- Email: [email protected]
Altered ventricular stretch contributes to initiation of cardiac memory
Altered ventricular stretch contributes to initiation of cardiac memory Eugene A. Sosunov, PhD, Evgeny P. Anyukhovsky, PhD, Michael R. Rosen, MD, FHRS From the Center for Molecular Therapeutics, Departments of Pharmacology and Pediatrics, College of Physicians and Surgeons of Columbia University, New York, New York. BACKGROUND Cardiac memory is a change in T-wave morphology induced by ventricular pacing or arrhythmias that persist after resumption of normal AV conduction. Changing the pacemaker site from atrium to ventricle alters ventricular activation and the mechanical pattern of ventricular contraction. Either or both alterations affect T-wave configuration. OBJECTIVE The purpose of this study was to study the role of altered contractile patterns on initiation of cardiac memory. METHODS Isolated rabbit hearts were immersed in Tyrode’s solution (37°C) and aortically perfused at a constant pressure of 70 mmHg. Three orthogonal quasi-ECG leads were recorded via six Ag–AgCl electrodes located on the walls of the bath. Hearts were paced at a constant cycle length from either the right atrial appendage or left ventricle lateral wall. The pulmonary artery was sealed, and both ventricles contracted isovolumetrically. Cardiac memory was quantified as T-wave vector displacement expressed as distance between T-wave vector peaks during atrial pacing before and after ventricular pacing.
Introduction Cardiac memory is a change in T-wave morphology induced by ventricular pacing or arrhythmias that persists after resumption of atrial pacing or sinus rhythm.1–3 Cardiac memory may be of short (lasting minutes to hours)1,3,4 or long (lasting weeks to months) duration,1,2,4 and the two types of cardiac memory may be induced by different mechanisms.5,6 Changing the pacemaker site from atrium to ventricle alters the pathway of ventricular electrical activation, and altered activation affects the spatial stress–strain relationships and contractile patterns of the ventricular wall.7,8 We previously demonstrated that a ventricular pacing protocol that induced cardiac memory was associated with altered wall motion, suggesting involvement of altered contractile patterns in cardiac memory.9 This possibility has
This study was supported by USPHS National Heart, Lung, and Blood Institute Grants HL-67101 and HL-28958. Address reprint requests and correspondence: Dr. Michael R. Rosen, Department of Pharmacology, College of Physicians and Surgeons of Columbia University, 630 West 168 Street, PH 7West-321, New York, New York 10032. E-mail address: [email protected]. (Received August 13, 2007; accepted September 7, 2007.)
RESULTS Five minutes of ventricular pacing induced significant T-wave vector displacement that returned to control in 5 to 10 minutes. No significant changes in intraventricular pressure occurred during and after ventricular pacing. Interventions that decreased ventricular load (shunting both ventricles to the bath) or contractility (excitation– contraction uncoupler blebbistatin) significantly decreased developed pressure and eliminated T-wave vector displacement. Neither intervention affected ventricular activation during ventricular pacing. Locally applied left ventricular epicardial stretch induced T-wave vector displacement similar to that induced by ventricular pacing. CONCLUSION Altered ventricular activation during ventricular pacing initiates cardiac memory via induction of altered contractile patterns and altered stretch. KEYWORDS Electrocardiography; Pacing-induced T-wave changes; Ventricles; Myocardial contraction; Myocardial stretch; Excitation– contraction uncoupler (Heart Rhythm 2008;5:106 –113) © 2008 Heart Rhythm Society. All rights reserved.
been suggested by others as well.10,11 The present study was undertaken to determine whether altering ventricular activation alone initiates cardiac memory (as originally suggested by Rosenbaum et al1) or if altered activation is important primarily because it initiates an altered contractile pattern. To this end, we studied short-term cardiac memory in isolated, Langendorff-perfused rabbit hearts. The method used in this study allowed us to intervene by altering stretch and/or contraction without affecting activation or coronary perfusion.
Methods The investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996) and the rules of the Columbia University Institutional Animal Care and Use Committee.
Modified Langendorff-perfused rabbit heart Male New Zealand white rabbits (age 90 –100 days, weight 2–2.5 kg) were anesthetized with sodium pentobarbital 40 mg/kg IV and heparinized with 1,000 U/kg. The chest was opened through a midsternal incision, and the heart was
1547-5271/$ -see front matter © 2008 Heart Rhythm Society. All rights reserved.
doi:10.1016/j.hrthm.2007.09.008
Sosunov et al
Initiation of Cardiac Memory
rapidly removed and chilled in ice-cold modified HEPESTyrode’s solution of the following composition (in mmol/ L): HEPES 10, NaCl 148, KCl 5, CaCl2 1.8, MgCl2 1.0, and dextrose 5.5. The solution was saturated with 100% oxygen. The cut aortic stump was cannulated and the heart transferred to a Langendorff apparatus. The coronary arteries were continuously perfused with warm (37° ⫾ 0.5°C, pH 7.4 ⫾ 0.2) HEPES–Tyrode’s solution at a constant pressure of 70 mmHg. The heart was immersed in a temperaturecontrolled chamber filled with warm HEPES–Tyrode’s solution that was continuously replenished via coronary sinus outflow. The pulmonary artery was sealed, and both ventricles contracted isovolumetrically. Right ventricular pressure was measured via the cannulated pulmonary artery. In some experiments, left ventricular pressure was monitored with a fluid-filled latex balloon inserted into the left ventricle via a left atrial incision. Balloon volume was adjusted to enddiastolic pressure of 0 mmHg. Bipolar pacing electrodes were inserted into the right atrial appendage and sewn to the left ventricular lateral wall. The heart was stimulated with 2-ms rectangular pulses of twice threshold amplitude at a constant cycle length from either the right atrium or left ventricle. Stimulus output could be switched between the two pacing sites without interrupting pacing or altering its frequency. Because electrical stimuli applied to the left ventricle to change the pathway of ventricular activation can induce local catechol-
107 amine release and affect repolarization, all solutions contained 0.4 M propranolol.12 We previously demonstrated that beta-blockade with propranolol has no effect on the evolution of short-term cardiac memory.5 To study a possible role of cardiac production of angiotensin II,9 the pacing protocol was performed in the presence of the AT-1 receptor blocker losartan 2 M in a subset of experiments.
ECG recording and determination of T-wave vector The chamber solution was used as a volume conductor for ECG recording. Three orthogonal ECG signals were recorded with a personal computer-based data acquisition system via six Ag–AgCl electrodes located in the chamber walls (Figures 1A and 1B). Three-dimensional vector images were reconstructed, and T-wave vector was defined as a line from the origin of the T-wave loop to its most remote point (Figure 1C).
Pacing protocols and quantification of T-wave vector displacement The pacing protocol was derived from that previously described for inducing short-term memory.3,9 Depending on the initial rhythm, the right atrium was paced at a cycle length of 350 or 400 ms to provide control records. After 1 hour of equilibration, ventricles were paced for 5 minutes or 1 hour and then atrial pacing was resumed (Figure 1A). Pacing cycle length was constant throughout each experi-
Figure 1 A: Schematic of isolated heart, ECG leads, and pacing protocol for inducing short-term memory (see Methods for description). B: Memory induced by 60 minutes of left ventricular (LV) pacing at the same rate. Representative recordings of three orthogonal ECG leads and right ventricular pressure (P) are shown. Shown from right to left are steady state at right atrial (RA) pacing, 10 seconds of LV pacing, 60 minutes of LV pacing, 10 seconds after resumption of RA pacing, and 15 minutes after resumption of RA pacing. C: Corresponding T-wave loops and vectors (arrows). D: T-wave vector displacement (TVD; measured in microvolts) is the distance between T-wave vector peaks before (RA) and 10 seconds after resumption of atrial pacing (10 s RA). By convention2, T-vector displacement during control atrial pacing ⫽ 0.
108
Heart Rhythm, Vol 5, No 1, January 2008 changes led to altered T-wave vector angles and amplitudes (Figure 1C) resulting in a prominent T-wave vector displacement (Figure 1D). Fifteen minutes after resuming atrial pacing, all three ECG leads and T-wave vectors had returned to their respective controls. Pressure measured in the right ventricle remained unchanged throughout the course of the experiment.
Interventions affecting ventricular contraction and stretch Ventricular contraction was altered using four different experimental designs. Experiment 1: Ventricular load was nearly completely removed by introducing latex catheters (2 mm inner diameter, 5-cm length) into the right and left ventricles through the tricuspid and mitral valves, respectively, thereby shunting both ventricles to the bath such that no pressure developed. Experiment 2: Ventricular load was decreased or increased by changing volume in a latex balloon inserted into the left ventricle.13 Experiment 3: To alter myocardial contractility, the excitation– contraction uncoupler blebbistatin was used. Blebbistatin is a cell-permeable inhibitor of myosin II ATPase activity14 that does not alter myocardial action potential morphology or the propagation of excitation.15 Experiment 4: To apply local stretch to the left ventricular myocardium, small (2 ⫻ 2 ⫻ 0.5 mm) plastic pads sliding along steel shafts were sutured to two epicardial sites. Initial distance between the sites was 10 mm. Calibrated mechanical manipulators allowed changing the distance and applying ⬃15% stretch.
Figure 2 Plots of linear regression analysis of relation between activation time (AT) and action potential duration to 90% repolarization (APD90) at various times of the pacing protocol. Shown from top to bottom are steady-state right atrial pacing, 5 minutes of left ventricular pacing, 120 minutes of left ventricular pacing, 5 minutes after resumption of right atrial pacing, and 60 minutes after resumption of right atrial pacing. See text for discussion.
ment. Cardiac memory was quantified as T-wave vector displacement, the distance between T-wave vector loop peaks during atrial pacing before and after ventricular pacing (Figure 1D). Figures 1B through 1D show the effects of 60 minutes ventricular pacing on T-wave morphology, vector angles, and T vector displacement, respectively. In panel B, note the prominent alteration of the QRS complex during ventricular pacing (compared with right atrial pacing), reflecting a change of the ventricular activation pattern. Ten seconds after resuming atrial pacing, the shapes of the T waves were different in all leads in comparison with those before ventricular pacing. Changes in T-wave amplitude in leads X and Z and T-wave inversion in lead Y were seen. These
Figure 3 A: Records of T-wave vector displacement (TVD) and systolic and diastolic right ventricular pressure (P) during right atrial (RA) pacing and at 5 and 60 minutes of left ventricular (LV) pacing. Arrows indicate TVD just after terminating 5 minutes (point d) and 60 minutes (point f) of LV pacing. B: Summary data for activation time (ATx), QRS duration (QRSx), QT interval (QTx measured in lead X) TVD, and systolic and diastolic right ventricular pressure (P) during RA pacing and at 5 minutes of LV pacing. In panels A and B: a ⫽ RA pacing before changing to LV pacing (control); b and c ⫽ 10 seconds and 5 minutes of LV pacing; d and e ⫽ 10 seconds and 15 minutes after resumption of RA pacing following 5 minutes of LV pacing; f ⫽ 10 seconds after resumption of RA pacing following 60 minutes of LV pacing. *P ⬍.05 vs a (n ⫽ 6).
Sosunov et al
Initiation of Cardiac Memory
Statistical analysis Data are expressed as mean ⫾ SEM. One-way analysis of variance for repeated or nonrepeated measures was used to compare means between groups. P ⬍.05 was considered significant.
Results Confirmation of the model Costard-Jackle et al16 previously used Langendorff-perfused rabbit heart to study initiation of cardiac memory, and we first tested whether we could reproduce their findings in our system. They recorded monophasic action potential from 12 to 20 different epicardial sites with a bipolar contact electrode. Activation time, action potential duration, and repolarization time in both ventricles were measured, first during right atrial pacing, then during 120 minutes of right ventricular pacing, and then again during 60 minutes of atrial pacing. Data were analyzed using linear regression analysis to consider action potential duration as a function of activation time. Four preliminary experiments were performed using the same methods for monophasic action potential recording and data analysis as those of Costard-Jackle et al16 (Figure 2). Our data were similar to theirs in, that during the initial atrial pacing period, action potential duration was inversely related to activation time (Figure 2A). With onset of ventricular pacing, this inverse correlation disappeared (Figure
109 2B). However, continuing ventricular pacing restored the inverse relation (Figure 2C). Returning to atrial pacing again perturbed the inverse correlation (Figure 2D), but after 60 minutes of atrial pacing an inverse correlation was reestablished (Figure 2E). This adjustment of ventricular myocardial repolarization sequence to a change in activation sequence was considered by Costard-Jackle et al16 as a manifestation of cardiac memory in isolated heart. We did not use the Costard-Jackle et al16 method of memory assessment in the present study because the time required to perform multisite monophasic action potential mapping does not permit detection of rapid (5–15 minutes) changes in ventricular repolarization and because recording the quasi-ECGs of isolated heart and quantifying memory as T-wave vector displacement permits us to relate our results to those reported in intact canine2,3,5,6,9 and human17–19 heart. In six hearts, we then studied the effects of two durations of ventricular pacing (5 and 60 minutes) in different sequences using the protocol shown in Figure 1. Figure 3A shows representative T-wave vector displacement and right ventricular pressure data. Because no differences in the magnitude of T-wave vector displacement and time course of its recovery were observed in the 5- and 60-minute pacing protocols (215 ⫾ 56 V at point d and 218 ⫾ 50 V at point f, n ⫽ 6, P ⬎.05), all further protocols were performed using 5 minutes of ventricular pacing.
Figure 4 A: Schematic diagram showing shunts introduced through the tricuspid valve into the right ventricle and through the mitral valve into the left ventricle. Representative recordings show ECG leads and right ventricular pressure at steady-state during right atrial (RA) pacing, 10 seconds of left ventricular (LV) pacing, 5 minutes of LV pacing, 10 seconds after resumption of RA pacing, and 15 minutes after resumption of RA pacing before (B) and after insertion of shunts in both ventricles (C). See text for discussion.
110
Heart Rhythm, Vol 5, No 1, January 2008
Figure 3B summarizes changes of the major ECG parameters, T-wave vector displacement, and right ventricular developed pressure during the 5-minute ventricular pacing protocol. Lead X activation time (which was measured from stimulus to Q wave), QRS duration, and QT interval changed significantly during ventricular pacing but recovered to control immediately on resumption of atrial pacing. Similar changes were seen in leads Y and Z (data not shown). On resumption of atrial pacing, T-wave vector displacement differed significantly from control and then returned to control in 15 minutes. Right ventricular pressure remained unaltered during the course of the protocol. To study the role of angiotensin II in memory initiation, we used the AT-1 receptor blocker losartan. The 5-minute pacing protocol was performed first during perfusion with control Tyrode’s solution. After complete T-wave vector displacement dissipation, the pacing protocol was repeated in the presence of 2 M losartan. The blocker had no effects on memory expression: 202 ⫾ 46 V and 196 ⫾ 45 V, respectively (n ⫽ 4, P ⬎.05). We then proceeded with the four experiments on altered activation and stretch. Experiment 1: A representative experiment in which shunts were inserted into the right and left ventricles to remove ventricular load is shown in Figure 4. Before shunting (Figure 4B), ventricles contracted isovolumetrically, and 5 minutes of ventricular pacing induced marked changes in T-wave shape in all three ECG leads (compare right atrial pacing before and 10 seconds after ventricular pacing). Shunt insertion led to a prominent drop of developed pressure, and ventricular pacing produced no changes in any ECG lead (Figure 4C). Summary data for five experiments showed that shunt insertion produced a significant reduction of developed pressure (from 21 ⫾ 2 mmHg to 5 ⫾ 1 mmHg, P ⬍.05) associated with a significant decrease of ventricular pacing-induced T-wave vector displacement (from 196 ⫾ 38 V before shunting to 11 ⫾ 8 V after shunt insertion, P ⬍.05). No significant differences in major ECG parameters occurred during ventricular pacing before and after shunt insertion (Table 1, Shunts), suggesting that shunting did not alter the overall ventricular Table 1 ECG parameters during left ventricular pacing in control (baseline) and after insertion of shunts into right and left ventricles (top) or perfusion with 0.4 ⌴ blebbistatin (bottom) ATx (ms) Shunts (n ⫽ 5) Baseline Shunts Blebbistatin (n ⫽ 10) Baseline Blebbistatin
QRSx (ms)
QTx (ms)
8⫾1 8⫾1
67 ⫾ 5 65 ⫾ 3
210 ⫾ 3 213 ⫾ 2
10 ⫾ 1 10 ⫾ 2
65 ⫾ 3 67 ⫾ 2
212 ⫾ 4 211 ⫾ 9
Values are given as mean ⫾ SEM values. No significant changes in ECG parameters in leads Y and Z were observed (data not shown). ATx ⫽ activation time (time interval between stimulus artifact and Q-wave); QRSx ⫽ QRS duration; QTx ⫽ QT interval measured in lead X.
Figure 5 Superposition of ECG leads during right atrial pacing before (trace C) and 10 seconds after 5 minutes of left ventricular pacing (trace CM). A–E: Pacing protocol was performed at different degrees of balloon inflation resulting in different left ventricular developed pressure (dev P). F: Dependence of T-wave vector displacement (TVD) on left ventricular developed pressure. See text for discussion.
activation pattern. Given that the same alteration in activation is observed with ventricular pacing in the setting of normal and of reduced pressure, we reasoned that a variable(s) other than altered activation must contribute to the onset of memory. The remaining experiments explored this possibility. Experiment 2: The purpose of this protocol was to assess the relationship between left ventricular developed pressure and pacing-induced T-wave vector displacement. Figure 5 shows results from 1 of 4 experiments in which the left ventricle was loaded with a fluid-filled latex balloon placed in the ventricular chamber. Ventricular load could be altered by changing the extent of balloon inflation. Note that 5-minute ventricular pacing performed at a high left ventricular developed pressure led to prominent T-wave changes in leads X and Y (Figure 5A: compare control trace [C] with the trace recorded 10 seconds after the end of left ventricular pacing [cardiac memory]). After the ECG had completely recovered, the extent of balloon inflation was decreased and the ventricular pacing protocol repeated (Figure 5B). T-wave changes were seen again, but they were less prominent than at higher developed pressure (compare Figures 5A and 5B). Further decreases in balloon inflation resulted in progressive reductions of ventricular developed
Sosunov et al
Initiation of Cardiac Memory
111
Figure 6 Effects of blebbistatin on ECG and right ventricular pressure. Representative recordings of ECG leads and right ventricular pressure at steady-state during right atrial (RA) pacing, 10 seconds of left ventricular (LV) pacing, 5 minutes of LV pacing, 10 seconds after resumption of RA pacing, and 15 minutes after resumption of RA pacing in control (A), after 30 minutes of perfusion with Tyrode’s solution containing 0.4 M of blebbistatin (B), and after 2 hours of blebbistatin washout (C). See text for discussion.
pressure and T-wave changes (Figures 5C and 5D), whereas partial restoration of balloon inflation led to an increase of developed pressure and T-wave changes (Figure 5E). Twave vector displacement was calculated for all these experimental conditions and plotted against left ventricular developed pressure (Figure 5F). Note the direct relationship between these values. A similar result was seen in all four hearts studied. Experiment 3: This protocol used blebbistatin to uncouple excitation and contraction. Before perfusion with blebbistatin, 5-minute ventricular pacing induced changes of T-wave shape in all ECG leads (compare “RA” with “10 s RA” in Figure 6A). Inhibition of myocardial contraction by blebbistatin resulted in a marked fall of right ventricular pressure (Figure 6B). Here, no T-wave changes were seen in any ECG lead after ventricular pacing. Two hours of blebbistatin washout partially restored both ventricular pressure and T-wave changes induced by ventricular pacing (Figure 6C). Summary data for 10 experiments demonstrated that blebbistatin significantly reduced developed pressure (from 20 ⫾ 2 mmHg to 5 ⫾ 2 mmHg, P ⬍.05) and ventricular pacing-induced T-wave vector displacement (from 213 ⫾ 25 V to 6 ⫾ 5 V, P ⬍.05). As in the experiments with shunts, blebbistatin did not alter ECG parameters during ventricular pacing, suggesting that it did not alter activation (Table 1, Blebbistatin). Experiment 4: Figure 7 shows typical results from 1 of 5 experiments in which ⬃15% stretch was applied locally to the left ventricle. The heart was continuously paced from the right atrium at a constant cycle length of 400 ms. Local
stretch applied for 5 minutes produced changes in T-wave morphology in all ECG leads (Figure 7B). Release of stretch led to gradual recovery of T-waves to control. At this time in each experiment, 5-minute ventricular pacing to induce cardiac memory was performed. This intervention produced T-wave vector displacement changes comparable to those of stretch (Figures 7C and 7D, respectively). The time course of T-wave vector displacement dissipation after release of stretch was similar to that seen after resumption of atrial pacing in the ventricular pacing protocol. These comparable patterns of stretch-induced T-wave vector displacement were observed in all five experiments.
Discussion The T-wave reflects epicardial potential gradients during ventricular repolarization that result from spatial heterogeneity in repolarization time.20 –22 Therefore, the altered T-wave morphology in cardiac memory implies that ventricular pacing induces spatial changes in ventricular repolarization that persist after resumption of supraventricular activation. Switching the pacemaker site from atrium to ventricle dramatically alters the sequence of ventricular electrical activation. It has been suggested that altered spatial electrotonic interactions occurring during cardiac activation create the stimulus for cardiac memory.1,23 An alternative proposal has been that a pacing-induced altered ventricular activation pattern modifies the spatial stress– strain relationships of the myocardial wall to initiate cardiac memory.9 –11 To discriminate between these two possible triggers, we used interventions that influenced one variable
112
Heart Rhythm, Vol 5, No 1, January 2008
Figure 7 Effects of local ventricular stretch on T-wave morphology and T-wave vector displacement (TVD). A: Schematic diagram showing application of stretching device on left ventricular surface. Left ventricular balloon is in place. B: Representative recordings show ECG leads and left ventricular pressure at various times during the stretching protocol. Heart was continuously paced from right atrium. Shown from left to right are steady-state before stretch, 10 seconds of local left ventricular stretch, 5 minutes of local left ventricular stretch, 10 seconds after release of stretch, and 15 minutes after release of stretch. C, D: Time course of TVD induced by 5 minutes of ventricular pacing or 5 minutes of local ventricular stretch, respectively.
(contractile pattern) while leaving another (activation pattern) unchanged. Experiments shown in Figures 4 through 6 show that decreasing load, contraction, or pressure all reduce the expression of cardiac memory. At the same time, developed pressure per se cannot be a trigger of cardiac memory because it remains constant when the site of pacing is moved from right atrium to left ventricle (Figure 1), suggesting by elimination of alternatives that the pattern of stretch must be the primary trigger. Indeed, we also demonstrated (Figure 7) that altering stretch in the absence of any change of load, pressure, or contraction induces T-wave vector displacement similar to that induced by ventricular pacing with respect to both magnitude and time course of dissipation. Because stretch is a final common pathway for load, contraction, or pressure changes, it appears that stretch is the ultimate determinant of cardiac memory. Jeyaraj et al11 compared acute effects of atrial, anterior, and posterior left ventricular pacing in anterior and posterior segments of canine left ventricle and found that left ventricular pacing induced immediate changes of circumferential strain localized in the segment distant from the pacing site. In another group of dogs, after a few weeks of similar pacing from the same two sites, action potential duration measured in wedge preparations of the respective left ventricular segments was longer in the segment in which the circumferential strain had increased in acute experiments. The authors suggested that long-term cardiac memory observed after a few weeks of left ventricular pacing could be triggered by mechanoelectrical feedback induced by the
change in cardiac activation sequence. In these experiments, spatial patterns of electrical activation and mechanical stress were not discriminated. That is, changes of strain were induced by altering activation sequence, and either activation or stress could be the primary trigger for cardiac memory. In the setting of short-term cardiac memory, our findings suggest that the mechanical factor is the primary trigger. Demonstration of a local change in strain within the left ventricular wall during short-term left ventricular pacing by Jeyaraj et al11 is consistent with our conclusion.
Study limitations A limitation of our study is seen with regard to the observation made by Rosenbaum et al1 that accumulation is a distinctive feature of cardiac memory. That is, in both human and in canine heart in situ, the longer that abnormal ventricular activation persists, the more memory is expressed and the greater time is necessary for recovery of T-wave morphology.1,3,23 In contrast, we did not observed accumulation in the rabbit heart. T-wave vector displacement magnitude and dissipation were the same after 5 minutes and 60 minutes of ventricular pacing. Importantly, in the same isolated rabbit heart model, Costard-Jackle et al16 did not report any memory accumulation. Thus, our study and that of Costard-Jackle et al suggest that memory accumulation in isolated rabbit heart is different from that in human and canine heart in situ3,9,17,19 and isolated perfused canine heart.9 One possible explanation for this difference is as follows. We previously demonstrated that transient out-
Sosunov et al
Initiation of Cardiac Memory
ward current Ito is important to the development of shortterm cardiac memory.3,24 There are species-specific differences in K⫹ channel expression. In both dogs and humans, Ito is determined by the pore-forming K⫹ channel subunit Kv4.3 and the accessory protein KChIP2.25,26 We previously showed downregulation of channel function in a gradient away from the pacing electrode in canine cardiac memory.27 However, in rabbit myocardium, Kv1.4 is a major contributor to Ito.28 Whether pacing to induce cardiac memory affects Kv1.4 in the same fashion as Kv4.3/ KChIP2 is not known. However, based on the results of the present study, we surmise that any importance of Kv1.4 to cardiac memory occurs via a very different mechanism. Another study limitation is in light of our prior finding that short-term cardiac memory in dog is angiotensin II– dependent.9 In contrast, in the present study, AT-1 receptor blockade did not influence the expression of memory in rabbit heart. Given that the action of angiotensin II in canine cardiac memory appears to derive from an effect on Ito,29,30 these data suggest different mechanisms of short-term cardiac memory in canine and rabbit hearts. Yet another possibility is that neither the pacing protocols we used nor angiotensin II is enough of a perturbation to displace the Ito in rabbit heart, which contributes more robustly to repolarization than does Ito in canine or human heart.
Conclusion Short-term cardiac memory (seen as ventricular pacing-induced T-wave vector displacement) occurs in isolated rabbit hearts. Initiation of cardiac memory is suppressed by interventions that decrease ventricular pressure or myocardial contraction without affecting activation pattern. Externally applied stretch in the absence of any change of load pressure or contraction induces T-wave vector displacement similar to that induced by ventricular pacing. The results suggest that spatial alterations in ventricular stretch during ventricular pacing are a key determinant to initiation of cardiac memory.
Acknowledgments We express our gratitude to Ms. Nimee Bhat for assistance in performing the studies and Ms. Eileen Franey for careful attention to manuscript preparation.
References 1. Rosenbaum MB, Blanco HH, Elizari MV, et al. Electrotonic modulation of the T wave and cardiac memory. Am J Cardiol 1982;50:213–222. 2. Shvilkin A, Danilo P Jr, et al. Evolution and resolution of long-term cardiac memory. Circulation 1998;97:1810 –1817. 3. Del Balzo U, Rosen MR. T-wave changes persisting after ventricular pacing in canine heart are altered by 4-aminopyridine but not by lidocaine: implications with respect to phenomenon of cardiac memory. Circulation 1992;85:1464 –1472. 4. Chatterjee K, Harris A, Davies G, et al. Electrocardiographic changes subsequent to artificial ventricular depolarization. Br Heart J 1969;31:770 –779. 5. Plotnikov AN, Yu HG, Geller JC, et al. Role of L-type calcium channels in pacing-induced short-term and long-term cardiac memory in canine heart. Circulation 2003;107:2844 –2849.
113 6. Patberg KW, Rosen MR. On the role of the cAMP response element binding protein in long-term cardiac memory. Circ Res 2003;93:e87. 7. Prinzen FW, Peschar M. Relation between the pacing induced sequence of activation and left ventricular pump function in animals. Pacing Clin Electrophysiol 2002;25:484 – 498. 8. Wyman BT, Hunter WC, Prinzen FW, et al. Effects of single- and biventricular pacing on temporal and spatial dynamics of ventricular contraction. Am J Physiol Heart Circ Physiol 2002;282:H372–H379. 9. Ricard P, Danilo P Jr, Cohen IS, et al. A role for the renin-angiotensin system in the evolution of cardiac memory. J Cardiovasc Electrophysiol 1999;10:545– 551. 10. Meghji P, Nazir SA, Dick DJ, et al. Regional workload induced changes in electrophysiology and immediate early gene expression in intact in situ porcine heart. J Mol Cell Cardiol 1997;29:3147–3155. 11. Jeyaraj D, Wilson LD, Zhong J, et al. Mechanoelectrical feedback as novel mechanism of cardiac electrical remodeling. Circulation 2007;115:3145– 3155. 12. Sosunov EA, Anyukhovsky EP, Rosen MR. Effects of exogenous neuropeptide Y on automaticity of isolated Purkinje fibers and atrium. J Mol Cell Cardiol 1996;28:967–975. 13. Uy RG, Ross-Ascuitto NT, et al. Recovery of the chronically hypoxic young rabbit heart reperfused following no-flow ischemia. Pediatr Cardiol 2006;27: 37– 46. 14. Kovacs M, Toth J, Hetenyi C, et al. Mechanism of blebbistatin inhibition of myosin II. J Biol Chem 2004;279:35557–35563. 15. Fedorov VV, Lozinsky IT, Sosunov EA, et al. Application of blebbistatin as an excitation-contraction uncoupler for electrophysiologic study of rat and rabbit hearts. Heart Rhythm 2007;4:619 – 626. 16. Costard-Jackle A, Goetsch B, Antz M, et al. Slow and long-lasting modulation of myocardial repolarization produced by ectopic activation in isolated rabbit hearts: evidence for cardiac memory. Circulation 1989;80:1412–1420. 17. Wecke L, Gadler F, Linde C, et al. Temporal characteristics of cardiac memory in humans: vectorcardiographic quantification in a model of cardiac pacing. Heart Rhythm 2005;2:28 –34. 18. Shvilkin A, Ho KKL, Rosen MR, et al. T-vector direction differentiates postpacing from ischemic T-wave inversion in precordial leads. Circulation 2005; 111:969 –974. 19. Wecke L, Rubulis A, Lundahl G, et al. Right ventricular pacing-induced electrophysiological remodeling in the human heart and its relationship to cardiac memory. Heart Rhythm 2007;4:1477–1486. 20. Spach MS, Barr RC, Lanning CF, et al. Origin of body-surface QRS and T-wave potentials from epicardial potential distributions in intact chimpanzee. Circulation 1977;55:268 –278. 21. Ghanem RN, Jia P, Ramanathan C, et al. Noninvasive electrocardiographic imaging (ECGI): comparison to intraoperative mapping in patients. Heart Rhythm 2005;2:339 –354. 22. Janse MJ, Sosunov EA, Anyukhovsky EP, et al. Repolarisation gradients in the canine left ventricle before and after induction of cardiac memory. Eur Heart J 2005;26:454 – 455. 23. Franz MR, Bargheer K, Costard-Jackle A, et al. Human ventricular repolarization and T-wave genesis. Prog Cardiovasc Dis 1991;33:369 –384. 24. Plotnikov AN, Sosunov EA, Patberg KW, et al. Cardiac memory evolves with age in association with development of the transient outward current. Circulation 2004;110:489 – 495. 25. Dixon JE, Shi W, Wang HS, et al. Role of the Kv4.3 K⫹ channel in ventricular muscle. A molecular correlate for the transient outward current. Circ Res 1996;79:659 – 668. 26. Schram G, Pourrier M, Melnyk P, et al. Differential distribution of cardiac ion channel expression as a basis for regional specialization in electrical function. Circ Res 2002;90:939 –950. 27. Yu H, McKinnon D, Dixon JE, et al. Transient outward current, Ito1, is altered in cardiac memory. Circulation 1999;99:1898 –1905. 28. Wang Z, Feng J, Shi H, et al. Potential molecular basis of different physiological properties of the transient outward K⫹ current in rabbit and human atrial myocytes. Circ Res 1999;84:551–561. 29. Yu H, Gao J, Wang H, et al. Effects of the renin-angiotensin system on the current Ito in epicardial and endocardial ventricular myocytes from the canine heart. Circ Res 2000;86:1062–1068. 30. Yu H, McKinnon D, Dixon JE, et al. Transient outward current, Ito1, is altered in cardiac memory. Circulation 1999;99:1898 –1905.
Introduction Cardiac memory is a change in T-wave morphology induced by ventricular pacing or arrhythmias that persists after resumption of atrial pacing or sinus rhythm.1–3 Cardiac memory may be of short (lasting minutes to hours)1,3,4 or long (lasting weeks to months) duration,1,2,4 and the two types of cardiac memory may be induced by different mechanisms.5,6 Changing the pacemaker site from atrium to ventricle alters the pathway of ventricular electrical activation, and altered activation affects the spatial stress–strain relationships and contractile patterns of the ventricular wall.7,8 We previously demonstrated that a ventricular pacing protocol that induced cardiac memory was associated with altered wall motion, suggesting involvement of altered contractile patterns in cardiac memory.9 This possibility has
This study was supported by USPHS National Heart, Lung, and Blood Institute Grants HL-67101 and HL-28958. Address reprint requests and correspondence: Dr. Michael R. Rosen, Department of Pharmacology, College of Physicians and Surgeons of Columbia University, 630 West 168 Street, PH 7West-321, New York, New York 10032. E-mail address: [email protected]. (Received August 13, 2007; accepted September 7, 2007.)
RESULTS Five minutes of ventricular pacing induced significant T-wave vector displacement that returned to control in 5 to 10 minutes. No significant changes in intraventricular pressure occurred during and after ventricular pacing. Interventions that decreased ventricular load (shunting both ventricles to the bath) or contractility (excitation– contraction uncoupler blebbistatin) significantly decreased developed pressure and eliminated T-wave vector displacement. Neither intervention affected ventricular activation during ventricular pacing. Locally applied left ventricular epicardial stretch induced T-wave vector displacement similar to that induced by ventricular pacing. CONCLUSION Altered ventricular activation during ventricular pacing initiates cardiac memory via induction of altered contractile patterns and altered stretch. KEYWORDS Electrocardiography; Pacing-induced T-wave changes; Ventricles; Myocardial contraction; Myocardial stretch; Excitation– contraction uncoupler (Heart Rhythm 2008;5:106 –113) © 2008 Heart Rhythm Society. All rights reserved.
been suggested by others as well.10,11 The present study was undertaken to determine whether altering ventricular activation alone initiates cardiac memory (as originally suggested by Rosenbaum et al1) or if altered activation is important primarily because it initiates an altered contractile pattern. To this end, we studied short-term cardiac memory in isolated, Langendorff-perfused rabbit hearts. The method used in this study allowed us to intervene by altering stretch and/or contraction without affecting activation or coronary perfusion.
Methods The investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996) and the rules of the Columbia University Institutional Animal Care and Use Committee.
Modified Langendorff-perfused rabbit heart Male New Zealand white rabbits (age 90 –100 days, weight 2–2.5 kg) were anesthetized with sodium pentobarbital 40 mg/kg IV and heparinized with 1,000 U/kg. The chest was opened through a midsternal incision, and the heart was
1547-5271/$ -see front matter © 2008 Heart Rhythm Society. All rights reserved.
doi:10.1016/j.hrthm.2007.09.008
Sosunov et al
Initiation of Cardiac Memory
rapidly removed and chilled in ice-cold modified HEPESTyrode’s solution of the following composition (in mmol/ L): HEPES 10, NaCl 148, KCl 5, CaCl2 1.8, MgCl2 1.0, and dextrose 5.5. The solution was saturated with 100% oxygen. The cut aortic stump was cannulated and the heart transferred to a Langendorff apparatus. The coronary arteries were continuously perfused with warm (37° ⫾ 0.5°C, pH 7.4 ⫾ 0.2) HEPES–Tyrode’s solution at a constant pressure of 70 mmHg. The heart was immersed in a temperaturecontrolled chamber filled with warm HEPES–Tyrode’s solution that was continuously replenished via coronary sinus outflow. The pulmonary artery was sealed, and both ventricles contracted isovolumetrically. Right ventricular pressure was measured via the cannulated pulmonary artery. In some experiments, left ventricular pressure was monitored with a fluid-filled latex balloon inserted into the left ventricle via a left atrial incision. Balloon volume was adjusted to enddiastolic pressure of 0 mmHg. Bipolar pacing electrodes were inserted into the right atrial appendage and sewn to the left ventricular lateral wall. The heart was stimulated with 2-ms rectangular pulses of twice threshold amplitude at a constant cycle length from either the right atrium or left ventricle. Stimulus output could be switched between the two pacing sites without interrupting pacing or altering its frequency. Because electrical stimuli applied to the left ventricle to change the pathway of ventricular activation can induce local catechol-
107 amine release and affect repolarization, all solutions contained 0.4 M propranolol.12 We previously demonstrated that beta-blockade with propranolol has no effect on the evolution of short-term cardiac memory.5 To study a possible role of cardiac production of angiotensin II,9 the pacing protocol was performed in the presence of the AT-1 receptor blocker losartan 2 M in a subset of experiments.
ECG recording and determination of T-wave vector The chamber solution was used as a volume conductor for ECG recording. Three orthogonal ECG signals were recorded with a personal computer-based data acquisition system via six Ag–AgCl electrodes located in the chamber walls (Figures 1A and 1B). Three-dimensional vector images were reconstructed, and T-wave vector was defined as a line from the origin of the T-wave loop to its most remote point (Figure 1C).
Pacing protocols and quantification of T-wave vector displacement The pacing protocol was derived from that previously described for inducing short-term memory.3,9 Depending on the initial rhythm, the right atrium was paced at a cycle length of 350 or 400 ms to provide control records. After 1 hour of equilibration, ventricles were paced for 5 minutes or 1 hour and then atrial pacing was resumed (Figure 1A). Pacing cycle length was constant throughout each experi-
Figure 1 A: Schematic of isolated heart, ECG leads, and pacing protocol for inducing short-term memory (see Methods for description). B: Memory induced by 60 minutes of left ventricular (LV) pacing at the same rate. Representative recordings of three orthogonal ECG leads and right ventricular pressure (P) are shown. Shown from right to left are steady state at right atrial (RA) pacing, 10 seconds of LV pacing, 60 minutes of LV pacing, 10 seconds after resumption of RA pacing, and 15 minutes after resumption of RA pacing. C: Corresponding T-wave loops and vectors (arrows). D: T-wave vector displacement (TVD; measured in microvolts) is the distance between T-wave vector peaks before (RA) and 10 seconds after resumption of atrial pacing (10 s RA). By convention2, T-vector displacement during control atrial pacing ⫽ 0.
108
Heart Rhythm, Vol 5, No 1, January 2008 changes led to altered T-wave vector angles and amplitudes (Figure 1C) resulting in a prominent T-wave vector displacement (Figure 1D). Fifteen minutes after resuming atrial pacing, all three ECG leads and T-wave vectors had returned to their respective controls. Pressure measured in the right ventricle remained unchanged throughout the course of the experiment.
Interventions affecting ventricular contraction and stretch Ventricular contraction was altered using four different experimental designs. Experiment 1: Ventricular load was nearly completely removed by introducing latex catheters (2 mm inner diameter, 5-cm length) into the right and left ventricles through the tricuspid and mitral valves, respectively, thereby shunting both ventricles to the bath such that no pressure developed. Experiment 2: Ventricular load was decreased or increased by changing volume in a latex balloon inserted into the left ventricle.13 Experiment 3: To alter myocardial contractility, the excitation– contraction uncoupler blebbistatin was used. Blebbistatin is a cell-permeable inhibitor of myosin II ATPase activity14 that does not alter myocardial action potential morphology or the propagation of excitation.15 Experiment 4: To apply local stretch to the left ventricular myocardium, small (2 ⫻ 2 ⫻ 0.5 mm) plastic pads sliding along steel shafts were sutured to two epicardial sites. Initial distance between the sites was 10 mm. Calibrated mechanical manipulators allowed changing the distance and applying ⬃15% stretch.
Figure 2 Plots of linear regression analysis of relation between activation time (AT) and action potential duration to 90% repolarization (APD90) at various times of the pacing protocol. Shown from top to bottom are steady-state right atrial pacing, 5 minutes of left ventricular pacing, 120 minutes of left ventricular pacing, 5 minutes after resumption of right atrial pacing, and 60 minutes after resumption of right atrial pacing. See text for discussion.
ment. Cardiac memory was quantified as T-wave vector displacement, the distance between T-wave vector loop peaks during atrial pacing before and after ventricular pacing (Figure 1D). Figures 1B through 1D show the effects of 60 minutes ventricular pacing on T-wave morphology, vector angles, and T vector displacement, respectively. In panel B, note the prominent alteration of the QRS complex during ventricular pacing (compared with right atrial pacing), reflecting a change of the ventricular activation pattern. Ten seconds after resuming atrial pacing, the shapes of the T waves were different in all leads in comparison with those before ventricular pacing. Changes in T-wave amplitude in leads X and Z and T-wave inversion in lead Y were seen. These
Figure 3 A: Records of T-wave vector displacement (TVD) and systolic and diastolic right ventricular pressure (P) during right atrial (RA) pacing and at 5 and 60 minutes of left ventricular (LV) pacing. Arrows indicate TVD just after terminating 5 minutes (point d) and 60 minutes (point f) of LV pacing. B: Summary data for activation time (ATx), QRS duration (QRSx), QT interval (QTx measured in lead X) TVD, and systolic and diastolic right ventricular pressure (P) during RA pacing and at 5 minutes of LV pacing. In panels A and B: a ⫽ RA pacing before changing to LV pacing (control); b and c ⫽ 10 seconds and 5 minutes of LV pacing; d and e ⫽ 10 seconds and 15 minutes after resumption of RA pacing following 5 minutes of LV pacing; f ⫽ 10 seconds after resumption of RA pacing following 60 minutes of LV pacing. *P ⬍.05 vs a (n ⫽ 6).
Sosunov et al
Initiation of Cardiac Memory
Statistical analysis Data are expressed as mean ⫾ SEM. One-way analysis of variance for repeated or nonrepeated measures was used to compare means between groups. P ⬍.05 was considered significant.
Results Confirmation of the model Costard-Jackle et al16 previously used Langendorff-perfused rabbit heart to study initiation of cardiac memory, and we first tested whether we could reproduce their findings in our system. They recorded monophasic action potential from 12 to 20 different epicardial sites with a bipolar contact electrode. Activation time, action potential duration, and repolarization time in both ventricles were measured, first during right atrial pacing, then during 120 minutes of right ventricular pacing, and then again during 60 minutes of atrial pacing. Data were analyzed using linear regression analysis to consider action potential duration as a function of activation time. Four preliminary experiments were performed using the same methods for monophasic action potential recording and data analysis as those of Costard-Jackle et al16 (Figure 2). Our data were similar to theirs in, that during the initial atrial pacing period, action potential duration was inversely related to activation time (Figure 2A). With onset of ventricular pacing, this inverse correlation disappeared (Figure
109 2B). However, continuing ventricular pacing restored the inverse relation (Figure 2C). Returning to atrial pacing again perturbed the inverse correlation (Figure 2D), but after 60 minutes of atrial pacing an inverse correlation was reestablished (Figure 2E). This adjustment of ventricular myocardial repolarization sequence to a change in activation sequence was considered by Costard-Jackle et al16 as a manifestation of cardiac memory in isolated heart. We did not use the Costard-Jackle et al16 method of memory assessment in the present study because the time required to perform multisite monophasic action potential mapping does not permit detection of rapid (5–15 minutes) changes in ventricular repolarization and because recording the quasi-ECGs of isolated heart and quantifying memory as T-wave vector displacement permits us to relate our results to those reported in intact canine2,3,5,6,9 and human17–19 heart. In six hearts, we then studied the effects of two durations of ventricular pacing (5 and 60 minutes) in different sequences using the protocol shown in Figure 1. Figure 3A shows representative T-wave vector displacement and right ventricular pressure data. Because no differences in the magnitude of T-wave vector displacement and time course of its recovery were observed in the 5- and 60-minute pacing protocols (215 ⫾ 56 V at point d and 218 ⫾ 50 V at point f, n ⫽ 6, P ⬎.05), all further protocols were performed using 5 minutes of ventricular pacing.
Figure 4 A: Schematic diagram showing shunts introduced through the tricuspid valve into the right ventricle and through the mitral valve into the left ventricle. Representative recordings show ECG leads and right ventricular pressure at steady-state during right atrial (RA) pacing, 10 seconds of left ventricular (LV) pacing, 5 minutes of LV pacing, 10 seconds after resumption of RA pacing, and 15 minutes after resumption of RA pacing before (B) and after insertion of shunts in both ventricles (C). See text for discussion.
110
Heart Rhythm, Vol 5, No 1, January 2008
Figure 3B summarizes changes of the major ECG parameters, T-wave vector displacement, and right ventricular developed pressure during the 5-minute ventricular pacing protocol. Lead X activation time (which was measured from stimulus to Q wave), QRS duration, and QT interval changed significantly during ventricular pacing but recovered to control immediately on resumption of atrial pacing. Similar changes were seen in leads Y and Z (data not shown). On resumption of atrial pacing, T-wave vector displacement differed significantly from control and then returned to control in 15 minutes. Right ventricular pressure remained unaltered during the course of the protocol. To study the role of angiotensin II in memory initiation, we used the AT-1 receptor blocker losartan. The 5-minute pacing protocol was performed first during perfusion with control Tyrode’s solution. After complete T-wave vector displacement dissipation, the pacing protocol was repeated in the presence of 2 M losartan. The blocker had no effects on memory expression: 202 ⫾ 46 V and 196 ⫾ 45 V, respectively (n ⫽ 4, P ⬎.05). We then proceeded with the four experiments on altered activation and stretch. Experiment 1: A representative experiment in which shunts were inserted into the right and left ventricles to remove ventricular load is shown in Figure 4. Before shunting (Figure 4B), ventricles contracted isovolumetrically, and 5 minutes of ventricular pacing induced marked changes in T-wave shape in all three ECG leads (compare right atrial pacing before and 10 seconds after ventricular pacing). Shunt insertion led to a prominent drop of developed pressure, and ventricular pacing produced no changes in any ECG lead (Figure 4C). Summary data for five experiments showed that shunt insertion produced a significant reduction of developed pressure (from 21 ⫾ 2 mmHg to 5 ⫾ 1 mmHg, P ⬍.05) associated with a significant decrease of ventricular pacing-induced T-wave vector displacement (from 196 ⫾ 38 V before shunting to 11 ⫾ 8 V after shunt insertion, P ⬍.05). No significant differences in major ECG parameters occurred during ventricular pacing before and after shunt insertion (Table 1, Shunts), suggesting that shunting did not alter the overall ventricular Table 1 ECG parameters during left ventricular pacing in control (baseline) and after insertion of shunts into right and left ventricles (top) or perfusion with 0.4 ⌴ blebbistatin (bottom) ATx (ms) Shunts (n ⫽ 5) Baseline Shunts Blebbistatin (n ⫽ 10) Baseline Blebbistatin
QRSx (ms)
QTx (ms)
8⫾1 8⫾1
67 ⫾ 5 65 ⫾ 3
210 ⫾ 3 213 ⫾ 2
10 ⫾ 1 10 ⫾ 2
65 ⫾ 3 67 ⫾ 2
212 ⫾ 4 211 ⫾ 9
Values are given as mean ⫾ SEM values. No significant changes in ECG parameters in leads Y and Z were observed (data not shown). ATx ⫽ activation time (time interval between stimulus artifact and Q-wave); QRSx ⫽ QRS duration; QTx ⫽ QT interval measured in lead X.
Figure 5 Superposition of ECG leads during right atrial pacing before (trace C) and 10 seconds after 5 minutes of left ventricular pacing (trace CM). A–E: Pacing protocol was performed at different degrees of balloon inflation resulting in different left ventricular developed pressure (dev P). F: Dependence of T-wave vector displacement (TVD) on left ventricular developed pressure. See text for discussion.
activation pattern. Given that the same alteration in activation is observed with ventricular pacing in the setting of normal and of reduced pressure, we reasoned that a variable(s) other than altered activation must contribute to the onset of memory. The remaining experiments explored this possibility. Experiment 2: The purpose of this protocol was to assess the relationship between left ventricular developed pressure and pacing-induced T-wave vector displacement. Figure 5 shows results from 1 of 4 experiments in which the left ventricle was loaded with a fluid-filled latex balloon placed in the ventricular chamber. Ventricular load could be altered by changing the extent of balloon inflation. Note that 5-minute ventricular pacing performed at a high left ventricular developed pressure led to prominent T-wave changes in leads X and Y (Figure 5A: compare control trace [C] with the trace recorded 10 seconds after the end of left ventricular pacing [cardiac memory]). After the ECG had completely recovered, the extent of balloon inflation was decreased and the ventricular pacing protocol repeated (Figure 5B). T-wave changes were seen again, but they were less prominent than at higher developed pressure (compare Figures 5A and 5B). Further decreases in balloon inflation resulted in progressive reductions of ventricular developed
Sosunov et al
Initiation of Cardiac Memory
111
Figure 6 Effects of blebbistatin on ECG and right ventricular pressure. Representative recordings of ECG leads and right ventricular pressure at steady-state during right atrial (RA) pacing, 10 seconds of left ventricular (LV) pacing, 5 minutes of LV pacing, 10 seconds after resumption of RA pacing, and 15 minutes after resumption of RA pacing in control (A), after 30 minutes of perfusion with Tyrode’s solution containing 0.4 M of blebbistatin (B), and after 2 hours of blebbistatin washout (C). See text for discussion.
pressure and T-wave changes (Figures 5C and 5D), whereas partial restoration of balloon inflation led to an increase of developed pressure and T-wave changes (Figure 5E). Twave vector displacement was calculated for all these experimental conditions and plotted against left ventricular developed pressure (Figure 5F). Note the direct relationship between these values. A similar result was seen in all four hearts studied. Experiment 3: This protocol used blebbistatin to uncouple excitation and contraction. Before perfusion with blebbistatin, 5-minute ventricular pacing induced changes of T-wave shape in all ECG leads (compare “RA” with “10 s RA” in Figure 6A). Inhibition of myocardial contraction by blebbistatin resulted in a marked fall of right ventricular pressure (Figure 6B). Here, no T-wave changes were seen in any ECG lead after ventricular pacing. Two hours of blebbistatin washout partially restored both ventricular pressure and T-wave changes induced by ventricular pacing (Figure 6C). Summary data for 10 experiments demonstrated that blebbistatin significantly reduced developed pressure (from 20 ⫾ 2 mmHg to 5 ⫾ 2 mmHg, P ⬍.05) and ventricular pacing-induced T-wave vector displacement (from 213 ⫾ 25 V to 6 ⫾ 5 V, P ⬍.05). As in the experiments with shunts, blebbistatin did not alter ECG parameters during ventricular pacing, suggesting that it did not alter activation (Table 1, Blebbistatin). Experiment 4: Figure 7 shows typical results from 1 of 5 experiments in which ⬃15% stretch was applied locally to the left ventricle. The heart was continuously paced from the right atrium at a constant cycle length of 400 ms. Local
stretch applied for 5 minutes produced changes in T-wave morphology in all ECG leads (Figure 7B). Release of stretch led to gradual recovery of T-waves to control. At this time in each experiment, 5-minute ventricular pacing to induce cardiac memory was performed. This intervention produced T-wave vector displacement changes comparable to those of stretch (Figures 7C and 7D, respectively). The time course of T-wave vector displacement dissipation after release of stretch was similar to that seen after resumption of atrial pacing in the ventricular pacing protocol. These comparable patterns of stretch-induced T-wave vector displacement were observed in all five experiments.
Discussion The T-wave reflects epicardial potential gradients during ventricular repolarization that result from spatial heterogeneity in repolarization time.20 –22 Therefore, the altered T-wave morphology in cardiac memory implies that ventricular pacing induces spatial changes in ventricular repolarization that persist after resumption of supraventricular activation. Switching the pacemaker site from atrium to ventricle dramatically alters the sequence of ventricular electrical activation. It has been suggested that altered spatial electrotonic interactions occurring during cardiac activation create the stimulus for cardiac memory.1,23 An alternative proposal has been that a pacing-induced altered ventricular activation pattern modifies the spatial stress– strain relationships of the myocardial wall to initiate cardiac memory.9 –11 To discriminate between these two possible triggers, we used interventions that influenced one variable
112
Heart Rhythm, Vol 5, No 1, January 2008
Figure 7 Effects of local ventricular stretch on T-wave morphology and T-wave vector displacement (TVD). A: Schematic diagram showing application of stretching device on left ventricular surface. Left ventricular balloon is in place. B: Representative recordings show ECG leads and left ventricular pressure at various times during the stretching protocol. Heart was continuously paced from right atrium. Shown from left to right are steady-state before stretch, 10 seconds of local left ventricular stretch, 5 minutes of local left ventricular stretch, 10 seconds after release of stretch, and 15 minutes after release of stretch. C, D: Time course of TVD induced by 5 minutes of ventricular pacing or 5 minutes of local ventricular stretch, respectively.
(contractile pattern) while leaving another (activation pattern) unchanged. Experiments shown in Figures 4 through 6 show that decreasing load, contraction, or pressure all reduce the expression of cardiac memory. At the same time, developed pressure per se cannot be a trigger of cardiac memory because it remains constant when the site of pacing is moved from right atrium to left ventricle (Figure 1), suggesting by elimination of alternatives that the pattern of stretch must be the primary trigger. Indeed, we also demonstrated (Figure 7) that altering stretch in the absence of any change of load, pressure, or contraction induces T-wave vector displacement similar to that induced by ventricular pacing with respect to both magnitude and time course of dissipation. Because stretch is a final common pathway for load, contraction, or pressure changes, it appears that stretch is the ultimate determinant of cardiac memory. Jeyaraj et al11 compared acute effects of atrial, anterior, and posterior left ventricular pacing in anterior and posterior segments of canine left ventricle and found that left ventricular pacing induced immediate changes of circumferential strain localized in the segment distant from the pacing site. In another group of dogs, after a few weeks of similar pacing from the same two sites, action potential duration measured in wedge preparations of the respective left ventricular segments was longer in the segment in which the circumferential strain had increased in acute experiments. The authors suggested that long-term cardiac memory observed after a few weeks of left ventricular pacing could be triggered by mechanoelectrical feedback induced by the
change in cardiac activation sequence. In these experiments, spatial patterns of electrical activation and mechanical stress were not discriminated. That is, changes of strain were induced by altering activation sequence, and either activation or stress could be the primary trigger for cardiac memory. In the setting of short-term cardiac memory, our findings suggest that the mechanical factor is the primary trigger. Demonstration of a local change in strain within the left ventricular wall during short-term left ventricular pacing by Jeyaraj et al11 is consistent with our conclusion.
Study limitations A limitation of our study is seen with regard to the observation made by Rosenbaum et al1 that accumulation is a distinctive feature of cardiac memory. That is, in both human and in canine heart in situ, the longer that abnormal ventricular activation persists, the more memory is expressed and the greater time is necessary for recovery of T-wave morphology.1,3,23 In contrast, we did not observed accumulation in the rabbit heart. T-wave vector displacement magnitude and dissipation were the same after 5 minutes and 60 minutes of ventricular pacing. Importantly, in the same isolated rabbit heart model, Costard-Jackle et al16 did not report any memory accumulation. Thus, our study and that of Costard-Jackle et al suggest that memory accumulation in isolated rabbit heart is different from that in human and canine heart in situ3,9,17,19 and isolated perfused canine heart.9 One possible explanation for this difference is as follows. We previously demonstrated that transient out-
Sosunov et al
Initiation of Cardiac Memory
ward current Ito is important to the development of shortterm cardiac memory.3,24 There are species-specific differences in K⫹ channel expression. In both dogs and humans, Ito is determined by the pore-forming K⫹ channel subunit Kv4.3 and the accessory protein KChIP2.25,26 We previously showed downregulation of channel function in a gradient away from the pacing electrode in canine cardiac memory.27 However, in rabbit myocardium, Kv1.4 is a major contributor to Ito.28 Whether pacing to induce cardiac memory affects Kv1.4 in the same fashion as Kv4.3/ KChIP2 is not known. However, based on the results of the present study, we surmise that any importance of Kv1.4 to cardiac memory occurs via a very different mechanism. Another study limitation is in light of our prior finding that short-term cardiac memory in dog is angiotensin II– dependent.9 In contrast, in the present study, AT-1 receptor blockade did not influence the expression of memory in rabbit heart. Given that the action of angiotensin II in canine cardiac memory appears to derive from an effect on Ito,29,30 these data suggest different mechanisms of short-term cardiac memory in canine and rabbit hearts. Yet another possibility is that neither the pacing protocols we used nor angiotensin II is enough of a perturbation to displace the Ito in rabbit heart, which contributes more robustly to repolarization than does Ito in canine or human heart.
Conclusion Short-term cardiac memory (seen as ventricular pacing-induced T-wave vector displacement) occurs in isolated rabbit hearts. Initiation of cardiac memory is suppressed by interventions that decrease ventricular pressure or myocardial contraction without affecting activation pattern. Externally applied stretch in the absence of any change of load pressure or contraction induces T-wave vector displacement similar to that induced by ventricular pacing. The results suggest that spatial alterations in ventricular stretch during ventricular pacing are a key determinant to initiation of cardiac memory.
Acknowledgments We express our gratitude to Ms. Nimee Bhat for assistance in performing the studies and Ms. Eileen Franey for careful attention to manuscript preparation.
References 1. Rosenbaum MB, Blanco HH, Elizari MV, et al. Electrotonic modulation of the T wave and cardiac memory. Am J Cardiol 1982;50:213–222. 2. Shvilkin A, Danilo P Jr, et al. Evolution and resolution of long-term cardiac memory. Circulation 1998;97:1810 –1817. 3. Del Balzo U, Rosen MR. T-wave changes persisting after ventricular pacing in canine heart are altered by 4-aminopyridine but not by lidocaine: implications with respect to phenomenon of cardiac memory. Circulation 1992;85:1464 –1472. 4. Chatterjee K, Harris A, Davies G, et al. Electrocardiographic changes subsequent to artificial ventricular depolarization. Br Heart J 1969;31:770 –779. 5. Plotnikov AN, Yu HG, Geller JC, et al. Role of L-type calcium channels in pacing-induced short-term and long-term cardiac memory in canine heart. Circulation 2003;107:2844 –2849.
113 6. Patberg KW, Rosen MR. On the role of the cAMP response element binding protein in long-term cardiac memory. Circ Res 2003;93:e87. 7. Prinzen FW, Peschar M. Relation between the pacing induced sequence of activation and left ventricular pump function in animals. Pacing Clin Electrophysiol 2002;25:484 – 498. 8. Wyman BT, Hunter WC, Prinzen FW, et al. Effects of single- and biventricular pacing on temporal and spatial dynamics of ventricular contraction. Am J Physiol Heart Circ Physiol 2002;282:H372–H379. 9. Ricard P, Danilo P Jr, Cohen IS, et al. A role for the renin-angiotensin system in the evolution of cardiac memory. J Cardiovasc Electrophysiol 1999;10:545– 551. 10. Meghji P, Nazir SA, Dick DJ, et al. Regional workload induced changes in electrophysiology and immediate early gene expression in intact in situ porcine heart. J Mol Cell Cardiol 1997;29:3147–3155. 11. Jeyaraj D, Wilson LD, Zhong J, et al. Mechanoelectrical feedback as novel mechanism of cardiac electrical remodeling. Circulation 2007;115:3145– 3155. 12. Sosunov EA, Anyukhovsky EP, Rosen MR. Effects of exogenous neuropeptide Y on automaticity of isolated Purkinje fibers and atrium. J Mol Cell Cardiol 1996;28:967–975. 13. Uy RG, Ross-Ascuitto NT, et al. Recovery of the chronically hypoxic young rabbit heart reperfused following no-flow ischemia. Pediatr Cardiol 2006;27: 37– 46. 14. Kovacs M, Toth J, Hetenyi C, et al. Mechanism of blebbistatin inhibition of myosin II. J Biol Chem 2004;279:35557–35563. 15. Fedorov VV, Lozinsky IT, Sosunov EA, et al. Application of blebbistatin as an excitation-contraction uncoupler for electrophysiologic study of rat and rabbit hearts. Heart Rhythm 2007;4:619 – 626. 16. Costard-Jackle A, Goetsch B, Antz M, et al. Slow and long-lasting modulation of myocardial repolarization produced by ectopic activation in isolated rabbit hearts: evidence for cardiac memory. Circulation 1989;80:1412–1420. 17. Wecke L, Gadler F, Linde C, et al. Temporal characteristics of cardiac memory in humans: vectorcardiographic quantification in a model of cardiac pacing. Heart Rhythm 2005;2:28 –34. 18. Shvilkin A, Ho KKL, Rosen MR, et al. T-vector direction differentiates postpacing from ischemic T-wave inversion in precordial leads. Circulation 2005; 111:969 –974. 19. Wecke L, Rubulis A, Lundahl G, et al. Right ventricular pacing-induced electrophysiological remodeling in the human heart and its relationship to cardiac memory. Heart Rhythm 2007;4:1477–1486. 20. Spach MS, Barr RC, Lanning CF, et al. Origin of body-surface QRS and T-wave potentials from epicardial potential distributions in intact chimpanzee. Circulation 1977;55:268 –278. 21. Ghanem RN, Jia P, Ramanathan C, et al. Noninvasive electrocardiographic imaging (ECGI): comparison to intraoperative mapping in patients. Heart Rhythm 2005;2:339 –354. 22. Janse MJ, Sosunov EA, Anyukhovsky EP, et al. Repolarisation gradients in the canine left ventricle before and after induction of cardiac memory. Eur Heart J 2005;26:454 – 455. 23. Franz MR, Bargheer K, Costard-Jackle A, et al. Human ventricular repolarization and T-wave genesis. Prog Cardiovasc Dis 1991;33:369 –384. 24. Plotnikov AN, Sosunov EA, Patberg KW, et al. Cardiac memory evolves with age in association with development of the transient outward current. Circulation 2004;110:489 – 495. 25. Dixon JE, Shi W, Wang HS, et al. Role of the Kv4.3 K⫹ channel in ventricular muscle. A molecular correlate for the transient outward current. Circ Res 1996;79:659 – 668. 26. Schram G, Pourrier M, Melnyk P, et al. Differential distribution of cardiac ion channel expression as a basis for regional specialization in electrical function. Circ Res 2002;90:939 –950. 27. Yu H, McKinnon D, Dixon JE, et al. Transient outward current, Ito1, is altered in cardiac memory. Circulation 1999;99:1898 –1905. 28. Wang Z, Feng J, Shi H, et al. Potential molecular basis of different physiological properties of the transient outward K⫹ current in rabbit and human atrial myocytes. Circ Res 1999;84:551–561. 29. Yu H, Gao J, Wang H, et al. Effects of the renin-angiotensin system on the current Ito in epicardial and endocardial ventricular myocytes from the canine heart. Circ Res 2000;86:1062–1068. 30. Yu H, McKinnon D, Dixon JE, et al. Transient outward current, Ito1, is altered in cardiac memory. Circulation 1999;99:1898 –1905.