REVIEW ARTICLE

CARDIAC CHANNELOPATHIES

Eduardo Marbán*, MD, Ph.D
 Institute of Molecular Cardiobiology, The Johns Hopkins University, Baltimore, Maryland, USA

Abstract
Introduction
Genesis of the heartbeat
The electrocardiogram
Mechanism of arrhythmias
Heritable channelopathies
    Disorders of repolarization
Loss of Na+ channel function and arrhythmia
Acquired channelopathies
    Heart Failure  
Drug-induced long-QT syndrome
The future in diagnosis and treatment
Gene therapy for cardiac arrhythmias?
References

Abstract

   Genetic alterations of various ion channels produce heritable cardiac arrhythmias that predispose affected individuals to sudden death.  The investigation of such ‘channelopathies’ continues to yield remarkable insights into the molecular basis of cardiac excitability.  The concept of channelopathies is not restricted to genetic disorders; notably, changes in the expression or post-translational modification of ion channels underlie the fatal arrhythmias associated with heart failure.  Recognizing the fundamental defects in channelopathies provides the basis for new strategies of treatment, including tailored pharmacotherapy and gene therapy.  (Heart Views,2003;4(3): 94-103

Introduction

   The heart pumps blood throughout the body and never rests, undergoing roughly three billion cycles in a typical lifetime.  To achieve this, the heart must first relax so that the atria and ventricles can fill with blood, and then contract to propel the blood throughout the body.  This cycle of relaxation and contraction occurs in a single heartbeat.

   Each heartbeat is initiated by a pulse of electrical excitation that begins in a group of specialized pacemaker cells and subsequently spreads throughout the heart.  This electrical impulse is made possible by the electrochemical gradient that exists across the surface membrane of each heart cell or myocyte.  At rest, the membrane is selectively permeable to K+ ions, and the electrochemical potential inside the myocyte is negative with respect to the outside.  During electrical excitation, the membrane becomes permeable to Na+ and the electrochemical potential reverses or “depolarizes”.  Ca2+ ions move into the cell and activate the contractile machinery - a process that, when it happens en masse, causes the atria and ventricles to contract and expel blood.  The wave of depolarization is self-limiting; as a negative membrane potential is restored, the heart relaxes and fills with blood for the next cycle.

   Because the heartbeat is so dependent on the proper movement of ions across the surface membrane, disorders of ion channels - or “channelopathies” - make up a key group of heart diseases.  Channelopathies predispose individuals to disturbances of normal cardiac rhythm.  If the heart beats too slowly (bradyarrhythmias) or so rapidly (tachyarrhythmias) that it cannot fill adequately, then this leads to circulatory collapse and, in the extreme case, death.  The incidence of arrhythmias is poorly defined, but conservative estimates are in the range of several million per year in the United States. Arrhythmias lead to more than 250,000 sudden deaths per year, countless lost work days, and financial costs related to treatment including the implantation of more than 250,000 electronic pacemakers and more than 60,000 defibrillators per year.  The numbers worldwide are certainly much greater.  Several different genetic and acquired channelopathies can cause such arrhythmias.

Genesis of the heartbeat

   Normal cardiac excitability results from a delicate balance of depolarizing and repolarizing ionic currents.  Each ionic current can be distinguished by its ionic selectivity and time course - properties that are conferred by specific transmembrane proteins called channels or transporters.  Figure 1a depicts the principal channels and transporters in heart cells.  Depolarization is mediated by channels or transporters that enable positively charged ions such as Na+ or Ca2+ to enter the cell (or negatively charged ions, usually Cl-, to exit); the reverse is true for repolarization.

   Many proteins have evolved to mediate the selective flux of ions across biological membranes; those that contain a high-throughput pore (more than 100,000 ions per second per molecule) are known as ion channels(1,2).  Ion channels are generally selective for one type of ion over all others in the physiological milieu.

Fig. 1.  Ion channels underlie cardiac excitability. (a) The key ion channels (and an electrogenic transporter) in cardiac cells. K+ channels (green) mediate K+ efflux from the cell; Na+ channels (purple) and Ca2+ channels (yellow) mediate Na+ and Ca2+ influx, respectively. The Na+/Ca2+ exchanger (red) is electrogenic, as it transports three Na+ ions for each Ca2+ ion across the surface membrane.  (b) Ionic currents and genes underlying the cardiac action potential. Top, depolarizing currents as functions of time, and their corresponding genes; centre, a ventricular action potential; bottom, repolarizing currents and their corresponding genes.

   When these channels open, they bias the membrane potential of the cell towards the equilibrium potential of the ion in question.  K+ channels steer the cell towards -90 mV; in contrast, the opening of Na+ or Ca2+ channels forces the potential to positive levels (+40 mV or greater).  Thus, one can visualize each cell as a dipole that is either positive or negative depending on the relative balance of the cell’s complement of ion channels, and whether or not those channels are open at any given time.  The resulting electrical signal is known as the “action potential.”

   A prototypical action potential from a ventricular myocyte is shown in the centre of Fig. 1b.  The shape is distinctive: a sharp depolarizing upstroke gives way to a sustained, slowly decaying plateau with eventual repolarization.  Above the action potential are shown the depolarizing ionic currents, which are generated by the channels (and a single electrogenic transporter, the Na+/Ca2+ exchanger3), that underlie the electrical signal; the nomenclature of the corresponding genes appears on the right.  It should be noted that channels allow the passive movement of ions down their respective concentration gradients; when K+ channels open during repolarization, K+ exits from the cell. Conversely, during depolarization Na+ and Ca2+ enter the cell.  The ionic gradients are maintained ultimately by energy-consuming processes such as the Na+/K+ ATPase (Nature 2002; 415: 198-205).

   Na+ channels and Ca2+ channels favor depolarization.  Each opens quickly in response to a voltage stimulus (perpetuating further depolarization), and then closes despite maintained depolarization in a process known as “inactivation.”  Under normal conditions, Na+ channels inactivate quickly and completely, and very rarely re-open(4).  Calcium channels inactivate less rapidly and less completely; they feature prominently in maintaining plateau depolarization(5).

   The molecules involved in the repolarizing mechanism consist of various types of K+ channels(6), which are depicted in the lower half of Fig. 1b.  The current known as “/K1” is active at negative potentials.  Its distinctive permeation properties cause /K1 to shut-off as depolarization progresses(2).   /K1 is, therefore, ideally suited to anchor the “resting” potential, and provides little opposition to depolarization once the balance has tilted in that direction.  The other K+ currents open in distinctive voltage- and time-dependent patterns. The transient outward current “/to” has two components:  /to1 activates and inactivates rapidly, in a conventional voltage-dependent manner(7,8);  /to2 is less well-characterized and varies from species to species, but the available evidence hints that it is activated by changes in the intracellular concentration of free Ca2+(9,10).  The KCND/Kv4 family of channels provides /to1(11), but the molecular identity of /to2 is not yet known.  Functionally, Ito underlies the “notch” during the initial phase of repolarization that is often evident in ventricular action potentials(12), as shown in Fig. 1b.  /to also influences the overall duration of the action potential, albeit indirectly(13,14).

Fig. 2. The action potential and the electrocardiogram (ECG).  (a) Temporal relationships between the ventricular action potential (top) and the ECG (bottom). The QRS complex, T wave and QT interval are indicated. Both signals are functions of the same timescale on the x axis; the y axis plots voltage, with a gain that is roughly 100-fold higher for the ECG than for the action potential (that is, the absolute signal amplitude is about 100-fold smaller for the ECG). (b) Normal rhythm on the ECG. P waves produced by atrial activity precede each QRS complex.  (c) Ventricular tachycardia. QRS complexes are broadened and irregular; independent or retrograde P wave activity may be evident (not shown here).

   The delayed rectifier /k, like Ito, consists of two components: a rapid component (/kr) and a slow component (/Ks).  Mutations in each of the four genes encoding IKr and IKs have been implicated in heritable long-QT syndrome(15), as discussed below.  Finally, /Kp is a time-independent background current(16) whose molecular identity is uncertain but may be related to two-pore K+ channels of the KCNK family(17,18).  Computational simulations have shown that /Kp is essential for repolarizing action potentials19, but it remains to be studied extensively at a biological level.

The electrocardiogram

   The mass of the heart is sufficient to allow electrodes placed on the body surface to register readily the electrical changes generated by cardiac myocytes.  The resultant signal - the electrocardiogram - represents an average of the electrical gradients at any given time.  Figure 2a shows the temporal relationships between the electrical activity of a typical ventricular myocyte, as would be measured using cellular recordings of the transmembrane potential, and the corresponding electrocardiogram.

Fig. 3. Explanation of the mechanism underlying arrhythmia in long-QT syndrome. (a) Time series of a cross-section of the ventricular wall in normal myocardium. Negatively polarized (resting) muscle is shown as blue; depolarized muscle is red.  Action potentials recorded at sites 1 and 2 are similar in timing and morphology. The resulting ECG is normal (compare Fig. 2a).   (b) Time series of a cross-section of the ventricular wall in long-QT myocardium. Part of the ventricular wall undergoes an afterdepolarization (site 1), whereas neighboring site 2 has repolarized. The afterdepolarization at site 1 prematurely re-excites site 2, initiating ventricular tachycardia.

   The upstroke of the action potential at the onset of depolarization produces the spiky QRS complex; repolarization is manifested on the body surface as the gently rolling T wave.  As a first approximation, the time between the beginning of the QRS complex and the end of the T wave - the QT interval - can be used to deduce the overall timing and duration of ventricular depolarization and repolarization.  The frequency of QRS complexes and their sequence relative to the smaller P waves produced by atrial activity allow the clinical detection of normal rhythm (Fig. 2b) or rhythm disorders (see below).

Mechanism of arrhythmias

   Repolarization of the heart cell is a precarious process(20).  Although many different currents orchestrate repolarization, only a few channels of each type are open at any given time; thus, small changes, corresponding to the opening or closing of a handful of individual channel molecules(21), can pervert repolarization.  The normally smooth trajectory of repolarization might then be interrupted by abnormal, secondary depolarizations(22).  Such “afterdepolarizations"(23) are innocuous in isolated cells, but in the syncytial heart, myocytes are coupled to their neighbors, enabling the electrical impulse to spread from cell to cell as shown in Fig. 3a(24).

   Imagine that one region of the heart undergoes afterdepolarizations while neighbouring regions have begun to repolarize (Fig. 3b).  The electrocardiogram would show a prolongation of the QT interval.  The aberrant zone may then re-excite its repolarizing neighbours, initiating a premature beat that spreads in wavelets throughout the heart(25-26).  Self-perpetuating wavelets of excitation in the ventricle undermine the normal, stereotyped progression of the electrical impulse, producing ventricular tachycardia (Fig. 2c, large wavelets) or ventricular fibrillation (small wavelets)(27).  The resultant incoordinate contraction and rapid rate lead quickly to circulatory collapse and death if the arrhythmia is sustained.

Heritable channelopathies

Disorders of repolarization

   Abnormal, inhomogeneous repolarization underlies the ventricular arrhythmias characteristic of many channelopathies.  The inherent instability of cardiac repolarization is accentuated when action potentials become prolonged23, leading to a characteristic prolongation of the QT interval.  Thus, the “long-QT syndrome” - the archetypical disease of repolarization - takes its name from the distinctive abnormality of the electrocardiogram in affected individuals.  Such individuals are predisposed to ventricular tachyarrhythmias caused by unstable repolarization, which often leads to sudden cardiac death as the first manifestation of the disease.

   Long-QT syndrome can be heritable or acquired(28).  The genetically transmitted form is caused by discrete mutations in genes that encode ion channels. Either the Na+ channel gene SCN5A or one of the four genes that make up /k can be affected(15).  (The existence of other long-QT genes is suggested by linkage analysis, but their identity is not yet known.)  Long-QT-associated mutations in SCN5A produce channels with increased Na+ flux(29), whereas those in the K⁺ channels lead to loss of function(30).  The common mechanistic thread is perturbation of the balance between inward and outward currents during the plateau of the action potential.  The inheritance pattern is most frequently autosomal dominant: alterations of a single allele are sufficient to produce the arrhythmogenic phenotype.  Functional studies of defined channel mutations in heterologous expression systems or in vivo have helped to rationalize the pathophysiology.  SCN5A mutations typically yield incompletely inactivating Na+ channels(31), leading to a tendency towards unbalanced depolarizing throughout the action potential plateau.  This gain of function explains why one abnormal allele is sufficient to undermine repolarization.

   In contrast, long-QT-associated mutations in the K⁺ channels decrease K+ flux through /Kr or /Ks by loss-of-function (for example, nonsense mutations that truncate the pore-forming subunit prematurely) or dominant-negative mechanisms(30,32).  Because K+ channels are multimeric (unlike Na+ channels, in which a single protein is sufficient to create a functional pore), the dominant-negative mutations cripple the healthy products of the wild-type allele, and thus provide a ready rationale for dominant transmission.  The fact that plain loss-of-function mutations also produce dominantly inherited long-QT syndrome implies, however, that two functional alleles are required for uneventful repolarization.  Assuming that the loss of only one functional allele will produce no more than a 50% reduction of either /Kr or /Ks, these findings illustrate vividly the precarious nature of the repolarization process.

   More rarely, long-QT syndrome is inherited in an autosomal recessive manner.  Such kindreds possess two dysfunctional K⁺ channel genes, which leads to a total absence of /Kr or /Ks(33).  Individuals also affected with associated deafness (Jervell and Lange-Nielsen syndrome) have mutations in the /Kr genes KCNQ1 or KCNE1; indeed, studies motivated by such individuals led to the recognition that /Ks is necessary for endolymph production in the inner ear(15).  Much has been made of other, more subtle genotype-phenotype correlations in long-QT syndrome, such as gene-specific variations in the precise electrocardiographic features(34); however, the paucity of people with these rare disorders weakens such correlations.  It should also be noted that there is extensive phenotypic variability among known gene carriers, even within the same family, hinting that there are modifier genes yet to be recognized.

   With the exception of Jervell and Lange-Nielsen syndrome, with its associated deafness, the classical long-QT syndrome is notable for its restriction to the heart. Although the affected K+ channel genes are expressed in various tissues, the phenotype hints that the physiological roles of these genes are most crucial in the heart.  In contrast, a rare genetic disease known as Andersen’s syndrome, in which long-QT syndrome is associated with multi-system pathology (periodic paralysis and dysmorphisms), has been attributed to mutations in KCNJ2(35).  Mutations in various kindreds and sporadic cases occur throughout the KCNJ2 coding sequence; two mutants that have been characterized functionally have dominant-negative properties, consistent with the observed autosomal dominant inheritance pattern.  These findings indicate that KCNJ2 is functionally important not only in stabilizing cardiac rhythm, but also in modulating the excitability of skeletal muscle and in morphogenesis(2).

Loss of Na+ channel function and arrhythmia

   Another uncommon but instructive cardiac channelopathy is that of idiopathic ventricular fibrillation, in which previously well individuals die suddenly of a tachyarrhythmia(36).  The electrocardiogram can be normal at baseline, although some individuals (with so-called Brugada syndrome) have associated electrocardiographic abnormalities (including a form of intraventricular conduction delay called right bundle branch block)(37).  The disease has been linked convincingly to SCN5A with dominant inheritance(36,38).  Interestingly, the functional abnormalities of the expressed mutant channels are opposite to those found in Na+ channel mutants associated with long-QT syndrome.  In idiopathic ventricular fibrillation, the Na⁺ channels show loss-of-function features such as enhanced inactivation.  Affected channels may not even make it to the surface membrane at physiological temperatures(39); thus, the available evidence suggests that the syndrome is one of pure or partial hemi-allelic Na+ channel insufficiency.

   There is no intuitive rationale for ventricular arrhythmias caused by partial Na+ channel deficiency.  The principal physiological role of Na⁺ channels is the fast conduction of the cardiac impulse(40).  What, then, is predisposing the heart to fatal tachyarrhythmias?  One hypothesis is that the key abnormality lies, once again, in the repolarization process. The premise is as follows: Na+ channels inactivate quickly, but before doing so they initiate the action potential plateau, and sustain it for several milliseconds until Ca2+ channels have a chance to function.  Ito, which is richly expressed in the outer layers (epicardium) of the human ventricle(12), simultaneously turns on and opposes the depolarization (normally producing the action potential notch).  Thus, the early plateau represents a tug of war between Na+ channels and /to.  A decreased density of Na+ channels in the face of a robust /to may allow premature repolarization, resulting in a very brief action potential.  If this occurs more readily in the epicardium, where Ito is prominent, than in the inner ventricular layers, then the normally depolarized inner layers can re-excite the prematurely repolarized epicardium.  The essential feature is the postulated inhomogeneity of repolarization across the thickness of the ventricular wall(37), which is brought about by an imbalance between /Na and /to.

   Although plausible, transmural inhomogeneity may not explain fully the ventricular arrhythmias associated with Na+ channel deficiency. Some individuals affected with complete hemi-allelic Na+ channel insufficiency present only with isolated slowing of myocardial conduction(41); other individuals with less severe Na+ channel gating defects than those typically seen in Brugada syndrome likewise show isolated conduction slowing(42).  Thus, although we recognize that the full spectrum of syndromes linked to Na+ channel deficiency includes conduction slowing and ventricular arrhythmias(43), much remains to be clarified about the detailed genotype-phenotype correlations.

Acquired channelopathies

Heart failure

   The heritable channelopathies have yielded important insights into the pathophysiology of some far more common, acquired diseases.  Heart failure is a case in point.  This disease afflicts hundreds of millions of people worldwide.  Whatever the initiating factors (for example, coronary atherosclerosis, hypertension or viral infection), the final common phenotype is one of a dilated, poorly contracting heart.  Affected individuals suffer from a decreased ability to exercise and shortness of breath caused by a decrease in cardiac pump function.  Mortality remains high despite the best current therapy, exceeding 10% per year in severely symptomatic individuals.  Although the name “heart failure” suggests that gradually dwindling cardiac output might be the most likely cause of death, it turns out that most individuals die suddenly of cardiac arrhythmias.

Fig. 4. Computational rationalization of heart failure arrhythmias on the basis of known cellular changes in ionic currents. (a) Simulation of normal rhythm in the heart, with the wave of excitation (red) spreading rapidly throughout the ventricles. The resulting virtual electrocardiogram closely resembles normal rhythm (bottom trace, solid red line). (b) Four snapshots during ventricular tachyarrhythmia in a virtual heart that contains a region of cells with properties of heart failure (a form of acquired long-QT syndrome). The calculated electrocardiogram reproduces ventricular tachycardia (superimposed on the normal rhythm shown in a as a dashed line). Results shown are from R. Winslow, Johns Hopkins University.

   We now know that heart failure represents a common, acquired form of the long-QT syndrome(44).  Myocytes from failing hearts show prolongation of action potentials(45,46), and repolarization in vivo is abnormally labile(47).  In human heart failure, the action potential prolongation reflects selective downregulation of two K⁺ currents, /to1 and /K1(45).  Much of the decrease in Ito1 occurs at the transcriptional level(11).  Such K+ channel downregulation may be adaptive in the short term: increased depolarization during the cardiac cycle means more time is available for excitation-contraction coupling, which mitigates the decrease in cardiac output.  Nevertheless, the downregulation of K⁺ channels becomes maladaptive in the long term, predisposing the individual to afterdepolarizations, inhomogeneous repolarization and ventricular tachyarrhythmias.  The arrhythmic tendency is aggravated by alterations in the cycling of intracellular calcium, including upregulation of the electrogenic Na+/Ca2+ exchanger(48).

   Factually based numerical models of electrical activity have begun to shed light on the mechanisms of cardiac arrhythmias.  The initial successes came in simulations carried out on a cellular level(19,49), which rationalized the mechanisms of long-QT-related action potential prolongation and afterdepolarizations(5).  More recently, massively parallel network simulations of whole-heart electrical activity have reproduced successfully polymorphic ventricular tachycardia - an arrhythmia commonly seen in heart failure.  The model(14) includes virtual cells coupled to each other in a geometry defined from an actual mammalian heart; each cell contains a biophysically detailed model of cardiac electrophysiology, including most of the ionic currents listed in Fig. 1.

   The results of the whole-heart simulation are highlighted in Fig. 4 and in video clips available at http://www.cmbl.jhu.edu/movies/normal.mpg and http://www.cmbl.jhu.edu/movies/ead.mpg.  Normal conduction and repolarization can be reproduced in normal hearts (Fig. 4a); in contrast, when an area of the ventricle is reprogrammed to reproduce the selective downregulation of K+ channels characteristic of heart failure, an arrhythmia can be induced readily in silico (Fig. 4b).  The initiating event is an afterdepolarization from the K⁺-channel-deficient zone; once triggered, the arrhythmia is perpetuated by rapid waves of depolarization spreading asynchronously throughout the heart.  Virtual electrocardiograms reveal the undulating waveform of polymorphic ventricular tachycardia (compare the simulated electrograms in Fig. 4a with those in Fig. 2b,c).  The biological hypothesis of repolarization-related arrhythmias has thus been validated numerically from first principles. Given the complexity of cardiac arrhythmias, such in silico simulations will undoubtedly feature more prominently in future investigations.

Drug-induced long-QT syndrome

   In addition to heart failure, acquired long-QT syndrome can also be induced by exposure to drugs that block K⁺ channels(28,50).  Selective blockers of IKr such as dofetilide have been developed for the treatment of various atrial arrhythmias; unfortunately, such drugs predictably evoke prolongation of the QT interval, which is sufficient to cause dangerous ventricular arrhythmias in 5-7% of recipients.  Many other drugs block K+ channels unintentionally.  IKr is a common target - a fact that can be interpreted through the unique structural features of the KCNH2 inner vestibule that render it rather promiscuous for small organic molecules(51).  Drug-induced arrhythmias occur more frequently in women than in men, and in a small percentage of those exposed to the drugs in question(52).  In these individuals, however, the arrhythmia can be lethal; such side-effects have led to the withdrawal of various prescription medications after their initial approval by regulatory agencies.

   It is not yet known why a subpopulation is particularly prone to drug-induced long-QT syndrome.  One notion is that the repolarization process has a certain “reserve” built in by virtue of the redundancy of K⁺ channels and their normal levels of expression.  A diminished repolarization reserve, perhaps caused by a channel mutation that alone does not cause symptoms, may potentially lead to arrhythmias in the presence of certain drugs(53).  In particular, otherwise innocent polymorphisms in ion channel genes may enhance drug binding and magnify the channel block(54).  The prototype for such polymorphisms may be a mutation in KCNE2 that has been reported to underlie arrhythmias triggered by an antibiotic(55), but the findings are so far restricted to a single proband.  Indeed, the mechanism whereby KCNE2 mutations produce repolarization instability remains uncertain(56).  The basis of gender differences is also unknown, but is consistent with the fact that the QT interval is longer, at baseline, in women than in men(52).

The future in diagnosis and treatment

   The clinical criteria for diagnosing long-QT syndrome and other cardiac channelopathies remain the traditional ones: history, physical examination and electrocardiography(57).  The alternatives for genetic diagnosis are linkage analysis, which is practicable only in large families, and a candidate gene approach.  No commercial high-throughput approach to genetic diagnosis is available as yet; indeed, such tests may be some time in coming, given the rarity of the classically inherited syndromes, the ever-increasing number of documented mutations, and the likelihood that several other culpable genes have yet to be recognized.  Nevertheless, those individuals who have been genotyped may benefit from tailored pharmacotherapy.

   Long-QT syndrome caused by mutations in SCN5A can be treated rationally using Na+ channel blockers of the local anesthetic class, such as mexilitine.  Such drugs preferentially block the non-inactivating mutant Na+ channel current; their clinical value has been documented in various case reports(58,59).  The existing pharmacopeia is less helpful in long-QT syndrome related to K⁺ channels.  Therapeutic measures centre on general measures such as oral potassium supplementation(60), strict avoidance of aggravating factors such as potassium-wasting diuretics, and the implantation of electronic devices to regularize rhythm (pacemakers and/or automatic cardioverter/defibrillators). Other avenues that merit further investigation include K⁺ channel agonist drugs(61) and gene therapy.

Gene therapy for cardiac arrhythmias?

   The currently available therapies for arrhythmias are limited by poor efficacy and the incidence of serious side-effects.  Options for treatment include pharmacotherapy, radiofrequency ablation and implanted devices.  Antiarrhythmic medications can sometimes reduce primary arrhythmic events, but their systemic effects are often poorly tolerated; in addition, their paradoxical ability to make some arrhythmias worse while treating others actually increases mortality in many situations.  Radiofrequency ablation cures a limited number of arrhythmias and has become the standard of care for people with congenital structural abnormalities (for example, Wolff-Parkinson-White syndrome), but more common atrial and ventricular tachycardias are less amenable to this form of therapy.  Device-based therapies (pacemakers and defibrillators), while palliative, can be quite effective.  However, this strategy does not prevent tachyarrhythmias and is associated with a lifetime commitment to repeated procedures, significant expense and potentially catastrophic complications (lung or heart perforation, lead dislodgement, or infection).

   The lack of effective therapeutic options motivates the pursuit of alternative strategies for cardiac arrhythmias - notably gene therapy.  The philosophy is very different from that of conventional gene therapy (Nature 2002; 415:234-239); here, the purpose of gene transfer is to effect electrical re-engineering.  The genes in question are those that encode either ion channels or modulators of ion channels such as G proteins (Nature 2002; 415:213-218).  The concepts can be generalized to ventricular arrhythmias, such as those discussed above.  In heart failure, for example, an overexpression of K⁺ channels can be used to antagonize acquired long-QT syndrome(62,63), and the attendant loss of contractility may be amenable to co-administration of a second gene to augment calcium cycling, in a strategy of dual gene therapy.  Although such work is conceptually attractive, widespread delivery with long-term expression will be required before human trials can be anticipated.

   More appealing targets for development in the short term are arrhythmias in which very local modifications of electrical properties are sufficient for effective treatment.  An example of such an arrhythmia is atrial fibrillation, which results from the rapid and uncoordinated firing of electrical impulses from several sites in the atria (Nature 2002; 415:219-226).  Many of these impulses then travel to the ventricles, resulting in irregular, erratic and rapid heart rhythm.  Atrial fibrillation affects more than two million people in the United States alone.  Initially, treatment is directed at maintaining normal sinus rhythm, but this is rarely successful in the long term.  When atrial fibrillation becomes persistent, therapy focuses on controlling the ventricular rate.  Such rate control can be achieved by local modification of electrical conduction in the atrioventricular node, because the atrioventricular node is the only pathway for conducting the electrical impulse from the rapidly activating atria to the ventricles.  Donahue et al.(64) have shown the feasibility of atrioventricular nodal modification by gene therapy in an animal model of atrial fibrillation. By locally overexpressing an inhibitory G protein (Gi) subunit, these investigators could slow atrioventricular conduction without affecting other electrical parameters.  The therapy resulted in a salutary reduction of the ventricular rate during atrial fibrillation.  Local gene delivery was highly enriched by selective infusion of the transgenes into the branch of the coronary arteries that supplies the atrioventricular node, and accomplished percutaneously using clinically available catheters.

   There are singular advantages of gene therapy for atrial fibrillation.  First, highly localized gene delivery is sufficient to treat the problem.  The amount of gene delivered can be reduced correspondingly, and potential problems owing to widespread dissemination can be averted more readily.  Second, treated cells remain responsive to endogenous nerves and hormones.  Such was the case with overexpression of G1 in the atrioventricular node: atrioventricular conduction remained responsive to adrenergic stimulation.  Third, implanted hardware is avoided, obviating long-term risks and the expense and morbidity associated with battery and lead replacements.  Fourth, the localized coronary circulation allows isolated delivery to the atrioventricular node.  Fifth, the proximity to the inner lining of the heart, the endocardium, allows access by intracardiac injection, providing a potential alternative delivery route.  Sixth, the therapeutic effects can be readily detected by electrocardio graphy.  Last, changes induced by gene transfer can be rescued by conventional electrophysiological methods (atrioventricular node ablation and pacemaker implantation).

   Similar advantages of gene therapy may be exploitable in the creation of genetically engineered pacemakers; such a possibility might provide the first viable alternative to electronic pacemakers for the treatment of bradyarrhythmias.  Thus, many arrhythmias may turn out to be reasonable targets for functional re-engineering by gene transfer.  This concept, and several others reviewed above - in silico modelling approaches, tailored drug therapy and pharmacogenomics - give ample reason to hope that channelopathies will not only be better understood in the future, but will also become increasingly amenable to rational therapy. ¨

Reprinted by permission. Nature 2002;415:219-226. 2002 Macmillan Publishers Ltd.

References

1.  Catterall, W. A. Molecular properties of sodium and calcium channels. J. Bioenerg. Biomembr.      1996; 28:219-230.

2.  Jan, L. Y. & Jan, Y. N. Voltage-gated and inwardly rectifying potassium channels. J. Physiol.      1997; 505:267-282.

3.  Philipson, K. D. & Nicoll, D. A. Sodium-calcium exchange: a molecular perspective. Annu. Rev.      Physiol. 2000; 62:111-133.

4.  Marban, E., Yamagishi, T. & Tomaselli, G. F. Structure and function of voltage-gated sodium      channels. J. Physiol. 1998; 508:647-657.

5.  Zeng, J. & Rudy, Y. Early afterdepolarizations in cardiac myocytes: mechanism and rate      dependence. Biophys. J. 1995; 68:949-964.

6.  Robbins, J. KCNQ channels: physiology, pathophysiology, and pharmacology. Pharmacol. Ther.      2001; 90:1-19.

7.  Nubauer, M., Beuckelmann, D. J. & Erdmann, E. Characteristics of transient outward current in      human ventricular myocytes from patients with terminal heart failure. Circ. Res. 1993;      73:386-394.

8.  Hoppe, U. C. et al. Manipulation of cellular excitability by cell fusion: effects of rapid       introduction of transient outward K+ current on the guinea pig action potential. Circ. Res. 1999;       84:964-972

9.  Papp, Z. et al. Two components of [Ca2+]i-activated Cl- current during large [Ca2+]i transients      in single rabbit heart Purkinje cells. J. Physiol. 1995; 483:319-330.

10. Koster, O. F., Szigeti, G. P. & Beuckelmann, D. J. Characterization of a [Ca2+]i-dependent      current in human atrial and ventricular cardiomyocytes in the absence of Na+ and K+.       Cardiovasc. Res. 1999:41;175-187.

11. Kääb, S. et al. Molecular basis of transient outward potassium current downregulation in human      heart failure: a decrease in Kv4.3 mRNA correlates with a reduction in current density.      Circulation 1998; 98:1383-1393.

12. Näbauer, M. et al. Regional differences in current density and rate-dependent properties of the      transient outward current in subepicardial and subendocardial myocytes of human left ventricle.      Circulation 1996; 93:168-177.

13. Hoppe, U. C., Marban, E. & Johns, D. C. Molecular dissection of cardiac repolarization by in      vivo Kv4.3 gene transfer. J. Clin. Invest. 2000; 105:1077-1084.

14. Winslow, R. L., Rice, J. & Jafri, S. Modeling the cellular basis of altered excitation-contraction      coupling in heart failure. Prog. Biophys. Mol. Biol. 1998; 69:497-514.

15. Keating, M. T. & Sanguinetti, M. C. Molecular and cellular mechanisms of cardiac arrhythmias.      Cell 2001; 104; 569-580.

16. Yue, D. T. & Marban, E. A novel cardiac potassium channel that is active and conductive at      depolarized potentials. Pflugers Arch. 1988; 413:127-133.

17. Patel, A. J., Lazdunski, M. & Honore, E. Lipid and mechano-gated 2P domain K+ channels.      Curr. Opin. Cell Biol. 2001; 13:422-428.

18. Kim, D. et al. Cloning and functional expression of a novel cardiac two-pore background K+      channel (cTBAK-1). Circ. Res. 1998; 82:513-518.

19. Luo, C. H. & Rudy, Y. A dynamic model of the cardiac ventricular action potential. I.      Simulations of ionic currents and concentration changes. Circ. Res.1994; 74:1071-1096.

20. Weidmann, S. The electrical constants of Purkinje fibres. J. Physiol. 1952; 118: 348-360.

21. Nichols, C. G., Ripoll, C. & Lederer, W. J. ATP-sensitive potassium channel modulation of the      guinea pig ventricular action potential and contraction. Circ. Res. 1991; 68:280-287.

22. Marban, E., Robinson, S. W. & Wier, W. G. Mechanisms of arrhythmogenic delayed and early      afterdepolarizations in ferret ventricular muscle. J. Clin. Invest.1986;78:1185-1192.

23. Cranefield, P. F. & Aronson, R. S. Torsades de pointes and early afterdepolarizations.      Cardiovasc. Drugs Ther. 1991; 5:531-537.

24. Saffitz, J. E., Laing, J. G. & Yamada, K. A. Connexin expression and turnover: implications for      cardiac excitability. Circ. Res.2000;86: 723-728.

25. Berenfeld, O. & Jalife, J. Purkinje-muscle reentry as a mechanism of polymorphic ventricular      arrhythmias in a 3-dimensional model of the ventricles. Circ. Res.1998;82:1063-1077.

26. Hoffman, B. F. & Rosen, M. R. Cellular mechanisms for cardiac arrhythmias. Circ. Res.      1981;49:1-15.

27. Samie, F. H. & Jalife, J. Mechanisms underlying ventricular tachycardia and its transition to      ventricular fibrillation in the structurally normal heart. Cardiovasc. Res.2001;50: 242-250.

28. Camm, A. J. et al. Congenital and acquired long QT syndrome. Eur. Heart J.2000;21:1232-      1237.

29. Wang, Q. et al. SCN5A mutations associated with an inherited cardiac arrhythmia, long QT      syndrome. Cell.1995;80: 805-811.

30. Sanguinetti, M. C. Dysfunction of delayed rectifier potassium channels in an inherited cardiac      arrhythmia. Ann. NY Acad. Sci. 1999;868:406-413.

31. Dumaine, R. et al. Multiple mechanisms of Na+ channel-linked long-QT syndrome. Circ.      Res.1996;78: 916-924.

32. Hoppe, U. C., Marban, E. & Johns, D. C. Distinct gene-specific mechanisms of arrhythmia      revealed by cardiac gene transfer of two long QT disease genes, HERG and KCNE1. Proc. Natl      Acad. Sci. USA 2001;98: 5335-5340.

33. Splawski, I. et al. Spectrum of mutations in long-QT syndrome genes. KVLQT1, HERG,      SCN5A, KCNE1, and KCNE2. Circulation 2000;102:1178-1185.

34. Schwartz, P. J. et al. Genotype-phenotype correlation in the long-QT syndrome: gene-specific      triggers for life-threatening arrhythmias. Circulation.2001;103: 89-95.

35. Plaster, N. M. et al. Mutations in Kir2.1 cause the developmental and episodic electrical      phenotypes of Andersen’s syndrome. Cell.2001;105: 511-519.

36. Chen, Q. et al. Genetic basis and molecular mechanism for idiopathic ventricular fibrillation.      Nature.1998;392: 293-296.

37. Antzelevitch, C. The Brugada syndrome: ionic basis and arrhythmia mechanisms. J.      Cardiovasc. Electrophysiol.2001;12: 268-272.

38. Rook, M. B. et al. Human SCN5A gene mutations alter cardiac sodium channel kinetics and      are associated with the Brugada syndrome. Cardiovasc. Res. 1999;44:507-517.

39. Dumaine, R. et al. Ionic mechanisms responsible for the electrocardiographic phenotype of the      Brugada syndrome are temperature dependent. Circ. Res.1999;85: 803-809.

40. Wang, Y. & Rudy, Y. Action potential propagation in inhomogeneous cardiac tissue: safety      factor considerations and ionic mechanism. Am. J. Physiol. Heart Circ. Physiol.2000;278:      H1019-H1029.

41. Schott, J. J. et al. Cardiac conduction defects associate with mutations in SCN5A. Nature      Genet. 1999;23:20-21.

42. Tan, H. L. et al. A sodium-channel mutation causes isolated cardiac conduction disease.      Nature. 2001; 409: 1043-1047.

43. Veldkamp, M. W. et al. Two distinct congenital arrhythmias evoked by a multidysfunctional       Na+ channel. Circ. Res.2000;86: E91-E97.

44. Marban, E. Heart failure: the electrophysiologic connection. J. Cardiovasc. Electrophysiol.       1999; 10:1425-1428.

45. Beuckelmann, D. J., Näbauer, M. & Erdmann, E. Alterations of K+ currents in isolated human      ventricular myocytes from patients with terminal heart failure. Circ. Res.1993;73: 379-385.

46. Kääb, S. et al. Ionic mechanism of action potential prolongation in ventricular myocytes from      dogs with pacing-induced heart failure. Circ. Res. 1996;78: 262-273.

47. Berger, R. D. et al. Beat-to-beat QT interval variability: novel evidence for repolarization lability      in ischemic and nonischemic dilated cardiomyopathy. Circulation 1997;96:1557-1565.

48. Studer, R. et al. Gene expression of the cardiac Na+-Ca2+ exchanger in end-stage human       heart failure. Circ. Res. 1994;75:443-453.

49. Winslow, R. L. et al. Mechanisms of altered excitation-contraction coupling in canine      tachycardia-induced heart failure, II: model studies. Circ. Res. 1999; 84:571-586.

50. Sanguinetti, M. C. et al. A mechanistic link between an inherited and an acquired cardiac      arrhythmia: HERG encodes the IKr potassium channel. Cell.1995:81;299-307.

51. Mitcheson, J. S. et al. A structural basis for drug-induced long QT syndrome. Proc. Natl Acad.      Sci. USA 2000; 97:12329-12333.

52. Liu, X. K. et al. Female gender is a risk factor for torsades de pointes in an in vitro animal      model. J. Cardiovasc. Pharmacol. 1999; 34:287-294.

53. Roden, D. M. & Spooner, P. M. Inherited long QT syndromes: a paradigm for understanding      arrhythmogenesis. J. Cardiovasc. Electrophysiol. 1999; 10:1664-1683.

54. Roden, D. M. Pharmacogenetics and drug-induced arrhythmias. Cardiovasc. Res. 2001:50;      224-231.

55. Abbott, G. W. et al. MiRP1 forms IKr potassium channels with HERG and is associated with      cardiac arrhythmia. Cell 1999; 97:175-187.

56. Mazhari, R. et al. Molecular interactions between two long-QT syndrome gene products, HERG      and KCNE2, rationalized by in vitro and in silico analysis. Circ. Res. 2001;89:33-38.

57. Priori, S. G. et al. Genetic and molecular basis of cardiac arrhythmias: impact on clinical      management parts I and II. Circulation.1999;99: 518-528.

58. Benhorin, J. et al. Effects of flecainide in patients with new SCN5A mutation: mutation- specific      therapy for long-QT syndrome? Circulation.2000; 101:1698-1706.

59. Windle, J. R. et al. Normalization of ventricular repolarization with flecainide in long QT      syndrome patients with SCN5A:DeltaKPQ mutation. Ann. Noninvasive Electrocardiol. 2001;      6:153-158.

60. Compton, S. J. et al. Genetically defined therapy of inherited long-QT syndrome. Correction of      abnormal repolarization by potassium. Circulation. 1996; 94:1018-1022.

61. Shimizu, W. et al. Improvement of repolarization abnormalities by a K+ channel opener in the      LQT1 form of congenital long-QT syndrome. Circulation.1998; 97:1581-1588.

62. Nuss, H. B. et al. Reversal of potassium channel deficiency in cells from failing hearts by      adenoviral gene transfer: a prototype for gene therapy for disorders of cardiac excitability and      contractility. Gene Ther. 1996; 3:900-912.

63. Nuss, H. B., Marban, E. & Johns, D. C. Overexpression of a human potassium channel      suppresses cardiac hyperexcitability in rabbit ventricular myocytes. J. Clin. Invest.1999; 103:      889-896 (1999)

64. Donahue, J. K. et al. Focal modification of electrical conduction in the heart by viral gene      transfer. Nature Med.2000; 6: 1395-1398.