Electrophysiological Remodeling in Heart Failure

Abstract

Heart failure affects nearly 6 million Americans, with a half-million new cases emerging each year. Whereas up to 50% of heart failure patients die of arrhythmia, the diverse mechanisms underlying heart failure-associated arrhythmia are poorly understood. As a consequence, effectiveness of antiarrhythmic pharmacotherapy remains elusive. Here, we review recent advances in our understanding of heart failure-associated molecular events impacting the electrical function of the myocardium. We approach this from an anatomical standpoint, summarizing recent insights gleaned from pre-clinical models and discussing their relevance to human heart failure.

Introduction

Heart failure is approaching epidemic proportions in industrialized societies[1]. Mortality in this syndrome, which derives from a host of disease-associated insults to the heart, is high, as much as 50% per year in advanced disease. A prominent mechanism of death in patients with heart failure is arrhythmia, especially tachyarrhythmia, where electrical activation of the heart occurs so rapidly that effective filling and pumping of blood cannot occur. To combat these arrhythmias, implantable devices (ICDs, implantable cardiodefibrillators) have emerged as an important therapeutic strategy designed to abort malignant arrhythmia and restore normal, sinus rhythm.

In parallel with the emergence of ICD therapy, the prominence of anti-arrhythmic drug therapy has declined. The reasons for this decline are several, but stem largely from inefficacy of existing antiarrhythmic pharmacotherapy in patients with structural heart disease. This deficiency, in turn, likely derives from inadequate understanding of fundamental mechanisms of arrhythmogenesis in the large number of disease states which culminate in heart failure.

A great deal of work is underway to decipher mechanisms of pathological remodeling of the ventricle[2]. Of necessity, the majority of this mechanistic work is conducted in preclinical models, both in vitro and in vivo. This fact, then, raises an important question: how relevant to clinical arrhythmias in humans are insights gleaned from cell culture or animal models of disease? This question is all the more relevant given that the fundamental electrophysiological anatomy of cardiac myocytes differs among mammals, from mouse to humans; differences in the electrical properties of cardiac myocytes between humans and non-mammalian species are yet more significant. How reliable are insights gleaned from preclinical models in identifying new therapeutic targets in humans? Can we extrapolate discoveries made in nonhuman models of electrical disease to our patients in the clinic? As a step toward addressing these near-intractable questions, we summarize here recent advances in preclinical studies of arrhythmia and discuss relevance to human heart disease. We address this topic from an anatomical standpoint, tracking electrical events as they course through the myocardium, and discussing heart failure-related remodeling events occurring in each tissue. As this field is vast, we do not discuss all of the major channels and transporters; for detailed discussion channels and currents not covered in this review, readers are referred to recent excellent reviews[3-5].

Normal electrical activation of the heart

Electrical activation of the heart normally commences at the sinoatrial (SA) node, a bundle of specialized cells in the right atrium that fire action potentials spontaneously and automatically (Figure 1). Signals arising in the SA node spread across the right and left atria, stimulating them to contract. Excitation ultimately travels to the atrioventricular (AV) node, a bundle of specialized cells between the atria and ventricles serving as an “electrical bridge” between these two tissues. After a delay, the electrical signal emerges on the ventricular side of the AV node and is conducted through specialized electrical cables (His bundle and Purkinje fibers) to rapidly spread across the endocardial surfaces of both ventricles. Just beyond the bundle of His, the cable splits into two branches, the left bundle branch and the right bundle branch. The left bundle branch is short, dividing into the left anterior fascicle and the left posterior fascicle. The left posterior fascicle is broad and fan-shaped with dual blood supply, making it relatively resistant to ischemic damage. Electrical excitation then enters the contractile myocardium, a functional syncytium where each cell is electrically coupled to its neighbors via gap junctions. As a result, electrical impulses propagate freely between cells in every direction, so that the myocardium functions as a single contractile unit. As electrical activation spreads rapidly across the endocardial surfaces of both ventricles via the His-Purkinje system, mechanical contraction occurs synchronously and efficiently.

Figure 1. Cardiac conduction system.

(© Carolina Biological Supply Company, Used with permission)

Throughout the myocardium, electrically activated cells return to their basal, resting state as their action potentials repolarize. In fact, resetting of electrical excitability in heart tracks the course of excitation with variable delay. One major exception to this rule occurs in the ventricle, where repolarization occurs first in the outer, sub-epicardial zone, even though these cells are the last to be excited. Disorders in the temporal and spatial patterns of repolarization are a major mechanism underlying heart failure-associated ventricular arrhythmias.

Heart failure-related electrophysiological remodeling

Nodal cells and specialized conduction system

• SA node

The SA node is a compact region within the right atrium located at the base of the superior vena cava[6, 7]. In the absence of extrinsic neural and hormonal control, cells in the SA node will naturally generate action potentials at a rate of approximately 100 times/minute. Rhythms driven by SA node pacemaker activity (“normal sinus rhythm”) occur at 60-100 beats per minute.

Within the SA node, a pacemaker current (I_f) generates spontaneous depolarization during diastole and consequent spontaneous pacemaker activity[8, 9]. The resulting electrical activity in several thousand cells is integrated via tight cell-cell coupling leading to nearly synchronous firing of action potentials[10]. Within the SA node, I_f manifests morphology-dependent variation, with a smaller I_f in spindle-like cells as compared with spider-like cells[11]. Cell size-dependent variation has been described[7], as well as age-related variation, with smaller I_f in adult compared with newborn heart[12]. It is estimated that only about 1% of the cells in the SA node (“P” cells localized to the center of the node) actually function as leading pacemaker cells[13]. Outside the center of the node, there is a gradual transition in cell type over several millimeters reaching the periphery of the node. Perinodal cells, also called transitional (T) cells, transmit the electrical impulse from the SA node to the right atrium. SA nodal dysfunction may result from abnormalities in either impulse generation by the P cells or in conduction across the T cells.

In SA node, the major currents generating phase 0 upstroke of the action potential are the L-type Ca²⁺ current, I_Ca, and sodium current, I_Na. I_Na is responsible for the action potential upstroke in SA node periphery cells, whereas I_Ca is responsible in the central cells[7]. In addition to generating the rapid upstroke, I_Ca is an important contributor to late stage phase 3 pacemaker depolarization. Thus, I_Ca appears to be important in determining both the action potential waveform and the slope of pacemaker depolarization[14].

Another distinct Ca²⁺ current present in mammalian SA node is the transient Ca²⁺ current, i.e. T-type Ca²⁺ current. Its basic properties include a more negative threshold potential than I_Ca and a more rapid rate of inactivation[15]. Although its amplitude is much smaller than I_Ca, its activation threshold lies near diastolic potentials and thus has been hypothesized to contribute to pacemaker activity[16]. Consistent with this, Hagiwara et al. have shown in rabbit SA node cells that abolishing T-type Ca²⁺ current with Ni²⁺ leads to a 14% increase in cycle length[15]. We now know that there are three T-type channel isoforms Ca_V3.1–3.3 (α_1G, α_1H and α_1I), two of which (Ca_V3.1 and Ca_V3.2) are expressed in the heart[17]. In contrast with Ca²⁺ current, I_Na has a negligible effect on SA node cycle length[14].

Abnormalities in sinoatrial node function are seen frequently in clinical and experimental heart failure[18, 19]. For example, patients with heart failure can present with bradycardia[20], and animal studies have pointed to increases in the intrinsic cycle length of the SA node with an amplified response to acetylcholine as a contributing mechanism[18, 21]. In a rabbit model of combined volume and pressure overload-induced heart failure, I_f was reduced significantly, accompanied by a decrease in the slow component of the delayed rectifier current (I_Ks). By contrast, T-type and L-type Ca²⁺ current, rapid and ultrarapid delayed rectifier K⁺ current (I_Kur), transient outward K⁺ current (I_to), and sodium-calcium exchange current (I_NCX) are unaltered[22], and HF-associated decreases in heart rate were attributed to remodeling of I_f. In addition to chronic reduction in heart rate[23], SA nodal function is altered in patients with persistent atrial fibrillation (AF)[24], a condition commonly seen in heart failure. Indeed, even transient episodes of atrial tachyarrhythmias[25] which occur commonly in heart failure, induce significant electrical remodeling in the SA node and consequent nodal dysfunction.

At a molecular level, chronic reductions in heart rate, as can be seen in heart failure, lead to declines in the transcript abundance of several ion channel subunits, including the voltage-gated Na⁺ channel β-subunit Navβ1 (-25%), T-type Ca²⁺ channel subunit Cav3.1 (-29%), K⁺-ATP channel α-subunits Kir6.1 (-28%), and K⁺ channel regulatory β-subunits Kvβ2 (-41%) and Kvβ3 (-30%)[23]. By contrast, some ion channel genes are up-regulated, including K⁺ channel α-subunits (Kv1.1, +30%; Kir2.1, +29%; Kir3.1, +41%), hyperpolarization-activated cation channels (HCN2, +24%; HCN4, +52%), and the gap junction channel connexin 43 (+26%)[23]. Down-regulation of HCN4 and HCN2 expression, however, was also reported in sinus node of canine ventricular tachypacing-induced HF[26]. Short bursts of rapid atrial pacing (10-15 minutes), simulating transient atrial tachyarrhythmias, alter sinus node function in humans, leading to increases in both sinoatrial conduction time (SACT) and corrected sinus node recovery time (CS-NRT)[25]. After electrical cardioversion of longstanding (>3 months) AF, CS-NRT was found to remain significantly increased, whereas impairments of sinus node automaticity were reversible[27].

• AV node

The atrioventricular (AV) node is an electrical relay station connecting the atria and ventricles. Conduction through the AV node is slow, affording a time delay that allows the ventricles to fill with blood that was mechanically pumped during atrial systole. The AV junctional tissue has its own intrinsic pacemaker activity at 40-60 beats per minute, such that the AV node can assume the role of cardiac pacemaker in the setting of SA node dysfunction. By virtue of the fact that the AV node transmits electrical impulses slowly, it limits the number of impulses conducted from the atria to the ventricles. This function is important during fast atrial rates (e.g. atrial flutter or fibrillation) where only a fraction of impulses are successfully conducted to the ventricles and mechanical performance of the ventricles is thereby preserved.

The AV node and perinodal area comprises at least three electrophysiologically distinct cell types: the atrionodal (AN), the nodal (N), and the nodal-His (NH) cells. In the N cells, an inward calcium channel current (I_Ca) is the basis of the action potential upstroke[28], whereas the upstroke of the action potential in AN and NH cells depends mainly on an inward sodium current (I_Na). Because I_Na in AN and NH cells is larger and has more rapid kinetics than I_Ca in N cells, conduction is faster through atrial transitional and NH regions than through the core of the AV node. Fast pathway conduction through the AV node bypasses many of the N cells by traversing transitional cells, whereas slow pathway conduction passes through the entire compact AV node[29].

Adrenergic activation is a hallmark feature of heart failure[30]. In the AV node, sympathetic nerve activation accelerates pacemaker function (due to increases in I_Ca), enhances excitability, heightens action potential amplitude, increases conduction velocity, and decreases the effective refractory period of AV nodal cells, which together facilitate both antegrade slow conduction and retrograde fast conduction through the AV node[31]. In failing human ventricle, adrenergic over-activation is accompanied by down-regulation of β1-adrenergic receptors with preservation of β2-adrenoceptors, which shifts the β1:β2 ratio towards the β2-subtype. And the resulting relative over-activation of β2-subtype receptors may play opposing roles in the failing heart, viz. maintaining cardiac contractility via activation of G_s with consequent increases in cAMP formation, or inhibition of contractility via enhanced coupling to G_i and thus decreased adenylyl cyclase activity[21, 32]. Whether a shift of the β1:β2 ratio also occurs in AV node in the setting of heart failure is not known.

Although β-adrenergic antagonists are effective in regulating AV nodal conduction in many patients, enhanced nodal conduction remains a significant issue in patients with atrial fibrillation and heart failure. In addition to the PKA/cAMP pathway, recent studies suggest that CaMKII, an enzyme which is significantly activated in heart failure, may play an important role in regulating heart rate and electrical conductivity. Khoo el al. reported that in a model of genetic CaMKII inhibition by cardiac-specific expression of autocamtide 3 inhibitory peptide (AC3-I), PR and AH intervals were significantly prolonged, manifesting enhanced Wenckebach-type conduction block[33]. CaMKII activity has also been implicated in electrical activity within the SA node via regulation of I_Ca[34] and I_f[35]. In line with these findings, our recent studies of CaMKIIδ knockout mice reveal heart rate declines in the setting of β-adrenergic stimulation or increased workload (unpublished results). Further, Wu et al. demonstrated that adrenergic regulation of heart rate depends on CaMKII-mediated regulation of SA nodal Ca²⁺ homeostasis; CaMKII inhibition had no effect on the isoproterenol response in SA nodal cells when SR Ca²⁺ release was disabled, and CaMKII inhibition was only effective at slowing heart rates during β-adrenergic stimulation[36]. Together, these observations raise important questions regarding whether CaMKII activity plays a role in failure-related alterations in heart rate and electrical conduction and highlight yet further the prospect of CaMKII as a therapeutic target in heart disease.

• Purkinje cells

In 1839, Jan Evangelista Purkinje described a network of fibers located on the endocardial surface of the left and right ventricles[37]. These fibers are specialized cables that conduct electrical impulses rapidly over the endocardial surface, thereby enabling coordinated mechanical activation of both ventricles. Purkinje myocytes have a higher density of sodium channels and mitochondria, and they contain fewer myofibrils than surrounding contractile myocardium. Transmission of electrical impulses to ventricular myocytes occurs by cell-to-cell communication via gap junctions. The Purkinje conducting system is capable of intrinsic pacemaker activity at a rate of 30-40 impulses per minute. If the SA and AV nodes are injured, the ventricular Purkinje conducting system can assume control of heart rate and rhythm.

Cardiac Purkinje cells play an important role in the generation of ventricular arrhythmias, particularly those related to triggered activity[38, 39]. In regions of myocardial infarction, ionic currents are significantly altered in subendocardial Purkinje cells[40-42]. In heart failure, the risk of drug-induced torsades de pointes is increased significantly, a phenomenon which is associated with the remodeling of ionic currents in cardiac Purkinje cells[43]. In rapid pacing-induced heart failure, I_to density in Purkinje cells is reduced without change in its voltage dependence or kinetics[44]. Heart failure is also associated with reduced inward-rectifier current density without changes in current-voltage relationship. In heart failure, densities of L- and T-type calcium currents, rapid and slow delayed rectifier K⁺ currents, and Na⁺-Ca²⁺ exchange currents are typically unaltered, but inactivation of I_Ca,L is slowed at positive potentials. Purkinje fiber action potentials from heart failure dogs manifest decreased phase 1 amplitudes and elevated plateau voltages, and they demonstrate twice as much prolongation on exposure to the rapid delayed rectifier K⁺ channel blocker E-4031 as control Purkinje fibers[44]. Down-regulation of I_to, along with slowed inactivation of I_Ca,L, is likely responsible for the positive shift in the plateau voltage of cardiac Purkinje cells. Together, these heart failure-induced ion channel changes in cardiac Purkinje cells may promote arrhythmogenic afterdepolarizations. In fact, slowed I_Ca,L inactivation would be expected to promote I_Ca,L-dependent early afterdepolarizations under conditions of delayed repolarization.

Contractile myocardium

• Atrial tissue

Atrial fibrillation (AF), the most common arrhythmia in heart failure, contributes substantially to morbidity and mortality in this syndrome[1]. AF is an independent risk factor for stroke and evidence suggests that atrial fibrillation promotes progression of heart failure[45]. AF tends to be self-maintaining by creating an electrophysiological substrate that facilitates AF, a concept which has been coined as “AF begets AF”[46]. Indeed, remodeling of atrial myocytes is a major contributor to the increased propensity to AF in patients with heart failure[47]. To date, therapy with antiarrhythmic drugs to maintain sinus rhythm has been disappointing, and there is concern that untoward side effects could offset the benefits of sinus rhythm maintenance. Thus, ventricular rate control and anticoagulation are recommended as mainstays of therapy for persistent AF[48].

In some patients, paroxysmal AF initiates from automatic foci localized within sleeves of musculature extending retrograde into the pulmonary veins. Indeed, catheter-based ablation of AF, often by electrically isolating the pulmonary veins, has emerged as a prominent treatment strategy[49] As a general rule, ablation is more effective for paroxysmal AF than for chronic AF[49], consistent with the concept that AF triggers remodeling events that promote its persistence. Presently, it is believed that changes in atrial electrophysiologic substrate, beyond structural heart disease, contribute to AF recurrence and/or perpetuation. Thus, a long-term objective of AF therapy is to prevent the initial development of arrhythmia-related ion channel remodeling and tissue fibrosis.

It was well established that atrial remodeling is a complex process that develops over time[4]. The type and extent of remodeling depends on the strength and duration of the stressor. The most common stressors of atrial myocytes are electrical (tachycardia, AF) or mechanical (volume or pressure overload). Some evidence suggests that the ionic remodeling that occurs in tachycardia-dependent disease is distinct from that occurring in heart failure[50, 51]

Ionic remodeling occurs early during AF. Loss of heart rate-dependent adaptation of atrial refractoriness, short atrial refractoriness, and AF duration have each been suggested to be a predictor of AF recurrence[52]. Decreases in I_Ca contribute to the shortening of the atrial action potential, whereas decreases in I_to likely contribute to loss of physiological rate adaptation of the action potential[53, 54]. These findings from preclinical models have been largely confirmed in humans[55, 56]. Ironically, declines in I_Ca could be viewed as an adaptive alteration that antagonizes rate-induced calcium overload[57]. Reduced Na⁺ current (I_Na) density has also been observed in a canine model of atrial fibrillation[58], which may contribute to declines in conduction velocity,

In one report, decreased density of L-type Ca²⁺ channels was consistently observed following long-term (several weeks) rapid atrial pacing, whereas inward T-type Ca²⁺ current remained unchanged[53]. This, combined with recent studies demonstrating that Ca²⁺ influx through T- and L-type channels have different effects on contractility and intracellular signaling [59], suggests that in chronic AF, T-type Ca²⁺ current may be a significant contributor to intracellular Ca²⁺ overload. Consistent with this, Lee et al[60] reported that verapamil attenuated shortening of refractoriness after 1-day of pacing but did not alter electrophysiological remodeling long term (1 and 6 weeks). Fareh et al[61, 62] reported that the T-type Ca²⁺ channel blocker, mibefradil, is superior to L-type Ca²⁺ channel blockers in attenuating tachycardia-dependent remodeling. Based on these findings, one could speculate that patients with paroxysmal AF may benefit from L-type Ca²⁺ blocker therapy, whereas the long-term remodeling which occurs in persistent AF might respond to T-type Ca²⁺ channel blockers.

Atrial remodeling in heart failure is difficult to study, owing to complex interplay between diseased tissue and disease-associated neurohumoral activation. In a canine model of pacing-induced heart failure, decreases in I_Ca,L, I_to, and I_Ks, and an increase in I_NCX were observed in atrial myocytes[47]. In atrial myocytes isolated from heart failure patients, by contrast, we have reported that I_to density is significantly larger with an accelerated time course of recovery[63]. In experimental heart failure, attenuated rate adaptation of atrial refractoriness (long action potential duration at high rates) is observed[53]. Similar findings have been reported in human heart failure[64]. Unlike the “AF begets AF” model, in which a decrease in effective refractory period (ERP) is critical to the development of sustained AF, in atrial myocytes of failing human heart, no difference or even an increase in ERP has been observed[65], and there is no significant change in action potential duration[63].

These differences may result from different remodeling mechanisms occurring between atrial tachycardia/AF and heart failure. For example, Yue et al reported that in a canine model, heart failure decreased both I_Ks and I_Ca,L by about 30%, whereas atrial tachycardia decreased I_Ca,L by 70% without affecting I_K[53]. Animal studies also suggest that electrical remodeling is reversible in both atrial tachycardia/AF and heart failure when the stressor is removed[66]. However, in both atrial tachycardia-dependent or in heart failure-dependent remodeling, irreversible changes can occur. For example, evidence suggests that aspects of structural remodeling, including interstitial fibrosis, cellular hypertrophy, and degeneration which develop concurrently with electrical alterations can be irreversible[66]. Fibrosis has been detected in the atria of patients with heart failure due to prior myocardial infarction or idiopathic heart failure[67-69] and in an animal model of ventricular pacing-induced HF[70]. Together, these irreversible changes may contribute to reduced efficacy of medical therapy for chronic AF in patients with heart failure.

• Ventricular tissue

Relative to atrial myocytes, electrical remodeling in failing ventricular myocytes has been intensively studied, and several excellent reviews are available[3-5, 71, 72]. Here, we focus on recent progress in a limited number of areas.

Cellular remodeling

Prolongation of the ventricular action potential is a hallmark of heart failure. Although many discrepancies exist regarding the specific ionic and molecular processes occurring in heart failure, studies in animal models and in humans with heart failure have consistently revealed action potential duration (APD) prolongation due to functional down-regulation of outward potassium currents[73, 74], functional up-regulation of inward calcium current and changes in Ca²⁺ current inactivation[75-77], or increases in late sodium currents[78]. APD prolongation may prolong calcium channel opening and thereby contribute to preservation of contractile force. However, it also increases the risk of Ca²⁺ overload which could contribute to abnormal triggered impulses and perturbed signaling events.

Although all myocardial cell layers exhibit significant APD prolongation in heart failure, such prolongation is typically heterogeneous. In a mouse model of pressure-overload heart failure, we found that APD was more prolonged in subepicardial myocytes than in subendocardial myocytes due to the more significant reduction of outward potassium currents[79]. In a canine model of rapid ventricular pacing-induced heart failure, APD prolongation of midmyocardial (M) cells was substantially greater than epicardial cells, eliciting a significant increase in the transmural APD gradient and consequent increases in the transmural dispersion of repolarization[80]. The magnitude of APD prolongation in M cells was found to be similar to QT interval prolongation, suggesting that it may contribute to QT interval changes observed in heart failure[80]. However, another study showed that, even though APD was increased, transmural dispersion of APD was diminished in pacing-induced heart failure relative to that seen in controls[81]. Although changes in transmural APD vary across species, leading to either an increase or a decrease in transmural dispersion, the heterogeneous and selective prolongation of repolarization between cell types across the ventricular wall underlies an electrophysiological mechanism for unidirectional block, reentry, and arrhythmogenicity.

• Ionic mechanisms underlying ventricular electrical remodeling

1) Sodium current

The cardiac voltage-dependent Na channel (SCN5A or NaV1.5) generates a large inward current (I_Na) within the first millisecond of excitation and thereby mediates rapid membrane depolarization (the upstroke of the cardiomyocyte action potential). As a result, it is a major determinant of conduction velocity in the atria and ventricles. Inactivation of I_Na has both fast (fast inactivation, which lasts several milliseconds) and slow (which can last hundreds of milliseconds) components[82]. The current associated with the slow inactivating phase has been referred to as late, sustained, or persistent I_Na (I_Na,L, I_Na,sus, I_Na,p) to distinguish it from the peak transient I_Na. Some evidence suggests I_Na,L arises from a fraction of the Na⁺ channels that intermittently lose their inactivation[83], as opposed to a distinct set of channel molecules.

The amplitude of late I_Na is less than 1% that of peak I_Na, but it is sufficient to prolong action potential duration[84]. In animal models of heart failure, peak I_Na has been reported to be decreased[85], unchanged [73], or increased[78, 86, 87]. Also, peak I_Na has been reported to be increased in human heart failure[88]. Maltsev et al[89] reported that late I_Na is present in ventricular myocytes from normal human mid-myocardium and from failing human hearts, contributing to 15–20% of APD in both species. Whereas peak I_Na density is often decreased in heart failure, I_Na,L as a percentage of peak I_Na can be significantly increased in both conditions (7-12-fold increase)[90]. Coincident with the increased density, I_Na,L decay was found to be slower in failing dog and human hearts[88]. Given this, increased I_Na,L likely plays an important role in APD prolongation in heart failure and may contribute to arrhythmogenesis. I_Na,L inhibition decreased beat-to-beat AP variability and eliminated early afterdepolarizations in failing cardiomyocytes[88]. Whether the changes in I_Na manifest transmural variations in the failing ventricle or contribute to changes in transmural dispersion is unknown.

2) Potassium currents

K⁺ channels are critical to the restoration of cardiac excitability, because they play a fundamental role in repolarization of the action potential. Over 50 genes encoding K⁺ channel subunits have been cloned in man, and many aspects of their biophysical properties, channel assembly and stoichiometry, and functional modulation by second messengers and ligands have been elucidated. Numerous studies have linked K⁺ channel gene mutations with a variety of diseases, and therapeutic approaches to target these channels have been formulated[91, 92]. Here we will focus on a limited set of K⁺ currents which play a major role in myocyte repolarization.

Transient outward current (I_to)

Prominent transient outward K⁺ current (I_to) has been recorded in ventricular myocytes isolated from the hearts of many species, including mice, rats, rabbits, cows, cats, dogs, ferrets and humans[93]. Depending on species and regions within the heart, there are at least two distinct I_to phenotypes that can be distinguished based on their molecular, biophysical and pharmacological properties: I_to,f which is characterized by fast recovery from inactivation, and I_to,s which is characterized by slow recovery from inactivation. I_to,f is mediated by Kv4.2 and/or Kv4.3 channels, whereas I_to,s is mediated by Kv1.4 channels. Their relative contributions to I_to vary between species and across regions of the heart[94]. Across the left ventricular free wall, a gradient of the two distinct I_to phenotypes and a gradient in peak current density of a single I_to phenotype are each present in many species, including humans.

Kv1.4, Kv4.3, and KChIP2 mRNAs are expressed in the human LV[95, 96], and based on the gradients of protein expression, both KChIP2 and Kv4.3 are thought to contribute to I_to,f [97]. However, only reductions in Kv4.2 and Kv4.3 abundances have been linked consistently to the diminished I_to densities typically seen in cardiac hypertrophy[98].

I_to, a major determinant of the early phase of AP repolarization, plays a crucial role in modulating AP plateau and repolarization profiles. In fact, down-regulation in I_to contributes to changes in AP morphology observed in many forms of heart disease[53, 99-101]. Studies designed to determine the contribution of I_to to AP duration, however, have produced conflicting results. In ventricular myocytes from failing heart, I_to was observed to be reduced and was postulated to promote prolongation of APD[73, 102]. However, inhibition of I_to by 4-aminopyridine (4-AP) in the absence of pipette Ca²⁺ buffers shortens APD in ventricular myocytes[103, 104]. Yue et al[53] reported that suppression of I_to could either prolong or shorten APD in canine atrial myocytes, depending on the density of I_Ca,L. We have reported that increased I_to in atrial myocytes from patients with heart failure did not alter APD[63]. One potential explanation for these inconsistencies is that in many common diseases, such as heart failure, myocardial infarction, and atrial fibrillation, down-regulation of I_to is accompanied by alterations in I_Ca,L. The complex interplay among these current alterations results in changes in plateau level, which, in turn, induces secondary changes in activity of the L-type calcium channel, Na⁺/Ca²⁺ exchangers, and other currents[105-107]. In aggregate, these complex changes make alterations in APD difficult to predict.

Although I_to is prominent in human heart, its role in ventricular APD is limited due to the presence of large, delayed rectifier currents. It is clear, however, that reduction in I_to will slow the early phase of AP repolarization (generally characterized by APD₃₀, the duration of time for the action potential to reach 30% repolarization). The slowed early repolarization is critical to the success of AP conduction in cardiac tissue. We first reported the crucial role of I_to in facilitation and maintenance of AP propagation at fast heart rates in rabbit atrial myocytes[108]. In these experiments, inhibition of I_to by fast pacing delayed early repolarization and increased APD₃₀. This lengthening of APD₃₀ allows the proximal (“leader”) cell to provide more coupling current to the distal (“follower”) cell to bring it to its excitation threshold, thereby facilitating AP propagation. In the normal LV, I_to density is larger in subepicardial myocytes than in subendocardial myocytes. This transmural gradient of I_to contributes to the transmural gradient of APD₃₀, which enables subendocardial myocytes to provide more coupling current to depolarize adjacent cells. As such, conduction of excitability is facilitated in the physiological direction of endocardium to epicardium but disfavored in the opposite direction (Figure 2). In heart failure, however, this directional preference of conduction is blunted due to disease-related declines in the transmural gradient of I_to[79]. Together, these changes may be arrhythmogenic, as conduction of abnormal impulses originating in the subepicardium is relatively facilitated. Consistent with this, epicardial-site pacing is associated with VT in patients with heart failure[109, 110]

Figure 2. Anisotropic conduction between subendocardial and subepicardial layers of LV.

In normal heart, impulse conduction is favored in the physiological direction of subendocardium (SEN) to subepicardium (SEP) and relatively disfavored from SEP to SEN. Electrical remodeling in heart failure, including AP prolongation and a diminished transmural gradient of APD30, abolishes directionally preferential conduction.

Delayed rectifier K⁺ current (I_K)

Delayed rectifier K⁺ current (I_K) has different functional components depending on the species. In human heart, I_K can be separated into ultrarapid (I_Kur), rapid (I_Kr) and slow (I_Ks) components. These currents exhibit different kinetics and pharmacological properties, are regulated by different intracellular signaling pathways, and are encoded by separate genes (hKv1.5, hERG, and KCNQ1/KCNE1, respectively)[111]. The delayed rectifier current is primarily responsible for initiating phase 3 repolarization, thereby governing action potential duration and tissue refractoriness.

Defects in I_K underlie certain forms of the hereditary long QT syndrome[112]. A number of I_Kr-blocking drugs can induce excessive action potential prolongation and cause acquired long QT syndrome with consequent increased risk of sudden death[113]. At fast heart rates, I_Kr contributes less to repolarization than I_Ks[114] leading to reverse use-dependence of I_Kr blockers, i.e. these agents manifest their greatest effects on APD at slow heart rates. In contrast, I_Ks blockers might be expected to exert their greatest effects on APD at relatively fast rates, with a reduced propensity for pro-arrhythmia at slow rates. However, some recent studies suggest that I_Ks plays only a limited role in action potential repolarization. For example, blocking I_Ks results in little or no prolongation of APD in normal ventricle, regardless of pacing frequency[115]. Furthermore, I_Ks contributes to repolarization only when the action potential is abnormally prolonged[116]. Thus, I_Ks serves predominantly to prevent excessive APD prolongation, and blocking I_Ks may remove this “safety mechanism” and thereby contribute to triggered arrhythmias.

Relative to I_to, less is known about delayed rectifier K⁺ currents in heart failure. Delayed rectifier K⁺ currents are reduced in ventricular myocytes from pacing-induced heart failure[117] and in models of cardiac hypertrophy[118-121]. However, other studies have reported no changes in delayed rectifier K⁺ current in cardiac hypertrophy[122-124]. In a rapid-pacing-induced HF model in rabbits, I_Ks was reduced, but I_Kr was unchanged[125]. In failing human ventricular myocytes, the delayed rectifier current has been reported to be small and unchanged[99].

ATP-sensitive potassium channels (K⁺-ATP channels)

K⁺-ATP channels play important roles in many cellular functions by coupling cell metabolism to electrical activity. Detected at high density first in cardiac sarcolemma[126], the K⁺-ATP channel is an octamer composed of four Kir6 subunits and 4 sulfonylurea receptor (SUR) subunits. The Kir6 subunits regulate the inwardly rectifying potassium channel pore, whereas the SUR subunits, ATP-binding cassette proteins, serve a regulatory role, modulating channel activity based on cellular ATP levels[127-129]. Heterologous reconstitution experiments have suggested that the cardiac K⁺-ATP channel comprises heteromultimerized Kir6.2 and SUR2A subunits[130].

Adenosine triphosphate (ATP) and nucleotide diphosphates (NDPs) are the major regulators for K⁺-ATP channel activity[131]. ATP inhibits channel activity, and MgADP induces channel opening[132]. In response to metabolic injury with associated declines in ATP and increases in MgADP, K⁺-ATP channels open, reducing cell excitability and protecting the tissue from damage. This occurs by a reduction in action potential duration, which leads to decreased inward Ca²⁺ flux, diminished contraction, and conservation of cellular energy stores.

In HF, poorly defined pathological events lead to disruption of the communication between cellular energetic signals and K⁺-ATP channel function. This breakdown in intracellular signaling results in an inability of the K⁺-ATP channel complex to appropriately recognize cellular metabolic stress and consequent failure to carry out its homeostatic functions[133]. Cardiomyocytes isolated from mice with heart failure fail to demonstrate either K⁺-ATP channel activation in response to cellular metabolic stress or hypoxia-induced shortening of the action potential[133]. Thus, heart failure-related dysregulation of K⁺-ATP channels appears to contribute to the overall pathogenesis of the syndrome.

Because of the heterogeneous effect of APD shortening in the epicardium versus endocardium, and in ischemic versus non-ischemic areas, concerns have been raised that potassium channel activation may further increase dispersion of refractoriness and promote arrhythmias. However, clinical trials have not uncovered increases in life-threatening arrhythmias for any of the K⁺-ATP channel openers[134]. Instead, some clinical studies even reported decreased incidence of malignant ventricular arrhythmias in subjects exposed to K⁺-ATP channel openers[135].

3) Other currents

L-type Ca²⁺ current

The L-type Ca²⁺ current (I_Ca,L) is the primary source of Ca²⁺ entry in cardiac myocytes, triggering release of Ca²⁺ from sarcoplasmic reticular stores, activating a number of Ca²⁺-sensitive signaling cascades, and initiating actin-myosin crossbridge cycling. These channels are comprised of heteromultimers of at least three different subunits (α1c, α2δ, and β2). The pore-forming α1c subunit specifies basic channel characteristics, and serves as a docking site for several regulatory molecules including calcineurin; the α2δ and β subunits are powerful modulators of channel expression, open probability, activation, and inactivation[136-140]. I_Ca,L is expressed differentially across the wall of the ventricle, manifesting a clear transmural gradient of channel density. Transmural variations of I_Ca,L have been reported in canine[141], rat[142], and mouse[143] LV, but not in guinea pig ventricle[144].

I_Ca,L density is altered in a number of animal models of ventricular hypertrophy and heart failure. Depending on the severity of disease, I_Ca,L density has been reported to be increased, decreased, or unchanged. In general, I_Ca,L density is increased in mild-moderate hypertrophy and decreased in severe hypertrophy and heart failure[138, 139]. The most common change in I_Ca,L in heart failure is a slowing of the decay of the whole-cell current[145, 146], a change which may contribute to APD prolongation and to elevation of intracellular Ca²⁺ levels.

The mechanism underlying the prolonged decay of whole-cell current is multifactorial. Cardiac L-type Ca²⁺ channels are known to inactivate through voltage- and Ca²⁺-dependent mechanisms[147, 148]. Pure voltage-dependent inactivation has a substantially slower time course of development than Ca²⁺-dependent inactivation and plays a minor role in inhibition of Ca²⁺ influx into the cell. Thus, the major determinants of the inactivation kinetics of Ca²⁺ current during depolarization are Ca²⁺-dependent mechanisms[149]. Furthermore, Ca²⁺-dependent inactivation includes Ca²⁺ current-dependent inactivation and Ca²⁺ release-dependent inactivation. The former causes a slow phase inactivation and the latter causes a fast phase inactivation. Although both Ca²⁺ released from the SR and Ca²⁺ permeating through channels each play a role, SR-released Ca²⁺ is the more powerful inactivation mechanism [150].

In heart failure, Ca²⁺-induced SR Ca²⁺ release is significantly diminished due to the diminished SR Ca²⁺ content that stems from both reduced SERCA function and RyR leak. As a result, I_Ca inactivation is significantly slowed in failing ventricular myocytes. In addition, recent work from our group showed that CaMKII-dependent mechanisms also contribute to the slowed I_Ca,L inactivation seen in load-induced heart failure[77]. Attenuation of frequency-dependent Ca²⁺-induced I_Ca facilitation has been described in failing human LV, pointing to a likely role of CaMKII in human heart failure[151]. In addition, spatial differences in remodeling of the transmural gradient of I_Ca,L have been reported, with reduced I_Ca,L in subendocardial cells and increased I_Ca,L in subepicardial and mid-myocardial cells[144]. In keeping with the findings reported in animal models with severe hypertrophy or failure, I_Ca,L in failing human ventricular myocytes also exhibits either no change[152, 153] or a decrease in density[154]. However, one report described increases in channel open probability in failing human ventricular myocytes[155]. Abundance of the pore-forming α1c subunit has been reported to be unchanged in failing human myocardium[156-159] or reduced[154, 160]. In addition, a switch from the IVS3A to the IVS3B isoform of the α1c subunit has been described in failing hearts[161].

Na⁺-Ca²⁺ exchanger (NCX)

The Na⁺-Ca²⁺ exchanger (NCX) is a surface membrane protein that transports one Ca²⁺ ion in exchange for 3 Na⁺ ions. Its activity is said to be “forward” when Na⁺ is transported into the cell and Ca²⁺ is extruded outwards and “reverse” when ions are transported in the opposite directions. Most studies from hypertrophied and failing human hearts have demonstrated an increase in both NCX mRNA and protein levels[162-165], which has been posited to preserve diastolic extrusion of cytosolic Ca²⁺. At the same time, increased NCX activity may impair systolic function by favoring transport of Ca²⁺ out of the cell rather than back into intracellular stores. However, direct studies of NCX function in failing hearts are limited. In fact, NCX mRNA and protein levels are increased in pressure-overload ventricular hypertrophy, but direct measurements of NCX transport function, i.e. the Na⁺/Ca²⁺ exchanger current (I_NCX), revealed significant decreases, which may reflect disease-related alterations in the targeting of NCX protein to the sarcolemma[165, 166]. In a rabbit model of HF induced by combined aortic insufficiency and aortic constriction, NCX mRNA level and the function were both found to be significantly increased[167, 168].

I_NCX can be arrhythmogenic, because inward current occurring during Ca²⁺ extrusion can give rise to delayed afterdepolarizations. Indeed, as a result of NCX up-regulation and inward rectifier potassium current (IK1) reduction, the propensity for triggered arrhythmias in HF is significantly increased[168]. Similar to animal models, increased NCX mRNA and protein abundances in failing human heart have been reported[162, 169]. The functional role of the NCX in failing human myocardium is of greater importance to Ca²⁺-homeostasis than in non-failing myocardium due to the altered function of SR[170]. However, the alterations of NCX in HF are controversial; unchanged levels of NCX protein in human heart failure have been reported[171, 172], and some have reported a HF-related decrease in NCX expression[173] [174].

NCX likely plays a significant role in shaping the action potential profile. Forward-mode exchanger function generates an inward depolarizing current, contributing to APD prolongation. Conversely, computer simulations suggest that augmentation of reverse mode exchanger function during the early plateau would tend to shorten the action potential. However, some data suggest that the combination of forward mode function and delayed I_Ca,L inactivation together sum to prolong action potential duration[175]. It has been reported that direct blocking of NCX leads to different changes in APD, including a shortening[176], no change[177], or both shortening and prolongation depending on [Na⁺]_i[178]. Also, intracellular Na⁺ concentration ([Na⁺]_i) is an important determinant of the effect of NCX on APD. In the presence of high [Na⁺]_i and low [Ca²⁺]_i, as seen in heart failure, there is a large NCX-dependent Ca²⁺ influx during the action potential[179], which ultimately mandates removal of this extra Ca²⁺ during the diastolic period. Given that the NCX is the most important efflux pathway, this would lead to a sustained inward current during diastole, contributing to APD prolongation, instability of the resting membrane potential, and triggered abnormal impulses[180].

Stretch-activated channels (SACs)

Biomechanical signals external to the myocyte are capable of altering electrical activity[181], a potential arrhythmogenic influence in patients with congestive heart failure[182, 183]. In fact, disordered mechano-electrical feedback is believed to contribute to arrhythmogenesis via mechanisms of membrane depolarization and modification of APD[184] [185].

Stretch-activated channels (SACs) are key mechanosensors of mechano-electrical signal transduction. SACs are nonselective cation channels[186] which allow Na⁺, K⁺, Cs⁺, and Ca²⁺ to permeate[187]. A linear current-voltage relationship has been described for SACs with reversal potentials between -6 and -15 mV in myocytes[186, 188, 189]. Stretch has been reported to depolarize resting membrane potential and to increase, decrease, or have a crossover effect on APD[190]. The inconsistent effect on APD may reflect differences stress-responsiveness of isolated myocytes versus tissue or intact heart. Additionally, the different effects on APD may reflect the variety of experimental techniques used to stretch the myocardium and to track electrical activity.

The effects of myocardial stretch on APD arise from complex interactions between the SAC current and the other intrinsic membrane currents (and channels) present in myocytes. Even if the SAC conductance were to be constant during stretch, the time dependence and polarity of the SAC current are functions of myocyte membrane potential. For instance, with a constant reversal potential of -10 mV, the SAC current would be inward during diastole but would switch to outward as the cell depolarizes. Further, the large variability in the repolarization process in ventricular myocytes across species (and in different regions of the ventricle) may significantly alter the effects of a given SAC current. In addition, the reversal potential of SAC varies across species. Thus, the role of SAC current in regulating myocyte depolarization and repolarization is complex and species-dependent. Furthermore, membrane potential measurements during dynamic axial stretching have shown that the action potential duration is prolonged when stretch is applied during the late phase of twitch contraction, and that membrane depolarization depends on the phase, amplitude, and speed of the applied stretch[191]. The amplitude of stretch may also modulate the ion selectivity of stretch-activated channels[191].

To isolate the effects of SAC current from the numerous other electrical events occurring in the cell, several mathematical models have been developed. An SAC simulation was incorporated into a guinea pig ventricular cell model, and SAC activation triggered EADs[192]. Recently, we modeled SAC in rat ventricular myocytes, finding that SAC activation elicited electrical changes in action potential repolarization similar to mechanical stretch, but it did not produce EADs[193]. However, in combination with oxidative stress, EADs were steadily induced by SAC activation, potentially implicating SACs in ventricular arrhythmias in cardiomyopathies with concomitant increased ventricular wall tension and oxidative stress. This SAC-related mechanism may explain, at least in part, the clinical findings that reducing myocardial wall tension and oxygen demand by mechanical unloading using intra-aortic balloon counterpulsation significantly reduces ventricular arrhythmias in heart failure patients with medically refractory ventricular arrhythmias[194]. In any event, owing to the technical difficulty of imposing physiologically relevant mechanical stretch and the nonselectivity of SAC ionic conductance, the arrhythmogenic role of SACs in heart failure is poorly understood.

Tissue level electrical remodeling

Gap-junction remodeling

Gap junctions are specialized regions of membrane with densely packed channels called connexons that dock end-to-end and directly connect the cytoplasmic compartments of two adjacent myocytes. The number, size, and spatial distribution of gap junctions play an important role in determining the conduction properties of different cardiac tissues. Connexons are comprised of six monomers termed connexins. There are 3 connexins expressed in ventricular myocytes, connexin 40 (Cx40), connexin 43 (Cx43), and connexin 45 (Cx45). Cx43 plays a major role in electrical conduction in myocytes[195, 196], whereas Cx45 is likely less important owing to its low level of expression[197]. Connexin 40, a connexin associated with high conductance channels[198], is expressed mainly in the bundle branches and Purkinje fiber system[199, 200].

In heart failure, gap junction expression is significantly reduced, redistributed, and disorganized[201-203]. As a result, the normal ordered pathways for cell-to-cell conduction are disrupted. A growing body of evidence from studies in animal models support the fact that gap junction remodeling is one of the key determinants of the pro-arrhythmic substrate in diseased heart. Altered distribution of connexin 43 gap junctions is a hallmark feature of infarct border zones in failing ventricle harvested at cardiac transplantation[201]. In these zones, connexin 43 gap junctions are scattered extensively over the cells, rather than clustered at regions of cell-cell contact. In experimental models of myocardial infarction, this redistribution has been shown to occur rapidly[204]. Similar alterations in gap junction distribution are found in ventricular hypertrophy and are associated with reduced longitudinal conduction velocity[205, 206]. In end-stage heart failure, connexin 43 transcript and protein levels are markedly reduced[207]. In addition, hypophosphorylation of Cx43 due to elevated PP2A activity has been documented in failing LV[208, 209], as well as spatially heterogeneous reductions in Cx43 expression (larger decreases in subepicardial than in deeper transmural layers)[210]. Thus, an expanding literature supports the hypothesis that gap junction remodeling is a key molecular feature contributing to both arrhythmia, and consequent sudden cardiac death (SCD), and exacerbating the ventricular dysfunction associated with acquired heart disease. Therefore, pathological remodeling of gap junctions is a therapeutic target of interest[211].

Conduction sequence remodeling

Normally, electrical activation of the ventricle initiates in the subendocardial Purkinje network and spreads outward through the ventricular wall. Although the epicardium is activated last, it repolarizes faster than the endocardium; as a result, repolarization proceeds in the opposite direction, from epicardium to endocardium. The combination of in-to-out depolarization and out-to-in repolarization produces an electrocardiographic T wave on the surface ECG with polarity similar to the QRS. As noted earlier, we have reported attenuation of the transmural gradient of repolarization in failing LV, which abolishes preferential directionality of AP propagation[79].

In addition to the direct roles of altered myocardial repolarization in electrical conduction, disease-related remodeling can indirectly affect electrical stability via exaggerated anisotropy and altered cell-to-cell coupling. Chung et al. showed in an in vitro model that cardiac cell arrangement can alter electrical stability via mechanisms that are both dependent and independent of the direction of wave propagation[212]. Notably, restitution of APD and conduction velocity were significantly steepened in the direction of cell alignment. Furthermore, prolongation of APD and calcium transient duration were seen in highly anisotropic cell networks, both for longitudinal and transverse propagation[212]. Patients with structural disease have a wider diastolic interval range over which APD alternans occurs with earlier onset and increased magnitude of APD alternans compared with patients without structural heart disease[213]. The occurrence of APD alternans during induced ventricular tachycardia and during rapid pacing could be derived from the dynamic restitution function. This suggests that tissue alignment must be taken into account in experimental and computational models of arrhythmia generation and in designing effective treatment therapies.

Sarcoplasmic reticulum malfunction

The sarcoplasmic reticulum (SR) surrounds each myofibril and is physically separate from the sarcolemma. The SR membrane contains high levels of Ca²⁺-ATPase which serves to uptake Ca²⁺ from the sarcoplasm, a process leading to mechanical relaxation. Ca²⁺ is released into the sarcoplasm via ryanodine receptors (RyRs) during systole to trigger contraction.

In mammalian striated muscles, the expression of the different RyR protein isoforms is tissue specific. The predominant RyR isoform in skeletal muscle is RyR1, commonly referred to as the skeletal RyR isoform[214, 215]. RyR2 protein is the most abundant isoform in cardiac muscle, while RyR3 is also found in mammalian striated muscles but at relatively low levels[216]. Ca²⁺ release from a cluster of ryanodine receptors results in a spatiotemporally-restricted rise in cytosolic Ca²⁺, which can be visualized as a calcium spark. Marx et al[217] showed that RyR2 assembles as a macromolecular signaling complex comprising FKBP12.6, the catalytic and regulatory subunit of PKA, PP1, PP2A, and a PKA-anchoring protein, mAKAP. The actual Ca²⁺-conducting channel is one of the constituents of an integrated macromolecular complex that specifically regulates channel Ca²⁺ gating. Studies in both heart failure models and in patients point to FKBP12.6 as a major regulatory element governing RyR2 function[217, 218]. The role of calcineurin on RyR function remains controversial and may be critically dependent on the RyR isoform.

Several groups have reported lack of significant alterations in RyR2 abundance in failing explanted human hearts compared with nonfailing controls[219-221], whereas other groups have reported a down-regulation of RyR2[222, 223]. However, spontaneous RyR leakage of Ca²⁺ has been observed consistently in failing ventricular myocytes resulting from increased RYR phosphorylation by CaMKII[224-226] and PKA[217]. Most recent studies have pointed to CaMKII, but not PKA, in HF-related RyR Ca²⁺ leak[227, 228].

Three distinct genes encode SERCA isoforms. The SERCA 1 gene is expressed in fast skeletal muscle, whereas the SERCA 2 gene gives rise to SERCA 2a and SERCA 2b isoforms by alternative splicing. The SERCA 2a isoform is expressed in cardiac and slow skeletal muscle, while SERCA 2b is ubiquitously expressed and is the dominant isoform found in the cerebellum. SERCA 3 is expressed in non-muscle tissues, such as platelets and lymphoid tissues[229, 230].

The rate at which SERCA moves Ca²⁺ across the SR membrane is controlled by phospholamban (PLB). PLB is a 52 amino acid integral membrane protein that regulates the Ca²⁺ pump in both cardiac and skeletal muscle. In vitro studies have shown that PLB can be phosphorylated at Ser¹⁰ by protein kinase C, at Ser¹⁶ by cAMP-dependent protein kinase (PKA), and at Thr¹⁷ by Ca²⁺-calmodulin-dependent protein kinase (CaMKII)[231-233]. However, in vivo studies have shown that only Ser¹⁶ and Thr¹⁷ are phosphorylated in cardiac myocytes[234, 235]. Each phosphorylation event appears to occur independently[236-239]. Some studies have reported additive effects of PKA and CaMKII phosphorylation of PLB on SR Ca²⁺ transport[236, 237, 240, 241]; others have proposed that maximal stimulation of the Ca²⁺ pump occurs by phosphorylation at a single site[239, 242], and additional phosphorylation events at other sites does not further stimulate pump activity[243]. In its unphosphorylated state, PLB associates with SERCA and inhibits the Ca²⁺-ATPase and its pumping of Ca²⁺. When phosphorylated PLB dissociates from the pump, Ca²⁺ movement increases. PP1 accounts for approximately 90% of phospholamban (Ser-16 or Thr-17) phosphatase activity[244].

Evidence now indicates that the levels of PLB protein remain unchanged in human heart failure, whereas the levels of SERCA2a protein decrease[220, 245-249]. This would be expected to lead to an increased functional stoichiometry of PLB to SERCA, facilitating inhibition of SERCA2a Ca²⁺ pumping activity, and prolonged relaxation times. In addition, the phosphorylation status of PLB at Ser16 and Thr17 is decreased in heart failure[250], indicating that there is a yet further increase in the inhibitory function of PLB. Indeed, studies of SR Ca²⁺ uptake have revealed decreases in both Vmax and Ca²⁺ affinity in failing myocardium[245]. Thus, alterations in the PLB:SERCA2a ratio and the degree of PLB phosphorylation together likely contribute to reduced SR Ca²⁺ uptake and increased diastolic Ca²⁺ levels in the cytoplasm. The resulting reduction of SR Ca²⁺ content will, in turn, have a negative inotropic effect on contraction, and at the same time, reduce Ca²⁺-induced Ca²⁺ inactivation, leading to a slowed inactivation time course for I_Ca. This will, together with a prolonged APD, further increase diastolic Ca²⁺ and the propensity for triggered ventricular arrhythmias.

Clinical relevance and perspective

Disease-related electrical remodeling is a fundamental mechanism underlying the proarrhythmic phenotype of heart failure. Reductions in both transient outward and delayed rectifier K⁺ currents contribute to prolongation of APD. Prolonged APD, in turn, promotes increased influx of Ca²⁺ during excitation. In addition, increases in NCX function and slowing of I_Ca,L inactivation contribute to Ca²⁺ overload. Further, prolongation of APD and abnormal handling of intracellular Ca²⁺ promote abnormal increases in focal activity and automaticity. In addition, heterogeneous APD prolongation within the ventricular wall amplifies dispersion of repolarization, an established mechanism contributing to re-entry. Finally, spatially different changes in I_to across the ventricular wall in heart failure alter cellular coupling current. Together, these changes, along with the alteration of gap junctions and tissue alignment, lead to significant changes in electrical conductivity and sequence, which are important mechanisms underlying the increased propensity to ventricular arrhythmia and SCD in heart failure.

Recent studies have also raised the possibility that alterations in membrane currents may be proximal events contributing to cardiomyopathy. For example, Ca²⁺ entering through the L-type voltage-dependent Ca²⁺ channel not only functions as a trigger for contraction but also transduces electrical activity into a series of intracellular signaling events. Numerous studies have demonstrated that increases in intracellular Ca²⁺ concentration may lead to contractile dysfunction, hypertrophy, and heart failure[76] and that “Ca²⁺ overload” may trigger downstream signaling cascades that initiate the hypertrophic gene program[251]. Enhanced sarcolemmal L-type Ca²⁺ channel activity can precipitate heart failure through myocyte necrosis[252]; in contrast, T-type currents can antagonize hypertrophy[253]. Also, whereas decreases in I_to density have been regarded as secondary to the hypertrophic or failure phenotype, some evidence suggests that suppression of I_to could result in cardiac hypertrophy[254, 255]. These data suggest that ion channel remodeling resulting from heart disease may contribute to the progression of heart disease. In addition, heart failure-related remodeling of one channel may trigger remodeling of other channels. For instance, we have reported that suppression of I_to facilitates I_Ca,L in cardiac myocytes[256].

Nearly 6 million Americans live with heart failure, a syndrome marked by substantial morbidity and mortality from arrhythmia[1]. Indeed, arrhythmic sudden death is a leading cause of mortality in the Western world, with an estimated 300,000 cases per year in the United States[1]. Whereas the mechanism of mortality in heart failure depends on disease severity, some estimates suggest that death stems approximately equally from progressive pump failure, SCD, and SCD during episodes of clinical exacerbations of heart failure[1]. In addition to ventricular tachyarrhythmias, patients with heart failure experience a variety of other arrhythmias. For example, AF is very common in heart failure (11.8%), is responsible for about 15-20% of all strokes[257], and contributes substantially to morbidity in heart failure.

The clinical efficacy of antiarrhythmic pharmacotherapy has proved disappointing in the majority of instances. However, recent insights into complex mechanisms of electrical remodeling have raised the prospect of targeting disease-related events contributing to the proarrhythmic substrate of the failing LV. Major challenges remain, but patients with heart disease are likely to benefit.