Cardiac memory new insights into molecular mechanisms
1 Departments of Pharmacology and Pediatrics, Center for Molecular Therapeutics, College of Physicians & Surgeons of Columbia University, New York, NY, USA
2 Department of Physiology and Biophysics, Institute of Molecular Cardiology, Stony Brook University, Stony Brook, NY, USA
Abstract
‘Cardiac memory’ describes an electrocardiographic T wave vector change, recorded during normal sinus rhythm that reflects the QRS complex vector during prior periods of ventricular pacing or arrhythmia. In this brief review we consider the mechanisms responsible for cardiac memory, which offer a unique window for relating molecular determinants of repolarization to their expression in the function of ion channels and in the electrophysiology of the heart. Understanding the steps that translate the molecular mechanisms for memory into clinical expression in this relatively straightforward model facilitates our comprehension of the complex pathways that order normal cardiac repolarization and repolarization changes.
, 百拇医药
Introduction to cardiac memory
‘Cardiac memory’ is a term used clinically to describe a specific form of remodelling seen as an altered electrocardiographic T wave (Rosenbaum et al. 1982). Following the return to sinus rhythm after an interval of ventricular pacing or arrhythmia, the T wave vector persists in tracking the vector angle and amplitude of the QRS complex that characterized the paced or arrhythmic state (Fig. 1). The nomenclature is generally accepted, and the phenomenon is thought to have no inherent pathological implications (Rosenbaum et al. 1982; Rosen et al. 1998).
, 百拇医药
Upper panels show canine ECG during control sinus rhythm, during ventricular pacing at about 5% faster than sinus rate, and – a few minutes after returning to sinus rhythm – on days 7, 14 and 21 of pacing. Note that the QRS complex is inverted during ventricular pacing and that the T wave in the subsequent panels in sinus rhythm becomes progressively inverted, following the direction of the paced QRS complex. The two left bottom panels show the vectorcardiogram of the same dog in control (sinus rhythm) and during ventricular pacing. Note the change in vector shape, angle and amplitude of the QRS complex. The right lower panel shows an enlargement of the T wave vector during control sinus rhythm and in sinus rhythm on days 14 and 21. Note that the T wave vector has moved in the direction of the paced QRS, and also shows an increased amplitude. Modified from Shvilkin et al. (1998).
, 百拇医药
Descriptions of such T wave changes induced by arrhythmia or ventricular pacing have been published since the early days of electrocardiography. The first large series of pacing-induced changes was provided by Chatterjee et al. (1969) and the phenomenon was termed ‘cardiac memory’ by Rosenbaum et al. (1982). The latter investigators noted that the magnitude of T wave change increased with repetitions of the pacing periods or arrhythmias – referred to as ‘accumulation’– and that the T wave changes persisted for long intervals (months or longer) after a return to sinus rhythm. According to Rosenbaum et al. (1982), the sinus T wave ‘remembered’ the paced QRS complex. As yet, only one study has examined the persistence of pacing-induced cardiac memory prospectively (Wecke et al. 2005). Here one week of right ventricular apical endocardial pacing induced memory that persisted for about a month after cessation of pacing. It is likely, however, that pacing from different sites will result in memory of varying magnitudes and durations.
, 百拇医药
Although some experimental studies of cardiac memory were published before 1990 (Costard-Jackle et al. 1989; and summarized in Katz, 1992; Rosen et al. 1998), the lion's share of experimentation has occurred since. The likely explanation for increased interest in the last 15 years is the heightened appreciation of remodelling in general as a physiological and a pathological phenomenon as well as recent insights into the clinical impact of cardiac memory. With regard to the clinic, the preponderance of examples of cardiac memory derive from cardiac pacing which, as long as it alters ventricular activation, consistently induces memory. Other forms of abnormal activation that induce memory include intermittent left bundle branch block, ventricular arrhythmias, and atrioventricular bypass tracts (see Katz, 1992, for detailed review). In addition, the T wave changes of memory can mimic those of ischaemic heart disease (Rosenbaum et al. 1982; Katz, 1992; Shvilkin et al. 2004) and can impact on and either mask or augment the effects on ECG of anti-arrhythmic drugs and of other drugs that alter K+ channel function (Plotnikov et al. 2001). More controversially, and therefore still speculative, cardiac memory may modulate the type of cardiac rhythm expressed, may be arrhythmogenic (although this holds largely for atrium, see Chandra et al. 2003, 2005) and may be an early portent of hypertrophy to come (Rosen et al. 1998).
, http://www.100md.com
Relationship to memory in other systems
Whether wisely or not, when we began to study cardiac memory in the late 1980s, we accepted the term literally and looked to work in other organ systems for clues regarding mechanism. At that time by far the greatest effort had been reported in CNS (summarized in Kandel, 2001), although even then (and certainly since) there was evidence for memory in the immune system (Mills et al. 1975) and the gastrointestinal tract (Gershon, 1998). The key to our understanding of memory and to the design of our early studies came from reports on long-term potentiation in the sea slug Aplysia. In brief, the application of electrical shocks to Aplysia results in retraction of the siphon for intervals directly proportional in duration to the number of shocks delivered: the more stimuli, the longer the retraction (and the longer the memory) (summarized in Kandel, 1989). However crude, the association between this observation and the fact that repetitive shocking (i.e. pacing) of the ventricle led to occurrence and persistence of a specific T wave change was obvious. Given this common behaviour, we decided to use pathways then in vogue for memory in CNS as templates for our work in heart. As such, we tested hypotheses that suggested memory induced by brief pacing periods and of short duration (referred to as ‘short-term memory’) might result from altered signal transduction affecting the behaviour of specific ion channels, while memory induced by long pacing periods and persisting for weeks or months (‘long-term memory’) might result from altered gene transcription. Hence we worked in a system incorporating (1) a common stimulus (pacing), and (2) multiple potential contributory factors including some that could translate a briefly occurring event into one of long duration. Important to our understanding of the system's operation was the observation that pacing rate per se was not the determinant of cardiac memory (although memory evolved faster at more rapid rates); far more important was the altered ventricular activation pathway imposed by the pacing site (Shvilkin et al. 1998).
, http://www.100md.com
Often in the setting of altered activation there is altered myocardial stretch, but there is no certainty regarding whether changing the activation path in its own right can induce memory or if stretch is the key component. It is clear that changing stretch can increase angiotensin II synthesis/release by cardiac cells (Sadoshima & Izumo, 1993a,b) and can turn on the cardiac immediate early gene programme (Meghji et al. 1997). Learning more about mechanism is essential here as this information impacts directly on the question of why cardiac pacing induces repolarization changes.
, http://www.100md.com
Methods for study
A variety of models running the ‘cell-to-bedside’ gamut have been used to study cardiac memory. Studies through the mid-1980s employed patient populations (Chatterjee et al. 1969; Rosenbaum et al. 1982) and animal models (Costard-Jackle et al. 1989). As the need to study mechanism became more cellularly and subcellularly orientated, canine models were paced from the ventricles (mimicking the use of pacemakers in humans) for hours to weeks to track the characteristics of memory (Shvilkin et al. 1998; Yu et al. 1999). Terminal experiments at various times permitted harvesting of cells for isolated tissue electrophysiology and ion channel and molecular experiments.
, 百拇医药
In all these studies a key intention was to use pacing protocols that altered ventricular activation pathways as this seemed to be essential to inducing cardiac memory. Moreover, although not directed uniquely at understanding memory, experiments using pacing and magnetic resonance imaging provided insights into the relationship between pacing site and stress/strain relationships on the myocardium (McVeigh et al. 1998; Prinzen et al. 1999).
The need to understand how specific signalling molecules affected ion channel properties and ion channel trafficking brought in the use of disaggregated cells in culture media (Yu et al. 2000) as well as mammalian cell lines (Doronin et al. 2004). Moreover, as information regarding mechanism and expression of memory in animals has accumulated, there has been renewed interest in patient studies (Alessandrini et al. 1997; Goyal et al. 1998; Haverkamp et al. 1998; Nahlawi et al. 2004; Shvilkin et al. 2004; Wecke et al. 2005).
, 百拇医药
What causes the T wave changes of cardiac memory
The T wave on ECG is the summed signal of action potential repolarization in the cells of the ventricle. More specifically, the gradient of repolarization within the ventricles gives rise to the electrocardiographic ST segment and T wave (reviewed in Wit & Rosen, 1989). The sources of gradients in the ventricle include: that between apex and base, such that repolarization time (incorporating activation time through the end of repolarization) is shorter basally than apically; that between left and right ventricle (Volders et al. 1999); and that which is transmural – at any site epicardial action potentials are shorter than endocardial. Regarding the transmural gradient, there is controversy regarding the presence of longer action potentials midmyocardially than either epi- or endocardially in the ventricles in situ. When cells are isolated such transmural differences are clearly seen; however, at physiological heart rates this transmural gradient is minimized in situ, probably due to the electrotonic effects of cell–cell coupling (Anyukhovsky et al. 1999).
, http://www.100md.com
In experiments using pacing to induce short-term cardiac memory, no transmural gradient in repolarization was seen before or after 2 h of pacing. However, the pacing protocol resulted in shortening of apico-basal repolarization times throughout the ventricles, with epicardial monophasic action potential duration decreasing by about 10 ms at the 90% repolarization level (Janse et al. 2005). Whether this change in apico-basal gradients is the sole determinant of T wave changes during short-term memory or other gradient changes occur as well to explain the T waves of short- and long-term memory is yet to be determined. For example, the possibility of changes in the right–left ventricular gradient has not been studied in memory.
, 百拇医药
Mechanisms determining short-term memory
Studies of isolated rabbit ventricle (Costard-Jackle et al. 1989) and of canine heart (del Balzo & Rosen, 1992) paced for minutes to hours have shown memory persisting for minutes to hours in proportion to the duration of stimulation. Magnetic resonance imaging of canine heart indicates that altering the pacing site alters the timing of contraction and relaxation in various regions of the heart (McVeigh et al. 1998; Prinzen et al. 1999). This knowledge together with the observation that altering stretch on cardiac cell cultures increases their synthesis and release of angiotensin II (Sadoshima & Izumo, 1993a,b) led to the hypothesis that angiotensin II synthesis and release in the heart as a result of altered stretch imposed by ventricular pacing would initiate cardiac memory via functional changes in ion channels (Rosen et al. 1998). We then found that angiotensin-converting enzyme inhibition, AT-1 receptor blockade and tissue chymase inhibition all prevented the development of cardiac memory over 1–2 h periods in canine heart, leading us to consider AT-1 receptor-related pathways in the genesis of memory (Ricard et al. 1999). L-type Ca2+ channel blockers also prevented memory development in this canine model, while -adrenergic and muscarinic blockers had no effect (Plotnikov et al. 2003).
, 百拇医药
How might angiotensin II initiate expression of cardiac memory We had observed that 4-aminopyridine, a blocker of the transient outward potassium current Ito, prevents the onset of cardiac memory during short periods of pacing canine heart (del Balzo et al. 1992). As shall be detailed below, Ito density in canine left ventricular epicardial myocytes is reduced by chronic pacing to induce long-term cardiac memory (Yu et al. 1999), and exposure of non-paced canine epicardial myocytes to angiotensin II comparably reduces Ito (Yu et al. 2000) (Fig. 2). In experiments exploring potential associations between angiotensin II and Ito, (Doronin et al. 2004) the -subunit of the channel, Kv4.3, colocalized with the AT-1 receptor; the two coimmunoprecipitated in canine ventricular myocytes and, in HEK 293 cells, the complex coimmunoprecipitated with the accessory subunit KChIP2. Importantly, exposure of the AT-1 receptor–Kv4.3 complex to angiotensin II resulted in internalization of the complex (Fig. 3). Moreover, the gating properties of those Kv4.3 channels remaining at the cell surface shifted to more positive voltages. This led to the proposal that the AT-1 receptor provides an internalization molecular scaffold via which angiotensin II initiates the trafficking of Kv4.3 (and perhaps KChIP2) to internal sites where their degradation might occur (Doronin et al. 2004) (see diagram in Fig. 4). The net result of the trafficking would be a loss of Ito density. The model system also showed a concomitant positive shift in activation for the remaining reduced transient outward current. This reduced current density and positive shift in activation are precisely what is seen in myocytes from canine hearts as cardiac memory evolves (Yu et al. 1999).
, http://www.100md.com
A and B, representative records of a control epicardial myocyte (Epi) and an angiotensin II-exposed myocyte (Epi + A-II), respectively: note the marked diminution of current in the latter. C shows a summary of current–voltage data for the entire series. There is a significant reduction in current throughout the voltage range studied. D and E show action potentials from comparable cells. In D the phase 1 notch attributable to Ito is indicated by an arrow. The notch disappears on exposure to angiotensin II (E). Modified from Yu et al. (2000).
, http://www.100md.com
HEK 293 cells were transiently transfected with Kv4.3–V5–KChIP2-myc and the HA-AT1 receptor. The distribution of AT1 receptors and Kv4.3 was visualized after cell fixation with 3.7% formaldehyde using anti-HA-TRITC and anti-V5-FITC antibody before and after treatment with 1 μM angiotensin for 1 h. The distribution of Kv4.3 and AT1 receptors is shown before (A) and after treatment with angiotensin II (B). Nuclei (blue colour) were counterstained with 4,6-diamidino-2-phenylindole (DAPI). Note that the AT1 receptor is located predominantly on the cell surface where it colocalizes with Kv4.3 in the absence of angiotensin II (A). Angiotensin II induces internalization of the AT1 receptor (B), and with this Kv4.3 is also removed from the cell surface. The majority of the internalized AT1 receptors colocalize with Kv4.3 in intracellular vesicles. Reproduced with permission from Doronin et al. (2004).
, 百拇医药
A, coassembled AT-1 receptor–Kv4.3–KChIP2 macromolecular complexes inserted in the cell membrane. Channel opening results in outward K+ current. In B, angiotensin II binds to a subset of receptors, resulting in the internalization of the macromolecular complex. The net result is a loss of functioning channels in the cell membrane and a reduction in current (C).
The role for angiotensin II in short-term memory is reinforced by experiments in which epicardial or endocardial myocytes from control dogs were maintained for 2 h to 2 days in culture media, respectively, with angiotensin II or an AT-1 receptor blocker (Yu et al. 2000). In epicardium, angiotensin II prolonged repolarization and reduced the action potential notch and Ito, while having no effect on Kv4.3 mRNA. These actions were blocked by the addition of an AT-1 receptor blocker to the culture medium. In contrast, in endocardium, angiotensin II alone had no effect, but AT-1 receptor blockade accelerated repolarization, induced an action potential notch and increased Ito. This phenomenon is consistent with a transmural ventricular gradient for angiotensin II levels, as well as with a non-transcriptional pathway whereby Ito function is modulated to alter the transmural repolarization gradient. Whether the mechanism here is the same trafficking mechanism as described for short-term memory is unknown.
, 百拇医药
Finally, it would be tempting to try to relate the actions of angiotensin II to changes in ventricular function that are induced by pacing clinically, but we believe such speculation to be premature. One should keep in mind, however, that pacing specific ventricular sites results in differing regional patterns of synchrony/dyssynchrony (McVeigh et al. 1998; Prinzen et al. 1999) and that angiotensin II might be a contributor here.
Transitions between short- and long-term memory
, 百拇医药
Studies of the transitional period have largely focused on time points 30 min to 2 h after onset of pacing. During this interval not only are the trafficking changes that putatively lead to short-term memory occurring, but transcriptional changes are beginning which initiate the cascade of events leading to long-term memory. The role of altered stretch in inducing transcriptional change was noted by Meghji et al. (1997) in porcine hearts subjected to altered regional ventricular stretch for 30 min which manifested increased expression of the immediate early gene programme.
, 百拇医药
Emphasis has been placed on the role of the cyclic AMP response element binding protein (CREB), central to neural memory transcription in the evolution of cardiac memory. Two hours of ventricular pacing down-regulates nuclear CREB (blocked by AT-1 receptor or Ca2+ channel blockade) (Patberg et al. 2003) (Fig. 5). Change is maximal near the pacing site and diminishes with distance. Because the genes for Kv4.3 and KChIP2 (- and accessory subunits for canine Ito, respectively) manifest CREB binding capability in their promoter regions, we suggested that CREB might regulate one or both genes by directly binding in these regions (Patberg et al. 2003; Patberg & Rosen, 2004). That CREB can modulate Ito expression was shown in experiments in which a CREB antisense virus packaged with green fluorescent protein was injected into the canine left ventricular epicardium (Patberg et al. 2005). Terminal studies several days later showed the absence of a monophasic action potential notch in the region of epicardial injection as well as the absence of Ito in the antisense-infected cells (but not in nearby control cells). Hence an argument can be made that a CREB-transcribed system regulates at least one of the ion channel changes associated with cardiac memory. However, roles for additional transcriptional factors in memory induction are highly likely.
, 百拇医药
In these experiments, anaesthetized dogs were subjected to atrioventricular pacing with a short PR interval to ensure 100% capture of the ventricles during the 2 h period of pacing. Individual panels show results from an atrioventricular paced dog (AVP), a sham control (instrumented but not paced), an AV paced dog that received the L-type Ca2+ channel blocker nifedipine (AVP + nif), and an AV paced dog that received the AT-I and AT-II receptor blocker saralasin (AVP + sar). Ref indicates reference biopsy; 2 h, biopsy taken after 2 h of AVP. Loading conditions were controlled using an antibody against histone 1 (hist 1). Note the marked reduction in CREB levels at 2 h in the AV paced dog. There was no change in CREB in the sham animal or in those treated with nifedipine or saralasin. Reproduced with permission from Patberg et al. (2003).
, 百拇医药
Long-term memory
Long-term memory has been the most completely studied state, as the stability of the memory process after cessation of pacing ensures steady-state function with regard to ion channels and their molecular determinants (Shvilkin et al. 1998). In general such experiments are performed terminally 24 h after pacing ceases. Transmembrane action potentials at this time show a prolongation of duration in epicardium and endocardium, with a loss of the epicardial action potential notch in the former: this would be consistent with an altered transmural repolarization gradient (although this has not been studied in situ) (Shvilkin et al. 1998; Yu et al. 1999). Long-term memory accumulation in situ is attenuated by protein synthesis inhibitors and by L-type Ca2+ channel blockers, but not by ACE inhibitors or AT-1 receptor blockade (Shvilkin et al. 1998; Plotnikov et al. 2003). The mechanism whereby the angiotensin II effect is altered in the long-term setting is not yet known: whether it reflects something as simple as up-regulation of AT-1 receptors awaits determination. Studies of CREB in the long-term setting demonstrate its continued down-regulation (Patberg et al. 2003).
, 百拇医药
In long-term memory, Ito density is decreased as described above, and Kv4.3 mRNA and KChIP2 message and protein are decreased as well (Yu et al. 1999; Patberg et al. 2003) (Fig. 6). Note that whereas the change in current density is the same here as that seen with 2 h of exposure of myocytes to angiotensin II, there was no change in Kv4.3 mRNA in the latter instance (Yu et al. 2000) (consistent with the hypothesis that altered trafficking explains the reduction in current). That there is a reduction in Kv4.3 mRNA in long-term memory is consistent with altered gene transcription in this setting. In addition, epicardial Ito activates at more positive potentials and recovers from inactivation more slowly than in control myocytes (Yu et al. 1999). Other channels are involved as well: the L-type Ca2+ current (ICa,L) shows no change in density but activation is more positive and time constants of inactivation are longer in memory than control hearts (Plotnikov et al. 2003). Preliminary studies of the rapidly activating delayed rectifier currents (IKr) suggest that the transmural gradient in control dogs (a larger current epicardially than endocardially) is reduced in memory (Obreztchikova et al. 2003). All these ion channel changes are consistent with and likely contributors to the absent phase 1 notch, the high plateau and the prolonged action potential duration of cardiac memory. Still other experiments have shown that the gap junctional protein connexin43 is lateralized in distribution and reduced in density in long-term memory, with the greatest changes in the region of the pacing electrode (Patel et al. 2001). Whether these changes are associated with altered gap junctional function and contribute to the occurrence of long-term memory or are an epiphenomenon remains to be determined.
, http://www.100md.com
Upper panel: epicardial action potentials recorded from slabs of left ventricular tissue removed from a dog paced for 3 weeks to induce cardiac memory, and from a sham control. Note that in the setting of memory, the phase 1 notch diminishes, the plateau is increased in height and action potential duration is prolonged. Lower left panel shows Kv4.3 mRNA for a control dog and one in cardiac memory. Note that message is reduced in the memory animal (Cyc = cyclophyllin). Lower right panel shows Ito conductance for a group of controls and a group of memory animals. There is a significant decrease in channel conductance in the setting of memory. Modified from Yu et al. (1999).
, http://www.100md.com
Is Ito the entire story
Categorically not: while Ito, because of its large magnitude and ease of study provides a useful window into mechanism, everything we have learned regarding memory argues against too simplistic an interpretation. That Ito is important is best illustrated by the observation that short-term memory cannot be induced in animals having no Ito. Specifically, in neonatal dogs having no epicardial Ito, no phase 1 action potential notch and no transmural repolarization gradient, memory is not inducible; as age increases and current, notch and gradient appear, the ability to use pacing to induce memory increases proportionally (Plotnikov et al. 2004). The importance of other currents is highlighted by the findings that in long-term memory not only the epicardial, but the endocardial action potential (at which site there is little or no Ito) manifests an elevated plateau and prolonged action potential duration. This information, together with the important changes described in ICa,L kinetics (Plotnikov et al. 2003) and IKr density and kinetics (Obreztchikova et al. 2003) argues in favour of multichannel involvement in the memory process.
, 百拇医药
Issues still to be considered
Cardiac memory is a readily quantifiable phenomenon that permits us to relate repolarization changes in situ to their molecular, ion channel and cellular determinants. Hence its applicability to general knowledge regarding the regulation of repolarization is great. Yet the relationship between cardiac and other forms of memory is complex as is the interface between memory and other forms of cardiac remodelling.
, 百拇医药
For example, neuronal memory is a synthetic process: transcriptional factors such as CREB are up-regulated and synaptic connections between cells are strengthened as memory develops (Kandel, 2001). In contrast, cardiac memory sees CREB levels down-regulated and the function of the ion channels thus far identified as involved down-regulated as well (Yu et al. 1999; Plotnikov et al. 2003; Patberg et al. 2003, 2004). Far from being a non-specific degradative process, cardiac memory creates action potential and cardiac repolarization phenotypes much like those in neonatal heart. In other words, memory, which is a specialized stress response in both CNS and heart, results in an advancement in structure/function in the former and a return to an earlier structure/function entity in the latter.
, 百拇医药
Is memory in heart, then, the ‘forgetting’ of an adult phenotype or the reacquisition/strengthening of a neonatal phenotype Far from a mere semantic quandary, this question impacts on our understanding of cardiac remodelling. This stress response in remodelling sees earlier developmental patterns of repolarization emerging as well. For example, in canine cardiac failure models, the action potential notch diminishes, repolarization is prolonged, and Ito density decreases (Spragg et al. 2003). But – and this is important – the changes in Ito activation and inactivation gating that characterize cardiac memory are not seen in cardiac failure. In other words, the remodelling processes while in some ways appearing similar are in other ways different.
, 百拇医药
Although this may seem a small point, we believe it to be key: despite the obvious attractions of extremes in generalization or reductionism, it is the understanding of common ground and of subtle differences that may best move knowledge ahead.
Conclusions
While much has been done to study cardiac memory, many steps central to its evolution remain to be determined (see Fig. 7 which illustrates the pathways studied). Ito has been the landmark current considered to date, and the thread of information suggesting the internalization of its molecular determinants by angiotensin II in a short-term model through its transcriptional regulation by CREB in a long-term model is a tantalizing story that requires further investigation. Beyond this, the likelihood that memory represents a monogenically determined function as opposed to a complex array of changes is small, requiring a great deal of effort to identify the totality of pathways involved.
, 百拇医药
Pacing alters activation and stretch, resulting in angiotensin II synthesis/release and the trafficking and internalization of the AT-1 receptor–Kv4.3–KChIP2 complex from its membrane site and a reduction in current. Angiotensin II also induces an increase in L-type Ca2+ current. Other potential sources for increased Cai2+ would be the Na+–Ca2+ exchanger and stretch-activated channels. It appears that Cai2+ may be a second messenger activating changes in transcriptional factors in the nucleus. The transcriptional factor thus far studied, CREB, is reduced. An association with KChIP2 reduction has been demonstrated here. The other factors and linkages that may be involved have not been identified. However, long-term changes in Ito, IKr and ICa,L have been demonstrated as well, all of which would be expected to contribute to the altered action potentials and ECG changes of cardiac memory.
, 百拇医药
References
Alessandrini RS, McPherson DD, Kadish AH, Kane BJ & Goldberger JJ (1997). Cardiac memory: a mechanical and electrical phenomenon. Am J Physiol 272, H1952–H1959.
Anyukhovsky EP, Sosunov EA, Gainullin RZ & Rosen MR (1999). The controversial M cell. J Cardiovasc Electrophysiol 10, 244–260.
Chandra P, Rosen TS, Herweg B, Danilo P & Rosen MR (2003). Left atrial pacing induces memory and is associated with atrial tachyarrhythmias. Cardiovasc Res 60, 307–314.
, 百拇医药
Chandra P, Rosen TS, Herweg B, Plotnikov AN, Danilo P & Rosen MR (2005). Atrial gradient as a potential predictor of atrial fibrillation. Heart Rhythm 2, 404–410.
Chatterjee K, Harris AM, Davies JG & Leatham A (1969). T-wave changes after artificial pacing. Lancet 1, 759–760.
Costard-Jackle A, Goetsch B, Antz M & Franz MR (1989). Slow and long lasting modulation of myocardial repolarization produced by ectopic activation in isolated rabbit hearts. Evidence of cardiac ‘memory’. Circulation 80, 1412–1420.
, http://www.100md.com
del Balzo U & Rosen MR (1992). 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 85, 1464–1472.
Doronin SV, Potapova IA, Lu ZJ & Cohen IS (2004). Angiotensin receptor Type 1 forms a complex with the transient outward potassium channel Kv4.3 and regulates Its gating properties and intracellular localization. J Biol Chem 279, 48231–48237.
, 百拇医药
Gershon MD (1998). Genes, lineages, and tissue interactions in the development of the enteric nervous system. Am J Physiol 275, G869–G873.
Goyal R, Syed ZA, Mukhopadhyay PS, Souza J, Zivin A, Knoght BP, Man KC, Strickberger SA & Morady F (1998). Changes in cardiac repolarization following short periods of ventricular pacing. J Cardiovasc Electrophysiol 9, 269–280.
Haverkamp W, Hordt M, Breithardt G & Borggrefe M (1998). Torsades de pointes secondary to d,1-sotalol after catheter ablation of incessant atrioventricular reentrant tachycardia – evidence for a significant contribution of the ‘cardiac memory’. Clin Cardiol 21, 55–58.
, http://www.100md.com
Janse MJ, Sosunov EA, Coronel R, Opthof T, Anyukhovsky EP, deBakker JMT, Plotnikov AN, Shlapakova IN, Danilo P, Tijssen JGP & Rosen MR (2005). Repolarization gradient in the canine ventricle before and after induction of short term cardiac memory. Circulation 112, 1711–1718.
Kandel ER (1989). Genes, nerve cells and the remembrance of things past. J Neuropsychol 1, 103–110.
Kandel ER (2001). The molecular biology of memory storage: a dialogue between genes and synapses. Science 294, 1030–1038.
, 百拇医药
Katz AM (1992). T wave ‘Memory’: Possible causal relationship to stress induced changes in cardiac ion channels J Cardiovasc Electrophysiol 3, 150–159.
McVeigh ER, Prinzen FW, Wyman BT, Tsitlik JE, Halperin HR & Hunter WC (1998). Imaging asynchronous mechanical activation of the paced heart with tagged MRI. Magn Reson Med 39, 507–513.
Meghji P, Nazir SA, Dick DJ, Bailey ME, Johnson KJ & Lab MJ (1997). Regional workload induced changes in electrophysiology and immediate early gene expression in intact in situ porcine heart. J Mol Cell Cardiol 29, 3147–3155.
, 百拇医药
Mills VA, Orbach-Arbouys S & Gershon RK (1975). Relation of DNA synthesis and suppression to development of memory by T cells. J Immunol 114, 200–205.
Nahlawi M, Waligora M, Spies SM, Bonow RO, Kadish AH & Goldberger JJ (2004). Left ventricular function during and after right ventricular pacing. J Am Coll Cardiol 44, 1883–1888.
Obreztchikova M, Plotnikov AN, Shlapakova IN, Danilo P Jr, Robinson RB & Rosen MR (2003). Cardiac memory inverts the gradient of IKr and induces a gradient for IKs in canine left ventricular myocytes. Circulation 108, IV–9.
, http://www.100md.com
Patberg KW, Obreztchikova MN, Giardina SF, Symes AJ, Plotnikov AN, Qu J, Chandra P, McKinnon D, Liou SR, Rybin AV, Shlapakova I, Danilo P Jr, Yang J & Rosen MR (2005). The cAMP response element binding protein modulates expression of the transient outward current: Implications for cardiac memory. Cardiovasc Res 68, 259–267.
Patberg KW, Plotnikov AN, Quamina A, Gainullin RZ, Rybin A, Danilo P Jr, Sun LS & Rosen MR (2003). Cardiac memory is associated with decreased levels of the transcriptional factor, CREB, modulated by angiotensin II and calcium. Circ Res 93, 474–478.
, http://www.100md.com
Patberg KW & Rosen MR (2004). Molecular determinants of cardiac memory and their regulation. J Mol Cell Cardiol 36, 195–204.
Patel PM, Plotnikov A, Kanagaratnam P, Shvilkin A, Sheehan CT, Xiong W, Danilo P Jr, Rosen MR & Peters NS (2001). Altering ventricular activation remodels gap junction distribution in canine heart. J Cardiovasc Electrophysiol 12, 570–577.
Plotnikov AN, Shvilkin A, Xiong W, de Groot JR, Rosenstraukh L, Feinmark S, Gainullin R, Danilo P & Rosen MR (2001). Interaction between antiarrhythmic drugs and cardiac memory. Cardiovasc Res 50, 335–344.
, 百拇医药
Plotnikov AN, Sosunov EA, Patberg KW, Anyukhovsky EP, Gainullin RZ, Shlapakova IN, Krishnamurthy G, Danilo P Jr & Rosen MR (2004). Cardiac memory evolves with age in association with development of the transient outward current. Circulation 110, 489–495.
Plotnikov AN, Yu H, Geller C, Gainullin RZ, Chandra P, Patberg KW, Friezema S, Danilo P Jr, Cohen IS, Feinmark SJ & Rosen MR (2003). Role of L-type calcium channels in pacing-induced short-term and long-term cardiac memory in canine heart. Circulation 107, 2844–2849.
, http://www.100md.com
Prinzen FW, Hunter WC, Wyman BT & McVeigh ER (1999). Mapping of regional myocardial strain and work during ventricular pacing: experimental study using magnetic resonance imaging tagging. J Am Coll Cardiol 33, 1735–1742.
Ricard P, Danilo P Jr, Cohen IS, Burkhoff D & Rosen MR (1999). A role for the renin-angiotensin system in the evolution of cardiac memory. J Cardiovasc Electrophysiol 10, 545–551.
Rosen MR, Cohen IS, Danilo PD & Steinberg SF (1998). The heart remembers. Cardiovasc Res 40, 469–482.
, http://www.100md.com
Rosenbaum MB, Blanco HH, Elizari MV, Lazzari JO & Davidenko JM (1982). Electrotronic modulation of the T wave and cardiac memory. Am J Cardiol 50, 213–222.
Sadoshima J & Izumo S (1993a). Mechanical stretch rapidly activates multiple signal transduction pathways in cardiac myocytes: potential involvement of an autocrine/paracrine mechanism. EMBO J 12, 1681–1692.
Sadoshima J & Izumo S (1993b). Molecular characterization of angiotensin II-induced hypertrophy of cardiac myocytes and hyperplasia of cardiac fibroblasts. Critical role of the AT1 receptor subtype. Circ Res 73, 413–423.
, 百拇医药
Shvilkin A, Danilo P Jr, Wang J, Burkoff D, Anyukhovsky EP, Sosunov EA, Hara M & Rosen MR (1998). Evolution and resolution of long-term cardiac memory. Circulation 97, 1810–1817.
Shvilkin A, Ho KK, Rosen MR & Josephson ME (2004). T-vector direction differentiates postpacing from ischemic T-waves in precordial leads. Circulation 111, 969–974.
Spragg DD, Leclercq C, Loghmani M, Faris OP, Tunin RS, DiSilvestre D, McVeigh ER, Tomaselli GF & Kass DA (2003). Regional alterations in protein expression in the dyssynchronous failing heart. Circulation 108, 929–932.
, 百拇医药
Volders PG, Sipido KR, Carmeliet E, Spatjens RL, Wellens HJ & Vos MA (1999). Repolarizing K+ currents ITO1 and IKs are larger in right than left canine ventricular midmyocardium. Circulation 99, 206–210.
Wecke L, Gadler F, Linde C, Lundahl G, Rosen MR & Bergfeldt L (2005). Temporal characteristics of cardiac memory in humans: Vectorcardiographic quantification in a model of cardiac pacing. Heart Rhythm 2, 28–34.
Wit AL & Rosen MR (1989). Cellular electrophysiological mechanisms of cardiac arrhythmias. In Comprehensive Electrocardiology, Theory and Practice in Health and Disease, vol. 2, section 6, ed. MacFarlane PW & Lawrie TDV, pp. 801–841. Pergamon Press, Elmsford, NY, USA.
, 百拇医药
Yu H, Gao J, Wang H, Wymore R, Steinberg S, McKinnon D, Rosen MR & Cohen IS (2000). Effects of the renin-angiotensin system on the current Ito in epicardial and endocardial ventricular myocytes from the canine heart. Circ Res 86, 1062–1068.
Yu H, McKinnon D, Dixon JE, Gao J, Wymore R, Cohen IS, Danilo P Jr, Shvilkin A, Anyukhovsky EP, Sosunov EA, Hara M & Rosen MR (1999). Transient outward current, Ito1, is altered in cardiac memory. Circulation 99, 1898–1905., 百拇医药(Michael R. Rosen and Ira S. Cohen)
2 Department of Physiology and Biophysics, Institute of Molecular Cardiology, Stony Brook University, Stony Brook, NY, USA
Abstract
‘Cardiac memory’ describes an electrocardiographic T wave vector change, recorded during normal sinus rhythm that reflects the QRS complex vector during prior periods of ventricular pacing or arrhythmia. In this brief review we consider the mechanisms responsible for cardiac memory, which offer a unique window for relating molecular determinants of repolarization to their expression in the function of ion channels and in the electrophysiology of the heart. Understanding the steps that translate the molecular mechanisms for memory into clinical expression in this relatively straightforward model facilitates our comprehension of the complex pathways that order normal cardiac repolarization and repolarization changes.
, 百拇医药
Introduction to cardiac memory
‘Cardiac memory’ is a term used clinically to describe a specific form of remodelling seen as an altered electrocardiographic T wave (Rosenbaum et al. 1982). Following the return to sinus rhythm after an interval of ventricular pacing or arrhythmia, the T wave vector persists in tracking the vector angle and amplitude of the QRS complex that characterized the paced or arrhythmic state (Fig. 1). The nomenclature is generally accepted, and the phenomenon is thought to have no inherent pathological implications (Rosenbaum et al. 1982; Rosen et al. 1998).
, 百拇医药
Upper panels show canine ECG during control sinus rhythm, during ventricular pacing at about 5% faster than sinus rate, and – a few minutes after returning to sinus rhythm – on days 7, 14 and 21 of pacing. Note that the QRS complex is inverted during ventricular pacing and that the T wave in the subsequent panels in sinus rhythm becomes progressively inverted, following the direction of the paced QRS complex. The two left bottom panels show the vectorcardiogram of the same dog in control (sinus rhythm) and during ventricular pacing. Note the change in vector shape, angle and amplitude of the QRS complex. The right lower panel shows an enlargement of the T wave vector during control sinus rhythm and in sinus rhythm on days 14 and 21. Note that the T wave vector has moved in the direction of the paced QRS, and also shows an increased amplitude. Modified from Shvilkin et al. (1998).
, 百拇医药
Descriptions of such T wave changes induced by arrhythmia or ventricular pacing have been published since the early days of electrocardiography. The first large series of pacing-induced changes was provided by Chatterjee et al. (1969) and the phenomenon was termed ‘cardiac memory’ by Rosenbaum et al. (1982). The latter investigators noted that the magnitude of T wave change increased with repetitions of the pacing periods or arrhythmias – referred to as ‘accumulation’– and that the T wave changes persisted for long intervals (months or longer) after a return to sinus rhythm. According to Rosenbaum et al. (1982), the sinus T wave ‘remembered’ the paced QRS complex. As yet, only one study has examined the persistence of pacing-induced cardiac memory prospectively (Wecke et al. 2005). Here one week of right ventricular apical endocardial pacing induced memory that persisted for about a month after cessation of pacing. It is likely, however, that pacing from different sites will result in memory of varying magnitudes and durations.
, 百拇医药
Although some experimental studies of cardiac memory were published before 1990 (Costard-Jackle et al. 1989; and summarized in Katz, 1992; Rosen et al. 1998), the lion's share of experimentation has occurred since. The likely explanation for increased interest in the last 15 years is the heightened appreciation of remodelling in general as a physiological and a pathological phenomenon as well as recent insights into the clinical impact of cardiac memory. With regard to the clinic, the preponderance of examples of cardiac memory derive from cardiac pacing which, as long as it alters ventricular activation, consistently induces memory. Other forms of abnormal activation that induce memory include intermittent left bundle branch block, ventricular arrhythmias, and atrioventricular bypass tracts (see Katz, 1992, for detailed review). In addition, the T wave changes of memory can mimic those of ischaemic heart disease (Rosenbaum et al. 1982; Katz, 1992; Shvilkin et al. 2004) and can impact on and either mask or augment the effects on ECG of anti-arrhythmic drugs and of other drugs that alter K+ channel function (Plotnikov et al. 2001). More controversially, and therefore still speculative, cardiac memory may modulate the type of cardiac rhythm expressed, may be arrhythmogenic (although this holds largely for atrium, see Chandra et al. 2003, 2005) and may be an early portent of hypertrophy to come (Rosen et al. 1998).
, http://www.100md.com
Relationship to memory in other systems
Whether wisely or not, when we began to study cardiac memory in the late 1980s, we accepted the term literally and looked to work in other organ systems for clues regarding mechanism. At that time by far the greatest effort had been reported in CNS (summarized in Kandel, 2001), although even then (and certainly since) there was evidence for memory in the immune system (Mills et al. 1975) and the gastrointestinal tract (Gershon, 1998). The key to our understanding of memory and to the design of our early studies came from reports on long-term potentiation in the sea slug Aplysia. In brief, the application of electrical shocks to Aplysia results in retraction of the siphon for intervals directly proportional in duration to the number of shocks delivered: the more stimuli, the longer the retraction (and the longer the memory) (summarized in Kandel, 1989). However crude, the association between this observation and the fact that repetitive shocking (i.e. pacing) of the ventricle led to occurrence and persistence of a specific T wave change was obvious. Given this common behaviour, we decided to use pathways then in vogue for memory in CNS as templates for our work in heart. As such, we tested hypotheses that suggested memory induced by brief pacing periods and of short duration (referred to as ‘short-term memory’) might result from altered signal transduction affecting the behaviour of specific ion channels, while memory induced by long pacing periods and persisting for weeks or months (‘long-term memory’) might result from altered gene transcription. Hence we worked in a system incorporating (1) a common stimulus (pacing), and (2) multiple potential contributory factors including some that could translate a briefly occurring event into one of long duration. Important to our understanding of the system's operation was the observation that pacing rate per se was not the determinant of cardiac memory (although memory evolved faster at more rapid rates); far more important was the altered ventricular activation pathway imposed by the pacing site (Shvilkin et al. 1998).
, http://www.100md.com
Often in the setting of altered activation there is altered myocardial stretch, but there is no certainty regarding whether changing the activation path in its own right can induce memory or if stretch is the key component. It is clear that changing stretch can increase angiotensin II synthesis/release by cardiac cells (Sadoshima & Izumo, 1993a,b) and can turn on the cardiac immediate early gene programme (Meghji et al. 1997). Learning more about mechanism is essential here as this information impacts directly on the question of why cardiac pacing induces repolarization changes.
, http://www.100md.com
Methods for study
A variety of models running the ‘cell-to-bedside’ gamut have been used to study cardiac memory. Studies through the mid-1980s employed patient populations (Chatterjee et al. 1969; Rosenbaum et al. 1982) and animal models (Costard-Jackle et al. 1989). As the need to study mechanism became more cellularly and subcellularly orientated, canine models were paced from the ventricles (mimicking the use of pacemakers in humans) for hours to weeks to track the characteristics of memory (Shvilkin et al. 1998; Yu et al. 1999). Terminal experiments at various times permitted harvesting of cells for isolated tissue electrophysiology and ion channel and molecular experiments.
, 百拇医药
In all these studies a key intention was to use pacing protocols that altered ventricular activation pathways as this seemed to be essential to inducing cardiac memory. Moreover, although not directed uniquely at understanding memory, experiments using pacing and magnetic resonance imaging provided insights into the relationship between pacing site and stress/strain relationships on the myocardium (McVeigh et al. 1998; Prinzen et al. 1999).
The need to understand how specific signalling molecules affected ion channel properties and ion channel trafficking brought in the use of disaggregated cells in culture media (Yu et al. 2000) as well as mammalian cell lines (Doronin et al. 2004). Moreover, as information regarding mechanism and expression of memory in animals has accumulated, there has been renewed interest in patient studies (Alessandrini et al. 1997; Goyal et al. 1998; Haverkamp et al. 1998; Nahlawi et al. 2004; Shvilkin et al. 2004; Wecke et al. 2005).
, 百拇医药
What causes the T wave changes of cardiac memory
The T wave on ECG is the summed signal of action potential repolarization in the cells of the ventricle. More specifically, the gradient of repolarization within the ventricles gives rise to the electrocardiographic ST segment and T wave (reviewed in Wit & Rosen, 1989). The sources of gradients in the ventricle include: that between apex and base, such that repolarization time (incorporating activation time through the end of repolarization) is shorter basally than apically; that between left and right ventricle (Volders et al. 1999); and that which is transmural – at any site epicardial action potentials are shorter than endocardial. Regarding the transmural gradient, there is controversy regarding the presence of longer action potentials midmyocardially than either epi- or endocardially in the ventricles in situ. When cells are isolated such transmural differences are clearly seen; however, at physiological heart rates this transmural gradient is minimized in situ, probably due to the electrotonic effects of cell–cell coupling (Anyukhovsky et al. 1999).
, http://www.100md.com
In experiments using pacing to induce short-term cardiac memory, no transmural gradient in repolarization was seen before or after 2 h of pacing. However, the pacing protocol resulted in shortening of apico-basal repolarization times throughout the ventricles, with epicardial monophasic action potential duration decreasing by about 10 ms at the 90% repolarization level (Janse et al. 2005). Whether this change in apico-basal gradients is the sole determinant of T wave changes during short-term memory or other gradient changes occur as well to explain the T waves of short- and long-term memory is yet to be determined. For example, the possibility of changes in the right–left ventricular gradient has not been studied in memory.
, 百拇医药
Mechanisms determining short-term memory
Studies of isolated rabbit ventricle (Costard-Jackle et al. 1989) and of canine heart (del Balzo & Rosen, 1992) paced for minutes to hours have shown memory persisting for minutes to hours in proportion to the duration of stimulation. Magnetic resonance imaging of canine heart indicates that altering the pacing site alters the timing of contraction and relaxation in various regions of the heart (McVeigh et al. 1998; Prinzen et al. 1999). This knowledge together with the observation that altering stretch on cardiac cell cultures increases their synthesis and release of angiotensin II (Sadoshima & Izumo, 1993a,b) led to the hypothesis that angiotensin II synthesis and release in the heart as a result of altered stretch imposed by ventricular pacing would initiate cardiac memory via functional changes in ion channels (Rosen et al. 1998). We then found that angiotensin-converting enzyme inhibition, AT-1 receptor blockade and tissue chymase inhibition all prevented the development of cardiac memory over 1–2 h periods in canine heart, leading us to consider AT-1 receptor-related pathways in the genesis of memory (Ricard et al. 1999). L-type Ca2+ channel blockers also prevented memory development in this canine model, while -adrenergic and muscarinic blockers had no effect (Plotnikov et al. 2003).
, 百拇医药
How might angiotensin II initiate expression of cardiac memory We had observed that 4-aminopyridine, a blocker of the transient outward potassium current Ito, prevents the onset of cardiac memory during short periods of pacing canine heart (del Balzo et al. 1992). As shall be detailed below, Ito density in canine left ventricular epicardial myocytes is reduced by chronic pacing to induce long-term cardiac memory (Yu et al. 1999), and exposure of non-paced canine epicardial myocytes to angiotensin II comparably reduces Ito (Yu et al. 2000) (Fig. 2). In experiments exploring potential associations between angiotensin II and Ito, (Doronin et al. 2004) the -subunit of the channel, Kv4.3, colocalized with the AT-1 receptor; the two coimmunoprecipitated in canine ventricular myocytes and, in HEK 293 cells, the complex coimmunoprecipitated with the accessory subunit KChIP2. Importantly, exposure of the AT-1 receptor–Kv4.3 complex to angiotensin II resulted in internalization of the complex (Fig. 3). Moreover, the gating properties of those Kv4.3 channels remaining at the cell surface shifted to more positive voltages. This led to the proposal that the AT-1 receptor provides an internalization molecular scaffold via which angiotensin II initiates the trafficking of Kv4.3 (and perhaps KChIP2) to internal sites where their degradation might occur (Doronin et al. 2004) (see diagram in Fig. 4). The net result of the trafficking would be a loss of Ito density. The model system also showed a concomitant positive shift in activation for the remaining reduced transient outward current. This reduced current density and positive shift in activation are precisely what is seen in myocytes from canine hearts as cardiac memory evolves (Yu et al. 1999).
, http://www.100md.com
A and B, representative records of a control epicardial myocyte (Epi) and an angiotensin II-exposed myocyte (Epi + A-II), respectively: note the marked diminution of current in the latter. C shows a summary of current–voltage data for the entire series. There is a significant reduction in current throughout the voltage range studied. D and E show action potentials from comparable cells. In D the phase 1 notch attributable to Ito is indicated by an arrow. The notch disappears on exposure to angiotensin II (E). Modified from Yu et al. (2000).
, http://www.100md.com
HEK 293 cells were transiently transfected with Kv4.3–V5–KChIP2-myc and the HA-AT1 receptor. The distribution of AT1 receptors and Kv4.3 was visualized after cell fixation with 3.7% formaldehyde using anti-HA-TRITC and anti-V5-FITC antibody before and after treatment with 1 μM angiotensin for 1 h. The distribution of Kv4.3 and AT1 receptors is shown before (A) and after treatment with angiotensin II (B). Nuclei (blue colour) were counterstained with 4,6-diamidino-2-phenylindole (DAPI). Note that the AT1 receptor is located predominantly on the cell surface where it colocalizes with Kv4.3 in the absence of angiotensin II (A). Angiotensin II induces internalization of the AT1 receptor (B), and with this Kv4.3 is also removed from the cell surface. The majority of the internalized AT1 receptors colocalize with Kv4.3 in intracellular vesicles. Reproduced with permission from Doronin et al. (2004).
, 百拇医药
A, coassembled AT-1 receptor–Kv4.3–KChIP2 macromolecular complexes inserted in the cell membrane. Channel opening results in outward K+ current. In B, angiotensin II binds to a subset of receptors, resulting in the internalization of the macromolecular complex. The net result is a loss of functioning channels in the cell membrane and a reduction in current (C).
The role for angiotensin II in short-term memory is reinforced by experiments in which epicardial or endocardial myocytes from control dogs were maintained for 2 h to 2 days in culture media, respectively, with angiotensin II or an AT-1 receptor blocker (Yu et al. 2000). In epicardium, angiotensin II prolonged repolarization and reduced the action potential notch and Ito, while having no effect on Kv4.3 mRNA. These actions were blocked by the addition of an AT-1 receptor blocker to the culture medium. In contrast, in endocardium, angiotensin II alone had no effect, but AT-1 receptor blockade accelerated repolarization, induced an action potential notch and increased Ito. This phenomenon is consistent with a transmural ventricular gradient for angiotensin II levels, as well as with a non-transcriptional pathway whereby Ito function is modulated to alter the transmural repolarization gradient. Whether the mechanism here is the same trafficking mechanism as described for short-term memory is unknown.
, 百拇医药
Finally, it would be tempting to try to relate the actions of angiotensin II to changes in ventricular function that are induced by pacing clinically, but we believe such speculation to be premature. One should keep in mind, however, that pacing specific ventricular sites results in differing regional patterns of synchrony/dyssynchrony (McVeigh et al. 1998; Prinzen et al. 1999) and that angiotensin II might be a contributor here.
Transitions between short- and long-term memory
, 百拇医药
Studies of the transitional period have largely focused on time points 30 min to 2 h after onset of pacing. During this interval not only are the trafficking changes that putatively lead to short-term memory occurring, but transcriptional changes are beginning which initiate the cascade of events leading to long-term memory. The role of altered stretch in inducing transcriptional change was noted by Meghji et al. (1997) in porcine hearts subjected to altered regional ventricular stretch for 30 min which manifested increased expression of the immediate early gene programme.
, 百拇医药
Emphasis has been placed on the role of the cyclic AMP response element binding protein (CREB), central to neural memory transcription in the evolution of cardiac memory. Two hours of ventricular pacing down-regulates nuclear CREB (blocked by AT-1 receptor or Ca2+ channel blockade) (Patberg et al. 2003) (Fig. 5). Change is maximal near the pacing site and diminishes with distance. Because the genes for Kv4.3 and KChIP2 (- and accessory subunits for canine Ito, respectively) manifest CREB binding capability in their promoter regions, we suggested that CREB might regulate one or both genes by directly binding in these regions (Patberg et al. 2003; Patberg & Rosen, 2004). That CREB can modulate Ito expression was shown in experiments in which a CREB antisense virus packaged with green fluorescent protein was injected into the canine left ventricular epicardium (Patberg et al. 2005). Terminal studies several days later showed the absence of a monophasic action potential notch in the region of epicardial injection as well as the absence of Ito in the antisense-infected cells (but not in nearby control cells). Hence an argument can be made that a CREB-transcribed system regulates at least one of the ion channel changes associated with cardiac memory. However, roles for additional transcriptional factors in memory induction are highly likely.
, 百拇医药
In these experiments, anaesthetized dogs were subjected to atrioventricular pacing with a short PR interval to ensure 100% capture of the ventricles during the 2 h period of pacing. Individual panels show results from an atrioventricular paced dog (AVP), a sham control (instrumented but not paced), an AV paced dog that received the L-type Ca2+ channel blocker nifedipine (AVP + nif), and an AV paced dog that received the AT-I and AT-II receptor blocker saralasin (AVP + sar). Ref indicates reference biopsy; 2 h, biopsy taken after 2 h of AVP. Loading conditions were controlled using an antibody against histone 1 (hist 1). Note the marked reduction in CREB levels at 2 h in the AV paced dog. There was no change in CREB in the sham animal or in those treated with nifedipine or saralasin. Reproduced with permission from Patberg et al. (2003).
, 百拇医药
Long-term memory
Long-term memory has been the most completely studied state, as the stability of the memory process after cessation of pacing ensures steady-state function with regard to ion channels and their molecular determinants (Shvilkin et al. 1998). In general such experiments are performed terminally 24 h after pacing ceases. Transmembrane action potentials at this time show a prolongation of duration in epicardium and endocardium, with a loss of the epicardial action potential notch in the former: this would be consistent with an altered transmural repolarization gradient (although this has not been studied in situ) (Shvilkin et al. 1998; Yu et al. 1999). Long-term memory accumulation in situ is attenuated by protein synthesis inhibitors and by L-type Ca2+ channel blockers, but not by ACE inhibitors or AT-1 receptor blockade (Shvilkin et al. 1998; Plotnikov et al. 2003). The mechanism whereby the angiotensin II effect is altered in the long-term setting is not yet known: whether it reflects something as simple as up-regulation of AT-1 receptors awaits determination. Studies of CREB in the long-term setting demonstrate its continued down-regulation (Patberg et al. 2003).
, 百拇医药
In long-term memory, Ito density is decreased as described above, and Kv4.3 mRNA and KChIP2 message and protein are decreased as well (Yu et al. 1999; Patberg et al. 2003) (Fig. 6). Note that whereas the change in current density is the same here as that seen with 2 h of exposure of myocytes to angiotensin II, there was no change in Kv4.3 mRNA in the latter instance (Yu et al. 2000) (consistent with the hypothesis that altered trafficking explains the reduction in current). That there is a reduction in Kv4.3 mRNA in long-term memory is consistent with altered gene transcription in this setting. In addition, epicardial Ito activates at more positive potentials and recovers from inactivation more slowly than in control myocytes (Yu et al. 1999). Other channels are involved as well: the L-type Ca2+ current (ICa,L) shows no change in density but activation is more positive and time constants of inactivation are longer in memory than control hearts (Plotnikov et al. 2003). Preliminary studies of the rapidly activating delayed rectifier currents (IKr) suggest that the transmural gradient in control dogs (a larger current epicardially than endocardially) is reduced in memory (Obreztchikova et al. 2003). All these ion channel changes are consistent with and likely contributors to the absent phase 1 notch, the high plateau and the prolonged action potential duration of cardiac memory. Still other experiments have shown that the gap junctional protein connexin43 is lateralized in distribution and reduced in density in long-term memory, with the greatest changes in the region of the pacing electrode (Patel et al. 2001). Whether these changes are associated with altered gap junctional function and contribute to the occurrence of long-term memory or are an epiphenomenon remains to be determined.
, http://www.100md.com
Upper panel: epicardial action potentials recorded from slabs of left ventricular tissue removed from a dog paced for 3 weeks to induce cardiac memory, and from a sham control. Note that in the setting of memory, the phase 1 notch diminishes, the plateau is increased in height and action potential duration is prolonged. Lower left panel shows Kv4.3 mRNA for a control dog and one in cardiac memory. Note that message is reduced in the memory animal (Cyc = cyclophyllin). Lower right panel shows Ito conductance for a group of controls and a group of memory animals. There is a significant decrease in channel conductance in the setting of memory. Modified from Yu et al. (1999).
, http://www.100md.com
Is Ito the entire story
Categorically not: while Ito, because of its large magnitude and ease of study provides a useful window into mechanism, everything we have learned regarding memory argues against too simplistic an interpretation. That Ito is important is best illustrated by the observation that short-term memory cannot be induced in animals having no Ito. Specifically, in neonatal dogs having no epicardial Ito, no phase 1 action potential notch and no transmural repolarization gradient, memory is not inducible; as age increases and current, notch and gradient appear, the ability to use pacing to induce memory increases proportionally (Plotnikov et al. 2004). The importance of other currents is highlighted by the findings that in long-term memory not only the epicardial, but the endocardial action potential (at which site there is little or no Ito) manifests an elevated plateau and prolonged action potential duration. This information, together with the important changes described in ICa,L kinetics (Plotnikov et al. 2003) and IKr density and kinetics (Obreztchikova et al. 2003) argues in favour of multichannel involvement in the memory process.
, 百拇医药
Issues still to be considered
Cardiac memory is a readily quantifiable phenomenon that permits us to relate repolarization changes in situ to their molecular, ion channel and cellular determinants. Hence its applicability to general knowledge regarding the regulation of repolarization is great. Yet the relationship between cardiac and other forms of memory is complex as is the interface between memory and other forms of cardiac remodelling.
, 百拇医药
For example, neuronal memory is a synthetic process: transcriptional factors such as CREB are up-regulated and synaptic connections between cells are strengthened as memory develops (Kandel, 2001). In contrast, cardiac memory sees CREB levels down-regulated and the function of the ion channels thus far identified as involved down-regulated as well (Yu et al. 1999; Plotnikov et al. 2003; Patberg et al. 2003, 2004). Far from being a non-specific degradative process, cardiac memory creates action potential and cardiac repolarization phenotypes much like those in neonatal heart. In other words, memory, which is a specialized stress response in both CNS and heart, results in an advancement in structure/function in the former and a return to an earlier structure/function entity in the latter.
, 百拇医药
Is memory in heart, then, the ‘forgetting’ of an adult phenotype or the reacquisition/strengthening of a neonatal phenotype Far from a mere semantic quandary, this question impacts on our understanding of cardiac remodelling. This stress response in remodelling sees earlier developmental patterns of repolarization emerging as well. For example, in canine cardiac failure models, the action potential notch diminishes, repolarization is prolonged, and Ito density decreases (Spragg et al. 2003). But – and this is important – the changes in Ito activation and inactivation gating that characterize cardiac memory are not seen in cardiac failure. In other words, the remodelling processes while in some ways appearing similar are in other ways different.
, 百拇医药
Although this may seem a small point, we believe it to be key: despite the obvious attractions of extremes in generalization or reductionism, it is the understanding of common ground and of subtle differences that may best move knowledge ahead.
Conclusions
While much has been done to study cardiac memory, many steps central to its evolution remain to be determined (see Fig. 7 which illustrates the pathways studied). Ito has been the landmark current considered to date, and the thread of information suggesting the internalization of its molecular determinants by angiotensin II in a short-term model through its transcriptional regulation by CREB in a long-term model is a tantalizing story that requires further investigation. Beyond this, the likelihood that memory represents a monogenically determined function as opposed to a complex array of changes is small, requiring a great deal of effort to identify the totality of pathways involved.
, 百拇医药
Pacing alters activation and stretch, resulting in angiotensin II synthesis/release and the trafficking and internalization of the AT-1 receptor–Kv4.3–KChIP2 complex from its membrane site and a reduction in current. Angiotensin II also induces an increase in L-type Ca2+ current. Other potential sources for increased Cai2+ would be the Na+–Ca2+ exchanger and stretch-activated channels. It appears that Cai2+ may be a second messenger activating changes in transcriptional factors in the nucleus. The transcriptional factor thus far studied, CREB, is reduced. An association with KChIP2 reduction has been demonstrated here. The other factors and linkages that may be involved have not been identified. However, long-term changes in Ito, IKr and ICa,L have been demonstrated as well, all of which would be expected to contribute to the altered action potentials and ECG changes of cardiac memory.
, 百拇医药
References
Alessandrini RS, McPherson DD, Kadish AH, Kane BJ & Goldberger JJ (1997). Cardiac memory: a mechanical and electrical phenomenon. Am J Physiol 272, H1952–H1959.
Anyukhovsky EP, Sosunov EA, Gainullin RZ & Rosen MR (1999). The controversial M cell. J Cardiovasc Electrophysiol 10, 244–260.
Chandra P, Rosen TS, Herweg B, Danilo P & Rosen MR (2003). Left atrial pacing induces memory and is associated with atrial tachyarrhythmias. Cardiovasc Res 60, 307–314.
, 百拇医药
Chandra P, Rosen TS, Herweg B, Plotnikov AN, Danilo P & Rosen MR (2005). Atrial gradient as a potential predictor of atrial fibrillation. Heart Rhythm 2, 404–410.
Chatterjee K, Harris AM, Davies JG & Leatham A (1969). T-wave changes after artificial pacing. Lancet 1, 759–760.
Costard-Jackle A, Goetsch B, Antz M & Franz MR (1989). Slow and long lasting modulation of myocardial repolarization produced by ectopic activation in isolated rabbit hearts. Evidence of cardiac ‘memory’. Circulation 80, 1412–1420.
, http://www.100md.com
del Balzo U & Rosen MR (1992). 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 85, 1464–1472.
Doronin SV, Potapova IA, Lu ZJ & Cohen IS (2004). Angiotensin receptor Type 1 forms a complex with the transient outward potassium channel Kv4.3 and regulates Its gating properties and intracellular localization. J Biol Chem 279, 48231–48237.
, 百拇医药
Gershon MD (1998). Genes, lineages, and tissue interactions in the development of the enteric nervous system. Am J Physiol 275, G869–G873.
Goyal R, Syed ZA, Mukhopadhyay PS, Souza J, Zivin A, Knoght BP, Man KC, Strickberger SA & Morady F (1998). Changes in cardiac repolarization following short periods of ventricular pacing. J Cardiovasc Electrophysiol 9, 269–280.
Haverkamp W, Hordt M, Breithardt G & Borggrefe M (1998). Torsades de pointes secondary to d,1-sotalol after catheter ablation of incessant atrioventricular reentrant tachycardia – evidence for a significant contribution of the ‘cardiac memory’. Clin Cardiol 21, 55–58.
, http://www.100md.com
Janse MJ, Sosunov EA, Coronel R, Opthof T, Anyukhovsky EP, deBakker JMT, Plotnikov AN, Shlapakova IN, Danilo P, Tijssen JGP & Rosen MR (2005). Repolarization gradient in the canine ventricle before and after induction of short term cardiac memory. Circulation 112, 1711–1718.
Kandel ER (1989). Genes, nerve cells and the remembrance of things past. J Neuropsychol 1, 103–110.
Kandel ER (2001). The molecular biology of memory storage: a dialogue between genes and synapses. Science 294, 1030–1038.
, 百拇医药
Katz AM (1992). T wave ‘Memory’: Possible causal relationship to stress induced changes in cardiac ion channels J Cardiovasc Electrophysiol 3, 150–159.
McVeigh ER, Prinzen FW, Wyman BT, Tsitlik JE, Halperin HR & Hunter WC (1998). Imaging asynchronous mechanical activation of the paced heart with tagged MRI. Magn Reson Med 39, 507–513.
Meghji P, Nazir SA, Dick DJ, Bailey ME, Johnson KJ & Lab MJ (1997). Regional workload induced changes in electrophysiology and immediate early gene expression in intact in situ porcine heart. J Mol Cell Cardiol 29, 3147–3155.
, 百拇医药
Mills VA, Orbach-Arbouys S & Gershon RK (1975). Relation of DNA synthesis and suppression to development of memory by T cells. J Immunol 114, 200–205.
Nahlawi M, Waligora M, Spies SM, Bonow RO, Kadish AH & Goldberger JJ (2004). Left ventricular function during and after right ventricular pacing. J Am Coll Cardiol 44, 1883–1888.
Obreztchikova M, Plotnikov AN, Shlapakova IN, Danilo P Jr, Robinson RB & Rosen MR (2003). Cardiac memory inverts the gradient of IKr and induces a gradient for IKs in canine left ventricular myocytes. Circulation 108, IV–9.
, http://www.100md.com
Patberg KW, Obreztchikova MN, Giardina SF, Symes AJ, Plotnikov AN, Qu J, Chandra P, McKinnon D, Liou SR, Rybin AV, Shlapakova I, Danilo P Jr, Yang J & Rosen MR (2005). The cAMP response element binding protein modulates expression of the transient outward current: Implications for cardiac memory. Cardiovasc Res 68, 259–267.
Patberg KW, Plotnikov AN, Quamina A, Gainullin RZ, Rybin A, Danilo P Jr, Sun LS & Rosen MR (2003). Cardiac memory is associated with decreased levels of the transcriptional factor, CREB, modulated by angiotensin II and calcium. Circ Res 93, 474–478.
, http://www.100md.com
Patberg KW & Rosen MR (2004). Molecular determinants of cardiac memory and their regulation. J Mol Cell Cardiol 36, 195–204.
Patel PM, Plotnikov A, Kanagaratnam P, Shvilkin A, Sheehan CT, Xiong W, Danilo P Jr, Rosen MR & Peters NS (2001). Altering ventricular activation remodels gap junction distribution in canine heart. J Cardiovasc Electrophysiol 12, 570–577.
Plotnikov AN, Shvilkin A, Xiong W, de Groot JR, Rosenstraukh L, Feinmark S, Gainullin R, Danilo P & Rosen MR (2001). Interaction between antiarrhythmic drugs and cardiac memory. Cardiovasc Res 50, 335–344.
, 百拇医药
Plotnikov AN, Sosunov EA, Patberg KW, Anyukhovsky EP, Gainullin RZ, Shlapakova IN, Krishnamurthy G, Danilo P Jr & Rosen MR (2004). Cardiac memory evolves with age in association with development of the transient outward current. Circulation 110, 489–495.
Plotnikov AN, Yu H, Geller C, Gainullin RZ, Chandra P, Patberg KW, Friezema S, Danilo P Jr, Cohen IS, Feinmark SJ & Rosen MR (2003). Role of L-type calcium channels in pacing-induced short-term and long-term cardiac memory in canine heart. Circulation 107, 2844–2849.
, http://www.100md.com
Prinzen FW, Hunter WC, Wyman BT & McVeigh ER (1999). Mapping of regional myocardial strain and work during ventricular pacing: experimental study using magnetic resonance imaging tagging. J Am Coll Cardiol 33, 1735–1742.
Ricard P, Danilo P Jr, Cohen IS, Burkhoff D & Rosen MR (1999). A role for the renin-angiotensin system in the evolution of cardiac memory. J Cardiovasc Electrophysiol 10, 545–551.
Rosen MR, Cohen IS, Danilo PD & Steinberg SF (1998). The heart remembers. Cardiovasc Res 40, 469–482.
, http://www.100md.com
Rosenbaum MB, Blanco HH, Elizari MV, Lazzari JO & Davidenko JM (1982). Electrotronic modulation of the T wave and cardiac memory. Am J Cardiol 50, 213–222.
Sadoshima J & Izumo S (1993a). Mechanical stretch rapidly activates multiple signal transduction pathways in cardiac myocytes: potential involvement of an autocrine/paracrine mechanism. EMBO J 12, 1681–1692.
Sadoshima J & Izumo S (1993b). Molecular characterization of angiotensin II-induced hypertrophy of cardiac myocytes and hyperplasia of cardiac fibroblasts. Critical role of the AT1 receptor subtype. Circ Res 73, 413–423.
, 百拇医药
Shvilkin A, Danilo P Jr, Wang J, Burkoff D, Anyukhovsky EP, Sosunov EA, Hara M & Rosen MR (1998). Evolution and resolution of long-term cardiac memory. Circulation 97, 1810–1817.
Shvilkin A, Ho KK, Rosen MR & Josephson ME (2004). T-vector direction differentiates postpacing from ischemic T-waves in precordial leads. Circulation 111, 969–974.
Spragg DD, Leclercq C, Loghmani M, Faris OP, Tunin RS, DiSilvestre D, McVeigh ER, Tomaselli GF & Kass DA (2003). Regional alterations in protein expression in the dyssynchronous failing heart. Circulation 108, 929–932.
, 百拇医药
Volders PG, Sipido KR, Carmeliet E, Spatjens RL, Wellens HJ & Vos MA (1999). Repolarizing K+ currents ITO1 and IKs are larger in right than left canine ventricular midmyocardium. Circulation 99, 206–210.
Wecke L, Gadler F, Linde C, Lundahl G, Rosen MR & Bergfeldt L (2005). Temporal characteristics of cardiac memory in humans: Vectorcardiographic quantification in a model of cardiac pacing. Heart Rhythm 2, 28–34.
Wit AL & Rosen MR (1989). Cellular electrophysiological mechanisms of cardiac arrhythmias. In Comprehensive Electrocardiology, Theory and Practice in Health and Disease, vol. 2, section 6, ed. MacFarlane PW & Lawrie TDV, pp. 801–841. Pergamon Press, Elmsford, NY, USA.
, 百拇医药
Yu H, Gao J, Wang H, Wymore R, Steinberg S, McKinnon D, Rosen MR & Cohen IS (2000). Effects of the renin-angiotensin system on the current Ito in epicardial and endocardial ventricular myocytes from the canine heart. Circ Res 86, 1062–1068.
Yu H, McKinnon D, Dixon JE, Gao J, Wymore R, Cohen IS, Danilo P Jr, Shvilkin A, Anyukhovsky EP, Sosunov EA, Hara M & Rosen MR (1999). Transient outward current, Ito1, is altered in cardiac memory. Circulation 99, 1898–1905., 百拇医药(Michael R. Rosen and Ira S. Cohen)