Arrhythmia Susceptibility Results from Cardiac Ion Channel Dysfunction
Like other excitable cells, cardiac myocyte excitability results from action potentials. Similar to action potentials in neurons, skeletal muscle, and smooth muscle, the action potential in cardiac myocytes is mediated by ion currents through membrane-bound channel proteins. The cardiac myocyte action potential, however, is distinctive in its duration, which is approximately 300 milliseconds. By contrast, the action potential in neurons occurs in a few milliseconds.
Figure 203-14 is a stylized ventricular action potential and correlates molecular events with the electrocardiographic findings. The action potential consists of five phases termed 0 to 4. Phase 0 corresponds to depolarization of the myocyte. This phase is initiated by the rapid opening (activation) of voltage-gated sodium channels encoded by SCN5A . On the electrocardiogram, depolarization of the ventricle is seen as the QRS complex. Phase 1 of the action potential occurs at the end of rapid depolarization, recognized as a partial repolarization of the membrane. This effect is primarily due to closure (inactivation) of cardiac sodium channels, and activation of a transient outward potassium current and, in some species, a calcium-activated chloride current. Phase 2 is the plateau phase. The relatively long duration of this phase is unique to ventricular and Purkinje fiber myocytes, and is maintained primarily by inward calcium currents through L-type calcium channels and gradually increasing outward potassium currents. The total amount of current during the plateau phase of the cardiac action potential is small. As a result, this phase of the action potential is very susceptible to small changes in ion currents, which can substantially shorten or lengthen the action potential. At this point in the cardiac cycle, the electrocardiogram has returned to the baseline. Phase 3 of the cardiac action potential is a period of cellular repolarization. Repolarization is caused by outward flow of potassium mediated by several different potassium channels. The repolarization phase correlates with the T wave on surface electrocardiogram. Thus, the duration of the QT interval is related directly to the length of ventricular action potentials. The final phase of the action potential is Phase 4, in which the membrane potential has returned to its baseline near −85 mV and correlates with ventricular diastole.
Cardiac action potential and underlying potassium currents that mediate repolarization. (From Zeng et al.198 Adapted with permission.)
Physiological and pharmacologic studies have defined two main repolarizing potassium currents, IKr and IKs (Fig. 203-14). IKr is the rapidly activating delayed rectifier potassium current (IKr) that is specifically blocked by methanesulfonamide drugs. When these drugs block the IKr current, IKs, the slowly activating delayed rectifier potassium current, remains.
Reduced Cardiac Repolarizing Currents Cause Long QT Syndrome
Herg Encodes IKr Channels.
Expression of HERG in heterologous systems led to the discovery that this gene encodes α subunits that form cardiac IKr potassium channels (Table 203-6). The characteristic biophysical properties of cardiac IKr were reproduced by expression of this gene in heterologous systems130,131 (Fig. 203-15). These properties include pronounced rectification of the current-voltage relationship and block by specific antiarrhythmic drugs. Rectification refers to the property whereby the whole-cell conductance decreases as the membrane is strongly depolarized, resulting in a bell-shaped relationship when current magnitude is plotted as a function of membrane voltage (Fig. 203-15). The rectification of IKr permits the plateau phase of cardiac myocytes to be very long, and explains the prominent role of this current during phase 3 repolarization of the action potential.
Table 203-6: Molecular and Cellular Mechanisms of Cardiac Arrhythmias |Favorite Table|Download (.pdf) Table 203-6: Molecular and Cellular Mechanisms of Cardiac Arrhythmias
|Disease ||Inheritance ||Gene ||Protein ||Function ||Mechanism |
|Long QT syndrome ||Autosomal dominant || KVLQT1 ||KvLQT1 ||IKs channel α subunit ||loss of function |
| || || HERG ||HERG ||IKr channel α subunit ||loss of function |
| || || SCN5A ||SCN5A ||sodium channel α subunit ||gain of function |
| || || LQT4 ||unknown ||? ||? |
| || || KCNE1 ||minK ||IKs channel β subunit ||loss of function |
| || || LQT6 ||unknown ||? ||? |
| ||Autosomal recessive || KVLQT1 ||KvLQT1 ||IKs channel α subunit ||loss of function |
| || || KCNE1 ||minK ||IKs channel β subunit ||loss of function |
| ||Syndactyl-associated ||? ||unknown ||? || |
| ||Acquired* || HERG || ||IKr channel α subunit ||acquired loss of function |
|Idiopathic ventricular fibrillation || || SCN5A || ||sodium channel α subunit ||loss of function |
HERG encodes IKr . A, HERG currents recorded from Xenopus oocyte at potentials of −50 to −10 mV, applied in 10 mV steps. B, Currents recorded at potentials of 0 mV (largest) to +40 mV (smallest). Bottom panels show current-voltage relationship (C) and voltage dependence of channel activation (D). (From Sanguinetti, et al.130 Adapted, with permission.)
The unusual rectification properties of HERG channels result from rapid inactivation and slow deactivation. HERG channels inactivate (close) much more rapidly than they activate (open). These properties vary at different membrane potentials, but the net result is that most HERG channels are inactivated during the depolarization and plateau phases (Phases 0 through 2) of the cardiac action potential. However, during repolarization the channels rapidly recover from the inactivate state to an open state. The subsequent transition from the open state to the closed state (deactivation) is much slower. As the membrane potential returns to its resting level, the current conducted by HERG channels actually increases, reaching a maximum near −50 mV.
Considerable progress has been made in determining the structural features of HERG channels that are responsible for rapid inactivation. Two distinct mechanisms have been described for inactivation of voltage-gated potassium channels. N-type inactivation is caused by binding of the N-terminal region of the protein to the mouth of the pore. This type of inactivation can be eliminated by removal of the N-terminus. However, this procedure has no effect on inactivation of HERG channels. C-type inactivation involves residues in the pore-loop of the channel and may be caused by collapse of the ion permeation pathway near the outer mouth of the channel. This type of inactivation can be eliminated by mutation of key residues near the outer mouth of the channel. Inactivation of HERG has properties most similar to C-type inactivation as it can be eliminated by mutation of two amino acids (G628C/S631C) in the pore-forming region of the channel.132 The current-voltage relationship of this mutant HERG channel is linear rather than rectifying, indicating the loss of fast inactivation. Moreover, the inactivation properties of HERG can be transferred to eag channels (which normally do not inactivate) by inserting the pore-loop and half of the S6 transmembrane domain of HERG into eag.133
The crystal structure of the eag domain in the N-terminus of HERG channels has been determined.134 This work demonstrates that the eag domain has a hydrophobic face that binds directly to an intracellular channel region, tentatively identified as the region that links the S4 and S5 transmembrane domains. The interaction of the eag domain with the S4-S5 linker slows the rate of channel deactivation. Removal of the N-terminal region, or mutation of specific residues in the S4-S5 linker, causes the channel to deactivate much faster than normal. The physiological relevance of this interaction is suggested by the finding that alternative splicing of HERG creates an isoform with a shortened N-terminal region. Channels formed by assembly of these subunits alone, or by coassembly with full length HERG subunits, produce a channel that deactivates much faster than those formed by homotetrameric assembly of full-length subunits.106,107
Many different types of HERG mutations have been identified in association with long QT syndrome. These mutations include large deletions, small deletions, frameshift mutations, and splice-site mutations. The most common HERG mutations identified thus far are missense mutations.77,96-99,101,135,136 As noted above, these mutations span the entire HERG coding region, but tend to be clustered in the pore domains, the membrane-spanning domains, and the intracellular N- and C-terminal domains. HERG is unusual in that there are many mutations in the intracellular N- and C-terminal domains, presumably due to the unusual properties of these domains in HERG.
At least two distinct molecular mechanisms account for reduced HERG function in the long QT syndrome. In the first, long QT syndrome-associated intragenic deletions of the HERG allele can result in the synthesis of aberrant subunits that do not coassemble with normal subunits into the functional tetrameric form. The net effect is a 50 percent reduction in a number of functional channels; that is, haploinsufficiency. In the second mechanism, missense mutations lead to the synthesis of HERG subunits with single amino acid substitutions. Channels formed from the coassembly of normal and mutant subunits have reduced or no function. The result is a greater than 50 percent reduction in HERG channel function, a dominant-negative effect (Fig. 203-16). The severity of dominant-negative effect of a specific mutation varies considerably. In some cases, the dominant-negative effect is relatively small. In others, the effect is complete, leading to marked reduction of IKr even in the heterozygous state. It is likely that the severity of the dominant-negative effect has an impact on the phenotype of an individual. However, there are many other factors that also impact on the phenotype, so an extensive study is required. Mutations in the pore region (e.g., G628S) lead to complete loss of function and a strong dominant-negative effect,137,138 presumably because the channel is processed normally139 and the pore of the channel has been destroyed. As expected, IKr was absent in a transgenic mouse model expressing high levels of G628S HERG.140 Mutations in other regions of the channel, however, have less pronounced effects. For example, mutations of the eag domain in the N-terminus of HERG generally disrupt slow deactivation.140a As a result, deactivation is accelerated, leading to reduced repolarizing potassium current. Some mutations (e.g., Y611H and V822M) cause defects in biosynthetic processing, resulting in retention of protein in the endoplasmic reticulum.139 Other mutations (e.g., N470D and T474I) form functional channels but alter gating properties.137,138 Three mutations in HERG (T474I, A614V, and V630L) cause loss of channel function, but when coexpressed with wild-type subunits, shift the voltage dependence of channel inactivation to more negative potentials and enhance inward rectification, thereby suppressing channel function.141
Dominant-negative mutation of HERG. Top three panels show currents recorded from a Xenopus oocyte expressing wild-type (WT) HERG channels, A561V HERG channels, or a combination of the two channel types. Bottom panel is a plot of currents as a function of voltage for the three conditions. (From Sanguinetti et al.137 Adapted with permission.)
Acquired Long QT Syndrome Results from Block of Herg Channels.
Dysfunction of HERG channels and reduced IKr are also the mechanisms of acquired long QT syndrome. For reasons that are not yet clear, many drugs, including many class III antiarrhythmics, antihistaminics, and some antibiotics, are very potent blockers of HERG channels.32,120,142-150 Block of IKr delays cardiac repolarization, lengthens QTc interval, and may induce torsade de pointes. 151
HERG channels are also paradoxically sensitive to extracellular potassium.130,152 Increasing the extracellular concentration of potassium actually increases outward potassium current, in opposition to the decreased chemical driving force (Fig. 203-17). The mechanism underlying this effect is not precisely known, but it is associated with a slowing153 and a positive shift in the voltage dependence of HERG inactivation.154 This shift reduces the degree of channel rectification and increases the magnitude of outward current within the voltage range traversed by a typical cardiac action potential. This effect is observed within the physiological range of serum potassium concentrations. For example, reducing extracellular potassium from 5 mM to 2 mM reduces HERG current by ~35 percent.130 This may explain the well-documented clinical association between hypokalemia, QT interval prolongation, and arrhythmia susceptibility.155
Potassium modulates magnitude of HERG channel function. Top panel shows current-voltage relationship for HERG for different external potassium concentrations. Bottom panel shows relationship between HERG magnitude and external potassium concentration. (From Sanguinetti, et al130 Adapted with permission.)
Several new drugs that block IKs have been described.156-159 These compounds prolong action potential, but it is unclear whether they will cause torsade de pointes.
KVLQT1 Subunits Coassemble with minK Subunits to Form Cardiac IKs Channels.
Heterologous expression of the KVLQT1 gene in mammalian cells and Xenopus oocytes confirm that this gene encodes α subunits that form voltage-gated potassium channels (Fig. 203-18). However, the biophysical properties of KvLQT1 channels were unlike known potassium channels recorded from cardiac myocytes. This led to the hypothesis that KvLQT1 α subunits might coassemble with other subunits to form potassium channels with recognizable biophysical characteristics.
KvLQT1 and minK coassemble to form cardiac IKs channels. (From Sanguinetti et al.199 Adapted with permission.)
MinK, the minimal potassium channel subunit, was initially cloned by functional expression in Xenopus oocytes in 1988.160 The biophysical properties of the current induced by expression of minK in Xenopus oocytes was similar to cardiac IKs . However, there were several problems. First, the structure of minK was unusual. As noted above, minK is 130 amino acids in length and is thought to contain only one membrane-spanning domain. The typical voltage-gated potassium channel subunit has six membrane-spanning domains, and four subunits are required for coassembly and formation of functional channels. Thus, investigators hypothesized that many minK subunits might be required to form functional channels, perhaps by aggregating in response to membrane depolarization. Although some experimental evidence supported this hypothesis, other evidence indicated that only two or four minK subunits were required for functional channels161 and that the channel was more likely a heteromultimeric complex.162 A second problem with minK was experimental; physiological studies indicated that heterologous expression of minK in Xenopus oocytes induced a functional potassium channel, but in most cases (except that described in reference163), expression in mammalian cells failed to induce a current. The third problem with minK was the observation of saturability; that is, injection of increasing amounts of KCNE1 cRNA into Xenopus oocytes did not lead to increasing current. These data all led to the hypothesis that minK might be a βsubunit that coassembled with another α subunit to form IKs channels.164,165 Heterologous expression of KvLQT1 subunits with minK subunits strongly support this hypothesis. As seen in Fig. 203-18, heterologous expression of minK alone in mammalian cells induced no current. However, heterologous expression of KvLQT1 and minK together in mammalian cells led to a large potassium current with the biophysical properties of cardiac IKs. Thus, KvLQT1 and minK coassemble to form cardiac IKs channels. Although the stoichiometry of coassembly is not yet known, it is likely that four KvLQT1 α subunits coassemble with four minK subunits to form these channels.
Mutations of KVLQT1 tend to cluster around the membrane-spanning domains and the pore region. In most cases, these mutations lead to a loss-of-channel function. Most often, missense mutations have a potential to a cause a dominant-negative effect. As with HERG, there is a spectrum of dominant-negative suppression of KvLQT1 function. For example, missense mutations (e.g., G306R, T312I, G314S, and D317N) in the pore region or elsewhere (e.g., A178P and L273F) of KvLQT1 are associated with dominant long QT syndrome and cause loss-of-function and a pronounced dominant-negative effect when coexpressed with wild-type KvLQT1 and minK.166,167 In general, mutations associated with recessive long QT syndrome, such as truncation of the C-terminus of KvLQT1,166 or missense mutations (e.g., W305S) cause loss-of-function and little or no dominant-negative effect.168 It should be noted, however, that these same mutations can be found in autosomal dominant long QT syndrome.
Two distinct isoforms of KvLQT1 have been characterized. One is the full-length clone that in heterologous systems coassembles with minK to form cardiac IKs channels. The other is a truncated form of KvLQT1.169,170 When coexpressed in vitro with full-length KvLQT1, this isoform has a dominant-negative effect on IKs channel function. The significance of this finding in vivo, however, is not known.
The findings that mutation of KVLQT1 caused long QT syndrome and that KvLQT1 coassembles with minK made it likely that mutation of KCNE1 would also cause long QT syndrome. This was confirmed using single-strand conformation polymorphism (SSCP) analyses of genomic DNA from affected members of a few small families.80 Two mutations in KCNE1 have been characterized by heterologous expression in Xenopus oocytes. These mutations (S74L and D76N) are located in the putative intracellular domain of the minK protein. Both mutations shift the voltage dependence of activation gating of minK channels to more positive potentials and increase the rate of channel deactivation. These changes in gating reduce the magnitude of outward current during the plateau and repolarizing phases of the action potential. In addition, D76N has a dominant-negative effect on IKs . That LQT-associated mutations of KCNE1 alter gating kinetics of IKs provides compelling evidence that minK forms an integral part of the IKs channel rather than simply serving as a chaperone, as was once hypothesized. KCNE1 mutations also lead to a loss-of-function; in some cases, to a dominant-negative loss-of-function.
Gain-of-Function Mutations in SCN5A Cause Long QT Syndrome.
As noted earlier, a spike of sodium current initiates the cardiac action potential. Cardiac sodium channels open for a very brief time in response to membrane depolarization (Fig. 203-19). Soon after opening, the channels inactivate and remain closed for the remainder of the action potential. An intracellular domain located between DIII and DIV mediates inactivation of the sodium channel. This is often referred to as the inactivation gate and is thought to physically occlude the inner mouth of the channel pore. SCN5A mutations associated with long QT syndrome were identified in this region.79 These mutations included an in-frame deletion of nine base pairs causing deletion of three amino acids (ΔKPQ). Physiological characterization of this mutant led to the discovery that the mutation destabilized the inactivation gate.75 That is, activation of these sodium channels is normal and the initial current at the whole-cell level appears normal. At the whole-cell level, the rate of inactivation of the ΔKPQ mutant appears a little faster than normal. However, a percentage of mutant sodium channels remain open, resulting in a small steady-state current during a long depolarization (Fig. 203-19). The net effect is a small, maintained, inward current during the plateau of the action potential. Recall that the plateau phase of the action potential is particularly susceptible to small changes in current. Maintained depolarizing sodium current during the plateau phase stalls the process of repolarization and lengthens action potential duration. Three missense mutations of SCN5A that cause long QT syndrome have also been characterized. N1325S and R1644H sodium channels exhibit dispersed reopenings after the usual initial opening, whereas ΔKPQ caused both dispersed reopenings and long-lasting bursts after membrane depolarization.171 Another missense mutation (R1623Q) in the S4 segment of domain IV, caused only minor abnormalities in channel activation, but significantly delayed inactivation. Single-channel recordings showed that mutant channels had prolonged open times with persistent bursting behavior.172
Long QT syndrome-associated mutation in SCN5A (ΔKPQ) alters single channel properties. Wild-type (WT) sodium channels open briefly upon membrane depolarization (left panel). Mutant channels (ΔKPQ) repetitively open, indicating dysfunctional inactivation gating (right panel). Lower panels show ensemble averages of single channel traces. (From Bennett et al.75 Adapted with permission.)
Loss-of-Function Mutation in the Cardiac Sodium Channel Can Cause Idiopathic Ventricular Fibrillation.
SCN5A mutations have been associated with idiopathic ventricular fibrillation. In some cases, these mutations are clearly loss-of-function. For example, a nonsense mutation of SCN5A has been associated with idiopathic ventricular fibrillation in an otherwise healthy individual who has borderline corrected QT interval prolongation and no other electrocardiographic abnormality.129 The functional consequences of the splicing mutation are not known, but are most likely loss-of-function.
Other mutations associated with prolongation of the QRS complex with a right-bundle-branch block pattern are less clear-cut in terms of physiological consequences. Unlike SCN5A mutations associated with long QT syndrome, no persistent inactivation-resistant current was observed in missense mutations. Biophysical analysis of the two missense mutations in SCN5A showed a shift in the voltage dependence of steady state inactivation toward more positive potentials, associated with an acceleration in recovery time from inactivation.129 It is not yet clear whether these biophysical properties are responsible for increased risk of idiopathic ventricular fibrillation in these individuals.
Nor is it clear how a reduction in the total number of functional sodium channels and expression of heterogeneous population of sodium channels leads to arrhythmia susceptibility. Inhibition of sodium channel current can cause heterogeneous action potentials in the right ventricular epicardium, leading to marked dispersion of repolarization and refractoriness.173,174 This creates a substrate for the development of reentrant arrhythmia. It is likely that a premature beat initiates ventricular tachycardia and fibrillation in this substrate, possibly through Phase 2 reentry produced by the underlying substrate.
The Mechanism of Arrhythmia at the Organ Level: Unidirectional Block and Reentry.
Multiple molecular mechanisms for arrhythmia susceptibility have been defined and more await discovery. Five loci and four genes are associated with long QT syndrome, and mutations in one gene are associated with idiopathic ventricular fibrillation. Gain-of-function mutations in the cardiac sodium channel, and loss-of-function mutations in cardiac potassium channels, lead to long QT syndrome by prolonging action potential duration in individual cardiac myocytes. Loss-of-function mutations in cardiac sodium channels can cause conduction abnormalities, increasing the risk of unidirectional block. As one might expect, these genes are not expressed uniformly throughout the myocardium but are instead differentially expressed. As a result, the effect of mutations varies from one region to another. Abnormal regional repolarization or conduction can lead to a substrate for arrhythmia. Under certain circumstances, the substrate can become critical when areas of unidirectional block are created (Fig. 203-20). Although necessary, this substrate is not sufficient for arrhythmogenesis. A trigger is still required. Many different triggers have been identified. In long QT syndrome, a common trigger is exercise or excitement, but rest and sleep have also been associated the syndrome.175 These triggers are thought to be mediated through autonomic tone. Increased autonomic tone increases cyclic AMP-mediated enhancement of calcium channel function, spontaneous secondary depolarizations resulting from reopening of calcium channels. These depolarizations can trigger an arrhythmia.
Unidirectional block and reentry. The mechanism of arrhythmia at the organ level involves unidirectional block of conduction through refractive or damaged tissue and development of a reentrant circuit of electrical activity. Solid lines represent normal conduction pathway in a bifurcated Purkinje fiber. Conduction is blocked in an area of refractory tissue (stippled area). If conduction velocity is slowed (dashed line), then reentry through area of refractory tissue can occur.
Several models have been developed to define the mechanism of arrhythmia at the level of the heart. For example, torsade de pointes tachyarrhythmias can be mimicked by administering anthopleurin A to dogs.176,177 This neurotoxin slows inactivation of sodium channels, roughly similar to the effect of mutations in SCN5A that cause long QT syndrome. Analysis of tridimensional activation patterns showed that the initial beat of toxin-induced tachyarrhythmia arose as focal activity from a subendocardial site. Subsequent reentrant excitation was due to infringement of a focal activity on the spatial dispersion of repolarization, resulting in functional conduction block and circulating wave fronts. Changing the site of origin of focal activity caused the resulting twisting pattern of the QRS configuration. The development of a functional block between the anterior and posterior right ventricular free wall and septum was responsible for initiating a bifurcation in a rotating scroll into two separate scrolls in each ventricle. The development of multiple reentry circuits within the heart is the likely cause of ventricular fibrillation and sudden death.
Deafness in Jervell and Lange-Nielsen Syndrome Results from Complete Loss of IKs Channel Function in the Inner Ear.
Individuals with Jervell and Lange-Nielsen syndrome carry two mutant copies of either KVLQT1 or KCNE1, gene symbol for potassium voltage-gated channel ISK-related subfamily member 1 encoding minK (see Chap. 254). These mutations lead to complete loss of IKs function. The result in the heart is substantial prolongation of the QT interval and marked arrhythmia susceptibility. Most individuals with two mutant copies of KVLQT1 or KCNE1 die in childhood unless treated. KVLQT1 and KCNE1 are also expressed in the inner ear. Here, the channel conducts potassium across the apical membrane of strial marginal cells into the lumen of the vestibular labyrinth of the inner ear.178 This generates the K+-rich fluid, endolymph. In the absence of IKs channels, inadequate endolymph is produced. As a result, Reisner's membrane, a thin membrane that covers the inner ear, collapses. This leads to degeneration of the organ of Corti, the organ that contains hair cells that sense movement of the endolymph, the mechanism responsible for hearing. As a result, individuals harboring these mutations have congenital neural deafness. This mechanism has been demonstrated in mice lacking minK.179