Role of L-Type Calcium Channels in Pacing-Induced Short-Term and Long-Term Cardiac Memory in Canine Heart
Background— We tested the hypothesis that ICa,L is important to the development of cardiac memory.
Methods and Results— The effects of L-type Ca2+ channel blockade and β-blockade were tested on acutely anesthetized and on chronically instrumented, conscious dogs. Short-term memory (STM) was induced by 2 hours of ventricular pacing and long-term memory (LTM) by ventricular pacing for 21 days. STM dogs received placebo, nifedipine, or propranolol, and LTM dogs received placebo, atenolol, or amlodipine. AT1 receptor blockade (candesartan) and ACE inhibition (trandolapril) were also tested in LTM. Microelectrodes were used to record transmembrane potentials from isolated epicardial and endocardial slabs using a protocol simulating STM in intact animals. Left ventricular epicardial myocytes from LTM or sham control dogs were dissociated, and ICa,L was recorded (whole-cell patch-clamp technique). Evolution of STM and LTM was attenuated by ICa,L blockers but not β-blockers. Neither AT1 receptor blockade nor ACE inhibition suppressed LTM. In microelectrode experiments, pacing induced an epicardial-endocardial gradient change mimicking STM that was suppressed by nifedipine. In patch-clamp experiments, peak ICa,L density in LTM and control were equivalent, but activation was more positive and time constants of inactivation longer in LTM (P<0.05).
Conclusions— ICa,L blockade but not β-adrenergic blockade suppresses cardiac memory. LTM evolution is unaffected by angiotensin II blockade and is associated with altered ICa,L kinetics.
Received December 20, 2002; revision received February 27, 2003; accepted March 6, 2003.
Altering ventricular activation by pacing or arrhythmia induces cardiac memory (CM), in which T-wave changes tracking the ectopic ventricular QRS complexes persist after normal atrioventricular activation resumes.1–3 CM has clinical impact in that its ST-T–wave changes may mimic those of ischemia,2 and it may modify antiarrhythmic drug efficacy4 and expression of rhythm and arrhythmia.5 We reported previously that CM incorporates an altered transmural gradient for ventricular repolarization contributed to by decreased density and altered kinetics of the transient outward potassium current, Ito.6 However, the elevated plateau and prolonged action potential duration (APD) of epicardial and endocardial myocytes in CM are not readily explained by Ito changes alone. Moreover, angiotensin II, the synthesis of which in heart cells accompanies the altered stress/strain patterns imposed by ventricular pacing, seems to be involved in initiating CM,7 perhaps by suppressing Ito.8
The cellular responses to repeated ventricular stimulation that characterize CM have a parallel in neurons. For example, in aplysia, rapid, transcription-independent and long-term, transcription-dependent mechanisms are involved in long-term potentiation learning and memory (for review, see Reference 9). The signaling pathway includes serotonergic activation of cAMP and protein kinase A and angiotensin II–associated mechanisms linking to protein kinase C.9–12 Both pathways incorporate a role for Ca2+ as a critical messenger facilitating both short-term neuronal plasticity and long-term information storage.9
Given our inability to explain the evolution of CM solely on the basis of Ito and using the central nervous system as a paradigm, we have now studied calcium modulation of CM. Experiments were performed in intact dogs, isolated tissues,13 and disaggregated ventricular epicardial myocytes.
All protocols were approved by the Columbia University and State University of New York at Stony Brook Animal Care and Use committees.
Intact Animal Study
Protocols for Short-Term Memory
Mongrel dogs of either sex weighing 22 to 26 kg were premedicated with morphine sulfate 3 mg/kg IM and anesthetized with α-chloralose 100 mg/kg IV supplemented as needed during the experiment. Animals were intubated and ventilated with humidified room air. The femoral artery and the cephalic vein were catheterized to monitor blood pressure and maintain drug or placebo infusion, respectively. A heating pad was used to maintain body temperature.
The heart was suspended in a pericardial cradle via a thoracotomy at the fifth left intercostal space. Platinum bipolar electrodes were sewn epicardially to the left atrial appendage and the inferolateral left ventricular wall.
Dogs were divided into 3 groups: control (n=6) (intravenous placebo infusion), intravenous nifedipine (n=6), and intravenous propranolol (n=5) and studied according to a previously described protocol,4 as follows: equilibration during atrial pacing at a cycle length (CL) of 400 ms for 15 minutes; drug or placebo infusion for 30 minutes; and two 1-hour periods of ventricular pacing (VP) from the inferolateral left ventricular wall site at a CL of 400 ms, with each run followed by 20 minutes of atrial pacing at a CL of 400 ms. ECGs were recorded online, and 3-mL blood samples were drawn before drug and before and at the end of the VP and atrial pacing periods.
Protocol for Long-Term Memory
Dogs were anesthetized with propofol 6 mg/kg IV and inhalational isoflurane (1.5% to 2.5%). After intubation and ventilation and under sterile conditions, a thoracotomy was performed as above, and epicardial unipolar leads (model 4965, Medtronic) were attached to the inferolateral left ventricular wall and left atrial appendage. The leads were connected to a dual-chamber pacemaker (Prodigy DR, Medtronic) affixed subcutaneously. One platinum bipolar electrode was sewn to the left atrial appendage epicardium to permit bipolar atrial pacing for ECG recordings. Animals recovered for 2 to 3 weeks, during which they were laboratory trained and the ECG was stabilized.
All animals underwent 2 DDD pacing protocols of 21 days each (PR interval=50 to 60 ms to minimize competing and fusion beats). A 3-week recovery period for CM “washout” separated the protocols. Lower and upper tracking rates were 120 and 150 bpm, respectively. The extent of capture was evaluated in the laboratory twice during the first week and once weekly thereafter. Rate histograms were monitored throughout each week. In all dogs, the number of paced beats exceeded 90%, and no fusion was seen during daily observations.
ECGs were recorded before pacing was initiated and on days 7, 14, and 21 of DDD pacing. For all ECG recordings of CM, DDD pacing was turned off for 15 minutes and recordings were made during atrial pacing (AOO mode) at a CL of 500 ms in conscious dogs resting on the right side. We then returned to the DDD mode to continue the protocol.
Dogs were assigned to one of the following drug groups: amlodipine (n=6), atenolol (n=6), candesartan (n=3), or trandolapril (n=6). Amlodipine and atenolol were substituted for nifedipine and propranolol because they could be administered BID rather than the QID that would otherwise have been necessary. Dogs in each group except candesartan were randomized to 1 of 2 protocols: 3 received drug treatment during the initial 21-day pacing protocol and were drug-free during the second. The other 3 were drug-free during the first 21-day protocol and received drug during the second. One dog in the candesartan group received drug during an initial 14-day pacing protocol, and 2 dogs received candesartan during a second 14-day pacing protocol, with the other pacing protocol being control.
ECGs were recorded and frontal plane vector images plotted with Dr Vetter PC-EKG software (Dr Vetter, Baden Baden, Germany) as we described previously.4 QRS and QT intervals were read by 2 independent investigators from at least 5 consecutive complexes at each experimental time point, and averaged values were analyzed. The QT interval was measured from the onset of the QRS complex to the end of the T wave (defined as its return to the T/P baseline). Flat T waves with an amplitude of <0.1 mV were considered unmeasurable. The maximum value across 6 leads was considered the QT interval.
CM was quantified as a function of T-wave vector amplitude and angle changes and expressed as the distance between frontal plane T-wave vector peaks recorded during atrial pacing at VP0 and at each subsequent time point (T-wave vector displacement, in mV) as described previously.4
Nifedipine and propranolol were purchased from Sigma Chemical Co. Placebo was isotonic 0.9% NaCl (Baxter Healthcare). Nifedipine was administered as a 30-μg/kg IV bolus followed by a 3-μg · kg−1 · min−1 infusion. For propranolol, we used a 0.75-mg/kg IV bolus followed by a 0.01-mg · kg−1 · min−1 infusion.
Amlodipine (Norvasc, Pfizer) was purchased through the Columbia-Presbyterian Medical Center pharmacy and atenolol from Sigma-Aldrich Inc. Candesartan and trandolapril were generous gifts of AstraZeneca and Knoll AG, respectively.
All drugs were given orally throughout the protocol. Amlodipine and atenolol administration started 3 to 5 days before initiation of VP, and candesartan and trandolapril, 7 days before VP. Drug doses were as follows: amlodipine, 1 mg/kg BID; candesartan, 3 mg · kg−1 · d−1; trandolapril, 10 mg · kg−1 · d−1; and for atenolol, 3 dogs received 200 mg BID and 3 dogs an initial dose of 400 mg BID that was increased to 800 mg BID during the protocol.
Blood was centrifuged at 10 000 rpm for 30 minutes, and plasma was stored at −70°C. Drug levels were determined by high-performance liquid chromatography. Propranolol was quantified as previously described14 with 4′-methylpropranolol as internal standard. Nifedipine was extracted on solid-phase C18 Bond-Elut columns and analyzed further on a 3-μm Microsorb C18 column eluted with acetonitrile/methanol/water/acetic acid (31/20/49/1, vol/vol, adjusted to pH 4 with NH4OH) running at 1 mL/min and 40°C. The internal standard was nisoldipine.
Isolated Tissue Studies
Adult mongrel dogs were anesthetized with sodium pentobarbital 30 mg/kg IV. Hearts were rapidly removed and placed in ice-cold Tyrode’s solution containing (in mmol/L) NaCl 131, KCl 4, NaHCO3 18, NaH2PO4 1.8, MgCl2 0.5, CaCl2 2.7, and dextrose 5.5, equilibrated with 95% O2–5% CO2. Left ventricular epicardial and endocardial slabs (2×3 to 2×4 cm) were prepared as described previously13 and placed in a Lucite chamber perfused at 15 mL/min with Tyrode’s solution at 37±0.2°C. Transmembrane potentials were recorded with glass microelectrodes as described previously.4
Muscles were equilibrated for 4 to 5 hours and stimulated with rectangular pulses 0.5 to 1.5 ms in duration and 1.5 to 2 times diastolic threshold. One stimulating electrode was positioned at the basal end of the preparation such that impulses initiated here could proceed parallel to fiber orientation, simulating normal activation. During control, the preparation was paced through this electrode at a CL of 650 ms. A second stimulating electrode was positioned at the lateral border of each slab such that impulses initiated here originated perpendicular to the fiber axis. Three 20-minute periods of pacing at a CL of 450 ms were initiated from the lateral site, thereby altering activation sequence and rate and simulating VP. Each rapid pacing period with changed activation was interrupted by pacing from the basal end of the preparation at a CL of 650 ms for 30 minutes. This protocol was identical to those in our original in vivo and in vitro studies,13,15 which consistently evoked CM.
Transmembrane potentials recorded simultaneously from subepicardial and subendocardial muscle slabs were led into one side of a differential amplifier (model DAM 50, WPI), and the difference signal was recorded concurrently with the transmembrane potentials on a strip-chart recorder. Because it is not possible to reproduce the physiological endocardial-to-epicardial activation sequence in vitro, both AP upstrokes were superimposed during simulated normal activation. The area of the repolarization (“T”) wave in the difference signal was planimetered and expressed in square centimeters.
Three controls and 3 dogs manifesting CM after 21 days of pacing were anesthetized, and left ventricular epicardial myocytes were dissociated as previously described.6 Cells were stored in KB solution at room temperature for >1 hour before electrophysiological experiments. KB solution contained (in mmol/L) KCl 83, K2HPO4 30, MgSO4 5, Na-pyruvic acid 5, β-hydroxybutyric acid (sodium salt) 5, taurine 20, glucose 10, EGTA 0.5, HEPES 5, Na2ATP 5, and creatine 5, titrated to pH 7.2 with KOH.
Isolated cells were maintained at 32°C to 35°C (±0.5°C in each experiment). ICa,L was recorded by whole-cell patch-clamp technique. ICa,L amplitude was defined as the difference between the peak inward value and the current at the end of the test pulse (500 ms). Capacitance was measured from the voltage response to a constant hyperpolarizing current of 100 pA. Pipettes were filled with solution containing (in mmol/L) NaCl 6, K-aspartate 130, MgCl2 2, EGTA 11, and HEPES 10 (pH adjusted to 7.2 with KOH). The external solution contained (in mmol/L) NaCl 137.7, NaOH 2.3, MgCl2 1, glucose 10, HEPES 5, KCl 5.4, BaCl2 5, and tetrodotoxin 0.02, pH 7.4.
To evaluate the ICa,L contribution to APD, we constructed the peak inward ICa,L-V relation and determined the voltage dependence of the time constant of ICa,L inactivation. To construct the I-V relationship, we held the membrane at −65 mV and depolarized to test voltages from −30 to +80 mV in 10-mV increments.
Data from intact animals and isolated tissue studies were analyzed by 2-way repeated-measures ANOVA. Subsequent analysis was performed with Bonferroni’s test or the Games-Howell test as appropriate. One-way repeated-measures ANOVAs were used to analyze the influence of one factor on a dependent variable. In ion-channel studies, Student’s t test was used. Data are presented as mean±SEM; a value of P<0.05 was considered significant.
Intact Animal Studies
CM evolution in control experiments and during nifedipine or propranolol administration is depicted in Figure 1. In the controls, the T-wave vector was displaced from baseline and then declined toward baseline during each recovery period. Nifedipine plasma levels were 209±80 and 209±92 ng/mL at the start and end of VP, respectively (P>0.05). Propranolol plasma levels were 315±50 and 370±132 ng/mL at the start and end, respectively (P>0.05). Nifedipine decreased T-wave vector displacement significantly, whereas propranolol had no effect. Neither drug altered QRS complexes or QT intervals (data not shown). A representative experiment is shown in Figure 2.
Amlodipine had no consistent effects on QRS duration or QT interval of atrially paced complexes (data not shown). Figure 3 demonstrates that amlodipine significantly reduced the accumulation of CM. In contrast, neither atenolol, trandolapril, nor candesartan altered CM accumulation (Figure 3B).
Isolated Tissue Model of Short-Term Memory
AP characteristics are summarized in Tables 1 and 2⇓. Control values for maximum diastolic potential (epicardium, −83±0.4 mV; endocardium, −86±0.3 mV), AP overshoot (12±1.7 and 31±1.0 mV, respectively), and (192±14 and 204±18 V/sec, respectively) did not change during the protocol and were not significantly affected by nifedipine; hence, they are not reported further.
In control epicardium, there were small increases in the epicardial notch and plateau height and in APD30 and APD90 and an increase in endocardial APD30 by the end of the pacing protocol (Table 1). Nifedipine significantly increased the depth of the AP notch while depressing the plateau in epicardium and reducing the plateau and APD30 in endocardium (Table 2). The changes described above in the presence of pacing were markedly attenuated by nifedipine (Table 2 and Figure 4).
Although the AP changes were small, their expression in the transmural gradient was significant (Table 3). With pacing alone, a significant change in repolarization evolved by the third postpacing period. With nifedipine, the major change occurred with drug alone, and no further alteration was seen with pacing.
Single-Myocyte Studies of Long-Term Memory
Figure 5 shows I/V relationships for ICa,L. Maximum ICa,L in control and CM myocytes did not differ significantly. However, the I-V relation was shifted to more positive potentials in memory, indicated by the overall position of the I-V relation on the voltage axis and also by the 10-mV positive shift of the potential at which peak ICa,L occurs (0 mV control, +10 mV memory).
Figure 6 compares the time constant of inactivation in control and CM myocytes. There was a 10-mV positive shift in the voltage at which the minimum time constant occurs in memory (0 mV control, +10 mV memory). Furthermore, the time constants of inactivation were longer at each membrane voltage in memory than in control (eg, at +10 mV, control=40±3 ms, n=12; memory=81±3 ms, n=12; P<0.05). Thus, at potentials in the plateau voltage range (>10 mV), peak ICa,L is greater and its inactivation slower in CM.
ICa,L blockade but not β-adrenergic blockade prevents induction of CM by both short- and long-term pacing protocols. Although we had previously shown that ACE inhibition or angiotensin II receptor blockade suppresses short-term CM induced by ≈2 hours of pacing,7 we have now found that ACE inhibition (trandolapril) and AT1 receptor blockade (candesartan) do not suppress long-term CM. These results suggest that modulation of calcium handling may be important in initiating and sustaining CM and that angiotensin II is involved in initiation but not sustenance. Previous studies have demonstrated that angiotensin II increases ICa,L,16–18 which may contribute to initiation of CM.
The fact that β-adrenergic blockade did not alter short- or long-term CM induction is important for 2 reasons: first, β-adrenergic stimulation is an important modulator of ICa,L. Second, the cAMP-CREB pathway is important to the evolution of neuronal memory.9 We have shown that CREB levels are altered when the heart is paced to induce CM, a process blocked by saralasin or nifedipine.19 This increases the likelihood that initiation of memory and involvement of CREB as a transcription factor are important in heart and is associated with a role for angiotensin II.
ICa,L kinetics rather than current density seems to be important in the evolution of long-term CM. Specifically, ICa,L activation threshold is more positive and its inactivation is slower in CM than in control myocytes. These changes may contribute to the AP of CM as follows: we have previously demonstrated that long-term CM is associated with several changes in AP configuration. In epicardial myocytes, these changes include a decreased phase 1 notch, increased plateau amplitude, and longer APD and plateau.3 We have also reported that the decreased notch results from a decreased Ito density associated with a positive shift in activation threshold and a prolonged time constant for recovery from inactivation.6 The Ito changes occur in epicardium but not endocardium (which has little Ito), whereas the changes in plateau height and APD occur in epicardium and endocardium. The changes described in ICa,L kinetics in the present study could easily contribute to the increased APD and plateau height in both tissues.
The finding that Ca channel blockade suppresses short- and long-term CM provides a messenger that is involved in a variety of control processes in heart and that plays a role in neuronal plasticity as well.9 Although the transcription and translation involved in the processing of CM remain to be investigated, it is clear that Ca will play a central role. Whether the process relates to angiotensin II action, stretch-activated receptors, or other mechanisms remains to be determined. However, the finding of a consistent role for calcium coupled with our previous finding3 that protein synthesis inhibition suppresses long-term CM argues in favor of fundamental changes in the signaling processes that control the ion channels determining repolarization.
This study was supported by US Public Health Service, National Heart, Lung, and Blood Institute, grants HL-28958, HL-67101, and HL-20558 and the Bawd Foundation. Dr Yu is the recipient of an AHA Scientist Development Award. We express our gratitude to Drs Natalia Egorova and Irina Golyakhovsky for their assistance in performing certain of the studies and to Eileen Franey for her careful attention to the preparation of the manuscript.
↵*The first 3 authors contributed equally to this work.
Chatterjee K, Harris AM, Davies JG, et al. Electrocardiographic changes subsequent to artificial ventricular depolarization. Br Heart J. 1969; 31: 770–779.
Shvilkin A, Danilo P Jr, Wang J, et al. Evolution and resolution of long-term CM. Circulation. 1998; 97: 1810–1817.
Plotnikov AN, Shvilkin A, Xiong W, et al. Interactions between antiarrhythmic drugs and CM. Cardiovasc Res. 2001; 50: 335–344.
Yu H, McKinnon D, Dixon JE, et al. Transient outward current, Ito1, is altered in CM. Circulation. 1999; 99: 1898–1905.
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.
Kandel ER. The molecular biology of memory storage: a dialogue between genes and synapses. Science. 2001; 294: 1030–1038.
Ocorr KA, Walters ET, Byrne JH. Associative conditioning analog selectively increases cAMP levels of tail sensory neurons in Aplysia. Proc Natl Acad Sci U S A. 1985; 82: 2548–2552.
Sossin WS, Sacktor TC, Schwartz JH. Persistent activation of protein kinase C during the development of long-term facilitation in Aplysia. Learn Mem. 1994; 1: 189–202.
Geller JC, Rosen MR. Persistent T-wave changes after alteration of the ventricular activation sequence: new insights into cellular mechanisms of “cardiac memory”. Circulation. 1993; 88: 1811–1819.
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.
Aiello EA, Cingolani HE. Angiotensin II stimulates cardiac L-type Ca2+ current by a Ca2+- and protein kinase C-dependent mechanism. Am J Physiol. 2001; 280: H1528–H1536.
De Mello WC. Intracellular angiotensin II regulates the inward calcium current in cardiac myocytes. Hypertension. 1998; 32: 976–982.
Patberg KW, Plotnikov AN, Quamina A, et al. Ventricular pacing-induced degradation of the cAMP response element binding protein (CREB) in the canine heart is blocked by nifedipine. Keystone Symposium: Molecular Pathology of Cardiac Arrhythmias, Santa Fe, NM. 2003; 49.