Experimental models of torsade de pointes1
Abstract
Torsade de pointes is a potentially life threatening form of polymorphic ventricular tachycardia typically seen in the setting of congenital and acquired abnormal QT-prolongation. Numerous in vitro studies have investigated basic ionic mechanisms underlying delayed repolarization. The role of different ion channels and the induction of early afterdepolarizations have been studied in various cardiac cells including M cells. In addition, isolated heart models with and without electrical stimulation and/or the use of drugs which prolong repolarization have been developed in recent years. Some of these models have simulated conditions likely to exist in the clinical setting of torsade de pointes, such as bradycardia and hypokalemia. In in vivo canine and rabbit models, torsade-like polymorphic ventricular tachyarrhythmias have been induced by the administration of different agents such as cesium, neurotoxins, e.g., anthopleurin or various class III drugs under conditions designed to mimic the clinical situation. In the context of recent advances in the molecular genetics of long QT syndrome, those models which have used sodium or potassium channel blockers have gained particular interest. Based on all experimental studies it seems probable that the first beats of torsade occur due to early afterdepolarizations and triggered activity. The development of subsequent beats is less clear. Reentry based on inhomogeneity of refractoriness has been suggested as the underlying mechanism.
Key words
Time for primary review 31 days.
1 Introduction
Torsade de pointes (TDP) is a term coined by Dessertenne [1]for a life threatening form of polymorphic ventricular tachycardia (VT) typically seen in the setting of abnormal QT-prolongation. TDP has become increasingly recognised as a common and important clinical phenomenon which may be associated with a wide variety of clinical entities including the congenital long QT syndrome (LQTS) [2]and acquired abnormal QT-prolongation (Table 1). The arrhythmia is typically characterised by QRS complexes of progressively changing amplitude and contour that appear to rotate around the isoelectric line. It usually ends spontaneously but may also degenerate into ventricular fibrillation. In congenital LQTS, TDP normally develops during exertion or emotional stress, and is exacerbated by beta-adrenergic agonists [3, 4]. The onset of TDP is not necessarily bradycardia-dependent, as is usually the case in the acquired form. The majority of cases of acquired TDP occur due to treatment with class IA and III antiarrhythmic agents which prolong the QT interval by delaying repolarization [5–9]. TDP related to antiarrhythmic drugs has been estimated to occur in 1 to 8% of treated patients, particularly in those with concomitant hypokalemia and bradycardia [5, 8, 10, 11]. The fact that only certain individuals are at risk for developing acquired TDP suggests a genetic predisposition toward excessive drug response [12, 13].
View this table:
Table 1
Clinical conditions associated with repolarization abnormalities and torsade de pointes
None of the different criteria used by various investigators for defining TDP is absolutely sensitive or specific for TDP. There is a lack of a uniform definition. In our opinion the term TDP should be defined as a polymorphic VT characterised by an onset with abnormal QT-prolongation and/or abnormal TU complexes, the electrocardiographic configuration of a progressively changing ventricular axis and spontaneous termination with the exception of rare degeneration into ventricular fibrillation.
2 Pathogenesis of TDP—major hypotheses
Several mechanisms for TDP have been suggested [14]. The mechanism, however, that initiates TDP might differ from that which maintains it. It appears possible that TDP has multiple causes serving as a final common pathway for the electrocardiographic pattern viewed as characteristic for TDP.
2.1 Early afterdepolarizations and triggered activity
Early afterdepolarizations (EADs) have been acknowledged as one or even the important mechanism underlying focal activity and the electrophysiological substrate of TDP [15, 16]. The relation between EADs, triggered activity and arrhythmogenesis has been extensively reviewed elsewhere [17–20]. In brief, EADs are cellular depolarizations occurring during phase 2 and 3 of the transmembrane potential before repolarization is completed. These depolarizations may give rise to premature action potentials or even trains of potentials that have been referred to as triggered activity [21]. The hypothesis that two or more automatic foci drive the heart at slightly different rates was first suggested by Dessertenne [1]whose paper on TDP was entitled “La tachycardie ventriculaire à deux foyers opposés variables”. Several foci discharging from different sites of the heart were considered to be the cause of the twisting peaks of the QRS [22]. The occurrence of multiform extrasystoles, often in opposite direction to the QRS complexes, before the onset of TDP was suggested to favour this hypothesis [23–25].
2.2 Reentry based on dispersion of repolarization
Nonuniform prolongation of action potentials is a substrate for reentry and has been proposed as the underlying mechanism of TDP (“dispersion hypothesis”) [26]. Two or more reentrant paths [5, 8]or reentry with interaction between two ventricular areas [27]may induce different waves of excitation that compete with each other and elicit the characteristic electrophysiologic pattern. Hence, the spontaneous termination of TDP can be explained by block of conduction in the pathway responsible for reentry. If the reentrant circuit contains more than one alternate pathway through which the surrounding tissue is excited, a single reentrant wavefront with a changing excitation pattern may also explain the undulating pattern.
3 Pathogenesis of TDP—experimental evidence
Numerous experimental models have been developed in order to study TDP and to simulate the electrocardiographic characteristics of TDP listed in Table 2. In our opinion, most of them have serious limitations and some are completely unsuitable for studying TDP. This review will focus on the different experimental approaches that have been developed so far. Our current understanding of the mechanisms initiating and maintaining TDP will be summarised and future goals in this field of research will be identified.
View this table:
Table 2
Electrocardiographic characteristics of torsade de pointes
3.1 In vitro studies
Several in vitro studies concerning the basic mechanisms of prolonged repolarization and TDP have been published (Table 3). Many of those dealing with basic ionic mechanisms of delayed repolarization were performed on isolated myocardial cells whereas studies investigating the electrophysiologic characteristics of TDP mainly used isolated whole heart preparations.
View this table:
Table 3
In vitro models related to torsade de pointes
3.1.1 Cellular mechanisms of early afterdepolarizations
Slowing of repolarization and critical lengthening of action potential duration (APD) which can result from an increase of one or more inward currents, a decrease of one or more outward currents, or from both, appear to be necessary, but not sufficient to evoke EADs and triggered activity. This is followed by a net depolarising current as the substrate of EADs. Currents implicated in the generation of EADs include Na+ “window” or slowly inactivating current, L-type Ca2+ current, T-type Ca2+ current, the transient inward current (ITI), Na+–Ca2+ exchange current, and depolarising K+ currents [19, 20, 28]. Damiano and Rosen [29]demonstrated two types of EADs by application of cesium in isolated canine Purkinje fibers. One type arises from voltages near the action potential plateau, from a prolonged phase 2 (takeoff potential around −30 mV). The second type arises from phase 3 at a takeoff potential of approximately −60 mV. Phase 2 and phase 3 EADs have been observed in Purkinje fibers in the presence of a variety of interventions including exposure to quinidine [30, 31].
Table 3 provides a nonexhaustive list of experimental approaches that result in prolonged repolarization and EADs. The majority of pharmaceutical interventions which prolong repolarization and are associated with EADs predominately act through one of three mechanisms: (1) Cesium [32, 33], quinidine [30, 34–36], and type III antiarrhythmic drugs (e.g., sotalol [37], clofilium [38]bretylium [39]), all delay repolarization by reducing K currents (Ik and/or Ik1). Quinidine mainly prolongs repolarization by a voltage-dependent block of Ik and at higher concentrations Ik1 [17, 34, 36, 40]. Noteworthy, the calcium channel blockers nifedipine and verapamil were demonstrated to eliminate quinidine-induced EAD without altering the action potential plateau duration [41]. Thus, EADs may occur via Ca channels regardless the underlying cause of action potential prolongation. Potassium channel openers [e.g., pinacidil that enhances membrane conductance of potassium by activating the adenosine triphosphate-sensitive K current] also suppressed EADs induced by quinidine, cesium, and sematilide in isolated canine Purkinje fibers [42]; (2) Bay K 8644 prolongs repolarization and produces EADs in sheep and canine Purkinje fiber segments by a voltage- and use-dependent increase of ICa, mainly by increasing mean open times of L-type Ca channels [43, 44]. In canine Purkinje fibers, d,l-sotalol and clofilium have recently been demonstrated to provoke EADs and triggered potentials only in the presence of Bay K 8644 or stimulation of Na–Ca exchange which facilitated L-type calcium current [28]. Thus, EADs and triggered activity may occur via calcium window current when L-type calcium current is increased by Bay K 8644 or via Na–Ca exchange current. Of note, calcium channel blockers abolished EAD-related triggered activity in several in vitro [43, 45]as well as a few in vivo [46, 47]studies; (3) aconitine as well as the neurotoxins anthopleurin-A (AP-A) and ATX-II are known to delay the inactivation of Na currents [48]through an enhanced Na “window” current [49]or a slowly inactivating Na current [50]without changing the kinetics of sodium or potassium activation [51, 52]. All three have demonstrated a bradycardia-dependent prolongation of APD [48, 53]. EADs caused by these drugs were found to be sensitive to Na channel blockers and rapid ventricular stimulation in vitro and in vivo [53].
In view of these basic ionic mechanisms it is important to note that recordings from a few cells may not represent all mechanisms and phenomena of the intact heart. Triggered activity can be induced easily in experimental conditions involving isolated small cardiac preparations in which reentry cannot be induced. Uncoupling Purkinje fibers from ventricular cells seems to facilitate triggered activity in the latter [31]. Moreover, heart rates favouring the arrhythmia in isolated cardiac tissue are generally much lower than the physiologic range and sustained triggered activity rarely exceeds 120 beats per min which is low compared to patients with TDP.
3.1.2 Role of different cardiac cells
The exact site of origin of EADs and triggered activity is still unclear. Several studies [54–56]demonstrated that M cells which are located in the deep subepicardium are primary targets of repolarization-prolonging agents in various animal models (Table 3) although a recent study by Anyukhovsky et al. [57]failed to demonstrate a significant difference in repolarization among myocardial layers of epi-, endo- and midmyocardial cells in the intact normal canine heart. A table of species in which M cells have been demonstrated has recently been published by Antzelevitch [58]. A smaller contribution of IKs participated in determining the repolarization properties of the M cells [59]. M cells, but not epicardial or endocardial cells have previously been shown to develop EADs and EAD-related triggered activity following exposure to potassium channel blockers such as quinidine, 4-aminopyridine, clofilium, cesium, amiloride, d-sotalol, d,l-sotalol, erythromycin, calcium agonists such as Bay K 8644, and agents that slow inactivation of sodium channels, including ATX-II and AP-A [54–56, 60–63]. Conversely, in a study by El Sherif et al. M cells failed to develop EADs at concentrations of AP-A and ATX-II that easily induced EADs in Purkinje fibers [64]. AP-A produced bradycardia-dependent lengthening of APD that was shown to be more pronounced in Purkinje fibers than ventricular muscle cells, and EADs were induced only in Purkinje fibers [53]. Furthermore, EADs induced in Purkinje fibers were able to excite an action potential in the attached ventricular muscle. Similar results were reported for clofilium [65], EDTA [66], and quinidine [41]. Therefore, EADs may preferentially occur in Purkinje fibers and be conducted across the cardiac Purkinje–muscle junction to induce activity in the surrounding ventricular tissue [19]. A process which may be facilitated by catecholamines [31].
3.1.3 The isolated heart and TDP
3.1.3.1 Electrical stimulation in the isolated heart
In isolated porcine hearts perfused by the Langendorff-technique, Naumann D'Alnoncourt et al. [23]induced a TDP-like VT by simultaneous electrical stimulation of the heart from two separated sites at slightly different frequencies. A focus in the right or left ventricle either showed left or right bundle branch block, respectively. Impulses spreading from the left over the heart were shown to interfere with a second accelerating pulse in the right ventricle. The subsequent electrocardiogram (ECG) showed a gradual change from right bundle branch block to left bundle branch block pattern, and resembled the characteristic appearance of TDP. Thus, this model favoured Dessertenne's [1]original hypothesis of focal activity as the underlying mechanism of TDP. Yet, this model is unlikely to be an adequate model of TDP as it does not include the typical initiation of TDP originating from a marked abnormal QT prolongation.
3.1.3.2 Isolated hearts without electrical stimulation
In order to find an isolated heart model to test repolarization-prolonging drugs for their ability to develop TDP, Zabel et al. [67]and our group [68, 69]have recently developed an experimental model using Langendorff-perfused rabbit hearts. In our experimental setup we aimed to reproduce conditions and circumstances that are clinically known to be associated with an increased propensity to develop TDP. Bradycardia was achieved by complete AV-block. The class III agent clofilium and d,l-sotalol as well as the antibiotic erythromycin which is also known to have proarrhythmic potential were infused in the presence of either normal (5.88 mM) or low potassium (1.5 mM) concentration. Under these conditions TDP spontaneously emerged in the clofilium, d,l-sotalol and erythromycin treated hearts (Fig. 1). Electrical stimulation at cycle lengths shorter than 600 ms and of MgSO4 suppressed arrhythmic activity. In the d,l-sotalol treated hearts no episodes were observed with sinus rhythm and normal potassium concentration. However, when the potassium concentration was lowered to 1.5 mM typical episodes of TDP were observed. Fig. 1 shows an example of TDP episodes after application of erythromycin in the presence of low potassium. Left endocardial as well as epicardial monophasic action potentials (MAP) recordings demonstrate EADs. The results of this and other studies [67, 70, 71]underline the importance of hypokalemia and bradycardia as factors which may contribute to the manifestation of TDP. Asano et al. [72]very recently investigated quinidine and E-4031-related EADs, triggered activity and polymorphic VTs resembling TDP in Langendorff perfused rabbit hearts. High resolution video imaging demonstrated that EADs occurred at multiple subendocardial sites which led to beat to beat changes in ventricular activation and the typical polymorphic ECG pattern. There was multifocal character of activation which occurred at concentric areas of subendocardial breakthroughs. Furthermore, the presence of several wave fronts at varying sites was suggested to initiate drifting spiral wave activity.
Fig. 1
Spontaneous episodes of torsade de pointes in the presence of erythromycin (100 μM) in a Langendorff-perfused isolated rabbit heart with complete AV-block in the presence of low potassium (K+=1.5 mM). Two volume conducted ECG leads as well as two left endo- and epicardial monophasic action potential recordings are demonstrated.
3.1.3.3 Computer models of TDP
Though ECG recordings from in vitro studies in isolated hearts have demonstrated the typical, characteristic undulating pattern of TDP, the mechanism of the changing QRS waveform has not been definitely established. It is very difficult to implicate EADs which are acknowledged as the initiating mechanism in TDP as the only mechanism underlying the maintenance of TDP. As already pointed out by Cranefield and Aronson [73]“there is no compelling evidence why rhythmic activity arising from afterdepolarizations need give rise to polymorphic ventricular tachycardia”. Abildskov and Lux [74, 75]developed a mathematical two-dimensional computer model of propagated excitation. Premature stimulation from a region with shorter refractoriness resulted in figure-of-eight circus movement reentry and a progressive migration of the site of the reentrant circuit. In this model, serial reentry was the most likely mechanism underlying the changing QRS waveform and the limited duration of the tachycardia. The effects of increased adrenergic activity (increased rate and decreased refractoriness) with respect to the onset of TDP have very recently been investigated in this model [76]. The adrenergic-dependent shortening of refractoriness and acceleration of rate facilitated the initiation of TDP by permitting earlier premature excitation and allowing reentry. This may be of particular importance for the occurrence of TDP in the setting of congenital QT syndrome where TDP is not necessarily associated with bradycardia. However, results in computer models do not prove that they occur in the heart as well. These models can only simulate a limited number of the properties of cardiac cells and represent a simplification of the anatomical and histological structure of the real heart [77].
3.2 In vivo models
Numerous studies of in vivo models of TDP have been published and are summarised in Table 4.
View this table:
Table 4
In vivo models related to torsade de pointes
3.2.1 Conscious hypokalemic dogs
Based on earlier studies by Roberts et al. [78], Weissenburger and coworkers [70, 79–82]developed a canine model with bradycardia and hypokalemia. Bradycardia was achieved by chemically- or electrically-induced complete AV-block whereas hypokalemia was induced over a period of weeks using high doses of diuretics. Quinidine and sotalol but not lidocaine, flecainide, or propanolol exhibited significant proarrhythmic effects with regard to TDP. Recently, mexiletine was shown to antagonise sotalol-related TDP in this model. Only one out of eight dogs treated with sotalol plus mexiletine as opposed to six of eight with sotalol alone, developed TDP [79]. The use of high doses of diuretics required to induce hypokalemia may be interpreted as a limitation of this canine experimental approach as a screening model of TDP. These drugs influence other electrolyte levels apart from potassium (e.g., magnesium), may induce renal failure, and may leading to metabolic alkalosis which itself has been shown to influence the incidence of TDP in man [83].
3.2.2 Canine models with cesium
Two other in vivo models have produced bradycardia-dependent QT-prolongation and polymorphic VTs in the dog. The two drugs used were cesium [33, 84]and AP-A [53]or ATX-II [53]. Despite the similarity of the studies—both used anaesthetized dogs made bradycardic by either AV-block [33, 84]or vagal stimulation [64]—these drugs induced prolonged repolarization and TDP through different ionic mechanisms (see above). In the anaesthetised open-chested dog with cesium-induced polymorphic VTs, EADs were demonstrated in MAP recordings [33, 85, 86]. The EAD-like deflections (that may just reflect electrotonic interactions between neighbouring cells) were more prominent in endocardial than in epicardial recordings [33, 53]. Brachmann et al. [84]induced cesium-related triggered activity in isolated tissues, and polymorphic VTs in dogs with bradycardia due to chronic AV-block after local injection of formaldehyde. The spontaneously occurring arrhythmias disappeared during pacing and were suppressed by magnesium. Although it is tempting to assume from these data that the successful suppression of the arrhythmia by magnesium may characterise an experimental approach as a useful TDP model, it should be remembered that magnesium has also been effective in the treatment of various other arrhythmias [87]. Sato et al. [85]induced sustained or nonsustained VTs in 12 dogs by administering cesium chloride as an intravenous bolus after complete AV-block had been produced by catheter ablation, and observed humps on MAP recordings which were interpreted to be EADs. However, a TDP-like arrhythmia occurred in only three dogs. Ventricular premature beats did not occur until the amplitude of the induced EADs increased sufficiently, suggesting the existence of a threshold response.
As pointed out earlier by Nayebpour et al. [88], it is important to exercise caution in interpreting data from models employing cesium for the following reasons. (1) Both, bolus administration [33, 84]as well as maintenance infusion [86, 89]of cesium in dogs cause arrhythmias that often do not resemble the morphological features of TDP. (2) Cesium gives rise to a marked increase in arterial pressure [88]. This may result from catecholamines released [90]and/or a direct effect on myocardial contractility [88]. Taking into account the fact that mechanical stretch can produce transient depolarizations and electrophysiological changes [91, 92], cesium-related hypertension may well contribute to ventricular arrhythmogenesis. (3) Cesium increases potassium concentration to 6–8 mM range. However, an increase in serum potassium concentration to 4–5.4 mM has been shown to suppress EADs caused by cesium in vitro [29]. The increase in potassium may explain the lack of a significant QT-prolongation in in vivo experiments. (4) Overdrive stimulation has often aggravated the arrhythmia in direct contrast to the clinical situation in which overdrive pacing suppresses TDP [5, 7, 8]. Thus, cesium-related VTs are in many respects different to TDP and this precludes the use of cesium for studying TDP in in vivo experimental models.
3.2.3 Canine models with AP-A and neurotoxins
In the canine AP-A-model, EAD-like deflections in endocardial MAP recordings corresponded to a prominent U wave in surface electrocardiographic leads [53]. Ventricular premature complexes arising from this U wave initiated polymorphic VTs that terminated spontaneously or degenerated into ventricular fibrillation. Recently, El Sherif et al. [64]performed high-resolution tridimensional isochronal mapping (64 plunge needles electrode) of both activation and repolarization in this model. Using activation recovery intervals, the continuation of TDP was demonstrated to be due to wandering foci or to foci alternating with reentrant pathways. The initiating beats of TDP were likely to arise from a subendocardial site. There was no evidence for the origin of the VTs in the M cell region. However, as pointed out by the investigators, the classic TDP morphology was uncommonly observed and the episodes of the VTs rarely terminated spontaneously which is different from the clinical situation.
3.2.4 Rabbit model with class III agents
Various class III agents have been investigated in experimental rabbit models [65, 93–97]. Low doses of clofilium, almokalant, dofetilide and sematilide induced a marked QT(U) prolongation in alpha-chloralose anaesthetised rabbits. This prolongation was followed by a TDP-like polymorphic VT [65, 94, 95]. Hallman and Carlsson [97]recently suppressed these arrhythmias with the type I antiarrhythmic drug flecainide. In anaesthetized rabbits, arrhythmias only occurred when the animals were concomitantly treated with an infusion of the alpha-1-agonist methoxamine. The resulting increased adrenergic stimulation may elevate the free cytosolic calcium level and may in combination with a drug that prolongs repolarization, lead to calcium-dependent EADs and triggered activity. In the clinical setting of drug-related TDP, the role of increased adrenergic stimulation is uncertain. In addition, in rabbits alpha-1-receptor density as well as the functional response to alpha-1-stimulation is very high compared to other species including man [98, 99]. Anaesthesia with alpha-chloralose also resulted in a potassium concentration of 2.87 mM in the arterial blood in contrast to 3.97 mM in conscious rabbits [94].
Hence, drugs cannot be studied in this model without the simultaneous presence of hypokalemia. The occurrence of TDP was shown to be concentration-dependent. A higher rate of infusion of almokalant may lead to an increased dispersion of repolarization, measured as QT interval dispersion, and more episodes of TDP than a slower rate of infusion [95].
3.2.5 Canine models and myocardial infarction
Bardy et al. [100]studied anaesthetised dogs with acute myocardial infarction and additional cardiopulmonary bypass instituted to maintain stable temperature, perfusion pressure and oxygenation. They analysed epicardial activation patterns during induced polymorphic VTs resembling TDP and demonstrated that each change in QRS morphology was associated with a change in the site of the earliest breakthrough. Inoue et al. [101]used a similar model in which acute ischemia was superimposed on quinidine toxicity and found nearly identical results.
Open-chested dogs are, however, prone to fluctuation in core temperature, blood pressure and cardiac output [100]. These factors may have influenced the incidence of arrhythmias. Moreover, the tachycardias nearly always degenerated into ventricular fibrillation whereas TDP typically terminates spontaneously. Arrhythmias were induced by premature stimulation in both studies in contrast to the initiation of TDP in the clinical situation. In addition, both studies used toxic doses of quinidine whereas quinidine-induced TDP are commonly associated with low or normal levels of quinidine concentration. Finally, myocardial infarction serves little relationship to most clinical situations involving drug-related TDP so that these models are also unsuitable for studying TDP.
3.2.6 Canine models with ventricular pacing
A number of studies focused on different pacing modes to facilitate the induction of TDP. The peculiar twisting pattern of TDP could be imitated by electrical stimulation of two distant endocardial sites at slightly different rates [100]or by application of aconitine soaked pledgets at two epicardial sites [102]. Vos and coworkers [71, 103]described an in vivo canine model with chronic AV-block using d-sotalol and different pacing modes mimicking sequences of short/long/short intervals which are an important finding in patients with acquired LQTS but have recently been shown to play a role in the genesis of TDP in patients with the congenital LQTS as well [104]. In this model, the incidence of polymorphic VT ranged between 52% and 89% depending on whether one or two d-sotalol doses were applied. A sudden rate change (often two or more cycle length changes) was shown to be necessary to induce VTs. Inducible dogs had a higher incidence of TDP. Magnesium, increased pacing rates, and isoproterenol all prevented the induction of TDP [71]. Ryanodine and flunarizine were shown to prevent the induction of TDP [103]. Dogs with induction of TDP have recently been demonstrated to have an increased interventricular dispersion of repolarization compared with noninducible dogs [105]. Although pacing is widely used to differentiate reentrant from nonreentrant arrhythmias, there is, at least so far, no role for pacing in testing patients with drug-related torsade or TDP due to LQTS. Thus, as pointed out by Verduyn et al. [105]pacing-induced TDP is not similar to spontaneously occurring TDP. One possible explanation is that no electrophysiological technique indicates the presence of triggered activity and differentiates it from other arrhythmogenic mechanisms [26].
4 Experimental models and congenital LQTS
In LQTS abnormalities of the sympathetic nervous system have been suspected because of the slow heart rates observed and the common precipitation of syncope by adrenergic stimulation. A few experimental studies demonstrated that the creation of sympathetic imbalance, by removing the right stellate ganglion or by stimulating the left, prolongs the QT interval and induces arrhythmias and characteristic T wave changes [2, 106, 107]. In the previously mentioned cesium chloride model, sympathetic stimulation increased the amplitude of EADs and the prevalence of VTs [89, 108]. However, total ablation of the right stellate ganglion and intensive stimulation of the left, did not lead to QTU abnormalities as severe, or spontaneous arrhythmias as dangerous, as in the clinical syndrome [5]. Malfatto and coworkers [109, 110]induced QT prolongation in a rat model by modulation of sympathetic neural development using nerve growth factor or its antiserum. However, in this study no episodes of sudden death and malignant arrhythmias were observed. Yet, abnormalities in sympathetic innervation may modify characteristics of cardiac membranes and/or change electrophysiological properties and the response to catecholamines. Moise et al. [111]established a colony of German shepherd dogs with inherited ventricular arrhythmias and sudden cardiac death in the absence of structural heart disease. Nonsustained polymorphic VTs were observed after long RR intervals. However, these dogs do not represent a congenital LQT model. The QT intervals from the routine ECGs for all affected dogs were within the normal range. So far, the exact mechanisms underlying these arrhythmias remain unknown but because they are age- (most prevalent in young dogs), behaviour- (most frequently during sleep or at rest after exercise), and heart rate-dependent, an imbalanced autonomic innervation to the heart may play an important role [111, 112]. In addition, EADs and triggered activity were shown to occur spontaneously in Purkinje fibers obtained from affected dogs [113].
4.1 Current knowledge of TDP and relevance of animal models for the patient
There is a high variability in occurrence of TDP in man with drug-related TDP as well as with congenital LQTS. This high variability is also observed in all animal models. There is no model that causes TDP with a 100% reproducibility. Nearly all in vivo studies had to use complex methods which indicates how difficult it is to induce TDP under experimental conditions. The rare occurrence of TDP in some experimental models suggests that other predisposing and, so far, unknown factors may contribute to the acquired LQTS. In our opinion only a few in vivo and/or in vitro models are suitable for studying TDP. These are mainly those that simulate the clinical situation most closely. This means for drug-related TDP the use of a drug that prolongs repolarization and leads to abnormal TU complexes plus possible additional factors like bradycardia and/or hypokalemia. Toxic doses of cesium or quinidine, myocardial infarction and various pacing protocols can only be used to mimic certain characteristics.
Prolongation of QT and APD appears necessary, but not sufficient to cause EADs, triggered activity and TDP. Noteworthy, Leenhardt et al. [114]described 14 young patients who presented with syncope and had at least one episode of polymorphic VT resembling TDP (in the absence of structural heart disease) in whom the coupling interval of the TDP was very short (245±28 msec). Arrhythmias were therefore termed “short-coupled variant” of TDP. These are the only clinical data which are not compatible with the above given definition of TDP. However, this form of arrhythmia may also represent polymorphic VTs of different nature. Besides, in ten patients the tachyarrhythmias deteriorated into ventricular fibrillation.
There is no assurance that EADs recorded with extracellular electrodes are not motion artefacts or a result of electrotonic interactions between neighbouring regions with disparate repolarization phases [8, 17, 26]. However, MAP recordings in patients with TDP have demonstrated that the late component of the T/U wave correlates with a secondary deflection near the end of repolarization [15, 16, 115]. Thus, it may well be that these deflections which are similar to EADs measured in vitro and in vivo represent a summation of EADs occurring throughout the myocardium. In addition agents that are useful in patients presenting with TDP (e.g., magnesium, isoproterenol [26, 116]) have been shown to suppress EAD-induced triggered activity in experimental studies.
The extent to which an individual current contributes to prolonged repolarization is not only determined by cell characteristics but also by species. Depending upon species and cell type, a number of K currents may be responsible for cardiac APD. The delayed rectifier current IK has been well characterised in several vertebrate cardiac cells [117–119]. In some species it is composed of the two components IKr and IKs [59, 120, 121]. These two components are, e.g., larger in guinea pig than in dog ventricular myocytes under control conditions [59]. In addition, the transient outward current is for example absent in guinea pigs, and the slow component of the delayed rectifier is more prominent than in most species [122]. Apart from basic electrophysiologic differences between species, variability is also present between individual animals. Heart size may also be of relevance although studies in rabbit hearts have proved to be large enough for studying TDP.
In the context of advances in the molecular genetics of congenital LQTS, the use of the neurotoxins AP-A and ATX-II has recently gained particular interest because of the report demonstrating a genetic mutation (LQT3) for the gene that encodes the voltage-gated sodium channel alpha–subunit (SCN5A) [123]. The mutant channels expressed in vitro were shown to generate a sustained inward current during depolarization [124]. This current is quite similar to a situation in which sodium channels are exposed to either two agents [64, 125]. Thus, the previously described animal model combines features of both a form of congenital LQTS (ion channel abnormality) and acquired LQTS (bradycardia-dependent onset of arrhythmia) [64]. The relevance of potassium channels has been demonstrated for the form of LQTS linked to chromosome 7 (LQT2) [126]. The corresponding gene HERG encodes subunits that form channels responsible for IKr [13, 127]. Interestingly, the potassium channel opener nicorandil shortened monophasic APD and decreased humps which were interpreted to be EADs on MAP recordings in a young patient with congenital LQTS [128]. Dofetilide [129]and sotalol by blocking potassium channels mimic the congenital ion channel abnormality of LQT2 in animal models. Of note, mexiletine as a sodium channel blocker has shown to antagonise sotalol related TDP in conscious hypokalemic dogs [79]. This is in contrast to the clinical situation in which patients with LQT3 may profit from blocking sodium channels. Schwartz et al. [130]were the first to show that LQT3 patients may be more likely to benefit from sodium channel blockers than LQT2 patients. Notably, this correlates to an in vitro study on guinea pig ventricular muscle cells in which mexiletine significantly reduced APD in anthopleurin exposed cells whereas it did not modify the APD prolongation induced by dofetilide [129].
5 Future experimental studies
Many aspects of TDP are still far from being completely understood. Why do the acquired LQTS and TDP only occur in certain individuals? Do those who demonstrate TDP have a subclinical genetic abnormality involving cardiac repolarization? If so, does it also exist in animals and are there age-related differences? Recent data suggest that some patients with acquired TDP may be asymptomatic LQTS gene carriers [131, 132].
It is well recognised that QT intervals are longer in women than in men, although the mechanism is not well understood. It is, therefore, interesting that clinical studies have documented a higher incidence of TDP among women [133–135]. No experimental TDP model has ever studied differences in gender. Yet, an association between sex hormones and QT prolongation as well as downregulated potassium channels has been shown recently in rabbit hearts [136].
The exact role of the autonomic component of LQTS needs to be clarified. The fact that alpha-1-adrenoeceptor stimulation increases, alpha-1 blockade decreases EAD amplitude and incidence of TDP in the canine cesium model [108], as well as the fact that alpha stimulation was necessary to induce TDP in Carlsson's conscious rabbit model [94], may point to a crucial role of alpha adrenoceptors in LQTS.
The role of hypertrophied myocardium remains unclear, too. Ventricular hypertrophy may alter calcium metabolism and thereby facilitate EADs. There are some data indicating that hypertrophied myocardium is more prone to develop TDP [137], but thorough experimental studies have not yet been performed. In addition, it is not well known whether an underlying heart disease affects the incidence of TDP. No in vitro or in vivo study on TDP has been performed on failing or ischemic myocardium. Pathological conditions may alter electrophysiology and thereby influence the propensity to develop TDP. At the cellular level, the role of M cells for the initiation of TDP has not been finally clarified. The relevance of regional (intra- and interventricular) differences in repolarization remains unknown.
New drugs that terminate, interrupt or even avoid proarrhythmia will have to be developed and tested in one or various of the described animal models. One of the most challenging approaches will be to maintain the beneficial effects of repolarization-prolonging agents without engendering proarrhythmia. As only some patients who are taking repolarization-prolonging drugs are suffering from TDP, the development of risk indicators in experimental models which could be extrapolated to the clinical situation would be of great value. Finally, it will be of interest to study the congenital ion channel abnormality of LQTS patients in animal models. With recent advances in molecular genetics, it should soon be possible to design such a model.
6 Conclusion
Thirty years after Dessertenne's first description, TDP seems to provoke as much interest as ever. Numerous animal studies have demonstrated the ability of several drugs to predispose animals to abnormally prolonged repolarization and TDP. Although extrapolation of experimental results obtained from animals to the clinical situation is uncertain it seems to be most likely that the first beats of TDP are due to EAD-induced triggered activity. Differences in ion channel constituents, rate-dependence, and pharmacological sensitivity within the ventricular wall may contribute to the occurrence of this arrhythmia and the site of origin. The genesis of subsequent beats during an episode of TDP remains contested. These ectopic beats are likely to be due to reentry caused by inhomogeneity of refractoriness. However, neither of the initially discussed hypotheses has been conclusively excluded as the cause of TDP. Each mechanism may be relevant under appropriate circumstances.
Animal models have been able to mimic drug-related TDP as well as ion channel abnormalities of the LQTS. They have provided an important insight into the pathophysiology of TDP. Further experimental efforts will improve our understanding, prediction, prevention and therapy of this picturesque but often deadly arrhythmia.
- Hospital of the Westfälische Wilhelms-University, Department of Cardiology and Angiology and Institute for Arteriosclerosis Research, D-48129 Münster, Germany
- * Corresponding author. Tel.: +49 (251) 834 7638, Fax: +49 (251) 834 7635, E-mail: l.eckardt@uni-muenster.de
- Received January 22, 1997.
- Accepted January 12, 1998.
Footnotes
-
↵1 Partly supported by the Franz Loogen Foundation, Düsseldorf, Germany.
- Copyright © 1998, European Society of Cardiology
References
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
This Article
- AbstractFree
- » Full Text (HTML)Free
- Full Text (PDF)Free
Classifications
Services
Responses
- submit an e-letter
- No responses published
Citing Articles
Related Content
Published on behalf of
The Journal
Impact Factor: 5.801
For Authors
- Services for authors
- Instructions to authors
- Submit now!
- Self Archiving Policy
- Special Issues - Call for Papers
Open access options for authors visit Oxford Open
