Electrophysiological Consequences of Dyssynchronous Heart Failure and Its Restoration by Resynchronization Therapy

Abstract

Background

Cardiac resynchronization therapy (CRT) is widely applied in patients with heart failure and dyssynchronous contraction (DHF), but the electrophysiological consequences of CRT in heart failure remain largely unexplored.

Methods and Results

Adult dogs underwent left bundle-branch ablation and either right atrial pacing (190 to 200 bpm) for 6 weeks (DHF) or 3 weeks of right atrial pacing followed by 3 weeks of resynchronization by biventricular pacing at the same pacing rate (CRT). Isolated left ventricular anterior and lateral myocytes from nonfailing (control), DHF, and CRT dogs were studied with the whole-cell patch clamp. Quantitative polymerase chain reaction and Western blots were performed to measure steady state mRNA and protein levels. DHF significantly reduced the inward rectifier K⁺ current (I_K1), delayed rectifier K⁺ current (I_K), and transient outward K⁺ current (I_to) in both anterior and lateral cells. CRT partially restored the DHF-induced reduction of I_K1 and I_K but not I_to, consistent with trends in the changes in steady state K⁺ channel mRNA and protein levels. DHF reduced the peak inward Ca²⁺ current (I_Ca) density and slowed I_Ca decay in lateral compared with anterior cells, whereas CRT restored peak I_Ca amplitude but did not hasten decay in lateral cells. Calcium transient amplitudes were depressed and the decay was slowed in DHF, especially in lateral myocytes. CRT hastened the decay in both regions and increased the calcium transient amplitude in lateral but not anterior cells. No difference was found in Ca_V1.2 (α1C) mRNA or protein expression, but reduced Ca_Vβ2 mRNA was found in DHF cells. DHF reduced phospholamban, ryanodine receptor, and sarcoplasmic reticulum Ca²⁺ ATPase and increased Na⁺-Ca²⁺ exchanger mRNA and protein. CRT did not restore the DHF-induced molecular remodeling, except for sarcoplasmic reticulum Ca²⁺ ATPase. Action potential durations were significantly prolonged in DHF, especially in lateral cells, and CRT abbreviated action potential duration in lateral but not anterior cells. Early afterdepolarizations were more frequent in DHF than in control cells and were reduced with CRT.

Conclusions

CRT partially restores DHF-induced ion channel remodeling and abnormal Ca²⁺ homeostasis and attenuates the regional heterogeneity of action potential duration. The electrophysiological changes induced by CRT may suppress ventricular arrhythmias, contribute to the survival benefit of this therapy, and improve the mechanical performance of the heart.

Nearly 5 million Americans have heart failure (HF), and >250 000 die annually. Up to 50% of deaths in patients with HF are sudden and unexpected, mainly because of lethal ventricular arrhythmias such as ventricular tachycardia and ventricular fibrillation¹; however, the mechanisms of ventricular tachycardia and ventricular fibrillation in patients with HF remain controversial.^1,2

As HF progresses, the heart adapts to intrinsic and extrinsic stresses through a complex process of chamber remodeling and molecular modifications of myocyte structure and function. The structural remodeling and alteration of cell-to-cell coupling are associated with heterogeneous conduction delays in the ventricular myocardium that can lead to dyssynchronous ventricular contraction.³ Dyssynchronous contraction adversely influences the electrical phenotype of the failing heart and has been associated with regional changes in calcium (Ca²⁺) handling protein expression⁴ and exaggerated heterogeneities in conduction⁵ and repolarization,^6-8 which enhances the susceptibility to ventricular arrhythmias in HF.⁹ Predictably, advanced HF with dyssynchronous cardiac contraction (DHF) is associated with a poor prognosis.¹⁰ Cardiac resynchronization therapy (CRT) with biventricular pacing improves symptoms,¹¹ cardiac function, and mortality,^12,13 presumably due to a reduction in stress-strain disparities and improvement in the efficiency of ventricular contraction. We have recently demonstrated that CRT reverses regional molecular remodeling and reduces apoptosis.¹⁴

Ion channel remodeling in HF underlies many of the changes in cellular electrophysiology and predisposes to atrial and ventricular arrhythmias in animal models and patients with HF.^1,2,15-20 A consistent electrophysiological consequence of HF is a prolongation of ventricular repolarization, which is due in part to functional downregulation of outward potassium currents.^{2,15-17,19-21} However, little information is available about regional differences in ion channel function and expression or action potential (AP) profiles between early- and late-activated regions of dyssynchronous contracting left ventricles (LVs). Furthermore, the extent to which CRT reverses the DHF-induced electrophysiological remodeling remains unexplored.

Here, we tested the hypothesis that dyssynchronous mechanical contraction in HF contributes significantly to pathological remodeling at the cellular and molecular levels. We studied whether and to what extent remodeling could be reversed by biventricular pacing. We characterized the regional differences of AP, ionic currents, and [Ca²⁺]_i transients (CaT) in myocytes isolated from the anterior and lateral LV myocardium of dogs with DHF and CRT and correlated the cellular electrophysiological changes with their subunit mRNA and protein expression in corresponding regions of the LV. These findings suggest distinct mechanisms of altered electrophysiological remodeling in DHF, as well as reverse remodeling by CRT. These data provide new mechanistic understanding of the therapeutic role of CRT and will help to identify new antiarrhythmic targets to prevent sudden death in patients with HF.

Methods

An expanded Methods section is available in the online-only Data Supplement.

Canine Tachypacing-Induced HF Model

All protocols followed US Department of Agriculture and National Institutes of Health guidelines and were approved by the Animal Care and Use Committee of the Johns Hopkins Medical Institutions. The canine models of DHF or CRT have been described previously.^5,14,22 In brief, adult male mongrel dogs underwent left bundle-branch radiofrequency ablation and right atrial pacing (190 to 200 bpm) for 6 weeks (DHF dogs; n=7) or 3 weeks of right atrial pacing followed by 3 weeks of resynchronization by biventricular pacing at the same pacing rate (CRT dogs; n=7).

ECG, Echocardiography, and Hemodynamic Recordings

To monitor LV function during tachypacing, 2-dimensional echocardiography with tissue Doppler imaging was performed periodically together with recordings of ECGs. Briefly, standard 2-dimensional and color Doppler image data, triggered to the QRS complex, were saved in cine-loop format. LV volumes (end-systolic and end-diastolic) and LV ejection fraction were calculated from the conventional apical 2- and 4-chamber views. LV dyssynchrony was quantified with speckle-tracking radial strain analysis (Figure 1) as described previously.²³ The LV dyssynchrony index was defined as the SD of time to peak systolic velocity of 12 segments. Standard 12-lead ECG and hemodynamic data (LV systolic and end-diastolic pressure, dP/dt_max) were recorded 6 weeks after the start of pacing.

Figure 1.

Representative ECGs and changes of LV septal (SEP) and lateral (LTR) strain from control (A), 6-week paced DHF (B), and CRT (C) dogs. ECGs from DHF and CRT dogs are shown during pacing, and control is shown during sinus rhythm. Biventricular pacing synchronized the strain patterns between LV septal and lateral walls and abbreviated the DHF-induced prolongation of QRS duration.

Patch Clamp and CaT

All electrophysiological and intracellular Ca²⁺ measurements were performed at physiological temperature (37°C). All methods for cell isolation, electrophysiological recording, and measurements are routine in our laboratory and have been described previously.^4,8,16,17,20 Isolated ventricular myocytes from the anterior and lateral mid myocardium were current and voltage clamped by the standard whole-cell patch-clamp technique for measurement of APs, transient outward K⁺ current (I_to), inward rectifier K⁺ current (I_K1), delayed rectifier K⁺ currents (I_K), and Ca²⁺ current (I_Ca). Borosilicate glass electrodes with tip resistances of ≈3.0 MΩ when filled with the pipette solution were used.

CaTs were measured by indo-1 fluorescence excited at 365 nm with a xenon arc lamp, and emitted light at 405 and 495 nm was collected with a 2-channel photomultiplier tube assembly. Fluorescence signals were digitized and stored with electrophysiological recordings for offline analysis with custom software.²⁴ The ratio of indo-1 fluorescence (R=F₄₀₅ nm/F₄₉₅ nm) was determined after subtraction of cellular autofluorescence. The rate of Ca²⁺ removal (τ_Ca) was determined by fitting a single exponential to the Ca²⁺ time course.

Molecular Analysis

Canine Kir2.1, K_v4.3, KChIP2, K_v11.1 (ether a-go-go related gene [ERG]), Kv7.1 (KvLQT1), minK, Ca_v1.2 (α1C), Ca_vβ1, Ca_vβ2, sarcoplasmic reticulum Ca²⁺ATPase (SERCA2), phospholamban, ryanodine receptor, and Na⁺-Ca²⁺ exchanger steady state mRNA levels were measured by reverse-transcription polymerase chain reaction in tissue isolated from the LV anterior and lateral walls in control, DHF, and CRT dogs (online-only Data Supplement Table I). Kir2.1, K_v4.3, KChIP2, ERG, KvLQT1, Ca_v1.2 (α1C), SERCA2, phospholamban, ryanodine receptor, and Na⁺-Ca²⁺ exchanger proteins were measured by Western immunoblotting. Detailed methods are provided in the online-only Data Supplement.

Statistical Analysis

Differences among multiple groups were compared by ANOVA with Bonferroni test. Two-group analysis was performed by t test (paired or unpaired as appropriate). Differences in serial studies were assessed by repeated-measures ANOVA. Data are expressed as mean±SD or mean±SEM as indicated in each of the Figures. A value of P<0.05 was considered significant.

The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.

Results

ECG and Hemodynamic Changes

Figure 1 shows representative standard ECG and LV wall strain plots by tissue Doppler from a control (A), 6-week paced DHF (B), and CRT (C) dog. In the DHF dog, the QRS interval was wider, with a left bundle-branch block pattern compared with the control, and the tissue Doppler image showed dyssynchronous contraction between the septum and the lateral walls. CRT shortened the QRS duration and resynchronized the LV wall strain pattern so that it was similar to the control.

The Table summarizes the ECG, echocardiographic, and hemodynamic changes in control, 6-week paced DHF, and CRT dogs. In sinus rhythm, the R-R interval was shorter and the corrected QT and QRS durations were longer in both DHF and CRT dogs than in control dogs, but no significant difference was observed in these parameters between DHF and CRT dogs. However, during pacing, QRS width was significantly shorter in CRT than DHF dogs, which was consistent with the larger dyssynchrony index by tissue Doppler echocardiography in DHF compared with control or CRT dogs. Moreover, CRT significantly decreased the corrected QT interval compared with DHF. Conversely, LV diastolic and systolic volumes, as well as hemodynamic parameters such as LV end-diastolic and end-systolic pressure and dP/dt_max, were not statistically different between DHF and CRT. A trend was observed toward increased stroke volume and ejection fraction in CRT versus DHF dogs.

Table. ECG, Echocardiography, and Hemodynamic Parameters.

	Control	DHF	CRT
ECG, ms
Sinus rhythm
RR	608±115	438±34*	428±34*
QT	278±31	286±43	300±15*
QTc	359±26	434±64*	467±38*
QRS	47±6	103±21*	102±10*
Pacing
RR	. . .	314±14	308±18
QT	. . .	280±23	239±39
QTc	. . .	489±35	429±58†
QRS	. . .	112±16	64±6†
Echocardiography
DI, ms	30±5	74±25*	29±23†
LVEDV, mL	52±6	87±18*	93±20*
LVESV, mL	17±4	64±17*	63±15*
SV, %	35±6	23±12*	29±15
LVEF, %	67±7	25±12*	30±11*
Hemodynamics
HR, bpm	105±26	151±42*	150±12*
LVEDP, mm Hg	6±5	35±8*	34±10*
LVESP, mm Hg	142±27	103±19*	101±10*
dP/dtmax mm Hg/s	1900±467	831±246*	846±133*

QTc indicates corrected QT; DI, dyssynchrony index; LVEDV, LV end-diastolic volume; LVESV, LV end-systolic volume; SV, stroke volume; LVEF, LV ejection fraction; HR, heart rate; LVEDP, LV end-diastolic pressure; and LVESP, LV end-systolic pressure.

Values are mean±SD.

^*P<0.05 vs control

^†P<0.05 vs DHF.

Cell Capacitance

Cell capacitance was larger in DHF and CRT than in control dogs in cells isolated from both the anterior and lateral walls, but no regional difference was found between anterior and lateral cells. The capacitance of cells isolated from CRT dogs was not different from that of cells isolated from DHF dogs (online-only Data Supplement Table II).

Inward Rectifier K⁺ Current

Figure 2A shows whole-cell I_K1 currents in lateral myocytes from control, DHF, and CRT dogs. The densities of I_K1 from control and DHF myocytes were similar to data previously reported from our laboratory.¹⁶ I_K1 density in DHF myocytes was reduced significantly compared with control, whereas CRT partially reversed the DHF-induced reduction of I_K1 over a wide voltage range (Figure 2B). On the other hand, no regional difference was found in I_K1 between anterior and lateral myocytes isolated from either DHF or CRT hearts (Figure 2C). The relevant component of I_K1 for AP repolarization is the outward current “hump” observed at potentials positive to the K⁺ reversal potential. The peak outward component of I_K1 at -70 and -60 mV was decreased modestly but significantly (P<0.05) in DHF myocytes (1.43±0.52 and 1.07±0.43 pA/pF, respectively) compared with control (1.78±0.54 and 1.26±0.51 pA/pF, respectively) and CRT (1.81±0.81 and 1.18±0.57 pA/pF, respectively). To investigate the molecular basis for changes in I_K1, we measured Kir2.1 mRNA and protein levels; these were significantly downregulated in DHF compared with control (P<0.05), whereas those in CRT were decreased (although not statistically significant) compared with control (Figures 2D and 2E). No regional differences in Kir2.1 mRNA and protein expression were found among the 3 groups, which is consistent with the cellular electrophysiology.

Figure 2.

I_K1 and Kir2.1 mRNA and protein levels in control, DHF, and CRT. A, Representative current traces in lateral myocytes isolated from control, DHF, and CRT canine ventricles elicited by the diagrammed voltage-clamp protocol (holding potential -40 mV, test pulse 500 ms in duration). B, Steady state current-voltage (I-V) relationship of I_K1 and outward current portion of the I-V (right) in each group. The voltage-clamp protocol is shown in the inset. C, I_K1 density at -140 and -70 mV in anterior (ANT) and lateral (LTR) myocytes of each group. D and E, Kir2.1 mRNA and protein expression in anterior and lateral myocytes of each group. A or ANT indicates anterior; L or LTR, lateral; and CSQ, calsequestrin. †P<0.05 vs control; #P<0.05 vs DHF. The values in parentheses are the number of cells or tissue samples studied in this and all remaining figures.

Transient Outward K⁺ Current

Figure 3A shows whole-cell I_to currents in lateral myocytes from control, DHF, and CRT dogs. The densities of I_to from control and DHF myocytes in this preparation are similar to those reported previously in our laboratory.¹⁶ I_to in DHF myocytes was reduced significantly compared with control (P<0.05), but CRT did not restore the DHF-induced reduction of I_to (Figure 3B). Interestingly, I_to current density was not different between anterior and lateral myocytes in any group (Figure 3C). To investigate the molecular basis for change in I_to, we measured the α-subunit (Kv4.3) and β-subunit (KChIP2) underlying this current (Figures 3D through 3G). Consistent with the I_to density, both Kv4.3 and KChIP2 mRNA and protein were significantly downregulated in both DHF and CRT, with no regional differences in each group.

Figure 3.

I_to, and its underlying subunit mRNA and protein expression in control, DHF, and CRT. A, Representative current traces in lateral myocytes from control, DHF, and CRT ventricles elicited by the voltage-clamp protocol shown in inset. B, Peak current-voltage relationship of I_to in each group. C, I_to density at +40 mV in anterior and lateral myocytes of each group. D through G, Kv4.3 and KChIP2 mRNA and protein expression in LV midmyocardial tissue isolated from control, DHF, and CRT dogs. A or ANT indicates anterior; L or LTR, lateral; and CSQ, calsequestrin. †P<0.05 vs control.

Delayed Rectifier K⁺ Currents

Figure 4A shows total I_K in lateral myocytes from control, DHF, and CRT dogs. As shown in the superimposed currents elicited by a step pulse to +50 mV, the peak and tail I_K currents were smaller in DHF than in control myocytes, and CRT partially restored the DHF-induced reduction of I_K. Figure 4B displays the current-voltage relationship of the tail current density of I_K in the 3 groups, showing that DHF reduced I_K density by ≈50% compared with control, whereas CRT partially restored the DHF-induced reduction of I_K (P<0.05). In addition, no regional difference was found in the reduction and restoration of I_K by DHF and CRT (Figure 4C).

Figure 4.

I_K and its underlying subunit mRNA and protein expression in control, DHF, and CRT hearts. A, Representative current traces from control, DHF, and CRT myocytes elicited by voltage-clamp protocol shown in the inset. B, Current-voltage relationship of the tail current of I_K (IK tail) fitted to the Boltzmann equation: IK,tail=1/{1+exp[(V_1/2-V_m)/k]}. Voltage-clamp protocol is shown in the inset. C, IK tail density at 40 mV between anterior and lateral myocytes in each group. D through H, KvLQT1, minK, and ERG mRNA and protein expression in mid myocardium of control, DHF, and CRT dogs. A or ANT indicates anterior; L or LTR, lateral; and CSQ, calsequestrin. †P<0.05 vs control; #P<0.05 vs DHF.

To investigate the molecular basis for changes in I_K, we measured the expression of the underlying α-subunit (K_v7.1 [K_vLQT1] for I_Ks and K_v11.1 [ERG] for I_Kr) and β-subunit (minK for I_Ks) mRNA and proteins. KvLQT1 mRNA and protein were reduced significantly in DHF compared with control myocytes; CRT did not alter KvLQT1 mRNA expression but increased the protein level such that it was not significantly different from control (Figure 4D and 4E). No significant differences were found in minK mRNA expression among the 3 groups (Figure 4F). ERG mRNA expression was reduced in DHF and completely restored by CRT (Figure 4G), but no significant difference was found in ERG protein expression in any of the groups (Figure 4H). In addition, consistent with the physiological data, no regional differences in gene or protein expression were found between anterior and lateral LV tissues from each group.

Inward Ca²⁺ Current (I_Ca) and CaT

Figure 5A shows I_Ca in LV midmyocardial myocytes isolated from control, DHF, and CRT hearts. In control, no difference was found in the peak I_Ca density or current decay between anterior and lateral myocytes; however, DHF reduced the peak I_Ca density and slowed the rate of current decay in lateral cells compared with anterior cells. Furthermore, in CRT, no regional difference was found in peak I_Ca density, but the rate of current decay was still slower in lateral than in anterior cells. The current-voltage relationships (Figure 5B) showed that peak I_Ca density was significantly less in lateral cells than in anterior cells in DHF at voltage steps from -15 to +20 mV (P<0.05). In contrast, CRT increased I_Ca density in lateral cells but not in anterior cells.

Figure 5.

I_Ca and CaT in control, DHF, and CRT dogs. A, Representative I_Ca recorded from anterior and lateral myocytes. B, Peak current-voltage relationships for I_Ca,L in control, DHF, and CRT myocytes. Voltage-clamp protocol is shown in the inset. C, Voltage dependence of I_Ca activation in control, DHF, and CRT myocytes was fit with a Boltzmann equation of the form G/G_max=1/[1+exp(V_1/2-V)/k]. The voltage at which G_Ca was half-maximal (V_1/2) did not differ among the 3 groups and regions. D, Plot of the time constant of the fast (major) component of current decay measured at 0 mV. E, Plot of the charge carried by I_Ca during the first 200 ms of the voltage step. F, Representative superimposed CaTs recorded in anterior and lateral myocytes from control, DHF, and CRT dogs stimulated at 0.5 Hz as measured with indo-1 fluorescence. G and H, Average peak fluorescence values and average decay τ of CaT measured at 0.5-Hz stimulation frequency in anterior and lateral myocytes from control, DHF, and CRT dogs. ANT indicates anterior; LTR, lateral. †P<0.05 vs control; #P<0.05 vs DHF by ANOVA; *P<0.05 by t test, anterior vs lateral.

Although the voltage dependence of activation of I_Ca (Figure 5C) was not altered significantly in either DHF or CRT compared with control in either anterior or lateral cells, the rate of current decay (Figure 5D) in DHF was significantly slower (41.7±9.5 versus 33.9±5.4 ms, P<0.05) in lateral cells and significantly faster (25.6±4.0 versus 32.1±5.3 ms, P<0.05) in anterior cells than in control cells, and this produced a robust difference in current decay between anterior and lateral cells in DHF. Moreover, CRT slowed the DHF-induced faster current decay in anterior cells (32.8±7.4 ms, P<0.05) but did not affect the rate of I_Ca decay in lateral cells (40.4±9.3 ms). Therefore, the total charge carried by I_Ca (Figure 5E) was not altered significantly by DHF in either anterior or lateral cells. On the other hand, CRT significantly increased the charge carried by I_Ca in lateral (from 0.14±0.04 to 0.19±0.06 pC/pF, P<0.05) but not anterior cells compared with control.

CaTs (Figure 5F) were not different in anterior and lateral cells from control hearts; however, DHF significantly reduced the CaT amplitude and slowed the rate of decay of the CaT most prominently in lateral myocytes. CRT hastened the DHF-induced slowing of the decay of the CaT without changing the amplitude in anterior cells, whereas CaT amplitude was increased and decay hastened significantly in lateral cells. The changes in amplitude and rate of decay of CaT in each group are shown in Figure 5G and 5H, respectively. Notably, CRT almost fully reversed the DHF-induced smaller amplitude and longer decay of CaT.

Molecular Basis for Abnormal Ca²⁺ Homeostasis

The molecular basis for changes in I_Ca and CaT was investigated by determining the steady state levels of Ca_v1.2 (Ca_vα1C), Ca_vβ1, Ca_vβ2, ryanodine receptor, phospholamban, SERCA2, and Na⁺-Ca²⁺ exchanger mRNA and protein (Figure 6). No significant differences in Ca_v1.2 mRNA and protein or Ca_vβ1 subunit mRNA expression were found among control, DHF, and CRT hearts (Figure 6A through 6C). On the other hand, Ca_vβ2 mRNA was decreased significantly in DHF but not CRT myocytes compared with control (Figure 6D). Steady state ryanodine receptor and phospholamban mRNA and protein expression were consistently lower in both DHF and CRT myocytes than in control (Figure 6E through 6H), and SERCA2 mRNA and protein levels were reduced significantly in DHF myocytes but were not decreased significantly in CRT (Figure 6I and 6J). On the other hand, Na⁺-Ca²⁺ exchanger mRNA expression (Figure 6K) was increased significantly in CRT (P<0.05) compared with control, and protein levels were increased in both DHF and CRT (Figure 6L). However, no regional difference was found in the steady state levels of Ca²⁺ channel or homeostasis-related genes and proteins in each group.

Figure 6.

Ca²⁺ handling-related mRNA and protein expression in control, DHF, and CRT myocytes. A and B, Cav1.2 (α1c) mRNA and protein. C and D, Ca_vβ1 and Ca_vβ2 mRNA. E and F, Ryanodine receptor (RyR2) mRNA and protein. G and H, Phospholamban (PLN) mRNA and protein. I and J, SERCA2 mRNA and protein. K and L, Na⁺-Ca²⁺ exchanger (NCX) mRNA and protein. A or ANT indicates anterior; L or LTR, lateral; and CSQ, calsequestrin. †P<0.05 vs control.

APs and Early Afterdepolarizations

Figure 7A shows superimposed APs elicited at pacing cycle lengths of 0.5, 1.0, 2.0, and 4.0 seconds in anterior and lateral myocytes from control, DHF, and CRT dogs. Compared with control, DHF prolonged the action potential duration (APD) at all pacing cycle lengths, more prominently in lateral than in anterior cells. In contrast, CRT partially reversed the DHF-induced prolongation of APD in lateral cells but not in anterior cells. The relationship between APD and pacing cycle length of each group (Figure 7B) revealed that APD in both DHF and CRT cells was similarly prolonged in anterior cells, and the slope of the APD-cycle length relationship was increased, whereas in lateral cells, a prominent prolongation of APD in DHF was found, particularly at long pacing cycle lengths, which resulted in a significant difference in APD between anterior and lateral cells (P<0.05). By selectively shortening the APD of lateral myocytes, CRT reduced the regional heterogeneity in APD compared with DHF. The resting membrane potential (Figure 7C) was not significantly different between control, DHF, and CRT dogs, whereas the phase 1 notch depth (Figure 7D) was significantly attenuated in DHF and CRT dogs, consistent with no restoration of I_to by CRT.

Figure 7.

APD in LV myocytes from control, DHF, and CRT hearts. A, Representative superimposed APs recorded at pacing cycle lengths (CL) of 0.5, 1.0, 2.0, and 4.0 seconds, B, Relationship between pacing CL and APD at 90% recovery (APD₉₀) from anterior and lateral myocytes in each group. C and D, Bar plot of resting membrane potential (RMP; C) and phase 1 notch amplitude (D) at pacing CL of 2.0 seconds. †P<0.05 vs control; #P<0.05 vs DHF; *P<0.05 vs anterior. ANT or A indicates anterior; LTR or L, lateral.

Early afterdepolarizations (EADs) were observed more frequently in DHF than in control myocytes (P<0.001), and CRT significantly reduced the frequency of EADs compared with DHF (P<0.05; Figure 8A); however, no regional differences were found in development of EADs between anterior and lateral cells in each group. Figure 8B shows representative APs with [EAD(+)] or without [EAD(-)] EADs in myocytes from DHF hearts. The APD at 90% recovery (APD₉₀) was modestly but significantly longer in EAD(+) myocytes than in EAD(-) myocytes; however, the APD at 20% recovery (APD₂₀) was markedly shorter in EAD(+) myocytes than in EAD(-) myocytes. Therefore, we examined the relationship between the APD₂₀, APD₉₀, and APD₂₀/APD₉₀ ratio and EAD development in the failing myocytes (Figure 8C; online-only Data Supplement Table III). APD₂₀ was shorter and APD₉₀ longer in EAD(+) cells than in EAD(-) cells, which resulted in a dramatically smaller APD₂₀/APD₉₀ ratio in EAD(+) myocytes.

Figure 8.

EADs in myocytes from control, DHF, and CRT hearts. A, Bar plot of frequency of EADs (%EADs indicates fraction of APs with EADs). B, Representative superimposed APs recorded in myocytes isolated from lateral wall of DHF hearts with [EAD(+)] or without [EAD(-)] EADs. C, Bar plots of APD₂₀, APD₉₀, and ratio of APD₂₀ to APD₉₀ (APD₂₀/APD₉₀) in failing myocytes. All data were obtained during pacing at 0.25 Hz. †P<0.05 vs control; #P<0.05 vs DHF. ANT or A indicates anterior; LTR or L, lateral.

Discussion

Our studies reveal the regional cellular electrophysiological consequences of synchronous and dyssynchronous ventricular contraction in the failing heart and include several novel insights into electrophysiological remodeling in HF. First, CRT partially reverses DHF-induced K⁺ channel remodeling (I_K1 and I_K)in both the anterior and lateral LV. An interesting divergence in the remodeling of K⁺ currents can be seen: CRT has no effect on DHF-induced downregulation of I_to or on the expression of Kv4.3 or KChIP2 mRNA or protein. Second, Ca²⁺ current remodeling and Ca²⁺ handling were significantly different in the anterior and lateral LV in DHF, and CRT significantly improved Ca²⁺ homeostasis, especially in the lateral wall. Third, the APD was significantly prolonged in DHF, especially in cells isolated from the lateral LV, and CRT abbreviated the APD in lateral cells and reduced the regional gradient of APD. Finally, EADs were more frequent in DHF, were significantly but not completely reduced to near control levels in CRT, and were associated with a modestly prolonged APD₉₀ and a markedly reduced APD₂₀/APD₉₀ ratio. It is important to recognize that the model that we used corrects dyssynchronous contraction but does not affect tachycardia-induced LV dysfunction. Thus, the reversal of DHF-induced electrical and Ca²⁺ handling remodeling by CRT is due to the effect of electrical resynchronization by biventricular pacing.

K⁺ Channel Remodeling

Downregulation of K⁺ currents is the most consistent ionic current change in animal models^1,16,20,21 and human HF.¹⁵ K⁺ current downregulation may promote ventricular tachycardia and ventricular fibrillation,²¹ either by direct prolongation of AP¹⁹ in the voltage range at which I_Ca,L reactivation occurs, which predisposes to the development of EADs,²⁵ or by heterogeneous reduction of the repolarization reserve and the promotion of functional reentry. Although expressed cardiac K⁺ channels vary in different species, I_to downregulation is the most consistent ionic current change in failing mammalian hearts.^15,16,20,21 In the present study, I_to and its related genes, Kv4.3 and KChIP2, were downregulated homogenously by DHF and were not affected by CRT. The data suggest that tachycardia, HF, or altered ventricular activation is more important in downregulation of I_to than mechanical synchrony. Reduced I_K1 density in HF^16,18,20 may contribute to prolongation of APD and enhanced susceptibility to spontaneous membrane depolarization.²⁶ Although small but significant changes in I_K1 density were observed in the outward current component, and the largest changes were observed at very negative voltages (Figure 2B and 2C), the resting membrane potential of myocytes did not differ in any of the groups (Figure 7C), because the major voltage range of altered I_K1 was beyond physiological potentials. On the other hand, in tachypacing-induced HF models, I_Ks is downregulated, but I_Kr is less consistent.^18,21 In the present study, both I_K1 and I_K densities were reduced in a regionally homogenous fashion by DHF in spite of the higher wall stress and myocyte stretch in the late-activated lateral wall than in the early-activated anterior wall²⁷; this was partially but significantly restored by CRT without a change in global LV function, which suggests that K⁺ channel remodeling was not directly associated with the mechanical stress caused by dyssynchronous ventricular contraction.

The dichotomy in the regulation of I_to compared with I_K1 and I_K in CRT is remarkable. These currents share regulatory mechanisms that are altered in the failing heart and are differentially remodeled by CRT.¹⁴ The detailed regulation by the autonomic nervous system, renin-angiotensin-aldosterone signaling, and reactive oxygen species are distinct for each of the K⁺ currents studied and may explain the differences in response to CRT. In addition, biventricular tachypacing in the present study improved the synchrony of mechanical contraction, but global LV function as assessed by LV ejection fraction or end-diastolic pressure was still significantly depressed. It is likely that the molecular mechanisms for the altered functional expression of each current are mixed. The steady state levels of Kv4.3 and KChIP2 mRNA are consistently downregulated by tachypacing in the presence and absence of mechanical synchrony; thus, the balance between transcription and RNA degradation is altered. Furthermore, altered ventricular activation by CRT may suppress I_to expression independent of the presence of HF.^28,29 Concordant changes in protein levels suggest the possibility of a pretranslational mechanism. However, the reduction in Kv4.3 and KChIP2 proteins in particular is not as pronounced as the magnitude of the current reduction, which suggests an additional posttranslational mechanism of functional downregulation of I_to. The changes in I_K1 and Kir2.1 mRNA and protein in DHF and CRT appear to be more consistent with but not proof of a pretranslational mechanism of downregulation. Changes in KvLQT1 and ERG mRNA and protein are more variable and appear to be partially altered by CRT. There appear to be distinct differences in the mechanism of down-regulation of individual K⁺ currents in DHF and variable degrees and mechanisms of restoration of expression in CRT.

Altered I_Ca and Ca²⁺ Handling

Changes in I_Ca functional expression in HF are variable.³⁰ Depending on the model and stage of HF, some studies showed a decrease in I_Ca density, whereas others reported no change.^2,16,31 In the present study, the peak I_Ca density and current decay in DHF were regionally different between myocytes isolated from the lower-stress anterior and higher-stress lateral walls. Furthermore, CRT regionally restored the DHF-induced changes in I_Ca, increasing the peak I_Ca density in lateral cells and slowing the decay in anterior cells; CRT thus served to mitigate the DHF-induced regional heterogeneities in I_Ca density and gating.

HF causes significant changes in Ca²⁺-handling proteins.^2,32 In failing human hearts, the steady state level of the Ca_v1.2(α1C) mRNA was reported to be decreased³³ or unchanged.¹⁷ Moreover, Vanderheyden et al³⁴ recently suggested the possibility of reverse molecular remodeling of SERCA2 in CRT responders. In the present study, Ca_v1.2(α1C) mRNA and protein were not different between control, DHF, and CRT dogs, consistent with our previous data in human myocardium.¹⁷

Ca_vβ-subunit expression has been positively correlated with peak Ca_v1.2 current density.³⁵ The present data showed that Ca_vβ1 mRNA levels were unchanged in DHF and CRT, whereas steady state Ca_vβ2 mRNA levels were reduced in DHF and partially restored by CRT. In DHF, ryanodine receptor, phospholamban, and SERCA2a mRNA and protein levels were downregulated, whereas Na⁺-Ca²⁺ exchanger was upregulated, consistent with previous studies.^4,31,36,37 No regional differences in mRNA and protein expression were found in any of these mediators of Ca²⁺ handling in DHF and CRT, which suggests that the regional differences of Ca²⁺ handling in DHF and its restoration by CRT are posttranslational.

The mechanisms underlying the differences in regional remodeling of K⁺ currents and Ca²⁺ handling in DHF remain obscure. Plotnikov et al⁷ reported that cardiac dyssynchrony by LV pacing (120 to 150 bpm for 3 weeks) produced a slower decay of I_Ca inactivation, consistent with the present results. Moreover, this phenomenon was suppressed by β-adrenergic blockade. These findings suggest that DHF-induced changes of I_Ca inactivation kinetics might be mediated by regionally heterogeneous β-adrenergic receptor stimulation. Furthermore, the Ca²⁺-handling proteins are functionally regulated by phosphorylation, prominently by the key intracellular enzymes protein kinase A and Ca²⁺-cal-modulin-dependent protein kinase II (CaMKII),^38-40 as well as a variety of phosphatases that may be regionally regulated.¹⁴

Prolongation of APD and Development of EADs

The APD is consistently prolonged in human and animal models of HF.^{2,15-17,19-21} A recent study using a canine model of dyssynchrony suggests that after 4 weeks, the APD is prolonged in the late-activated compared with the early-activated regions.⁶ In the present work, CRT produced a partial but statistically significant shortening of DHF-induced prolongation of APD selectively in lateral cells. Although the peak I_Ca density in DHF was decreased in lateral cells compared with anterior cells, the decrease was modest and not significant compared with control lateral myocytes (Figure 5B). On the other hand, I_Ca decay was slowed significantly, which might contribute to regional AP prolongation in the setting of homogenous K⁺ current down-regulation, although Ca²⁺ flux through L-type channels during a square pulse protocol does not fully reflect that during an AP. These data provide some insight into the mechanism of regional AP remodeling in DHF and CRT.

Moreover, on a molecular level, we have shown that tumor necrosis factor-α and CaMKII were increased in DHF, prominently in the lateral wall, and these differences were absent in CRT.¹⁴ Tumor necrosis factor-α decreases I_to and prolongs APD in rat ventricular myocytes.⁴¹ CaMKII influences Ca²⁺ current and sarcoplasmic reticulum function^39,40 and increases persistent Na⁺ current,^42,43 which results in prolongation of APD.⁴⁴ It is possible, and indeed likely, that other regional alterations in Ca²⁺ handling (Figure 5F) or an increased persistent Na⁺ current contribute to regional differences in the APD and AP profile in DHF and to the regionally specific effects of biventricular pacing on this phenotype. Furthermore, EADs were observed more frequently in DHF than in control myocytes, and CRT reduced the occurrence of EADs in both anterior and lateral cells.

Prolongation of APD₉₀ was associated with an increased frequency of EADs, but notably, a smaller APD₂₀/APD₉₀ ratio was even more strongly correlated with the appearance of EADs (Figure 8C). Thus, APD prolongation alone may not be sufficient to generate EADs in failing myocytes.⁴⁴ The APD₂₀/APD₉₀ ratio is an empiric metric but suggests that long APs with a reduced plateau voltage are the most likely to exhibit EADs. This type of AP profile that is generated by the reduction in K⁺ current density and reduced I_Ca density with slowed kinetics is highly susceptible to EADs that result from reactivation of I_Ca.

Study Limitations

The model of DHF and CRT used in the present study is a limitation. Six weeks of tachypacing (200 bpm) reproducibly induces dilated cardiomyopathy with LV enlargement, increased LV end-diastolic pressure, and decreased dP/dt_max; however, the study was designed to examine the effects of CRT with ongoing HF, and thus, tachypacing was maintained. This differs from CRT in patients, which is performed at lower heart rates. Therefore, the processes of both remodeling and the reversal of remodeling in this circumstance are likely to be different from those in human DHF.

The second limitation is that the present study focused on the cellular and molecular bases of electrophysiological remodeling in DHF and restoration by CRT. It did not evaluate the susceptibility to and frequent development of arrhythmia in vivo or in a whole-heart model.

Clinical Implications

CRT has emerged as an effective pacing/mechanical therapy for patients with HF and a prolonged QRS duration. CRT has been associated with improved cardiac function, symptomatology, and exercise capacity and, when combined with defibrillator therapy, reduced mortality.^12,13 The role of CRT in preventing arrhythmias^45,46 or reversing adverse electrical remodeling remains controversial. The present study provides novel information on the altered expression and function of ionic currents and CaTs in DHF and after CRT. Understanding the fundamental mechanisms of the altered ion channel function and Ca²⁺ homeostasis, as well as the electrophysiological remodeling in DHF, and the capacity for restoration by CRT will not only help define the therapeutic role of CRT but will help to identify antiarrhythmic targets that can be exploited by other therapeutic strategies designed to prevent sudden death.