The Calcium/Calmodulin/Kinase System and Arrhythmogenic Afterdepolarizations in Bradycardia-Related Acquired Long-QT SyndromeCLINICAL PERSPECTIVE
Background— Sustained bradycardia is associated with long-QT syndrome in human beings and causes spontaneous torsades de pointes in rabbits with chronic atrioventricular block (CAVB), at least partly by downregulating delayed-rectifier K+-current to cause action potential (AP) prolongation. We addressed the importance of altered Ca2+ handling, studying underlying mechanisms and consequences.
Methods and Results— We measured ventricular cardiomyocyte [Ca2+]i (Indo1-AM), L-type Ca2+-current (ICaL) and APs (whole-cell perforated-patch), and Ca2+-handling protein expression (immunoblot). CAVB increased AP duration, cell shortening, systolic [Ca2+]i transients, and caffeine-induced [Ca2+]i release, and CAVB cells showed spontaneous early afterdepolarizations (EADs). ICaL density was unaffected by CAVB, but inactivation was shifted to more positive voltages, increasing the activation-inactivation overlap zone for ICaL window current. Ca2+-calmodulin–dependent protein kinase-II (CaMKII) autophosphorylation was enhanced in CAVB, indicating CaMKII activation. CAVB also enhanced CaMKII-dependent phospholamban-phosphorylation and accelerated [Ca2+]i-transient decay, consistent with phosphorylation-induced reductions in phospholamban inhibition of sarcoplasmic reticulum (SR) Ca2+-ATPase as a contributor to enhanced SR Ca2+ loading. The CaMKII-inhibitor KN93 reversed CAVB-induced changes in caffeine-releasable [Ca2+]i and ICaL inactivation voltage and suppressed CAVB-induced EADs. Similarly, the calmodulin inhibitor W7 suppressed CAVB-induced ICaL inactivation voltage shifts and EADs, and a specific CaMKII inhibitory peptide prevented ICaL inactivation voltage shifts. The SR Ca2+-uptake inhibitor thapsigargin and the SR Ca2+ release inhibitor ryanodine also suppressed CAVB-induced EADs, consistent with an important role for SR Ca2+ loading and release in arrhythmogenesis. AP-duration changes reached a maximum after 1 week of bradypacing, but peak alterations in CaMKII and [Ca2+]i required 2 weeks, paralleling the EAD time course.
Conclusions— CAVB-induced remodeling enhances [Ca2+]i load and activates the Ca2+-calmodulin-CaMKII system, producing [Ca2+]i-handling abnormalities that contribute importantly to CAVB-induced arrhythmogenic afterdepolarizations.
Received August 19, 2008; accepted April 8, 2009.
Torsade de pointes (TdP) is a polymorphic ventricular tachycardia associated with QT-interval prolongation and sustained bradycardia, potentially leading to sudden cardiac death.1 Action potential duration (APD) prolongation is ubiquitous in TdP syndromes, and the importance of APD prolongation and transmural APD heterogeneity are well recognized.2 Rabbits and dogs with sustained bradycardia caused by chronic atrioventricular block (CAVB) are prone to TdP,3,4 occurring spontaneously in rabbits3 and on IKr-blocking challenge in dogs.4 Recent studies in rabbits with CAVB-related TdP emphasized the importance of downregulation of subunits underlying rapid (IKr) and slow (IKs) components of the delayed-rectifier system in APD prolongation and spontaneous TdP generation.3,5 Delayed-rectifier K+-current downregulation is also central in the CAVB dog model.6 Early afterdepolarizations (EADs) are crucial for the generation of TdP arrhythmias.1,2 There is evidence that Ca2+-handling abnormalities contribute to the generation of EADs7 and that Ca2+-related signaling promotes EADs in hypertrophic mouse models via Ca2+-calmodulin dependent kinase-II (CaMKII).8 In studies of ventricular cardiomyocytes from CAVB rabbits, we noted that cellular contractility increases, suggesting alterations in cell Ca2+ handling. The present study was therefore designed to (1) assess potential Ca2+-handling abnormalities in CAVB rabbits; (2) define potential underlying mechanisms; and (3) evaluate a possible role in arrhythmogenic afterdepolarizations.
Clinical Perspective on p 295
CAVB Rabbit Model
All animal-handling protocols were approved by the Montreal Heart Institute animal research ethics committee. Rabbits were anesthetized with ketamine/xylazine, intubated, and ventilated with room air supplemented with oxygen. CAVB was induced by injecting 37% formalin into the atrioventricular junction. A unipolar pacing lead fixed to the right ventricular free wall was connected to a subcutaneous pacemaker. Ventricular pacing at 110 bpm was instituted immediately after creating CAVB to allow recovery from surgery. One week after surgery, the pacemaker rate was adjusted to 90 bpm for 1 or 2 additional weeks.
Rabbits were euthanized by cervical dislocation. Hearts were removed, immersed in normal Tyrode solution, and perfused retrogradely (36° C, 100% O2-saturated solutions) through the aorta with Tyrode solution containing 200 μmol/L Ca2+ for 3 to 5 minutes, then with Ca2+-free Tyrode solution for 5 minutes, followed by Ca2+-free Tyrode solution containing collagenase (0.8 mg/mL CLS II, Worthington Biochemical) for ≈40 minutes. The hearts were subsequently washed with 200 μmol/L Ca2+-containing Tyrode solution for 2 to 3 minutes, and left ventricular cells dispersed by trituration. Isolated cardiomyocytes were stored in 200 μmol/L Ca2+-containing Tyrode solution. Only Ca2+-tolerant rod-shaped cells with clear cross-striations and without spontaneous contractions were used for experiments.
Whole-Cell Perforated Patch
Whole-cell perforated-patch methods were used to record action potentials (APs) in current-clamp and ICaL in voltage-clamp mode. ICaL measurements in the presence and absence of the peptide CaMKII inhibitor AC3-I and its inactive congener AC3-C were performed with tight-seal patch-clamp to permit intracellular dialysis. Borosilicate glass electrodes (Sutter Instruments) filled with pipette solution were connected to a patch-clamp amplifier (Axopatch 200A, Axon) and had tip resistances of 2 to 4 MLΩ. Nystatin-free intracellular solution was back-filled into pipette tips by capillary action (30 seconds), and pipettes were then filled with nystatin-containing (600 μg/mL) internal solution. Cell capacitance and series resistance (Rs) were compensated by 75% to 85%. Rs was calculated by dividing the capacitive-transient decay time constant by the calculated membrane capacitance. Capacitance was assessed by integrating current elicited by 5-mV, 10-ms hyperpolarizing steps from a holding potential of −60 mV and dividing by the voltage drop. Before compensation, Rs averaged 6.3±0.9 MΩ, and capacitive time constants averaged 710±56 μs. Mean Rs and capacitive time-constant values after compensation averaged 1.8±0.3 MΩ and 119±9 μs. Leakage compensation was not used. Currents are expressed as densities (pA/pF). Junction potentials between the bath and pipette solution averaged 15.9 mV and were corrected for APs only. All recordings were performed at 35±0.5°C.
Tyrode solution contained (mmol/L): NaCl 136, CaCl2 1.8, KCl 5.4, MgCl2 1, NaH2PO4 0.33, dextrose 10, and HEPES 5 (pH 7.4; NaOH). Li+ was used to replace Na+ for Na+-free Tyrode solution. The pipette solution for AP recording contained (mmol/L): GTP 0.1, potassium aspartate 110, KCl 20, MgCl2 1, ATP-Mg 5, HEPES 10, Na-phosphocreatine 5, and EGTA 0.05 (pH 7.4; KOH). The extracellular solution for ICaL measurement contained (mmol/L): n-methyl-d-glucamine 137.0, CsCl 25.0, HEPES 10.0, glucose 10.0, CaCl2 1.8, MgCl2 0.5 (pH 7.4; HCl). Niflumic acid (50 μmol/L) was added to inhibit Ca2+-dependent Cl− current, and 4-aminopyridine (2-mmol/L) was added to suppress transient outward K+-current. The pipette solution for ICaL recording contained (mmol/L): CsCl 120, tetraethylammonium-chloride 20, MgCl2 1, EGTA 20 (EGTA 20 for perforated-patch or 0 for tight-seal studies), ATP-Mg 5, HEPES 10, and Li-GTP 0.1 (pH 7.4; CsOH). AC3-I and AC3-C concentrations in pipette solutions were 20 μmol/L.
Cell Shortening, Ca2+ Transients, and AP Clamp
Isolated cardiomyocytes were field-stimulated by 10 ms 1.5×threshold square-wave pulses delivered through 2 platinum electrodes separated by 2 cm in the experimental chamber. Cell shortening was measured relative to diastolic cell length with a video edge detector (Crescent Electronics) coupled to a charged-coupled device camera (digitization at 200 Hz, TL-1 A/D Converter, Axon). Edge-detection cursors were positioned at both cell ends to measure whole-cell shortening.
To record Ca2+ transients, ventricular cardiomyocytes were incubated with Indo-1 AM (5 μmol/L, Molecular Probes) in 100 μmol/L pluronic F-127 (Molecular Probes) and 0.5% DMSO (Sigma) for 3 to 5 minutes, then superfused with Tyrode solution at 36±1°C for at least 20 minutes to wash out extracellular dye and allow for deesterification. Ultraviolet light from a 100-W mercury arc lamp passing through a 340-nm interference filter (±10 nm bandwidth) was reflected by a dichroic mirror into a ×40 oil-immersion fluor objective for excitation of intracellular Indo-1 (excitation beam ≈15 μm diameter). Exposure to ultraviolet light (5 to 10 seconds of every 30 to 60 seconds) was controlled by an electronic shutter (Optikon, model T132) to minimize photobleaching. Emitted light (<550 nm) was reflected into a spectral separator, passed through parallel filters at 400 and 500 nm (±10 nm), detected by matched photo-multiplier tubes (Hamamatsu R2560 HA) and electronically filtered at 60 Hz. The ratio of fluorescence signals (R400/500) was digitized (1 kHz) and recorded, with subsequent conversion to [Ca2+]i values as previously described.9 Before each measurement, background fluorescence was removed by adjusting the 400- and 500-nm channels to zero over an empty field of view near the cell.
The AP voltage-clamp technique was used to study AP waveform-dependent effects on Ca2+ transients. Voltage waveforms recorded at 2 Hz in control and 2-week CAVB rabbits were applied under perforated-patch conditions to control cardiomyocytes for 2 minutes each at 2 Hz from a holding potential of −80 mV, and Ca2+ transients were recorded as described above. Control and CAVB waveforms were applied in randomized order to control for time-related changes.
Left and right ventricular tissue samples were fast-frozen in liquid nitrogen and stored at −80°C. Tissue samples were homogenized in RadioImmuno Precipitation Assay (RIPA) buffer as previously described.10 The homogenate was centrifuged (15 000 rpm, 20 minutes, 4°C). The supernatant was used for protein concentration measurement by Bradford assay (Bio-Rad) with bovine serum albumin (BSA) as a standard. Protein samples (10, 20, or 40 μg) were denatured with Laemmli buffer and fractionated on 6%, 8%, or 15% SDS-polyacrylamide gels, then transferred electrophoretically to Immobilon-P polyvinylidene fluoride (PVDF) membranes (Millipore) in 25 mmol/L Tris-base, 192 mmol/L glycine, and 20% methanol at 0.3 A for 1 hour or overnight at 4°C. Membranes were blocked in Tris-buffered saline containing 0.1% Tween-20 (TBS-T) with 5% nonfat dry milk for 1 hour and incubated with primary antibodies (anti-SERCA2a 1:2500, anti-NCX 1:2500, Affinity BioReagents [ABR]; antiphospho-CaMKII, 1:5000, Promega; anti-CaMKIIδ, 1:1000, Santa Cruz; anti-calsequestrin, 1:2500, ABR; antiphospho-phospholamban-thr17, 1:5000, Badrilla; antiphospho-phospholamban-ser16, 1:2000, Badrilla; anti-phospholamban total, 1:10,000, ABR; anti-ryanodine receptor, 1:1000, ABR) overnight at 4°C. After washing and reblocking, membranes were incubated with horseradish peroxidase–conjugated goat anti-rabbit or donkey anti-goat IgG secondary antibody (1:10 000, Jackson) or goat anti-mouse IgG secondary antibody (1:10 000, Santa Cruz). Antibodies were detected with Western-Lightning Chemiluminescence Reagent Plus (Perkin-Elmer Life Sciences). Later, the same membranes were also probed with anti-GAPDH (1:15 000, ABR) at room temperature for 2 hours to control for protein loading. Band intensities for all proteins studied were normalized to GAPDH intensity on the same lane/blot.
Clampfit 9.2 (Axon), GraphPad Prism 4.0, IgorPro5.04B, SPSS and Origin 5.0 were used for data analysis. All continuous data are expressed as mean±SEM. Comparisons involving more than 2 groups were obtained by 2-way ANOVA, 2-way ANOVA for repeated-measures, and 1-way ANOVA as appropriate. Student t tests were used for comparisons involving only 2 groups, and t tests with Bonferroni correction were used to compare individual group differences when multiple-comparison ANOVA was significant. Categorical data were assessed by Fisher exact test. A 2-tailed P<0.05 was considered statistically significant. The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.
APs and Cellular Contractility
CAVB cardiomyocytes were hypertrophied after 2-week (but not 1-week) bradypacing, with a 24% increase in mean cell capacitance (Figure 1A). Figure 1B shows examples of AP recordings. CAVB did not change resting potential (−88±1, −85±4, −86±2 mV for control, 1-week, 2-week CAVB, respectively) or AP amplitude (129±2, 128±2, 126±2 mV respectively) but significantly delayed repolarization (Figure 1C). APD values reached their maximum at 1-week CAVB, remaining significantly increased relative to control at 2-week CAVB. Figure 1D shows recordings of cell shortening at 1 Hz, with corresponding mean data in Figure 1E. Cell shortening did not change appreciably with 1-week CAVB but a statistically significant ≈65% increase was seen at 2 weeks.
Changes in Ca2+ Transients
Figure 2A through 2C shows examples of steady-state Ca2+ transient recordings. CAVB greatly increased Ca2+ transient systolic levels after 2-week bradypacing, with much less effect at 1 week. Diastolic [Ca2+]i changed relatively little (Figure 2D). Total [Ca2+]i transient amplitude changes increased slightly but significantly at 1 week, with striking changes for 2-week CAVB (Figure 2E). The Ca2+ transient decay time constants (τs) obtained with monoexponential curve fits were significantly decreased by 2-week CAVB (Figure 2F), indicating accelerated Ca2+ transport out of the cytosol.
To obtain a more direct assessment of sarcoplasmic-reticulum (SR) Ca2+ load changes, we measured local caffeine (10 mmol/L) puff-induced Ca2+ release (caffeine Δ[Ca2+]i) with the use of a linear-flow rapid delivery system. Figure 3A shows examples of caffeine-induced Ca2+ transients in control and CAVB cardiomyocytes. CAVB significantly increased caffeineΔ[Ca2+]i by >2-fold at 2 weeks (Figure 3B). The decay time constant of the caffeine-induced Ca2+ transient, which is dominated by the Na+,Ca2+-exchanger (NCX), was similar in control and CAVB cells (Figure 3C).
Potential Mechanisms of CAVB-Induced Ca2+ Loading
The results shown in Figures 2 and 3⇑ document increased Ca2+ loading in 2-week CAVB rabbit cardiomyocytes. We therefore addressed potential underlying mechanisms. The first mechanism that we considered was enhancement of ICaL conductance. Figure 4A shows ICaL recordings from control and 2-week CAVB. There are no apparent differences, an impression supported by similar mean ICaL densities over the full test-pulse voltage range, as shown in Figure 4B. Thus, increased CAVB cell Ca2+ loading was not due to increased ICaL conductance. Closer study of ICaL voltage dependence (Figure 4C) provided evidence for a potentially significant, albeit subtle, change in ICaL properties with CAVB. ICaL inactivation voltage dependence was determined with 1000-ms prepulses to voltages between −100 and +30 mV, followed by 300-ms test pulses to 0 mV. Activation voltage dependence was obtained from data obtained as illustrated in Figure 4B, based on the relation Iv=Imax(V−Vr)(Gv/Gmax), where Iv and Gv are current and conductance at voltage V; Imax and Gmax are maximum current and conductance; and Vr is the reversal potential. Vr was determined from the horizontal axis intercept of the ascending limb of the ICaL-voltage relation. ICaL activation voltage was unaffected by CAVB, but ICaL inactivation V1/2 was positively shifted in 2-week CAVB cells (Figure 4C). Inactivation V1/2 averaged −25.5±1.1 mV (n=17) in control versus −21.3±0.9 mV in CAVB (n=11, P<0.05). Consequently, the inactivation-activation overlap “window area” for ICaL was substantially increased in CAVB cells (checked area under the corresponding curves in Figure 4C) compared with controls (stippled area). The time constants obtained by biexponential curve-fits of ICaL inactivation were not altered by CAVB (supplemental Figure IA). Time-dependent ICaL recovery was assessed with a paired-pulse protocol (supplemental Figure IB), providing time constants that were not significantly different between control and CAVB cells (supplemental Figure IC).
We then turned our attention to the potential impact of APD prolongation on Ca2+ homeostasis in CAVB rabbits. Figure 5A shows Ca2+ transients from control rabbit cardiomyocytes that were exposed alternately and in randomized order to a series of 2-Hz AP-clamp waveforms recorded from control and 2-week CAVB cardiomyocytes at the same frequency (inset). Application of the CAVB waveform led to significantly increased Ca2+ transients, as indicated in Figure 5B, without affecting diastolic [Ca2+]i or the time constant of [Ca2+]i decay. These results suggest that CAVB-induced increases in Ca2+ entry related to APD prolongation probably contribute to the Ca2+-loaded state of CAVB cardiomyocytes.
We then considered alterations in the expression and/or phosphorylation of key Ca2+-handling proteins as explanations for the Ca2+-loaded state of CAVB cardiomyocytes. Figure 6A shows immunoblots for a number of important proteins that regulate cellular Ca2+ handling, separated according to left ventricular and right ventricular findings. Corresponding mean data for 5 individual hearts per group are shown in Figure 6B. CAVB hearts showed significantly increased CaMKII autophosphorylation, without any significant change in total CaMKII expression. Consistent with the increased CaMKII activation indicated by enhanced CaMKII autophosphorylation, phosphorylation of phospholamban at the CaMKII-specific threonine-17 site was significantly increased in CAVB hearts. Statistically significant changes in CaMKII autophosphorylation and threonine-17 phospholamban phosphorylation occurred at 1-week CAVB, with maximum changes at 2 weeks. We also observed significant serine-16 phospholamban hyperphosphorylation (mediated principally by protein-kinase A, PKA), with 1- and 2-week CAVB. Overall phospholamban expression was not altered. Similarly, SERCA2a, NCX, calsequestrin, and ryanodine receptor type-2 expression were unchanged by CAVB (supplemental Figure II).
Arrhythmogenic Role of Altered Ca2+ Handling
Mechanistic studies of acquired TdP have tended to focus on APD prolongation and underlying ionic current alterations, particularly K+-current modifications.2–6 To assess the potential importance of changes in the Ca2+ handling system, particularly those associated with sarcoplasmic-reticulum (SR) Ca2+-loading, in arrhythmogenesis produced by CAVB, we studied afterdepolarizations observed during AP recording with the perforated patch method to minimize alteration of the intracellular macromolecular milieu.
Figure 7A shows spontaneous EADs in CAVB cardiomyocytes paced at 0.1 Hz. Spontaneous EADs were seen in 62 of 92 2-week CAVB cardiomyocytes (67%) from 20 rabbits, compared with none of 34 cardiomyocytes (0%, P<0.001 versus CAVB) from 4 control rabbits (supplemental Figure IIIA) and 3 of 28 1-week CAVB cardiomyocytes (11%, P=NS versus control) from 5 rabbits. When cardiomyocytes were paced at 2 Hz, they rarely showed spontaneous afterdepolarizations. However, isoproterenol commonly induced EADs in both 1-week and 2-week CAVB (but not control) cardiomyocytes lacking them under baseline conditions (Figure 7B). Isoproterenol also induced delayed afterdepolarizations (DADs) in all CAVB-cardiomyocytes (supplemental Figure IV). Isoproterenol-induced afterdepolarizations were rare in control rabbit cardiomyocytes (supplemental Figure IIIB), occurring in only 3 of 34 (8.4%, P<0.01 versus CAVB) and consisting in all cases of DADs. Isoproterenol commonly induced spontaneous Ca2+ transient events in CAVB cells after a 1-minute period of 3-Hz pacing to cause further Ca2+ loading, but this was rarely seen in control cells (supplemental Figure V).
Figure 8A illustrates the effect of thapsigargin (1 μmol/L), which suppresses SERCA function, on spontaneous EADs in CAVB rabbits. In this and 6 other cells studied in a similar fashion, thapsigargin fully suppressed EAD generation. Figure 8B shows that preventing SR Ca2+ release by adding ryanodine (10 μmol/L) to the superfusate similarly suppressed CAVB-associated EADs in all 7 cells tested. Suppression of NCX function by changing the bath solution to a modified Tyrode solution lacking Na+ (Li+ substitution) and with Ca2+ removed suppressed EADs in 5 cells (eg, see supplemental Figure VI). These results implicate SR Ca2+ stores, SR Ca2+ release through ryanodine-receptors, and NCX function as essential contributors to EAD generation in CAVB cardiomyocytes.
Evidence for Involvement of the Ca2+-Calmodulin/CaMKII System
Our results are consistent with a central role for Ca2+/calmodulin-activated CaMKII in CAVB-related cellular arrhythmogenesis. We sought to explore this role further with the use of the calmodulin antagonist W-7, the CaMK-II inhibitor KN-93, and the CaMK-II inhibitory peptide AC3-I. First, we assessed whether suppressing CaMKII-phosphorylation could reduce SR Ca2+ loading in CAVB cells, which would be the case if CaMKII phosphorylation of phospholamban is important in enhancing Ca2+ stores. Figure 3D shows that exposure of cells to 1 μmol/L KN-93 significantly reduced the caffeine-induced Ca2+-transient in CAVB cardiomyocytes without affecting its decay time constant (Figure 3E). We then examined the positive shift in ICaL inactivation observed in CAVB cells. Both KN-93 and W-7 (1 μmol/L) significantly shifted the ICaL inactivation voltages of CAVB cells in a hyperpolarizing direction, restoring them to values not significantly different from control (Figure 4D). In control cells, neither KN-93 nor W-7 significantly affected inactivation voltage, arguing against a nonspecific effect. Intracellular dialysis with the CaMKII inhibitory peptide AC3-I also reversed the inactivation voltage shift in CAVB cells, whereas the inactive analogue AC3-C had no effect (Figure 4E). Finally, we examined the ability of W-7 and KN-93 to suppress EADs in CAVB cardiomyocytes. The examples shown in Figure 8C and 8D indicate the ability of these agents to suppress CAVB-associated afterdepolarizations. Similar results were seen in 9 cells with W-7 and 7 cells with KN-93. In contrast, the inactive congener of KN-93, KN-92 (1 μmol/L), had no effect on CAVB-related EADs in 5 cells (Figure 8E).
Previous studies of the rabbit CAVB model of spontaneous TdP have appropriately emphasized ionic current remodeling, particularly changes in K+-channel expression, that underlie pathological QT prolongation. In this report, we provide evidence for CAVB-induced alterations in cellular Ca2+ handling, along with their underlying molecular basis, that appear central to the generation of arrhythmogenic afterdepolarizations.
Mechanisms of Cellular Ca2+ Handling Changes and Role in Early Afterdepolarizations
We identified 2 abnormalities in Ca2+-handling systems that could contribute to arrhythmogenic EADs. The most prominent is an increased SR Ca2+ load caused by APD prolongation and phospholamban hyperphosphorylation. Phospholamban hyperphosphorylation disinhibits SERCA and enhances SR Ca2+ uptake, leading to enhanced Ca2+ loading and accelerated [Ca2+]i transient decay. A second is a depolarizing shift in ICaL inactivation, which increases the calculated ICaL window current conductance and voltage range. Evidence for an important role of CaMKII activation in CAVB derives from the demonstration of enhanced CaMKII autophosphorylation (indicating CaMKII activation) and phospholamban hyperphosphorylation at the CaMKII-specific threonine-17 site, along with the ability of the CaMKII inhibitor KN-93 to normalize the caffeine-induced Ca2+ transient and ICaL inactivation voltage dependence. Support for the role of CaMKII in arrhythmogenic EADs derives from the ability of KN-93, and not its inactive congener KN-92, to suppress EADs in CAVB cells.
To our knowledge, our study is the first to indicate a role for Ca2+ handling abnormalities and the CaMKII system in the CAVB rabbit model of spontaneous TdP and the first to detail the molecular basis of CAVB induced Ca2+ handling disturbances. In the CAVB dog model of drug-induced TdP, altered Ca2+ handling has been implicated but underlying molecular mechanisms have not been defined. Antoons et al11 found changes in ICaL properties of CAVB dogs similar to those in our CAVB rabbits (unchanged overall ICaL density-voltage relations but a depolarizing shift in ICaL inactivation) and implicated enhanced ICaL window current in isoproterenol-induced EADs. CaMKII activation is known to be able to shift ICaL inactivation voltage.12 Our observations (reversal of CAVB-induced ICaL-inactivation shifts with the CaMKII blocker KN-93, the inhibitory peptide AC3-I, and the calmodulin inhibitor W-7) implicate Ca2+/calmodulin-dependent CaMKII activation in CAVB-induced ICaL window current enhancement. Sipido et al13 reported increased SR Ca2+ loading, enhanced NCX function, and Ca2+ transients in CAVB dogs. We did not directly examine NCX function, but unchanged NCX protein expression and caffeine-induced Ca2+ transient decay kinetics argue against intrinsic NCX changes. Nevertheless, the ability of Na+-free, Ca2+-free solution to suppress EADs indicates that intact NCX function is required for EAD generation in CVB rabbits. Ryanodine and flunarizine suppress drug-induced TdP in CAVB dogs,14 consistent with the role for SR-Ca2+ loading suggested by our data. Our observations of CaMKII hyperphosphorylation of phospholamban, which increases SR Ca2+ uptake by removing phospholamban inhibition of SERCA,15 along with enhancement of [Ca2+]i loading by CAVB-AP waveforms, explain CAVB-induced Ca2+ loading.
A proarrhythmic role of CaMKII has been suggested in a variety of models.8 CaMKIIδ-overexpressing mice exhibit ventricular tachyarrhythmias and sudden death.16 Calmodulin inhibition suppresses TdP caused by IKr-inhibiting drugs in the rabbit,17 and CaMKII-inhibitory peptides suppress EADs induced by prolonged current-clamp waveforms in rabbit ventricular myocytes.18 Our work extends the role of the CaMKII system to arrhythmogenic mechanisms occurring in a clinically relevant acquired long-QT model with spontaneous TdP.
What Activates CaMKII in This Model?
Our results show that CaMKII is required for the generation of EADs in the CAVB rabbit TdP model. The effectiveness of W-7 in suppressing arrhythmogenesis in a Ca2+-loaded context points to Ca2+-calmodulin activation as the signal for CaMKII activation. CAVB waveforms increase cell Ca2+ loading (Figure 5), suggesting that APD prolongation resulting from K+-current downregulation in CAVB rabbits5 may initiate Ca2+-calmodulin activation and CaMKII stimulation as previously described.19 However, other components of the adaptive response to CAVB could also be involved. The volume-overloaded and/or neurohumoral state resulting from CAVB may promote CaMKII activation, for example, through β-adrenergic pathways.20,21 Tissue oxidation also promotes CaMKII activation.22 Further work is needed to define specific contributors to CAVB-induced CaMKII activation.
Contributions of K+-Current Remodeling and Ca2+-Related Changes to Arrhythmogenesis in Bradycardia-Dependent TdP
We previously showed a role for combined downregulation of IKr- and IKs-encoding subunits (KCNH2, KCNQ1, and KCNE1) in spontaneous TdP of CAVB rabbits.5 In the present study, we demonstrate that CaMKII activation is required for EAD generation. Our time-course data showed that APD was maximally increased at 1-week CAVB but that cellular hypertrophy, Ca2+ loading, CaMKII activation, and EAD generation were maximal at 2-weeks. Along with the W-7 and KN-93 data in Figure 8, these results suggest that APD prolongation is a necessary but not sufficient condition for spontaneous EADs. Although APD prolongation may provide an early signal for CAVB-related arrhythmogenic remodeling, it appears insufficient to generate EADs on its own. The ability of isoproterenol to unmask latent EAD predisposition in 1-week CAVB rabbit cells (Figure 7B) indicates that Ca2+ loading and/or CaMKII activation can dynamically trigger EAD generation in the presence of APD prolongation.
Our experiments with thapsigargin and ryanodine implicate SR Ca2+ loading and release in EAD generation, and our Na+/Ca2+ replacement studies indicate NCX involvement. Forward-mode NCX generates inward currents that contribute to APD prolongation and EAD generation in CAVB dogs.13 In our CAVB rabbits, SR Ca2+ loading resulting from APD prolongation and enhanced SR Ca2+ uptake caused by phospholamban hyperphosphorylation increased systolic Ca2+ release. Subsequent removal of the cytoplasmic Ca2+ load by NCX generates enhanced inward currents, promoting EAD generation.
There is evidence for participation of ICaL in arrhythmogenic EADs associated with drug-induced APD prolongation,23,24 primarily through increased ICaL window current.25 Mathematical modeling suggests that APD prolongation associated with K+-current downregulation plays a particularly important role in generating the ICaL-mediated component of EADs and in determining their properties.26
Adrenergic stimulation may also contribute to arrhythmia generation in the CAVB rabbit model. In addition to phospholamban phosphorylation at the threonine-17 CaMKII site, we also found substantial hyperphosphorylation at the PKA-related27 serine-16 site. PKA phosphorylation is primarily driven by β-adrenergic stimulation and synergistically enhances the effects of CaMKII phosphorylation on phospholamban function.27 A recent mathematical modeling study suggested that spontaneous SR Ca2+ release prolongs APD but only generates EADs when K+-current density is decreased and that reentry reinitiation by EADs requires enhancement of ICaL density and SR Ca2+ cycling to simulate adrenergic stimulation effects.28
Novel Findings and Potential Significance
Acquired long-QT syndrome is a significant clinical problem,29 and bradyarrhythmias are an important precipitator of TdP.30 Our results provide direct evidence for a significant role of Ca2+ homeostasis alterations, along with a delineation of underlying molecular mechanisms, in a bradycardia-related animal model of spontaneous TdP. Work in this area traditionally focused on ion-channel alterations and associated APD prolongation, which are undoubtedly very important. The present study adds to growing evidence indicating that APD prolonging ion-channel dysfunction is only part of the pathophysiology of TdP and that Ca2+ handling alterations and CaMKII activation play important roles.8,31 Improved understanding of the interplay between various pathophysiological contributors to TdP-related arrhythmogenic mechanisms in different clinically relevant paradigms should lead to new mechanistic insights and improved clinical management.
We used a pharmacological approach to inhibit KN-93 for AP studies, and all pharmacological probes have potential nonspecific actions. We selected a concentration of KN-93 that is the lowest concentration producing substantial CaMKII inhibition12 to minimize nonspecific actions. We tried to repeat AP studies with cell dialysis of AC3-I.18 However, APD and EADs decreased rapidly with tight-seal patch-clamp, in contrast to the recording stability seen with perforated-patch methods (supplemental Figure VII). We did succeed in recording stable ICaL with tight-seal patch-clamp and found that AC3-I-dialysis reversed ICaL inactivation shifts, confirming the role of CaMKII.
We used right ventricular bradypacing, which can induce ventricular activation dyssynergy that may contribute to ventricular remodeling over and above the effect of bradycardia per se.32 There are regional differences in AP and ionic-current properties, with left ventricular APD being more prolonged than right ventricular in CAVB dogs.6 The present study focused on left ventricular cardiomyocytes: Further work on left-right and regional differences would be of interest.
We found APD to be more prolonged at 1- versus 2-week CAVB, and it is unclear why this should be the case. One possible explanation relates to the high sudden death mortality rate (≈50%) that occurred in the last week before intended euthanasia in 2-week CAVB rabbits. Animals with longer APDs may have been more likely to have malignant TdP and die, leaving a population with shorter mean APDs (albeit still much greater than control) because of data dropout.
We thank Chantal St-Cyr for technical help, France Theriault for secretarial support with the manuscript, and Mark Anderson of the University of Iowa for providing AC3-I/AC3-C.
Sources of Funding
This work was supported by funds from the Canadian Institutes of Health Research (MOP 68929), the ENAFRA network grant (07/CVD-03) from Fondation Leducq, and the Dutch Organization for Scientific Research (NWO program-grant 916.46.043).
Antzelevitch C. Ionic, molecular, and cellular bases of QT-interval prolongation and torsade de pointes. Europace. 2007; Suppl 4: iv4–15.
Tsuji Y, Opthof T, Yasui K, Inden Y, Takemura H, Niwa N, Lu Z, Lee JK, Honjo H, Kamiya K, Kodama I. Ionic mechanisms of acquired QT prolongation and torsades de pointes in rabbits with chronic complete atrioventricular block. Circulation. 2002; 106: 2012–2018.
Vos MA, Verduyn SC, Gorgels AP, Lipcsei GC, Wellens HJ. Reproducible induction of early afterdepolarizations and torsade de pointes arrhythmias by d-sotalol and pacing in dogs with chronic atrioventricular block. Circulation. 1995; 91: 864–872.
Tsuji Y, Zicha S, Qi XY, Kodama I, Nattel S. Potassium channel subunit remodeling in rabbits exposed to long-term bradycardia or tachycardia: discrete arrhythmogenic consequences related to differential delayed-rectifier changes. Circulation. 2006; 113: 345–355.
Volders PG, Sipido KR, Vos MA, Spätjens RL, Leunissen JD, Carmeliet E, Wellens HJ. Downregulation of delayed rectifier K+ currents in dogs with chronic complete atrioventricular block and acquired torsades de pointes. Circulation. 1999; 100: 2455–2461.
Anderson ME. Multiple downstream proarrhythmic targets for calmodulin kinase, II: moving beyond an ion channel-centric focus. Cardiovasc Res. 2007; 73: 657–666.
Coutu P, Chartier D, Nattel S. Comparison of Ca2+-handling properties of canine pulmonary vein and left atrial cardiomyocytes. Am J Physiol. 2006; 291: H2290–H2300.
Brundel BJ, van Gelder IC, Henning RH, Tuinenburg AE, Deelman LE, Tieleman RG, Grandjean JG, van Gilst WH, Crijns HJ. Gene expression of proteins influencing the calcium homeostasis in patients with persistent and paroxysmal atrial fibrillation. Cardiovasc Res. 1999; 42: 443–454.
Vinogradova TM, Zhou YY, Bogdanov KY, Yang D, Kuschel M, Cheng H, Xiao RP. Sinoatrial node pacemaker activity requires Ca2+/calmodulin-dependent protein kinase II activation. Circ Res. 2000; 87: 760–767.
Sipido KR, Volders PG, de Groot SH, Verdonck F, Van de Werf F, Wellens HJ, Vos MA. Enhanced Ca2+ release and Na/Ca exchange activity in hypertrophied canine ventricular myocytes: potential link between contractile adaptation and arrhythmogenesis. Circulation. 2000; 102: 2137–2144.
Zhang T, Maier LS, Dalton ND, Miyamoto S, Ross J Jr, Bers DM, Brown JH. The deltaC isoform of CaMKII is activated in cardiac hypertrophy and induces dilated cardiomyopathy and heart failure. Circ Res. 2003; 92: 912–919.
Mazur A, Roden DM, Anderson ME. Systemic administration of calmodulin antagonist W-7 or protein kinase A inhibitor H-8 prevents torsade de pointes in rabbits. Circulation. 1999; 100: 2437–2442.
Wu Y, Roden DM, Anderson ME. Calmodulin kinase inhibition prevents development of the arrhythmogenic transient inward current. Circ Res. 1999; 84: 906–912.
Erickson JR, Joiner ML, Guan X, Kutschke W, Yang J, Oddis CV, Bartlett RK, Lowe JS, O'Donnell SE, Aykin-Burns N, Zimmerman MC, Zimmerman K, Ham AJ, Weiss RM, Spitz DR, Shea MA, Colbran RJ, Mohler PJ, Anderson ME. A dynamic pathway for calcium-independent activation of CaMKII by methionine oxidation. Cell. 2008; 133: 462–474.
Nattel S, Quantz MA. Pharmacological response of quinidine induced early afterdepolarisations in canine cardiac Purkinje fibres: insights into underlying ionic mechanisms. Cardiovasc Res. 1988; 22: 808–817.
January CT, Riddle JM, Salata JJ. A model for early afterdepolarizations: induction with the Ca2+ channel agonist Bay K 8644. Circ Res. 1988; 62: 563–571.
January CT, Riddle JM. Early afterdepolarizations: mechanism of induction and block: a role for L-type Ca2+ current. Circ Res. 1989; 64: 977–990.
Anderson ME, Braun AP, Wu Y, Lu T, Wu Y, Schulman H, Sung RJ. KN-93, an inhibitor of multifunctional Ca++/calmodulin-dependent protein kinase, decreases early afterdepolarizations in rabbit heart. J Pharmacol Exp Ther. 1998; 287: 996–1006.
The association between bradycardia, particularly severe bradycardias occurring with atrioventricular block (AVB), and clinical long-QT syndrome is well recognized. In rabbits, experimental AVB causes ventricular remodeling and acquired long-QT syndrome with spontaneous torsades de pointes (TdP). Combined decreases in slow (IKs) and rapid (IKr) delayed-rectifier current are known to be important for QT prolongation and TdP in AVB rabbits. AVB rabbit cells are also hypercontractile, suggesting changes in cell Ca2+ handling. We therefore measured single-cell Ca2+ concentrations with a Ca2+-sensitive dye and found that heart cells from AVB rabbits have increased Ca2+ stores. Looking for causes, we found that prolonged action potentials caused by K+-channel downregulation increased Ca2+ entry during the plateau phase. Increased cell Ca2+ activated the Ca2+-sensitive phosphorylating enzyme Ca2+/calmodulin-dependent protein kinase-II (CaMKII), as shown by CaMKII autophosphorylation and enhanced phosphorylation of the sarcoplasmic-reticulum (SR)-associated protein phospholamban. Phospholamban phosphorylation enhances Ca2+ uptake into SR stores, aggravating cell Ca2+ loading. In addition, we found functional evidence of CaMKII Ca2+ channel phosphorylation, which causes increased inward Ca2+ current during the action potential plateau and early phase-3. Cellular Ca2+ overloading enhances Na+,Ca2+ exchange (NCX) activity in an attempt to eliminate the extra Ca2+, but enhanced NCX activity generates depolarizing inward currents. We found that arrhythmogenic early afterdepolarizations were much more common in AVB cells than in control cells and that early afterdepolarizations could be prevented by interventions that suppress SR Ca2+ loading, SR Ca2+ release, NCX function, or activation of CaMKII, despite the persistence of repolarization abnormalities. These results indicate that changes in cell Ca2+ signaling are important for AVB-induced long-QT/TdP arrhythmogenesis and suggest that targeting Ca2+-dependent abnormalities caused by repolarization deficiency could be a useful approach to treating acquired TdP.
Drs Qi and Yeh contributed equally to this work.
The online-only Data Supplement is available at http://circep.ahajournals.org/cgi/content/full/CIRCEP.108.815654/DC1.