chelation of cytosolic Ca²⁺ with BAPTA, supporting the role of SSCE in the generation of early afterdepolarizations in that model.

The data reported here demonstrate that in rabbit heart perfused at physiological temperature, the secondary Ca²⁺ rise corresponding to early afterdepolarization is indeed cell-synchronous. However, Ca²⁺ release from overloaded sarcoplasmic reticulum rather than I_Ca,L reactivation is responsible, since SSCEs are suppressed by K201 (1 μΜ). At this concentration, K201 has no effect on I_Ca,L.³⁰ Given the importance of the latter result, we repeated the voltage-clamp measurements of I_Ca,L from rabbit base myocytes and demonstrate that K201 at 1 μΜ, does not alter I_Ca,L. Hence, K201 abolished early afterdepolarizations by stabilizing ryanodine receptors and not by inhibiting I_Ca,L. In addition, the rise of CaT consistently precedes the rise of V_m at the site of the ectopic focus and multiple Ca²⁺-elevations are frequently observed during a single AP. In contrast to SSCEs, diastolic Ca²⁺ waves in LQT2 were not synchronous within or between adjacent cells. The diastolic waves and were typically of lower amplitude than SSCEs and more difficult to detect at low resolution. Nevertheless, K201 eliminated both SSCEs and diastolic Ca²⁺ waves, suggesting a common mechanism of SR Ca²⁺ release in both phenomena.

Different experimental setting may explain why others reached a different conclusion regarding mechanism of early afterdepolarizations. We performed experiments in perfused heart at physiological temperature, as opposed to single-cells or superfused tissue. We used a highly selective I_Kr inhibitor to elicit early afterdepolarization at a high concentration, dictated by the need to reliably induce Torsade de Pointes within minutes due to the life-time of perfused hearts (3–4 hours). We acknowledge that the details of arrhythmogenesis may differ in patients, who may experience a single Torsade episode in their lifetime. Non-confocal microscopy was used to improve light throughput and temporal resolution. The resulting decrease in depth resolution to about 5 μm improves the detection of Ca²⁺ waves and is still considerably less than the myocyte thickness.

Diastolic Ca²⁺ waves and delayed afterdepolarizations in LQT2

Our data indicate that diastolic Ca²⁺ waves classically associated with sarcoplasmic reticulum Ca²⁺ overload and delayed afterdepolarizations also occur in LQT2 and may contribute to the initiation of arrhythmias. Ectopic beats occurring during diastole are preceded by a rise of Ca²⁺ (Figure 4S). Epicardial regions exhibiting diastolic Ca²⁺ waves and SSCEs are well correlated (Fig. 3S), suggesting that sarcoplasmic reticulum Ca²⁺ overload triggers both SSCEs and diastolic waves. Visualization of diastolic Ca²⁺ waves in LQT2 requires high spatial and temporal resolution imaging; their role in LQT2 arrhythmogenesis may thus have been underappreciated.

Synchronization of seemingly random diastolic Ca²⁺ waves in individual cells is required to overcome the source-sink mismatch and to elicit an ectopic beat. A mechanism of synchronization was demonstrated during a pause following rapid pacing in a rat heart perfused with elevated extracellular Ca²⁺ at room temperature.³⁵ Despite questions regarding the clinical relevance of the model, the mechanism proposed by Wasserstrom, i.e. increasing degree of temporal overlap among diastolic waves in adjacent cells with increasing sarcoplasmic reticulum Ca²⁺ load, may well apply to LQT2.³⁵

K201 effect on APD

The precise mechanism for the striking shortening of CaTD and APD by K201 in LQT2 is uncertain. Our published data suggest that in the absence of dofetilide, K201 has little effect on APD during 2 Hz pacing in rabbit hearts, but partially prevents APD prolongation in right ventricular base.²⁶ Given the lack of effect of K201 on I_CaL in the rabbit heart at this concentration,³⁰ and lack of effect on late sodium current,³⁶ it is likely that K201-dependent shortening of APD in LQT2 could be mediated by RyR2 stabilization. It is possible that a complete I_Kr block initially prolongs APD only to a moderate degree, which is nevertheless sufficient to increase myocyte Ca²⁺ load and SSCEs. Higher cytosolic Ca²⁺ during phase 3 of the AP caused by SSCE would augment I_NCX, slow the AP downstroke and increase Ca²⁺ influx via I_CaL. Under certain conditions, the increase in Ca²⁺ influx during phase 3 might exceed the increase in Ca²⁺ efflux via NCX. In such a case, APD would increase gradually with increasing Ca²⁺ load, even during full I_Kr block. Hence, it is possible that K201 reduces APDs in LQT2 by inhibiting SSCEs. Additional experiments will be required to address this hypothesis directly.

Mechanism of arrhythmogenesis in LQT2

Based on the above considerations, we propose the following sequence of events to explain the role of Ca²⁺ dynamics in LQT2-related arrhythmias:

1. Prolonged APD results in increased Ca²⁺ entry into myocyte because I_Ca,L remains partially activated during prolonged plateau phases and because Ca²⁺ efflux via sodium-calcium exchange is impaired due to shorter diastolic intervals.

2. The initial CaT upstroke is caused by Ca²⁺-induced Ca²⁺-release, and empties the junctional sarcoplasmic reticulum. Ca²⁺ is subsequently pumped to the non-junctional sarcoplasmic reticulum by SERCA, and diffuses towards the junctional sarcoplasmic reticulum, increasing the Ca²⁺ load in junctional sarcoplasmic reticulum and eliciting a second reopening of ryanodine receptors during the AP plateau phase.

3. The open probability of cardiac ryanodine receptor is known to increase with both higher cytoplasmic and luminal Ca²⁺ concentrations.^37–39 The high cytosolic Ca²⁺ during the AP plateau in LQT2 may explain why systolic, but not diastolic sarcoplasmic reticulum Ca²⁺-release is cell-synchronous. We hypothesize that when junctional sarcoplasmic reticulum is reloaded before cytoplasmic Ca²⁺ recovers to baseline, the threshold of luminal Ca²⁺ needed to re-open cardiac ryanodine receptors is low. Reloading the terminal cisternae may occur in most junctions at approximately the same time. In contrast, the luminal threshold for release would be higher during diastole, and reached in only one or a few regions of a myocyte.

4. The amplitude of SSCEs differs in different regions, resulting in marked spatial heterogeneity of CaT. The reasons for this large-scale spatial heterogeneity are uncertain and likely include spatial variations in expression of Ca²⁺ handling proteins.²⁶

5. Spatial heterogeneity of CaT drives the spatial heterogeneity of AP by means of the electrogenic sodium-calcium exchanger, i.e. regions of higher cytoplasmic Ca²⁺ during AP phases 2 and 3 experience higher inward sodium-calcium exchange current.

6. Steep spatial gradients of V_m cause cell membrane depolarization and trigger propagated early afterdepolarizations,⁴⁰ perhaps through I_CaL reactivation.

7. Chaotic intracellular Ca²⁺ dynamics may reflect a marked degree of myocyte Ca²⁺ overload, with failure of cytoplasmic Ca²⁺ to reach a low concentration in the entire cell even transiently. The myocyte or ventricular region with such dynamics would be expected to remain at least partially depolarized and serve as a source of depolarizing current for adjacent cells. Clearly, I_CaL window current cannot account for this form of Ca²⁺ dynamics.

In summary, we describe Ca²⁺ dynamics in LQT2 at whole-heart and subcellular resolutions. The abnormal Ca²⁺ handling appears to involve store overload-induced Ca²⁺ release and directly participates in the generation of triggered activity. Compared to acute bradycardia without I_Kr block, well-developed secondary CaT peaks were observed in all LQT2 experiments and triggered early afterdepolarizations, likely reflecting a higher degree of sarcoplasmic reticulum load.

Improved understanding of arrhythmogenesis due to delayed repolarization may lead to novel therapeutic and diagnostic strategies. For example, it is possible that targeting SSCEs may be protective with respect to polymorphic ventricular tachycardia in the clinical setting. Spatial heterogeneity of CaT is the likely mechanism of regional heterogeneity in ventricular contraction timing detected in LQT2 patients with echocardiography.⁴¹ More generally, our findings underscore the role of store overload-induced Ca²⁺ release in generation of triggered activity, which is not limited to delayed afterdepolarizations.