Cardiac Resynchronization Therapy Improves Altered Na Channel Gating in Canine Model of Dyssynchronous Heart Failure

Canine Tachypacing-Induced Heart Failure Model

All protocols followed the 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 previously described and represent models of heart failure with dyssynchronous ventricular contraction (DHF) and synchronous contraction (CRT).^17–19 In brief, adult male mongrel dogs underwent left bundle branch (LBB) radiofrequency ablation and the right atrial (RA) pacing (200 bpm) for 6 weeks (DHF dogs: n=7), or 3 weeks of RA pacing followed by 3 weeks of resynchronization by biventricular pacing at same pacing rate (CRT dogs: n=7).¹⁵ Control (non-failing: NF) dogs (n=4) underwent no tachypacing and no ablation.

Whole-cell patch clamp recording

Midmyocardial LV, LV anterior (ANT) and lateral (LTR) myocytes from normal (NF), DHF and CRT dogs were studied using the whole cell patch clamp. Ventricular myocytes were current and voltage clamped in the whole-cell configuration of the patch-clamp as previously described.^{15, 20} Voltage and current clamp control and data acquisition were performed using custom-written software. For the measurement of APs, ventricular myocytes were current clamped with borosilicate glass electrodes with tip resistances of 3.0–4.0 MΩ when filled with pipette solution, and the stimulation frequency was varied over cycle lengths of 2.0, 1.0, 0.5 and 0.25 sec. The steady-state APs were recorded and analyzed at least 1 min after pacing at each cycle length in standard Tyrode’s solution (37 °C).

I_Na recording was performed at room temperature with patch pipettes with tip resistances of 1–1.5 MΩ tip resistance when filled with pipette solution containing (in mM): NaCl 5, CsCl 40, glutamate 80, CsOH 80, Mg-ATP 5, EGTA 5, HEPES 10, CaCl2 1.5 (free [Ca²⁺] = 100 nM, pH 7.2 with CsOH, LJP = +5mV). The bath solution contained (in mM) NaCl 10, MgCl₂ 2, CsCl₂ 5, TEA-Cl 125, HEPES 20 (pH 7.4 with CsOH). In all experiments, recording was begun 10–15 min after establishment of the whole-cell mode to permit stabilization of current amplitude, voltage dependence and kinetics of gating. Standard voltage protocols were used for assessment of the voltage-dependence of activation/inactivation, recovery from inactivation and entry into inactivation. To determine the membrane potential of half inactivation (V_1/2 ) and the slope factor k, steady state inactivation data were fit with a Boltzmann function of the form: I/I_max={1+exp[(V-V_1/2)/k]}⁻¹. Recovery from inactivation data were fit with a biexponential function of the form: I(t)/Imax=A1·exp(−t/ 1)+A2·exp(−t/ 2)+A3, using a nonlinear least squares minimization. The decay phase of the current during a voltage step was fit with biexponential function of the form: I(t)=A_f·exp(−τ/ f)+ A_s·exp(−τ/ s), where A_f and A_s are the fractions of fast and slow inactivating components, respectively. Persistent (late) I_Na was the tetrodotoxin (TTX) (30µM)-sensitive current elicited by 800 ms depolarizations to −20 mV (from −140 mV), from 100–500 ms after a depolarizing voltage step normalized to the peak I_Na.

Molecular Analysis

Canine Na_V1.5 steady state mRNA levels were measured by kinetic real time PCR of RNA in tissue isolated from the midmyocardial layer of LV anterior and lateral walls in NF, DHF and CRT dogs. Na_V1.5, and CaMKII proteins were measured by Western immunoblotting. Detailed methods are provided in the online-only supplement.

Computational Model

Simulations were performed using a modified version of the canine coupled model incorporating local control of Ca²⁺-induced Ca²⁺ release ²¹. This model was modified by replacing its description of I_Na with a new Na⁺ channel representation, based on that of Grandi et al ²² (Figure S1). The model is further described in the on-line supplement.

Statistical Analysis

Differences among the three groups were compared by one-way ANOVA with Bonferroni's test. Bonferroni-corrected comparisons were performed only if the overall comparison was statistically significant. Two-group analysis was performed by unpaired or paired t-test as appropriate (Figure 3D and Figure 4B, respectively). Data were expressed as mean values ± 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.

Changes of I_Na availability by DHF and CRT

There is no statistically significance in the gating parameters between ANT and LTR cells in I_Na, this study focused on the difference in Na⁺ currents between NF, DHF and CRT conditions.

Figure 1A shows representative whole cell I_Na in ventricular myocytes from NF, DHF and CRT dogs. The current-voltage (I-V) relationship of cardiac I_Na revealed no significant difference in the peak and reversal potential of the I_Na between NF, DHF and CRT conditions (Figure 1B). Furthermore, no significant difference was found in the voltage-dependence of Na⁺ current activation in each group (Figure 1C, Table 1). On the other hand, DHF produced approximately a −3 mV shift in I_Na steady state inactivation compared with the NF dog (V_1/2; − 74.5±3.2 vs. −71.1±3.8 mV, p<0.05) (Figure 1D, Table 1), whereas CRT normalized the DHF-mediated hyperpolarizing shift of steady state inactivation relationship (−70.6 ±.2.9 mV, p<0.05 vs. DHF).

In DHF, I_Na had a slower recovery from inactivation, and with a hastened entry into inactivation as compared to NF (Figure 1E, Table 1). The fast time constant of recovery from inactivation (τ_fast) was significantly larger in DHF than in NF (τ_fast; 7.59 ± 2.88 vs. 4.54 ± 0.95 ms, p<0.05). Overall CRT did not significantly alter τ_fast compared to DHF (Table 1). Table 2 and Figure 1F summarizes the effect of DHF or CRT on the entry of cardiac Na⁺ channels into inactivated states. DHF increased the fraction of channels undergoing slow inactivation compared to NF cells (y0: 0.74 ± 0.08 DHF versus 0.83 ± 0.05 NF; p<0.05). Myocytes from CRT hearts exhibited intermediate entry into slow inactivated states with statistically significantly larger fraction of current entering intermediate inactivated states compared with cells from NF hearts. Consistent with a global rather than regional alteration, CRT-induced changes in I_Na kinetics and gating were similar in cells from the ANT and LTR walls

Changes of Decay time constant by DHF and CRT

As shown in Figure 2A, DHF slowed I_Na decay compared to NF, and CRT normalized the DHF- induced change of I_Na decay. The initial (τ_fast) and late (τ_slow) time constants of I_Na decay over a range of voltages are shown in Figure 2B and Table 1. There was no significant difference in the τ_fast between the groups, whereas τ_slow was significantly increased in DHF compared to NF (13.91 ± 6.80 vs. 4.44 ± 3.62 ms; p<0.05) and CRT normalized the DHF-induced change of I_Na decay (5.02 ± 4.37 ms). No significant difference was observed in these time constants between ANT and LTR cells.

Change of late-I_Na (I_Na-L ) by DHF and CRT

Figure 3A shows the superimposed persistent I_Na-L in myocytes from NF, DHF and CRT canine ventricles ([Na⁺]_o= 10mM). I_Na-L was elicited by 800 ms depolarizations to −20 mV (from −140 mV), and tetrodotoxin (TTX)-sensitive currents were normalized to peak I_Na. The current integral was calculated between 100 and 500 ms after the depolarizing pulse. As shown in Figure 3B and Table 3, DHF significantly increased I_Na-L compared to NF (1.33 ± 0.66 % vs. 0.09 ± 0.10 %, p<0.05), but this increase was virtually eliminated in CRT myocytes (0.42 ± 0.21 %, p<0.05 vs. DHF).

We then studied I_Na-L in physiological bath solution ([Na⁺]_o=140 mM). I_Na-L was elicited by 2000 ms depolarizations to −40 mV (from −140 mV). I_Na-L was measured as the average amplitude of the current between 200 and 220 ms. DHF significantly increased I_Na-L in physiological Na⁺ conditions (0.16 ± 0.08 to 0.82 ± 0.26 pA/pF; p<0.05 vs. NF), and CRT dramatically reduced the DHF-induced increase of I_Na-L (0.23 ± 0.16 pA/pF; p<0.05 vs. DHF). To demonstrate the I_Na-L increase in heart failure, we measured I_Na in [Na⁺]_o=140 mM at baseline and after ranolazine which is known to block I_Na-L. Ranolazine (1µM) reduced I_Na-L in myocytes from DHF (0.82±0.25 to 0.37±0.15 pA/pF; p<0.05) but not in myocytes isolated from NF or CRT hearts (Figure 3C and 3D). There were no regional differences in I_Na-L in ANT and LTR cells in either [Na⁺]_o= 10mM and 140mM conditions.

APD prolongation and late-I_Na

Normal impulse formation and conduction depend on the fast inward I_Na. Also, an increase in I_Na-L can markedly prolong APD and promote polymorphic VT. Action potential prolongation in DHF is highly arrhythmogenic with frequent EADs that were not observed in myocytes isolated from NF hearts. CRT partially abbreviated the DHF-induced APD prolongation and reduced the frequency of EADs in cells isolated from both the anterior and lateral LV.^{4, 15} In order to understand the contribution of I_Na-L to prolongation of the AP in DHF and CRT, we added ranolazine to the bath solution during continuous AP recording. Shown in Figure 4A, ranolazine (1µM) did not change the APD in myocytes from NF, however it significantly shortened APD in cells isolated from DHF hearts (Figure 4B and 4C), consistent with an increase of I_Na-L having a substantial role in APD prolongation in DHF. Myocytes from CRT hearts had a reduced I_Na,L density compared with DHF thus ranolazine had less effect on APD and EADs in cells from these hearts compared with DHF (Figure 4D).

CRT has a number of effects on electrical remodeling in the failing heart including reversal of K⁺ channel down regulation and improvement of SR Ca handling.¹⁵ To determine if the changes in I_Na-L could explain the changes in APD, we used a mathematical model of the canine AP ^{21, 22}. Increasing normalized I_Na-L from 0.11 to 0.88% of the peak I_Na reproduced APD prolongation similar to that observed in myocytes from DHF hearts. An intermediate normalized I_Na-L of 0.35% of peak current shortened the APD compared to DHF, similar to that observed in CRT. The simulations are consistent with changes of I_Na-L in heart failure contributing to APD prolongation in DHF and its abbreviation by CRT (Figure 4E).

CaMKII

Heart failure is associated with global²³ and regional^{19, 24} increases in calcium (Ca²⁺) -calmodulin dependent protein kinase II (CaMKII) activity that are partially reversed by CRT.¹⁹ CaMKII modulates the function of Na currents producing an increase in I_Na-L,^{20, 25} therefore, we propose that changes in CaMKII expression in myocytes are linked to alteration in the Na current phenotype in NF, DHF and CRT myocytes. As shown in Figure 5, DHF increased total and phosphorylated (pThr287) CaMKII compared to NF. Notably CRT significantly reduced the DHF-induced increase of CaMKII pThr287 with a trend in a reduction in total CaMKII. No significant regional differences between ANT and LTR regions of the LV were found in any group.

I_Na remodeling in DHF and CRT

Normal impulse formation and conduction as well APD depend on the inward I_Na. Studies of I_Na regulation in heart failure have been somewhat variable and model dependent. In this study we found peak current was not significantly altered, though others have reported down regulation in a similar model.²⁷ However, we did observe slowing of the decay and rise in late current, and this has been observed in a canine myocardial infarction model of heart failure and in failing human hearts.^{26, 28, 29} An increase in the late component of Na⁺ current (I_Na-L) can markedly prolong APD and facilitate the generation of EADs and promote polymorphic VT.³⁰ Therefore, changes in I_Na density and kinetics may predispose to arrhythmias either by disrupting conduction and/or prolonging repolarization. In this study, the increase of I_Na-L by DHF prolonged APD, while CRT partially normalized I_Na-L, and abbreviated the APD.

Ranolazine is an inhibitor of a number of ion channels with relative selectivity for I_Na-L and antiarrhythmic effects in vitro.^{31, 32} In an ischemic canine model of HF, ranolazine improved abnormal repolarization and contraction in LV myocytes.³³ Consistent with CRT modifying I_Na-L as key mechanism for electrical reverse remodeling, we found that ranolazine suppresses I_Na-L and, at a low dose, completely normalized the APD and prevented EADs in acutely isolated DHF myocytes. In cells from NF and CRT hearts with significantly less I_Na-L, the drug had little effect on the APD. Simulations of the experimentally determined changes in the magnitude of I_Na-L in a canine AP model for each condition predicted prolongation of the APD in DHF, albeit larger than that observed experimentally. The quantitative contribution of the increase in I_Na-L to APD is complex, but reproduces the AP prolongation in DHF and the partial reversal of the APD in CRT myocytes. The data, which are consistent with a global reduction in APD and hastening of repolarization by CRT, suggest a mechanism for the antiarrhythmic properties of this pacing therapy. The data do not allow us to speak to the role of ranolazine as an antiarrhythmic therapy in this model since we used a concentration of the drug designed to affect only I_Na-L and not other changes in the Na current or other currents that are remodeled in heart failure.

CaMKII and Oxidant Stress

Among the many signaling changes in the failing heart, an increase in CaMKII protein and activity has been regularly observed.³⁴ A mechanistic link between the increase in CaMKII activity and I_Na-L that is observed in human and canine heart failure is suggested by recent studies showing that intracellular CaMKII signaling increases I_Na-L and Na⁺ influx in failing myocytes by slowing inactivation kinetics and shifting steady state inactivation.^{20, 25,35} In our study, consistent with these previous results, DHF is associated with an increase CaMKII activity; moreover, CRT partially normalizes CaMKII activity indexed by a reduction in the autophosphorylated forms of the enzyme (Figure 5), while at the same time reducing the amplitude of I_Na-L. The data support a critical role of CaMKII and its regulation in the remodeling and reverse electrical remodeling of ventricular myocyte electrophysiology by DHF and CRT respectively.^{19, 36–38}

Oxidative stress is linked to the progression of heart failure, and mitochondria are critical source of ROS in failing myocardium. Blocking I_Na-L reduces hydrogen peroxide-induced arrhythmogenic activity and contractile dysfunction.^{9, 39} Oxidative stress-induced afterdepolarizations are directly linked to CaMKII activity.^{40, 41} Furthermore, Kohlhaas et al suggest that increased [Na⁺]_i in heart failure promotes ROS formation by reducing mitochondrial Ca²⁺ uptake.^{42, 43} Cardiac resynchronization of the failing heart potently alters the mitochondrial proteome with an associated improvement in mitochondrial function reflected in an increase in the respiratory control index.⁴⁴ Moreover, DHF is associated with global and region activation of stress-related proteins, and CRT reduces stress kinase activation.^{19, 24} We did not directly measure ROS in these studies; however, our findings suggest a mechanism of remodeling in DHF characterized by a ROS and CaMKII mediated increase of I_Na-L, and that is reversed by a reduction in CaMKII activity decreasing I_Na-L and suppressing EADs in CRT.

Limitations

One limitation in this study is the absence of a direct measurement of oxidant levels in the tissue; however, our recent study demonstrating that resynchronization improved mitochondria oxidative phosphorylation coupling⁴⁴ suggests that CRT may suppress ROS level in the heart failure. Second, as we have shown previously, CRT improves multiple ion channel and Ca handling abnormalities in DHF¹⁵ and the functional effects on cardiac electrophysiology are complex and interrelated. Therefore, we used computer simulations to quantify the effect of reverse remodeling of Na⁺ current in the context of the changes in other ionic currents and transporters in our models of DHF and CRT. An isolated change in I_Na-L in the simulated AP is sufficient to dramatically alter the duration consistent with a significant contribution of I_Na-L to prolongation of AP in the failing heart but is certainly not only change that influences the APD in DHF and CRT.

Clinical Implications

CRT is a remarkable therapy that improves symptoms and survival in a subset of patients with heart failure associated with dyssynchronous mechanical contraction. The mechanical act of resynchronizing contraction by pacing has significant effects on cellular signaling in the failing heart that are associated with not only improved mechanical performance but also enhanced energetic efficiency. This provides an explanation for the mechanism of action of CRT in heart failure and identified pathways that are central to improving cardiac function and which could be approached pharmacologically in patients that do not response to CRT. In addition, a role for I_Na-L in the AP prolongation associated with DHF and its reversal by CRT in the setting of improved mechanical and electrical performance of the heart suggests a pathogenic role of the late Na⁺ current in arrhythmias in heart failure that could be addressed not only by CRT, but pharmacological agents that target this current component.