Transient Outward Current, Ito1, Is Altered in Cardiac Memory
Background—Cardiac memory refers to an altered T-wave morphology induced by ventricular pacing or arrhythmias that persist for variable intervals after resumption of sinus rhythm.
Methods and Results—We induced long-term cardiac memory (LTM) in conscious dogs by pacing the ventricles at 120 bpm for 3 weeks. ECGs were recorded daily for 1 hour, during which time pacing was discontinued. At terminal study, the heart was removed and the electrophysiology of left ventricular epicardial myocytes was investigated. Control (C) and LTM ECG did not differ, except for T-wave amplitude, which decreased from 0.12±0.18 to −0.34±0.21 mV (±SEM, P<0.05), and T-wave vector, which shifted from −37±12° to −143±4° (P<0.05). Epicardial action potentials revealed loss of the notch and lengthening of duration at 20 days (both P<0.05). Calcium-insensitive transient outward current (Ito) was investigated by whole-cell patch clamp. No difference in capacitance was seen in C and LTM myocytes. Ito activated on membrane depolarization to −25±1 mV in C and −7±1 mV (P<0.05) in LTM myocytes, indicating a positive voltage shift of activation. Ito density was reduced in LTM myocytes, and a decreased mRNA level for Kv4.3 was observed. Recovery of Ito from inactivation was significantly prolonged: it was 531±80 ms (n=10) in LTM and 27±6 ms (n=9) in C (P<0.05) at −65 mV.
Conclusions—Ito changes are associated with and can provide at least a partial explanation for action-potential and T-wave changes occurring with LTM.
Cardiac memory” refers to a change in the T wave induced by and persisting for minutes to months after cessation of a period of altered ventricular activation.1 2 3 4 Canine studies have shown that short-term T-wave memory is prevented by intravenous 4-aminopyridine (4-AP), suggesting that the memory might depend on pacing-induced changes in the transient outward current, Ito.5 Studies of an isolated tissue model of cardiac memory supported this view, demonstrating that an altered epicardial-endocardial voltage gradient contributes to memory and that 4-AP abolishes it.6
The expression of 4-AP–sensitive Ito is heterogeneous through the canine ventricular wall, with current density higher in epicardial than endocardial myocytes.7 8 We hypothesized that the 4-AP–sensitive Ito might be altered during the evolution of cardiac memory and used a chronic canine model to study it. The persistence of memory long after cessation of pacing permitted us to perform our experiments without fear that decay of the memory phenomenon would occur. We focussed on the 4-AP–sensitive Ito1 in normal canine epicardial myocytes and those from animals with “long-term memory.” For convenience, we use the term Ito instead of Ito1 throughout.
Mongrel dogs weighing 20 to 26 kg premedicated with propafol 5 mg/kg IV were anesthetized with 3% isoflurane and subjected to a left thoracotomy under sterile conditions. A Medtronic 5069 screw-in pacemaker (contact diameter, 0.020 inches) was inserted transepicardially into the anteroapical left ventricle (LV) and attached to a Medtronic 8340 programmable pacemaker embedded in a subcutaneous pocket. Animals recovered for 2 to 3 weeks, during which time ECGs were recorded every 2 to 3 days with the animal resting comfortably in a sling. Within 3 weeks, all postsurgical ECG changes had disappeared, and stable, reproducible recording of standard limb leads was demonstrated over a period of 1 week. LV pacing was then initiated at 120 bpm and maintained for 23 hours of each day. Holter monitoring was performed for 2 to 3 days of the pacing period. In no animal was there <75% capture. While pacing was discontinued, the ECG was recorded continuously to permit determination of ST-T waves in sinus rhythm. After 20 to 22 days of pacing, we performed terminal studies, which began within 3 hours of discontinuation of ventricular drive.
Isolated Tissue and Single Myocyte Studies
Animals were anesthetized with pentobarbital sodium 30 mg/kg IV, and the heart was rapidly excised and placed in iced Tyrode’s solution, containing (in mmol/L) NaCl 131, KCl 4, CaCl2 2.7, NaHCO3 18, MgCl2 0.5, NaH2PO4 1.8, and dextrose 5.5, gassed with 95% O2/5% CO2. For microelectrode studies, strips of LV free wall epicardium obtained 1 cm below the mitral ring and measuring ≈1 cm×1 cm×0.5 to 1.5 mm were filleted from control dogs and those studied at 20 to 22 days of pacing and having cardiac memory. Tissues were placed in a chamber perfused with Tyrode’s solution at 37±0.1°C and stimulated by standard techniques6 via bipolar silver electrodes insulated with Teflon. The tissue bath was connected to ground with a 3 mol/L KCl-Ag-AgCl bridge. Glass microelectrodes filled with 3 mol/L KCl were used to impale the tissues. Standard means for calibration and recording were used.6
Single cells were isolated from control hearts and those with cardiac memory. The free-wall epicardial layer (1 to 3 mm) near the LV base, ≥2 cm from the pacemaker, was excised as above and subjected to trituration as reported previously.9 Details of the transport solution and disaggregation have been described.10 Disaggregated cells were kept in KB medium11 at room temperature for ≥1 hour before electrophysiological experiments.
To avoid Ca2+ current contamination, Ito was recorded by the whole-cell patch-clamp technique in modified Tyrode’s solution containing (in mmol/L) NaCl 137.7, NaOH 2.3, MgCl2 1, glucose 10, HEPES 5, KCl 5.4, CaCl2 1.8, MnCl2 2, and CdCl2 0.2, pH 7.4). Pipettes were filled with solution containing (in mmol/L) NaCl 6, potassium aspartate 130, MgCl2 2, CaCl2 5, EGTA 11, HEPES 10, Na2-ATP 2, and Na-GTP 0.1; pH was adjusted to 7.2 by KOH. Pipette resistance was 2 to 4 MΩ. The pH of the Tyrode’s solution containing 4-AP was titrated to 7.4 with HCl. Temperature was maintained at 30°C to 32°C. An Axopatch 1B amplifier (Axon Instruments, Inc) was used.
We used Tyrode’s solution containing NaCl 137.7 mmol/L to facilitate comparison with action potential studies. Because INa was not blocked, we performed control experiments to determine whether the overlap of INa distorted our estimate of Ito threshold. No change in threshold occurred (ie, in 5 control myocytes, Ito threshold was −24±4 mV in both Na 140 and 10 mmol/L. At the same [Na+]o, the threshold in all 3 myocytes from memory animals was −10 mV.)
Data were collected by FM recording (Hewlett-Packard Co; 3964a, speed 7/8 in s-1, 600-Hz bandwidth) and pClamp software (Axon Instruments, Inc).
Preparation of RNA and RNase Protection Assay
Tissue samples were quick-frozen in liquid N2 and homogenized in guanidinium thiocyanate. Total RNA was prepared by pelleting the homogenate over a CsCl step gradient. All RNA samples were quantified spectrophotometrically.
Canine Kv4.3 and cyclophilin probes were prepared as previously.12 Significant nonhybridizing sequence (≈50 bp) was included in the probes to facilitate distinction between the probe and the specific protected band. There was no evidence for unwanted cross-reaction between probes and nonspecific transcripts.
RNase protection assays were performed as previously.13 For each sample point, 5 or 10 μg total RNA was used. A cyclophilin probe was included as an internal control to confirm that the sample was not lost or degraded during the assay. Yeast RNA 35 μg was a negative control to test for the presence of probe self-protection bands. RNA expression was quantified directly from dried RNase protection gels by use of a PhosphorImager (Molecular Dynamics). There was no significant change in cyclophilin expression between the control and memory samples (average cyclophilin expression in long-term memory samples was 96% of control, P>0.05).
Data are presented as mean±SEM. Comparisons of control and cardiac memory ECG from the same animals were done by paired t test. Comparisons of control and cardiac memory action potentials and ion currents were by t test for grouped data. A value of P<0.05 was considered significant.
ECGs and Epicardial Action Potentials
Cardiac memory was seen as an altered T-wave vector angle and amplitude (Table 1⇓). No other significant ECG changes occurred. Persistence of memory for 3 days after cessation of pacing is shown in Figure 1⇓. Action potential characteristics for tissue slabs from control and memory epicardial muscle are presented in Table 2⇓. The maximum diastolic potential and overshoot do not differ significantly in control and memory. However, the latter shows a smaller phase 1 notch and a longer action potential duration (APD) at day 20 (see example in Figure 2⇓).
Capacitance of Myocytes
To test whether the pacing protocols might alter cell size, we determined the capacitance of disaggregated myocytes from control and memory epicardium. The capacitance of control myocytes was 104±8 pF (n=26) and that of memory myocytes, 107±9 pF (n=19) (P>0.05).
Properties of Ito in Normal and Memory Epicardial Myocytes
The action potential results suggested that alterations in Ito could be important to the genesis of long-term memory. Figure 3⇓ demonstrates that Ito activates at a more positive potential in a myocyte from a dog with memory than in a control. The voltage thresholds for Ito in Figure 3⇓ are −30 mV for the control and 0 mV for memory. The changes in Ito threshold are summarized in Table 3⇓ and include an 18-mV positive shift in threshold. We could not construct complete activation curves because even at extremely positive voltages, saturation was not reached.
We next examined steady-state dependence of inactivation on voltage. We held the cell at a given potential for 400 ms and then stepped to the same test potential of 0 mV (see protocol, Figure 4⇓, inset). Results and a Boltzmann 2-state model fit to the data are provided in Figure 4⇓, and midpoint values and slope factors are in Table 3⇑. There is an 8-mV positive shift of the inactivation-versus-voltage curve in myocytes from hearts with memory, and no change in slope factor. Six additional experiments were performed with memory myocytes, with 2-second prepulses. No change in midpoint for inactivation voltage occurred (−35±3 mV, n=6).
Kinetic Properties of Inactivation.
Figure 5⇓ shows results from a protocol demonstrating the inactivation kinetics of Ito from normal and memory myocytes. Each cell was held at −65 mV and depolarized to the values indicated on the figure. There were no significant differences in inactivation time constants, although a positive shift in their voltage dependence was observed (for average values for time constants of inactivation at +20 mV, see Table 3⇑).
Despite the lack of significant difference in inactivation kinetics in memory, recovery from inactivation was dramatically altered (see Figure 6⇓). From a holding potential of −65 mV, we depolarized to +5 mV in control and +20 mV in memory (to maintain approximately equal activation), returned to the holding potential for a variable time, and again depolarized to the test potential. The ratio of Ito amplitude in the second pulse compared with the amplitude in response to the first was plotted against time along with the best fit to a monoexponential function in Figure 6C⇓. In Figure 6⇓, the time constant of recovery was 33 ms in control and 531 ms in memory. As summarized in Table 3⇑, the recovery time constant is more than an order of magnitude slower in cells from animals with memory.
Studies of hearts from several species have shown that Ito recovery from inactivation proceeds more rapidly with hyperpolarization.14 We investigated whether a similar voltage dependence exists for control and memory myocytes (Figure 6D⇑). Although Ito recovery from inactivation is an order of magnitude slower in memory than control, there is still the same speeding of recovery with hyperpolarization.
The changes in activation threshold and inactivation recovery kinetics provide a potential explanation for the reduction in the action potential notch recorded during cardiac memory. We also asked whether Ito conductance was reduced. Ideally, to estimate this conductance, one would saturate activation, measure tail currents at the same voltage, and normalize the tail currents to the capacitance of the cells. This was not possible because we could not fully activate the current at depolarized potentials. We therefore quantified Ito amplitude at each test potential and then divided current magnitude by the driving force to estimate the conductance at each potential. We corrected the measured conductance by the difference in inactivation between control and memory myocytes that we measured in Figure 4⇑ at the holding potential of −65 mV. Finally, we divided the measured values by the capacitance to obtain the normalized conductance against membrane potential (Figure 7⇓). Because activation threshold differs in the 2 preparations, we could not compare conductance at the same voltage. Instead, we compared conductance by fitting the data by linear regression. The slope of the line gives the increase in conductance normalized to capacitance per change in membrane potential. This slope was 2.3 pS/(pF · mV) in control and 1.5 pS/(pF · mV) in memory. Thus, the slope of conductance versus potential for Ito in memory is ≈65% of control.
Ito Block by 4-AP
Figure 8⇓ shows the effects of 4-AP on Ito in control (A, B, and C) and memory (D, E, and F). The holding potential was −65 mV for both cells, although the test potentials differed to allow for the differences in activation threshold. In both cells, 1 mmol/L 4-AP blocked ≈50% of the Ito at potentials ≥20 mV positive to voltage threshold. We found little difference in percentage block by 1 mmol/L 4-AP at +15 mV in control and memory (see Table 3⇑).
Changes in Kv4.3 mRNA
The Kv4.3 gene encodes a potassium channel that underlies the bulk of Ito in canine ventricular myocytes.12 In rat heart, regulation of Ito expression appears to be determined predominantly at the level of transcription.13 15 Therefore, we were interested in ascertaining whether changes in the level of expression of the Kv4.3 gene could contribute to the decline in Ito in memory. The value of Kv4.3 expression in memory was 65±7% (n=6, P<0.05) of control (Figure 9⇓), a change similar in magnitude to that in Ito conductance with voltage. This result suggests that expression of the Kv4.3 gene is downregulated in ventricular muscle during memory and that this downregulation may contribute to changes in the Ito magnitude observed in the isolated cell recordings.
Explanations accounting for cardiac memory have included an altered repolarization pattern resulting in a changed electrical gradient across the myocardium,1 16 ischemia,2 17 18 changes in mechanoelectrical feedback,19 protein phosphorylation, and gene expression.3 The hypothesis that ion channel alterations induce changes in the voltage-time course of repolarization and myocardial voltage gradients contributing to the evolution of memory is based on the following:
First, the T wave is generated by the voltage gradient during ventricular repolarization. Action potential configurations vary transmyocardially, one significant difference being the phase 1 “notch” seen in epicardium and midmyocardium but not endocardium.14 20 Ito appears to be largely responsible for this notch14 20 ; when Ito is blocked, the plateau is elevated and the time course of repolarization changes.
Second, epicardial action potentials recorded during memory in isolated tissue models6 and the present experiment lack a notch; they resemble those from endocardium. This observation led to the hypothesis that cardiac memory is largely due to a change in the voltage gradient that normally exists between epicardium and endocardium.6 It was also hypothesized that if this voltage gradient were minimized by pharmacological interventions, memory might be prevented. This hypothesis gained support when 4-AP (an Ito blocker) prevented development of memory in the intact canine heart5 and an isolated tissue model.6
The above-described work was performed in models of short-term cardiac memory, lasting minutes. Such models are of limited value in evaluating ion channel contributions to memory, because the electrophysiological changes do not persist long enough to ensure a stable substrate for myocardial disaggregation and whole-cell study. The use of a long-term memory model that we have demonstrated to be unassociated with cardiac failure21 or hypertrophy (cell capacitance, reflecting cell size, is unchanged) provides long-term T-wave changes that render the study of disaggregated myocytes practicable. Moreover, we have previously shown that the statistically insignificant variations in sinus rate occurring before and after pacing cannot be the cause of these T-wave changes.21
The action potentials we recorded are entirely consistent with results from earlier isolated tissue studies in that the notch of the epicardial action potential in T-wave memory is smaller than in controls.6 In short-term pacing, however, the epicardial APD did not manifest the prolongation seen with long-term pacing and cardiac memory in our present or previous study. The fact that the APD is not significantly prolonged in a memory model of short duration6 suggests that in the long-term setting, repolarizing currents other than Ito may be involved (possibly ICa,L, IK, Na/K pump current, IK1). This possibility is also suggested by the study of Shvilkin et al.21 Here, epicardial, endocardial, and midmyocardial action potentials were studied with microelectrodes in animals paced for ≥20 days. Significant prolongation of epicardial and endocardial action potentials as well as reduction of the difference in duration between the 2 types of action potentials occurred, such that epicardial action potentials prolonged to a greater extent than endocardial. In contrast, APDs of midmyocardial cells were unchanged. Hence, with memory, there was a reorientation of the interrelationships of APDs of each myocardial tissue. A prolongation of endocardial action potentials, which have a small notch and little Ito, and no change in duration of M cell action potentials, which have significant Ito, provides a further argument that currents other than Ito are involved in the memory process.
Kinetics of Ito
Despite the need to expand our studies to include other ion channels, our observations concerning Ito provide an understanding of the fundamental alterations in one ion channel that contribute to the memory phenomenon. Specifically, the shift in long-term memory Ito activation toward more positive potentials by almost +20 mV (Table 3⇑) is consistent with the absence of a notch in ventricular epicardial action potentials from dogs with cardiac memory. Ito activation at relatively positive potentials will contribute reduced outward current to phase 1 repolarization. In the physiological voltage range, the memory-induced positive shift of Ito activation along the voltage axis in epicardium mimics the activation of Ito in endocardium and therefore would decrease the voltage gradient from epicardium to endocardium.
A significantly longer time course of Ito recovery from inactivation was found in memory myocytes. This phenomenon is an additional contributor to the loss of the action potential notch recorded in epicardial myocytes from animals with cardiac memory; eg, at a ventricular pacing rate 20% greater than sinus rate, the basic cycle length was ≈500 ms. Here, a 531-ms time constant of recovery from inactivation will not allow Ito to recover to a significant degree (even if it is activated at the end of phase 0).
A major question relates to the identity of the signaling mechanism responsible for the change in Ito. Our work in progress22 23 and the published literature3 19 24 suggest that altered stress-strain relationships induced by ventricular pacing activate the endogenous cardiac renin–angiotensin II system; this protein kinase C–linked signaling cascade both alters channel phosphorylation (perhaps explaining the changes in Ito in short-term memory) and induces the immediate early gene program, which alters new protein (channel) synthesis. Although we hypothesize that new gene expression and protein synthesis are involved in long-term memory, we do not know which genes are expressed and which new proteins are synthesized. Even though the sensitivity to 4-AP is not significantly changed in long-term memory myocytes, this result does not completely rule out the possibility of an altered isoform, which remains to be investigated.
The clinical importance of cardiac memory initially was said to lie in its ECG patterns, mimicking those of ischemia.1 4 17 However, the potential importance of the memory phenomenon appears to be greater, as follows: first, any event that alters repolarization might alter refractoriness as well, and changes in the effective refractory period can be permissive of or suppress arrhythmias. Second, in the setting of tachycardia-induced CHF when heart rate is also rapid, abnormalities in Ito and in the T wave occur.25 The possibility exists that the rate- and activation-induced changes in Ito induced by the memory phenomenon are important contributors to the T-wave anomalies in such pathological conditions. Third, recent studies by Allessie et al26 27 have demonstrated pacing-induced atrial fibrillation in the goat. This, too, is associated with accelerated repolarization and loss of the action potential notch. Here, an atrial analog to the memory phenomenon might be considered. The statement by Wijffels et al that “fibrillation begets fibrillation”27 leads us to suggest that any change in rhythm that alters activation pathway and rate might alter the action potential and render persistence/recurrence of the arrhythmia more likely. Similarly, maintenance of sinus rhythm and normal activation and/or interventions that maximize Ito might turn out to be antiarrhythmic. Finally, these initial results demonstrate that this stable canine model for cardiac memory can provide a “window” through which the alterations of ion channel characteristics that underlie altered repolarization and their relationship to signal transduction pathways can be investigated intensively.
These studies were supported by US Public Health Service–NHLBI grants HL-28958, HL-53956, and HL-20558 and a grant from the Wild Wings Foundation. We express our great appreciation to Joan Zuckerman for her excellent cell dissociation; to Judy Samarel, Eileen Franey, Rachel Rosen, and Ester Chin for their patience in data entry and in preparing the manuscript; and to Dr Natalia Egorova for her assistance with the intact animal and microelectrode studies.
- Received August 4, 1998.
- Revision received November 23, 1998.
- Accepted November 30, 1998.
- Copyright © 1999 by American Heart Association
Del Balzo U, Rosen MR. T wave changes persisting after ventricular pacing in canine heart are altered by 4-aminopyridine but not by lidocaine. Circulation. 1992;85:1464–1472.
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.
Litovsky SH, Antzelevitch C. Transient outward current prominent in canine ventricular epicardium but not endocardium. Circ Res. 1988;62:116–126.
Liu DW, Gintant GA, Antzelevitch C. Ionic bases for electrophysiological distinctions among epicardial, midmyocardial, and endocardial myocytes from the free wall of the canine left ventricle. Circ Res. 1993;72:671–687.
Dixon JE, Shi W, Wang H-S, MacDonald C, Yu H, Wymore R, Cohen IS, McKinnon D. Role of the Kv4.3 K+ channel in ventricular muscle: a molecular correlate for the transient outward current. Circ Res. 1996;79:659–668.
Dixon JE, McKinnon D. Quantitative analyses of potassium channel mRNA expression in atrial and ventricular muscle of rats. Circ Res. 1994;75:252–260.
Campbell DL, Rasmussen RL, Comer MB, Strauss HC. The cardiac calcium-independent transient outward potassium current: kinetics, molecular properties, and role in ventricular repolarization. In: Zipes DP, Jalife J, eds. Cardiac Electrophysiology: From Cell to Bedside. 2nd ed. Philadelphia, Pa: WB Saunders Co; 1995:83–96.
Xu H, Dixon JE, Barry DM, Trimmer JS, Merlie JP, McKinnon D, Nerbonne JM. Developmental analysis reveals mismatches in the expression of K+ channel alpha subunits and voltage-gated K+ channel currents in rat ventricular myocytes. J Gen Physiol. 1996;108:405–419.
Costard-Jäckle A, Goetsch B, Antz M, Franz MR. Slow and long-lasting modulation of myocardial repolarization produced by ectopic activation in isolated rabbit hearts. Circulation. 1989;80:1412–1420.
Chatterjee K, Harris A, Davies G, Leatham A. Electrocardiographic changes subsequent to artificial ventricular depolarization. Br Heart J. 1969;31:770–779.
Lab MJ. Contraction-excitation feedback in myocardium. Circ Res. 1982;50:757–766.
Antzelevitch C, Sicouri S, Lukas A, Nesterenko VV, Liu DW, Di Diego JM. Regional differences in the electrophysiology of ventricular cells: physiological and clinical implications. In: Zipes DP, Jalife J, eds. Cardiac Electrophysiology: From Cell to Bedside. 2nd ed. Philadelphia, Pa: WB Saunders Co; 1995:228–245.
Shvilkin A, Daniel P Jr, Wang J, Burkhoff D, Anyukhovsky EP, Sosunov EA, Hara M, Rosen MR. The evolution and resolution of long-term cardiac memory. Circulation. 1998;97:1810–1817.
Rosen MR. T wave memory. J Cardiovasc Electrophysiol. 1996;7:787–788. Letter.
Ricard P, Yu H, Daniel P Jr, Gao J, Wymore R, Cohen IS, Rosen MR. The role of the renin-angiotensin II system in cardiac T-wave memory. Circulation. 1996;94(suppl I):I-6. Abstract.
Kääb S, Nuss B, Chiamvimonvat N, O’Rourke B, Pak PH, Kass DA, Marban E, Tomaselli GF. Ionic mechanism of action potential prolongation in ventricular myocytes from dogs with pacing-induced heart failure. Circ Res. 1996;78:262–273.
Allessie M, Konings K, Wijffels M. Electrophysiological mechanisms of atrial fibrillation. In: DiMarco JP, Prystowsky EN, eds. Atrial Arrhythmias: State of the Art. Armonk, NY: Futura Publishing Co; 1995:155–161.
Wijffels MCEF, Kirchhof CJHJ, Boersma LVA, Dorland R, Allessie MA. Atrial fibrillation begets atrial fibrillation. In: Allessie M, Campbell R, Olsson B, eds. Atrial Fibrillation: Mechanisms and Therapeutic Strategies. Armonk, NY: Futura Publishing Co; 1994:195–202.