Molecular Pathophysiology of Congenital Long QT Syndrome

Abstract

Ion channels represent the molecular entities that give rise to the cardiac action potential, the fundamental cellular electrical event in the heart. The concerted function of these channels leads to normal cyclical excitation and resultant contraction of cardiac muscle. Research into cardiac ion channel regulation and mutations that underlie disease pathogenesis has greatly enhanced our knowledge of the causes and clinical management of cardiac arrhythmia. Here we review the molecular determinants, pathogenesis, and pharmacology of congenital Long QT Syndrome. We examine mechanisms of dysfunction associated with three critical cardiac currents that comprise the majority of congenital Long QT Syndrome cases: 1) I_Ks, the slow delayed rectifier current; 2) I_Kr, the rapid delayed rectifier current; and 3) I_Na, the voltage-dependent sodium current. Less common subtypes of congenital Long QT Syndrome affect other cardiac ionic currents that contribute to the dynamic nature of cardiac electrophysiology. Through the study of mutations that cause congenital Long QT Syndrome, the scientific community has advanced understanding of ion channel structure-function relationships, physiology, and pharmacological response to clinically employed and experimental pharmacological agents. Our understanding of congenital Long QT Syndrome continues to evolve rapidly and with great benefits: genotype-driven clinical management of the disease has improved patient care as precision medicine becomes even more a reality.

I. INTRODUCTION

A genetic disorder disrupting electrical activity in the heart, congenital Long QT Syndrome (LQTS) can lead to life-threatening arrhythmias and sudden cardiac death. In the first cases of congenital LQTS, described in 1957, several children in one family presented with prolongation of the QT interval on the electrocardiogram (ECG) and congenital deafness (189). This came to be known as Jervell and Lange-Nielsen syndrome (JLNS), the autosomal recessive form of LQTS. The more common, autosomal dominant form of congenital LQTS that presents without deafness was first described in 1963 and 1964 in two separate cases and became known as Romano-Ward syndrome (346, 459). Since these initial patient descriptions, advances in our understanding of the mechanisms of cardiac electrical excitability at the tissue, cellular, and molecular level have yielded much insight into the pathophysiology of congenital LQTS.

Clinically, congenital LQTS patients often first present after episodes of syncope and/or seizure, and the ECG reveals a prolonged QT interval. The ECG measures electrical activity of the heart over time, at the patient's body surface. The primary electrical signals observed include the P wave, which signifies atrial depolarization; the QRS complex, which arises from ventricular depolarization; and the T wave, due to ventricular repolarization (Figure 1A). The QT interval, therefore, reflects the time elapsed from the initiation of ventricular depolarization to the end of ventricular repolarization. The QT interval shortens with increasing heart rate thus requiring a normalization, or “correction,” for heart rate. For a diagnosis of LQTS, this rate-corrected QT (QTc) interval prolongation on a 12-lead ECG generally is referenced as >470 ms for males and >480 ms for females. QTc also varies with age, and thus, an age-appropriate prolonged QTc interval in a patient aids the diagnosis of LQTS (209). However, diagnosis of LQTS based on absolute QT interval cutoffs can be challenging, since there is considerable overlap in the QTc distribution of affected patients and otherwise healthy individuals (193). Asymptomatic patients can have intervals beyond these cutoff and develop no arrhythmias; similarly, QTc intervals below this cutoff can be seen in patients with established LQTS (with clinical arrhythmias and positive genetic testing) (16, 17, 338). Clinical scoring systems (368), as well as genetic testing, can be helpful to assist with the diagnosis of congenital LQTS (193), particularly when QT intervals are on the borderline (within 20 ms of these cutoffs) or when clinical history is equivocal. This review will focus primarily on the various forms of congenital LQTS.

FIGURE 1.

ECG to cellular ionic currents. A: membrane depolarization and the rapid upstroke of the ventricular action potential give rise to the QRS complex. The duration of the QT-interval corresponds to the time to ventricular repolarization. The relatively stable membrane potential during the plateau phase of the action potential gives rise to a brief isoelectric period. Ventricular repolarization gives rise to the T-wave. B: time course of several ionic currents that underlie ventricular action potential morphology (currents not to scale). The rapidly activating and inactivating I_Na drives membrane depolarization. Two K⁺ currents, I_Ks and I_Kr, contribute most to the repolarizing current necessary to drive membrane potential back to rest.

Ion channels are the molecular entities underlying most ionic currents in the heart, allowing passive diffusion of ions across the cell membrane's electrochemical gradients (Figure 2). A selectivity filter in the channel pore, determined by distinct atomic components (152, 315), endows selective permeation of ions, such as Na⁺, K⁺, and Ca²⁺ (246). Some ion channels exhibit voltage-dependent gating, where voltage-sensing domains respond to changes in membrane potential to cause channel opening or closing (247).

FIGURE 2.

Schematic of a generic K⁺ and Na⁺ ion channel. Ion channels allow for selective permeation of ions through the plasma membrane down their electrochemical gradient. The classic K⁺ channel consists of four identical pore-forming subunits, whereas each Na⁺ channel is formed by a single polypeptide with four homologous domains.

The ventricular cellular action potential results from the summation of a large number of ion channel currents and electrogenic pumps that control the cellular membrane potential (see Nerbonne and Kass review, Ref. 299), but in this review we will focus on three key ion channels that are well-established to be linked to LQTS and that are illustrated in Figure 1B. The cellular resting membrane potential is approximately −85 mV, determined largely by inwardly rectifying K⁺ channels. Inward rectification is a property that hinders the outward flow of K⁺ as membrane potentials become positive, but passes K⁺ more efficiently at potentials negative to the K⁺ equilibrium potential where the flow of K⁺ would be inward. Hence, the term “inward rectification” is used to describe these channels (304). Activation of Na_v1.5, the primary voltage-gated sodium channel in the heart, leads to sodium influx (I_Na) and membrane depolarization. As the cell reaches approximately −40 mV, voltage-gated L-type calcium channels begin to open, leading to calcium influx (I_Ca) (299). Concurrently during the upstroke of the action potential, potassium channels, including those carrying the I_Ks and I_Kr delayed rectifier currents, begin to activate slowly. As the cell reaches +30 mV, I_Na inactivates almost completely. At this time, a brief and small repolarization of the membrane potential occurs via fast activation of the transient outward K⁺ current, I_to. In the plateau phase that follows, influx of Ca²⁺ through voltage-gated calcium channels is balanced mainly by I_Ks and I_Kr. The plateau phase ceases as calcium channels inactivate and outward potassium efflux persists, leading to a net outward membrane current, and cell repolarization back to the cellular resting potential. During this process, the Na⁺-K⁺-ATPase helps maintain intracellular concentrations of these key ions (304).

In LQTS patients, the QT is prolonged presumably due to prolongation of underlying action potential durations (APDs), most often caused by decreased repolarizing I_Ks or I_Kr activity, or persistent sodium influx that extends through the plateau phase. A loss of I_Ks or I_Kr function, or a gain of I_Na function, predisposes ventricular myocytes to early afterdepolarizations (EADs), and in some cases to delayed afterdepolarizations (DADs) which may underlie degeneration into a characteristic sinusoidal wave pattern on the ECG, referred to as torsades de pointes, which may further regress into ventricular fibrillation and sudden cardiac death. EADs are driven in large part by calcium entry via L-type calcium channels during prolonged action potential plateau phases, whereas DADs, which occur over the diastolic range of potentials after action potential repolarization, are caused by intracellular calcium overload, also a consequence of action potential prolongation (126). Additionally, arrhythmic activity may result from altered refractoriness and impulse block, also putative consequences of prior APD prolongation. Thus treatment in LQTS aims to prevent malignant ventricular arrhythmia by shortening the QTc interval to minimize cardiac event rates.

Ion channels may interact with a variety of molecular entities that contribute to their trafficking, stabilization, signaling, and function (299). Coordination among different ion channel types facilitates the ionic balance necessary for the generation of an action potential and normal electrical propagation through the heart. LQTS mutations may cause an increase or decrease in ion channel function, disrupting normal ionic balance leading to pathological electrical activity in the heart.

There are 15 subtypes of congenital LQTS, each associated with mutations on a different gene (15) (Table 1). The most common subtypes, LQT1, LQT2, and LQT3, account for the vast majority of congenital LQTS. LQT1 and LQT2 are associated with mutations in KCNQ1 and KCNH2 (which encodes hERG), respectively, while LQT3 is associated with mutations in SCN5A, the gene coding for the Na_v1.5 sodium channel alpha subunit. Disease association for variants in these three proteins is supported by genome-wide association studies (300) and functional electrophysiological characterization of mutant channels. In addition, LQTS-associated mutations exist less frequently in other ion channels, modulatory channel subunits, and signaling- or cytoskeleton-associated proteins. Understanding the molecular mechanisms that cause LQTS allows for optimization of genotype-specific treatments. In this review, we discuss the molecular physiology, biology, and pathophysiology underlying congenital LQTS, and the cellular and molecular underpinnings of genotype-driven clinical management of LQTS.

Table 1.

Subtypes of congenital LQTS and their associated genes, proteins, and effects on cardiac currents

LQT Subtype	Gene	Protein	Current
LQT1	KCNQ1	KCNQ1 (Kv7.1)	↓IKs
LQT2	KCNH2	hERG (Kv11.1)	↓IKr
LQT3	SCN5A	Nav1.5	↑INa
LQT4 (ankyrin-B syndrome)	ANK2	Ankyrin-B	Multichannel interactions
LQT5	KCNE1	KCNE1 (minK)	↓IKs
LQT6	KCNE2	KCNE2 (MiRP1)	↓IKr
LQT7 (Andersen-Tawil syndrome type 1)	KCNJ2	Kir2.1	↓IK1
LQT8 (Timothy syndrome)	CACNA1C	Cav1.2	↑ICa
LQT9	CAV3	Caveolin 3	↑INa
LQT10	SCN4B	Nav1.5 β4	↑INa
LQT11	AKAP9	AKAP-9 (yotiao)	↓IKs
LQT12	SNTA1	α1-Syntrophin	↑INa
LQT13	KCNJ5	Kir3.4 (GIRK4)	↓IKACh
LQT14	CALM1	Calmodulin	Multichannel interactions
LQT15	CALM2	Calmodulin	Multichannel interactions

II. I_Ks DYSFUNCTION IN CONGENITAL LQTS

A. Introduction

Of the different subtypes of inherited LQTS, subtypes 1, 5, and 11 are associated with mutations in proteins that participate in the macromolecular complex which generates and modulates the slow delayed rectifier potassium current (I_Ks), which plays a critical role in the repolarization of the cardiac action potential. Among all LQTS subtypes, LQT1 is the most common, representing 30-35% of all congenital LQTS (8). Upregulation of I_Ks during β-adrenergic stimulation is critical to normal physiology by shortening ventricular APD and allowing for adequate diastolic filling in the context of an elevated heart rate (354). Cardiac events in patients with I_Ks-associated LQTS are often triggered by stress and exercise, consistent with the role of adrenergic stimulation in the regulation of I_Ks (350, 371). Insight into the molecular mechanisms of disease mutations have greatly improved our understanding of LQTS pathophysiology and helped provide a first step to the future development of targeted therapies.

B. Physiology

I_Ks is an outward potassium current with unique kinetics and voltage dependence and plays a key role in the repolarization of the cardiac action potential (299). In 1969, the delayed rectifier potassium current in sheep Purkinje fibers was studied and shown to comprise two kinetically distinct components (305). These currents were later subjected to pharmacological dissection in guinea pig ventricular myocytes and identified as I_Ks and I_Kr (360).

I_Ks is slowly activating and most prominent during the plateau and repolarizing phases of the cardiac action potential, where it contributes to counterbalancing calcium influx and repolarization (Figure 1B). Expression of I_Ks has been demonstrated in both human atrial and ventricular myocytes (198, 235, 457). In addition, it has also been measured in cardiomyocytes from a variety of non-human mammalian species including dogs (241, 406, 437, 442, 493) and rabbits (352). On the other hand, I_Ks is expressed at very low levels or absent in mouse hearts (478), most likely because at very high heart rates in mouse heart (∼500 beats/min) this channel would have little or no time to activate and not affect cardiac electrophysiology in the mouse.

Importantly, I_Ks is subject to upregulation by β-adrenergic stimulation to control APD in the face of sympathetic nerve activity (210, 444). During sympathetic activation, adrenergic stimulation increases the outward I_Ks current, which counterbalances the concomitant increase in inward calcium current, prevents prolongation of the cardiac APD, and allows for adequate diastolic filling times between heart beats (386). However, insufficient I_Ks activation such as that seen in LQT1 results in failure to counterbalance the calcium influx, prolonging the action potential and increasing susceptibility to arrhythmia. This is consistent with exercise being a key trigger of cardiac events seen in LQT1 patients (371).

C. Molecular Biology

The first known subtypes of inherited LQTS, LQT1-3, were distinguished by mapping to distinct chromosomes, with LQT1 mapping to chromosome 11 (191, 212, 213). Eventually it was found that the KCNQ1 (KvLQT1) gene is responsible for LQT1 and encodes a potassium channel (451). The current conducted by this channel was rapidly activating and minimally inactivating, unlike any previously known current in the heart, but soon it was shown that KCNQ1 together with the accessory protein KCNE1 (minK) generates I_Ks (36, 358). While KCNE1 had previously been thought to be an independent potassium channel (167, 412), it was confirmed that KCNQ1 is actually the α- or pore-forming subunit of I_Ks while KCNE1 is a critical β- or modulatory subunit. Coexpression of KCNQ1 and KCNE1 generates the hallmark I_Ks current with slow activation. KCNQ1 and KCNE1 have been shown to be expressed in all four chambers in the heart (45), as well as the inner ear (301, 351), where I_Ks is thought to play a role in K⁺ secretion into the endolymph (68). This explains the observation that congenital deafness is a key feature of JLNS. In addition, KCNQ1 and KCNE1 are expressed elsewhere in the body, including the pancreas, the kidneys, and the brain (1).

1. KCNQ1, the pore-forming subunit

KCNQ1, like most voltage-gated potassium channels, consists of six transmembrane helices (451) (Figure 3A). Four subunits of KCNQ1 come together to form a channel that is capable of voltage-dependent gating (Figure 3B). On each subunit, the helices S1–S4 comprise the voltage-sensing domain, where the S4 helix, with its positively charged arginine residues, senses changes in membrane potential (310). Following the voltage-sensing domain is the pore domain, which comprises the pore-loop, an extracellular segment containing the selectivity filter, and helices that line the pore, S5 and S6. Furthermore, the cytoplasmic loop between S4 and S5 plays important roles in the voltage sensor-to-pore coupling and in voltage-dependent gating (82, 229, 498), which has been demonstrated in other voltage-gated channels as well (75, 122, 355). The cytoplasmic loop between S2 and S3 also plays a role in channel gating (498). The COOH-terminal domain (CTD) of KCNQ1 is large and contains four intracellular α-helices referred to as A-D. A wide range of functions has been attributed to the CTD including calmodulin binding, interaction with β-subunits and scaffolding proteins, as well as channel assembly and trafficking (159, 469).

FIGURE 3.

Molecular biology of I_Ks and regulation by PKA-mediated signaling. A: the I_Ks macromolecular complex, including KCNQ1, KCNE1, and associated scaffolding and signaling proteins. B and C: single pulse voltage-clamp recordings of KCNQ1 expressed alone or coexpressed with KCNE1 in Xenopus oocytes. D: dialysis with 200 μM cAMP and 0.2 μM okadaic acid (OA) increases I_Ks amplitude and slows deactivation when heterologously expressed in CHO cells. [From Chen et al. (78).]

2. KCNE1, the β-subunit

Coexpression of KCNE1 with KCNQ1 leads to a drastic change in channel function to generate I_Ks. Most prominently, assembly with KCNE1 leads to a delay in the onset of activation, an increase in channel amplitude (Figure 3C), as well as a depolarizing shift in the current-voltage relationship (not illustrated) (36, 358). This results in a channel that, compared with most other voltage-gated potassium channels, activates at more positive voltages and with slower kinetics. KCNE1 is a 129-amino acid protein that consists of a single transmembrane helix, with an extracellular NH₂ terminus and intracellular COOH terminus (412). It is thought to have extensive contact with KCNQ1 including the voltage-sensing domain (37, 84, 203, 309, 378, 479), the pore domain (84, 262, 311), as well as the CTD (161, 510). Previous crosslinking studies suggest that KCNE1 is located in a cleft between the voltage-sensing domain and pore domain of different KCNQ1 subunits (84, 479), underlying its ability to modulate KCNQ1 gating. With respect to the stoichiometry of KCNE1 to KCNQ1, some studies suggest a fixed 2:4 ratio (278, 326), while others suggest a flexibility in stoichiometry (293, 456) that allows for modulation of kinetics of assembled channels to provide another level of flexibility in channel function. Although three other members of the KCNE family, KCNE2-KCNE4, also are expressed in the heart (45) and are capable of modulating KCNQ1 activity (45, 154, 367, 422), whether they associate with KCNQ1 in vivo to contribute to potassium currents in the heart remains to be explored.

3. Molecular components of adrenergic stimulation

There have been considerable efforts to elucidate the molecular pathway for the β-adrenergic regulation of I_Ks. In 1988 it was shown that stimulation of PKA activity by a cAMP analog can upregulate the delayed rectifier current (444). Later it was shown that the scaffolding protein A-kinase anchoring protein 9 (AKAP-9), also known as yotiao, plays a central role in adrenergic regulation of I_Ks by compartmentalizing key elements of the PKA signaling pathway, allowing for spatiotemporal control. AKAP-9 binds to the CTD of KCNQ1 and recruits signaling proteins including protein kinase A (PKA), protein phosphatase 1 (PP1) (255), adenylyl cyclase 9 (AC9) (238), and the phosphodiesterase PDE4D3 (419) (Figure 3A). Together these proteins form the I_Ks macromolecular complex that can tightly control the phosphorylation state of the channel in response to adrenergic stimulation. PKA phosphorylates KCNQ1 at the S27 residue, adding a phosphate group and hence a change in charge to this residue, which leads to increased channel activation and slower deactivation (226, 255) (Figure 3D). In addition, phosphorylation of AKAP-9 itself contributes to the PKA-mediated upregulation of I_Ks (78).

4. Role of PIP₂ and calmodulin in I_Ks function

The lipid molecule phosphatidylinositol 4,5-bisphosphate (PIP₂) is critical for KCNQ1 function. PIP₂ is found in the inner leaflet of plasma membranes (260) and can regulate a variety of ion channels (51, 83, 172, 179, 250, 475). PIP₂ mainly binds to KCNQ1 at its cytoplasmic loops and the COOH-terminal region near S6 (80, 208, 498), which are thought to form the interface between the voltage-sensing domain and the pore domain (82, 497). PIP₂ binding is mediated by electrostatic interactions between the anionic lipid headgroup and positively charged channel residues (418). Rundown of PIP₂ in the membrane leads to a drastically reduced I_Ks amplitude and accelerated deactivation, suggesting that PIP₂ stabilizes the open state of I_Ks (250). Furthermore, PIP₂ rundown reduces the current amplitude of KCNQ1 in the absence of KCNE1 (239). In addition, PIP₂ plays a critical role in the coupling between voltage sensor movement and pore opening/closing (498). Interestingly, one modulatory effect of KCNE1 on KCNQ1 is a large (100-fold) increase in channel sensitivity for PIP₂, which contributes to augmentation of KCNQ1 current amplitude (239). Whether PIP₂ level is under control by physiological mechanisms to modulate I_Ks in cardiomyocytes remains to be demonstrated.

Calmodulin is a calcium-binding protein that serves in myriad calcium signaling pathways, some of which regulate ion channels (399). Calmodulin has been linked to I_Ks function (141, 379). Multiple studies have shown that calmodulin binding to the CTD of KCNQ1 is important for normal channel trafficking and folding (141, 379).

5. Channel trafficking

A variety of molecular entities are involved in the trafficking of KCNQ1KCNE1. The small GTPase-RAB11 plays an important role in the exocytosis of KCNQ1-KCNE1 to the plasma membrane, while RAB5 mediates endocytosis of KCNQ1-KCNE1 from the plasma membrane into endosomes (374). In a physiological stress response, serum and glucocorticoid-regulated kinase 1 (SGK1) upregulates KCNQ1-KCNE1 expression by increasing RAB-11-dependent channel exocytosis (374). There is also evidence to suggest that KCNQ1-KCNE1 can be internalized from the plasma membrane via clathrin-mediated endocytosis, a process facilitated by KCNE1 (480). Furthermore, KCNQ1 expression is regulated by ubiquitination, which can label membrane proteins for internalization and degradation. For example, the ubiquitin ligase Nedd42 can ubiquitinate KCNQ1 (190), while ubiquitin-specific protease 2 (USP2) can prevent channel ubiquitination (221). Both processes allow for control of KCNQ1 expression.

D. Molecular Pathophysiology

There are more than 530 disease-causing mutations associated with the I_Ks complex (15). While most of these are missense mutations, they also include nonsense mutations, splice site mutations, frameshifts, as well as deletions. The vast majority are mutations in KCNQ1, which cause LQT1, but a number of LQTS-causing mutations have also been found in KCNE1 and AKAP-9, which are classified as LQT5 and LQT11, respectively. JLNS, an autosomal recessive form of LQTS with severe bilateral sensorineural deafness, has so far only been associated with mutations in KCNQ1 or KCNE1 (423). Nonetheless, the majority of LQTS arising from mutations in I_Ks are autosomal dominant in a form known as Romano-Ward syndrome, which is not associated with deafness.

Congenital LQTS patients with dysfunction in I_Ks present with prolongation of the QT interval on ECG. An ECG characteristic associated with LQT1 patients more specifically is a broad based T wave (Figure 4A). Dysfunction of I_Ks reduces repolarizing current during the plateau phase, thereby leading to APD prolongation (Figure 4B). With a prolonged APD, the cardiomyocyte is vulnerable to EADs (441, 500), where a second depolarization occurs prior to the complete repolarization of the first due to recovery of voltage-gated Na⁺ or Ca²⁺ channels from inactivation, triggering life-threatening arrhythmia such as torsades de pointes (26, 464). As previously mentioned, during adrenergic stimulation I_Ks plays an especially important role in controlling the cardiac APD. This manifests clinically in LQT1 patients as a susceptibility to arrhythmia during exercise (371).

FIGURE 4.

I_Ks dysfunction leading to congenital LQTS. A: ECG from a LQT1 patient demonstrates a characteristic broad-based T wave (unpublished data). B: simulated action potential (top) and I_Ks (bottom) in WT (black) and heterozygous LQT1 (gray) conditions, demonstrating the effect of 50% reduction in I_Ks. C: single pulse voltage-clamp I_Ks recording in Xenopus oocytes demonstrating loss of channel function in an I_Ks-associated LQTS mutation. D: topology of KCNQ1 and KCNE1 in the plasma membrane. Several examples of LQT1- and LQT5-associated mutations are highlighted in blue and teal, respectively. These mutations represent a variety of mechanisms of loss-of-function including disruption of permeation (yellow-filled square), gating (yellow-filled circle), trafficking (yellow-filled triangle), PKA-mediated signaling (yellow-filled star), KCNQ1-KCNE1 interaction (white-filled square), PIP₂ affinity (white-filled circle), and calmodulin affinity (white-filled triangle).

1. LQT1

Missense mutations are responsible for the majority of LQT1 cases and can cause channel loss of function through a variety of molecular mechanisms, including defects in ion permeation (altering the pathway through which ions flow through open channels), channel gating (mechanisms that regulate the opening and/or closing or channels), trafficking, KCNQ1-KCNE1 interaction, PKA-mediated signaling pathway, PIP₂ binding, and calmodulin binding (Figure 4, C AND D) (Table 2). Non-missense mutations can also cause LQT1. Mutations belonging to certain groups may bear implications on patient phenotype, severity of arrhythmia, as well as response to therapy.

Table 2.

Representative LQT1-associated mutations classified by mechanism

Mechanism	Mutations	Reference Nos.
K+ permeation	G314S, Y315C	52, 237
	T322A, T322M, G325R	13, 61
Gating	D202H	113, 114
	S225L	52, 170
	R231C	41
	A344V	391
Trafficking	Y111C, L114P, P117L	98, 375
	T587M, R591H, R594Q	205
PKA-mediated signaling	G189R, R190Q, R243C, V254M	39
	A341V	169
	G589D	255
KCNQ1-KCNE1 interaction	S546L, K557E	110
PIP2 affinity	R539W, R555C	312
	S546L, K557E	110
Calmodulin affinity	S373P, W392R	379
Large-scale defects
Nonsense	R518X, Q530X	181
Deletion-insertion, frameshift	Δ544	81
Splice site	1032G>A	286

A) PERMEATION DEFECT.

A number of LQT1 mutations that lead to defects in permeation are found in the pore region of KCNQ1. Pore mutations are thought in general to carry greater risk of cardiac events than other mutations. One study shows that mutations in highly conserved residues, many of which are located in the pore-loop, are associated with higher risk of cardiac events (61, 197). Three LQT1 mutations near the selectivity filter, T322M, T322A, and G325R, have been shown to abolish channel conductance (13, 61). They exert dominant negative effects on wild-type (WT) I_Ks current in heterologous expression systems (61), suggesting that mutant subunits coassemble with WT subunits to form disrupted channels. Molecular dynamic simulations suggest that these mutations disrupt the conformation of the selectivity filter, leading to diminished K⁺ permeation. A pair of adjacent LQT1 pore mutations, G314S and Y315C, located in the selectivity filter, dramatically reduce I_Ks current amplitude and exert dominant negative effects on WT currents (52, 237). Immunofluorescence studies show that Y315C traffics to the membrane normally, suggesting that the mutation results in trafficking of non-conducting channels.

B) GATING DEFECT.

LQT1 mutations that cause defects in KCNQ1 gating can be found in the pore domain (S5–S6). Similar to other voltage-gated potassium channels, the COOH-terminal region of the S6 helix plays a key role in KCNQ1 gating (55, 157, 247). Scanning mutagenesis and heterologous expression studies show that a number of LQT1-linked residues in this region, such as F351 and L353, control KCNQ1 gating properties. For example, F351A leads to a drastic slowing of channel activation and a depolarizing shift in voltage dependence of activation, while L353K leads to a constitutively open channel (55). To further elucidate the mechanism of F351A, a technique known as voltage clamp fluorometry (VCF), which utilizes fluorophore labeling to allow simultaneous measurement of voltage sensor movement and channel current, has been used (37, 292, 308, 309, 348, 498). It was shown using VCF that F351A alters the coupling between voltage sensor and the pore (309), resulting in a slowly activating channel that partially resembles I_Ks. In addition, a LQT1 mutation in on the S6 helix, A344V, also affects channel gating by shifting the current-voltage relationship of I_Ks in the depolarizing direction by 30 mV, thereby destabilizing channel opening (391).

In addition to those in the pore domain, a number of LQT1 mutations that alter channel gating are found on the voltage-sensing S4 helix of KCNQ1. For example, the mutation S225L exerts dominant negative suppression of WT I_Ks current (52). This mutation alters channel gating, shifting the current-voltage relationship of I_Ks toward more depolarized membrane potentials (170). The effect of this mutation on voltage sensor movement remains to be determined, but given its location it is possible that it disrupts the movement of the S4 helix in response to changes in voltage. Another LQT1-associated mutation on S4, R231C, decreases peak I_Ks amplitude, but it also leads to constitutive activation at the same time (41). Interestingly, in one family, this mutation causes familial atrial fibrillation, which is more consistent with action potential shortening than prolongation (41, 296). To explain this finding, the study has used a computational model to show that the atrial action potential is more susceptible to shortening than ventricular action potential when I_Ks is made constitutively active (41). Structural and mutational studies suggest that R231 forms a salt-bridge interaction with E160 in S2 that stabilizes the channel in its closed state (340, 393, 472). Mutating R231 may disrupt this interaction and cause a defect in channel gating, resulting in constitutive activation. One tool that can be used to elucidate the effects of these S4 mutations on voltage sensor movement is VCF. The technique is capable of providing insights on I_Ks channel gating not possible with current measurement alone.

LQT1 gating mutations are also present in other transmembrane regions (S1–S3) of the voltage-sensing domain. For example, the mutation D202H leads to biophysical defects in I_Ks, shifting the current-voltage relationship to more positive potentials, slowing activation, and accelerating deactivation, all leading to reduced channel opening (114). Single-channel recordings of I_Ks utilized as a tool to study effects of select mutations (113) shows that D202H causes a decrease in channel open probability, a decrease in open states dwell time, and an increase in closed states dwell time, while maintaining single-channel conductance.

C) TRAFFICKING DEFECT.

In addition to altering channel gating, LQT1 mutations can also disrupt channel trafficking. Several LQT1 mutations in helix D of the CTD, including T587M, R591H, and R594Q, have been found to diminish channel trafficking to the membrane (205). These mutations are thought to disrupt a coiled-coil motif in helix D that plays an important role in channel trafficking (205). Yet trafficking mutations are not limited to the CTD of KCNQ1. LQT1 mutations in the NH₂-terminal region of KCNQ1, including Y111C, L114P, and P117L, have also been found to reduce surface expression of KCNQ1 and increase retention in the endoplasmic reticulum (98). This region appears to be a conserved trafficking determinant across the KCNQ family. In particular, Y111C and L114P disrupt SKG1's ability to increase channel trafficking to the plasma membrane, suggesting that the NH₂ terminus may be important in RAB-mediated exocytosis of channels (375).

D) DEFECT IN PKA-MEDIATED SIGNALING.

Given the critical importance of adrenergic stimulation and PKA-mediated signaling in the regulation of I_Ks and repolarization of the cardiac action potential during stress, disruption of this stimulation is expected to prolong the QT interval. For example, the mutation G589D is thought to disrupt a leucine zipper motif to which AKAP-9 binds, resulting in failure of I_Ks to be stimulated by cAMP (255). A mutation on S6, A341V, exhibits dominant suppression of I_Ks that fails to respond to adrenergic stimulation with cAMP (169). This effect is mediated by a reduction in the phosphorylation S27 on KCNQ1 and is not due to disruption in AKAP-9's interaction with KCNQ1. This result implicates a role of S6 in the regulation of the phosphorylation state of KCNQ1.

Mutations that lead to defective adrenergic stimulation of I_Ks in LQT1 patients are suggested to be associated with higher risk of cardiac events during exercise and greater response to β-blocker therapy. One study has identified missense mutations in the cytoplasmic loops between S2/S3 and S4/S5 of KCNQ1 to be associated with an elevated risk of aborted cardiac arrest and sudden cardiac death (39). Four mutations in these cytoplasmic loops, G189R, R190Q, R243C, and V254M, all diminish I_Ks upregulation in response to forskolin, an activator of the PKA signaling pathway, suggesting that patients with these mutations are expected to be especially susceptible to arrhythmic events during stress. This study bears implications on better risk-stratification for LQT1 patients and predicting response to β-blocker therapy. In addition, it suggests that the cytoplasmic loops are involved in adrenergic stimulation of KCNQ1.

E) DISRUPTED KCNQ1-KCNE1 INTERACTION.

A number of LQT1 mutations disrupt the KCNQ1-KCNE1 interaction. Two such mutations, S546L and K557E, are located in the helix C of the CTD of KCNQ1 and disrupt its interaction with the COOH terminus of KCNE1, as demonstrated by GST pulldown assays (110). The same mutations also decrease channel affinity for PIP₂, leading to decreased current amplitude and a depolarizing shift in the current-voltage relationship. The decreased PIP₂ affinity may be due to disruption of a potential PIP₂ binding site on helix C.

F) DECREASED PIP₂ AFFINITY.

Some LQT1 mutations have been shown to decrease channel PIP₂ affinity. In addition to the mutations described above, two LQT1 mutations on helix C of the CTD, R539W and R555C, both lead to decreased PIP₂ affinity, decreased I_Ks amplitude, and a depolarizing shift in the current-voltage relationship, underscoring the importance of the CTD in PIP₂ binding (312). The cytoplasmic loops of KCNQ1 between S2/S3 and S4/S5 also play important roles in PIP₂ binding and channel gating (208, 498). Several LQT1-linked residues in these cytoplasmic loops, including residues 195, 258, and 259, are thought to form a binding site for PIP₂. Disease mutations in this binding site may result in decreased PIP₂ affinity, leading to altered channel function (80, 498).

G) DECREASED CALMODULIN AFFINITY.

A number of LQT1 mutations have been found to weaken calmodulin binding to KCNQ1 as the underlying mechanism of disease. For example, the mutations S373P and W392R, located in the CTD, reduce calmodulin binding both in the absence and presence of KCNE1 (379). Both of these mutations cause decrease in surface channel expression and dramatic reduction in I_Ks amplitude. Overexpression of calmodulin is able to increase S373P mutant channel expression in the membrane. These results are consistent with the role calmodulin plays in channel assembly and trafficking (141, 379).

H) NON-MISSENSE MUTATIONS.

Non-missense mutations can also cause LQT1. For example, nonsense LQT1 mutations such as R518X and Q530X introduce a stop codon, leading to early termination of channel transcription and loss of I_Ks function (181). These mutations are mostly associated with autosomal recessive LQTS, although autosomal dominant cases have also been reported for R518X. One study shows that these mutant channels only mildly affect WT I_Ks current, consistent with their mostly recessive mode of inheritance (349). It is thought that the nonsense transcripts are selectively degraded and do not interfere with WT channel production. Interestingly, the same study suggests that LQT1 patients with nonsense mutations have reduced risk of cardiac events compared with patients with missense noncytoplasmic loop mutations, although the explanation remains to be determined.

Another non-missense LQT1 mutation is the deletion-insertion at residue 544, denoted Δ544, which leads to a frameshift that alters subsequent 107 amino acids and introduces an early stop codon (81). It is an autosomal recessive mutation occurring in the CTD. The mutation has been shown to disrupt channel assembly in vitro (364), underscoring the role of the CTD of KCNQ1 in channel assembly.

LQT1 mutations can also disrupt transcript splicing. A base substitution at a consensus splice donor site the end of exon 7, 1032G>A, can lead to a dropped exon 7 or exons 7 and 8 (286). In terms of channel regions affected, exon 7 encodes parts of the pore-loop and S6, while exon 8 encodes the rest of S6 and part of the CTD. Transcripts lacking one or both of these exons therefore are not expected to produce functioning channels. A study shows that this splicing mutation exerts a dominant-negative suppression of WT I_Ks, likely through a direct interaction between mutant and WT channels preventing trafficking to the membrane (428). In addition to splice donor sites, splice acceptor sites can also be mutated in LQT1, such as the base substitution 922-1 G→C at the end of intron 6, which leads to the loss of exons 7 and 8 (286).

2. LQT5

Coassembly of KCNQ1 with the KCNE1 subunit is key to the generation of the I_Ks current. Thus it follows that KCNE1 mutations can alter the physiologically critical current of the assembled channel and lead to LQT5. Similar to LQT1, the mode of inheritance for LQT5 can be autosomal dominant (RW) or recessive (JLNS) (107, 430). Mutations in KCNE1 can lead to defects in gating, trafficking, KCNQ1-KCNE1 interaction, as well as adrenergic stimulation (Figure 4D) (Table 3).

Table 3.

Representative LQT5-associated mutations classified by mechanism

Mechanism	Mutations	Reference Nos.
Gating	D76N	76, 405
Trafficking	L51H	53
KCNQ1-KCNE1 interaction	T58P/L59P	166, 180
	P127T	110
PKA-mediated signaling	D76N, P127T	110, 226

A) GATING DEFECT.

One LQT5 mutation that affects channel gating is D76N. It is an autosomal dominant mutation in the COOH terminus of KCNE1 that has been shown to drastically suppress I_Ks amplitude, accelerate deactivation, and cause a depolarizing shift in the voltage dependence of I_Ks activation when expressed in Xenopus oocytes and CHO cells (76, 405). Overexpression of the mutant KCNE1 in guinea pig ventricular myocytes leads to APD prolongation and early afterdepolarizations (176). The mutant appears to traffic normally to the cell surface (53) and does not disrupt KCNE1 binding to the COOH terminus of KCNQ1 (510). However, it reduces I_Ks upregulation secondary to stimulation of the PKA signaling pathway, suggesting a role for the COOH terminus of KCNE1 in adrenergic stimulation of I_Ks (226).

B) TRAFFICKING DEFECT.

In addition to channel gating, LQT5 mutations can lead to defects in channel trafficking. The JLNS mutation L51H, located in the transmembrane helix of KCNE1, results in diminished KCNE1 trafficking to the cell surface (53). Furthermore, when coexpressed with KCNQ1 in HEK cells, the mutant KCNE1 decreases KCNQ1 trafficking to the surface (53, 220). In addition, coexpression of the mutant KCNE1 with KCNQ1 in CHO cells leads to a diminished current amplitude and channel biophysical properties that resemble KCNQ1 alone rather than I_Ks, consistent with a drastic reduction of functional KCNE1 in the membrane. These results together suggest that the mutant KCNE1 interacts with KCNQ1 to disrupt the trafficking of both proteins to the membrane surface, leading to a reduction in current. The transmembrane region of KCNE1 may therefore be important to channel trafficking.

C) DISRUPTED KCNQ1-KCNE1 INTERACTION.

LQT5 mutations can disrupt the interaction between KCNE1 and KCNQ1 required to generate I_Ks. For example, the double mutation T58P/L59P, located in the transmembrane region of KCNE1, results in near-complete loss of I_Ks amplitude but has minimal effect when coexpressed with WT I_Ks in Xenopus oocytes (181). The mutation leads to a diminished ability for KCNE1 to associate with KCNQ1 in coimmunoprecipitation studies, suggesting that the transmembrane region of KCNE1 is important for the KCNQ1-KCNE1 interaction (166). Furthermore, the mutation P127T, located in the COOH-terminal region of KCNE1, appears to disrupt the interaction of KCNE1 with helix C in the CTD of KCNQ1 (110). Interestingly, the mutation was also found to diminish PKA-stimulated upregulation of I_Ks by decreasing phosphorylation at the S27 residue. Since PKA phosphorylation at this site was previously shown to be independent of KCNE1 (226), it is possible that the disruption in adrenergic stimulation by P127T is independent of the mutation's disruption of KCNQ1-KCNE1 interaction (110).

3. LQT11

AKAP-9 is a scaffolding protein and part of the I_Ks macromolecular complex that plays a critical role in the compartmentalization of adrenergic-stimulated PKA signaling pathway leading to I_Ks upregulation. That a mutation in AKAP-9, S1570L, can cause LQTS is a testament to the critical role of PKA signaling in the regulation of I_Ks macromolecular complex (79). This mutation is located near the COOH-terminal binding domain of AKAP-9, disrupting its interaction with KCNQ1, reducing cAMP-stimulated phosphorylation of KCNQ1, and abolishing I_Ks upregulation in response to cAMP-mediated stimulation. Computational modeling suggests that a disruption in the basal phosphorylation state of I_Ks alone can alter I_Ks function sufficiently to prolong APD (79).

E. Molecular Pharmacology

1. β-Blockers

β-Blockers have been demonstrated as a particularly effective form of therapy for LQT1 patients, who are more sensitive to stress- and exercise-induced arrhythmia than other LQT subtypes (282, 371). Insufficient upregulation of I_Ks by adrenergic stimulation to counterbalance concomitant rise in inward calcium current is thought to underlie APD prolongation in LQT1 during sympathetic activation (386). β-Blockers antagonize adrenergic receptors and helps prevent this imbalance between potassium and calcium currents, decreasing predisposition for cardiac arrhythmic events.

2. Channel activators

While I_Ks plays a critical role in cardiac repolarization, it is not a direct target of drugs currently used to treat LQTS. However, conceptually, I_Ks activators could allow for more precise rescue of disease phenotype. In this section we briefly review I_Ks activators and refer readers to other studies in which activators of the ATP-sensitive K⁺ channel, such as nicorandil, have been used in LQTS patients and model systems (14, 388). Currently there are no I_Ks activators being used in clinical trials or therapy, but a number of small molecules that activate I_Ks have been identified at the benchside and may guide future development of therapeutic agents. For example, the compounds DIDS and mefenamic acid both increase I_Ks current amplitude (5). In addition, DIDS has been shown to drastically slow I_Ks deactivation. Interestingly, the effects of these drugs appear to be dependent on the presence of KCNE1. Compared with KCNQ1 alone, current augmentation by these compounds is much greater in the presence of KCNE1, suggesting their effects are mediated by the β-subunit. Indeed, deletion of the residues 39–43 of KCNE1 leads to a diminished response to these compounds. To demonstrate the potential for activators as a class of therapeutic agents for LQTS in in vitro studies, DIDS and mefenamic acid have been shown to rescue I_Ks function in a LQT5 mutation, D76N. Future studies will be required to better understand their mechanisms of action and to develop I_Ks-specific activators that can be effective and used safely in patients.

Another KCNQ1 activator, ML277, has been shown to augment current for KCNQ1 alone more effectively than KCNQ1 with KCNE1 (492). In fact, its activating effect diminishes with progressive increase in KCNE1:KCNQ1 stoichiometry, suggesting that the β-subunit may act to preclude the drug from binding to the channel (491). While a drug with such properties is expected to have minimal effectiveness in human cardiomyocytes, surprisingly the same study showed that the drug can augment I_Ks and shorten action potential in human iPSC-derived cardiomyocytes from a healthy control. The mechanism underlying this effect requires further elucidation.

3. Channel blockers

I_Ks-specific blockers are not useful as therapeutic agents for LQT1, but serve as useful tools for research by allowing for pharmacological dissection of I_Ks currents. Chromanol 293B is the first known I_Ks-specific blocker, with an IC₅₀ of 6.9 μM in Xenopus oocytes (62, 139). A more effective blocker, HMR1556, with an IC₅₀ of 0.12 μM, was developed using Chromanol 293B as the lead compound (139). These drugs inhibit KCNQ1 with KCNE1 more effectively than KCNQ1 alone. Blockers of I_Ks are speculated to serve useful roles in the treatment of disease conditions resulting from I_Ks gain of function, such as familial atrial fibrillation.

III. I_KR DYSFUNCTION IN CONGENITAL LQTS

A. Introduction

Alterations in I_Kr, the rapid component of the delayed rectifier current in the cardiomyocyte action potential (Figure 1), underlies congenital Long QT (LQT) syndromes type 2 and 6, which arise from mutations in the KCNH2 and KCNE2 genes, respectively. LQT2 is the second most common cause of congenital LQTS. KCNH2 mutations lead to defective hERG protein, resulting in a decrease in I_Kr. Mutations in KCNE2 cause defects in the KCNE2 (or MiRP1) protein leading to LQT6, which also results in a decrease in I_Kr. Irrespective of the underlying cause, a decrease in I_Kr delays repolarization of the cardiac action potential prolongs the QT interval on the ECG, and predisposes patients to lethal arrhythmia. This section reviews I_Kr dysfunction leading to congenital and I_Kr-mediated drug-induced LQTS. [For a detailed summary of hERG channel structure, molecular biology, and basic electrophysiology, see Vandenberg et al. (436).]

B. Physiology

Heterologous expression of hERG reveals a strong, near identical resemblance to I_Kr in cardiomyocytes (359, 427). I_Kr is distinguished based on its relatively slow activation and deactivation kinetics, combined with rapid inactivation and recovery from inactivation (Figure 5). Inactivation refers to a conformation of the channel protein in which the channels cannot conduct ions even if the activating machinery is in a conformation that would promote conduction of ions. Channels will not conduct ions when in an inactivated state, but will conduct ions after the channels recover from the inactivated state, a recovery that takes place at negative voltages (174). Deactivation refers to the transition to a conformation in which channels return to a closed, nonconducting state, a transition that also occurs at negative (diastolic) voltages. hERG channels undergo voltage-dependent and C-type inactivation. Because channel activation is slow relative to the rapidly occurring inactivation process, the hERG I–V curve takes on a bell-shaped relationship, as shown in Figure 5C. Upon repolarization of the membrane, channels recover from inactivation much faster than they deactivate. This crucial channel property results in a marked outward K⁺ current during the repolarization phase of the cardiomyocyte action potential. By this time, a large percentage of hERG channels have recovered from inactivation, such that outward potassium efflux helps return the cell to its resting potential, despite the fact that the electrochemical gradient for potassium efflux decreases as repolarization progresses (Figure 1B).

FIGURE 5.

hERG structure and electrophysiology. A: schematic of the I_Kr channel complex. Four hERG1 subunits tetramerize to comprise the pore-forming alpha subunit of I_Kr. hERG1 contains a voltage-sensing domain (purple), including the S4 helix which contains positively charged gating residues, and a pore domain (gray). KCNE2, an accessory subunit of the I_Kr channel complex, consists of a single transmembrane helix (blue). B: voltage-clamp protocol (top panel) and heterologously expressed hERG1 ionic currents (bottom panel) recorded from a Xenopus oocyte. Currents were recorded at potentials that ranged from −70 to +50 mV; deactivating (“tail”) currents were measured at −70 mV. C: current-voltage (I-V) relationship for hERG1 currents measured at the end of test pulses, as indicated by red circle in B. D: voltage dependence of hERG1 current activation. The peak of tail currents measured at −70 mV (indicated by blue square in B) were normalized to the largest value and plotted as a function of the test potential. E: voltage dependence of hERG1 inactivation. Channel availability is decreased at positive potentials, resulting in a decreased magnitude of peak outward currents and the bell-shaped I-V relationship depicted in C. [B–E from Sanguinetti (356), with permission of Springer.]

Since hERG inactivation and recovery from inactivation proceed more rapidly than activation or deactivation of hERG (365, 394, 401), I_Kr contributes to prolongation of the plateau phase duration and thus cardiomyocyte contraction, in addition to cardiac repolarization (361, 394). As hERG recovers from inactivation during cell repolarization, the repolarization itself promotes greater hERG recovery from inactivation due to the voltage dependence of hERG inactivation gating (359, 394, 427). As the cell continues to repolarize and return to its resting membrane potential, the slower process of channel deactivation progresses, leading to closure of hERG (436). The fraction of channels remaining open near the resting potential (see Figure 5E) acts to oppose cell depolarization (242, 394), which helps prevent premature heart beats from leading to an early action potential and a tachyarrhythmia. Loss of hERG function thus predisposes to arrhythmia in the setting of premature beats(46).

C. Molecular Biology

The KCNH2 gene, located on chromosome 7q35-36, encoding the human ether a go-go-related K⁺ channel protein (hERG), was first discovered in 1994 (460). Mutations in KCNH2 associated with congenital LQTS were discovered a year later in 1995 (97), and soon after, it was determined that hERG represents the α-subunits of the K⁺ channel responsible for I_Kr (359, 427). hERG may be referred to as “hERG1,” since other hERG proteins (hERG2 and hERG3) have since been discovered (385). hERG1a, the main hERG isoform present in cardiomyocytes, is 1,159 amino acids in length, with a predicted molecular weight of 127 kDa. As with most voltage-gated K⁺ channels, functional hERG channels are composed of four hERG α-subunits, forming either a hERG1a homomeric channel, or a hERG1a/hERG1b heteromeric channel that contributes to native cardiac I_Kr activity. The heteromeric assembly of hERG isoforms 1a and 1b has distinctively altered kinetics compared with hERG1a monomeric channels, including faster activation and deactivation. hERG1b is expressed in smaller amounts at the mRNA level in the heart compared with hERG1a, but nevertheless adds to the complexity of hERG channel regulation and potential therapeutic modalities (195, 196, 231, 323, 353, 426).

Each of the four hERG α-subunits consists of six transmembrane helices, S1–S6, as shown in Figure 5A. The voltage-sensing domain spans from S1 to S4; the S4 transmembrane segment of hERG contains the primary positively charged amino acids required for voltage-sensing and opening of the activation gate (325, 408, 505). The pore domain of the channel, spanning from S5–S6, harbors the K⁺ permeation pathway necessary for K⁺ conduction (Figure 5A) (105). The long cytoplasmic NH₂ terminus contains the PAS domain with a PAS-cap, and together these sequences make up the “EAG” domain that is conserved among EAG-related voltage-gated potassium channels(274). The long cytoplasmic COOH terminus of the channel contains the cyclic nucleotide-binding domain (cNBD) (56), as well as a more distal RXR endoplasmic reticulum retention signal (224), and a coiled-coil domain (188).

The NH₂-terminal PAS domain accelerates deactivation of hERG and plays a role in channel trafficking (77, 274, 436). Slow deactivation of hERG may involve interaction of the PAS domain in the NH₂ terminus of the channel, with the S4–S5 linker (450), which couples movement of the transmembrane voltage sensor to activation gate movement (247). Experimental hERG mutants with a deleted PAS domain have faster deactivation kinetics (274, 401). Proper folding of the PAS domain also leads to trafficking of hERG from the endoplasmic reticulum to the plasma membrane (314). The cNBD contributes to channel trafficking and gating as well, while cAMP binding to the cNBD domain results in changes in the gating kinetics of the channel, which are altered in the presence of KCNE2, an accessory hERG subunit (see below) (94). Furthermore, the PAS domain and cNBD appear to bind, working in concert to modulate hERG gating by positioning the NH₂-terminal residues in close proximity to the cytosolic side of S6 (23, 100, 160, 287).

hERG can coassemble with two different β-subunits in heterologous systems: KCNE1, encoded for by the KCNE1 gene; and KCNE2 (or the MiRP1 protein), encoded for by the KCNE2 gene (Figure 5A) (2, 259). KCNE1 and KCNE2 are single-pass transmembrane subunits that can interact with the hERG channel (2, 3, 20, 259). While the precise physiological role of the KCNEs in regulating hERG and I_Kr remains unclear (21, 257, 462), mutations in KCNE1 and KCNE2 lead to LQT5 and LQT6, respectively, and mutation of either KCNE1 or KCNE2 can predispose patients to drug-induced LQT syndrome, possibly via modulation of hERG activity (2, 9, 53, 295, 376, 402).

KCNE1 and hERG associate in heterologous systems and may contribute to regulation of I_Kr; however, KCNE1 is better characterized physiologically as the β-subunit of the KCNQ1 complex that produces I_Ks (36, 53, 259, 358). KCNE2 was reported to alter gating properties and drug responses of hERG channels(2), and mutations in KCNE2 leading to LQT6 likely result from changes in hERG and thus I_Kr current activity, producing a pro-arrhythmic state, further supported by the KCNE2 T10M mutation that confers an arrhythmogenic substrate to auditory stimuli, a known trigger of LQT2-associated arrhythmia (see Figure 7) (2, 151, 182, 251). It is possible that physiologically relevant KCNE2 expression exists only in the Purkinje fibers and pacemaker cells of the human heart (297, 330), which further confounds the exact role of KCNE2 as a β-subunit of hERG in vivo. Moreover, KCNE2 coassembles with KCNQ1, which decreases I_Ks current(192), adding to the complexity and diversity of I_Ks and I_Kr channel subunit interactions.

FIGURE 7.

Topology of hERG and KCNE1 in the plasma membrane, representative LQT2- and LQT6-associated mutations highlighted. Different mechanisms of loss of function in hERG or KCNE2, including gating (yellow-filled circle), K⁺ permeation (black-filled square), trafficking (white-filled triangle), or combined defects (green-filled star) are categorized.

D. Molecular Pathophysiology

A decrease in I_Kr results in LQT2 via mutations in the KCNH2 gene that encodes hERG, and LQT6 via mutations in the KCNE2 gene that encodes KCNE2 (or MiRP1) protein. The mechanisms underlying disease pathogenesis are described below.

1. LQT2

Patients with congenital LQT2 often present clinically after syncope or seizures not explained by a known medical condition. Prolongation of the QT interval on the ECG supports the diagnosis, and in LQT2 in particular, the T-wave may appear as a classic notched, or “bifid” T-wave (Figure 6A) (503). Syncope occurs following a common pathological process in the LQT syndromes, wherein delayed ventricular repolarization (and resultant QT interval prolongation on the ECG) precipitates an EAD, in some cases DADs, and ventricular tachycardia. EADs trigger the torsades de pointes sinusoidal waveform on the ECG, which may progress to ventricular fibrillation and sudden cardiac death (356).

FIGURE 6.

I_Kr dysfunction leading to congenital LQTS. A: ECG from a LQT2 patient demonstrates a characteristic “notched,” or bifid, T wave with QTc prolongation (unpublished data). B: simulated action potential (top) and I_Kr (bottom) in WT (black) and heterozygous LQT2 (gray) conditions, demonstrating the effect of 50% reduction in I_Kr.

Homozygous mutations in KCNH2 leading to LQT2 are extremely rare in humans, resulting in either death in utero, or very severe prolongation of the QT interval upon birth (175, 194). Heterozygous mutation leading to one defective hERG copy is far more common and may result in significant prolongation of the cardiac action potential duration (see Figure 6B). hERG mutations leading to LQT2 occur by a variety of genetic mechanisms. To date, ∼500 mutations in KCNH2 have been identified in association with LQT2 (23). A study analyzing 226 different LQT mutations in genotype-confirmed LQT2 patients reported that 62% of mutations were missense, 24% were frameshift, while 14% were a combination of nonsense mutations, inframe insertions/deletions, or splice site mutants in the KCNH2 gene. Of the combined 226 mutations, 32% resided in transmembrane and pore-pore domains; 29% in the NH₂ terminus, including 8% in the PAS/PAC domains; and 31% in the COOH terminus, including 8% in the c-NBD (207). In terms of molecular mechanism, a small proportion of LQT2 mutations result in nonsense-mediated mRNA decay (150). More commonly, decreased protein trafficking to the cell surface leads to eventual endoplasmic reticulum-associated degradation (ERAD) of mutant hERG (22, 149, 443, 512).

Several important general characteristics of KCNH2 mutations associated with congenital LQT2 were recently highlighted in a large-scale functional study of 167 different LQT2-associated missense mutations (23). Functional analysis was performed by voltage clamp after coexpression of WT and mutant subunits in HEK293 cells. Anderson et al. (23) found that 1) hERG trafficking defects comprised the most common (88%) mechanism of loss of function overall, derived from mutations in the PAS domain, pore domain, and C-linker/cyclic nucleotide-binding domain. Only the distal COOH-terminal region did not yield trafficking defects as the primary mechanism of loss-of-function. 2) Greater than 70% of pore mutations led to dominant negative suppression of hERG, whereas the other intracellular domain mutation locations did not yield dominant negative suppression of current. 3) All mutant channels, regardless of mutation location within the channel, were rescued pharmacologically by E4031, a hERG pore blocker that stabilizes channel structure and folding, thus promoting channel retention at the cell surface (148). Coexpression of WT with pore mutant channel rendered heteromeric pore mutant-WT channels especially responsive to pharmacological correction compared with coexpression of WT with non-pore mutant channels. Overall, these data suggest that pharmacological recovery of hERG channel function is feasible for a large proportion of LQT2 mutations(345) and raise the possibility of this approach as a therapeutic strategy for LQT-2 patients, but to our knowledge this has not been implemented in the clinic.

A) MUTATION SEVERITY: HAPLOTYPE INSUFFICIENCY VS. DOMINANT-NEGATIVE LOSS OF FUNCTION.

Heterozygous mutation resulting in defective KCNH2 gene expression or hERG protein that does not impact normal functioning of the WT hERG protein that remains yields haplotype insufficiency. Only the gene product from the mutant KCNH2 allele is negatively affected, while WT hERG subunits still homomerize to form functional channels. In contrast, some hERG mutations have increased pathogenicity by exerting a dominant-negative loss of channel function, wherein the mutant and defective KCNH2 gene product reduces function of the WT hERG protein encoded in the patient's genome via heteromerization of mutant with WT channel, rendering the healthy hERG subunits nonfunctional when combined with mutant hERG, either by decreasing forward trafficking of mutant WT channels, or decreasing function of the heteromers at the cell surface. In a study of LQT2 patients with 44 unique mutations in different regions of the hERG channel, it was discovered that those patients harboring mutations in the pore region of the channel were more susceptible to cardiac events than patients with non-pore region hERG mutations, likely due to a greater dominant-negative suppression of I_Kr current exerted by pore vs. non-pore mutations (187, 283).

Perhaps counterintuitively, more harmful mutations in one KCNH2 allele, resulting in a premature stop codon for instance, prevents formation of a hERG protein product from the mutant KCNH2 allele, leading to haplotype insufficiency and the possibility of a less severe phenotype compared with some dominant-negative mutations, as WT hERG remains unaffected(356).

Studies of specific hERG mutants have greatly enhanced our understanding of the underlying mechanisms of loss of function leading to LQT2. Figure 7 provides a schematic of representative LQT-associated hERG and KCNE2 mutations, categorized by mechanism of loss of function (see Table 4). Representative LQT-associated hERG mutations are divided into four general classes: I) decreased hERG synthesis, II) trafficking defect, III) gating defect, and IV) decreased K⁺ permeability (436).

Table 4.

Representative LQT2- and LQT6-associated mutations classified by mechanism

Mechanism	Mutations	Reference Nos.
hERG LQT2 mutations
K+ permeation	G628S	59, 116, 357
Gating	D16A, G584S, T613A, I711V, R835W	23, 329, 509
Trafficking	Y611H, V822M	512
Decreased hERG synthesis	R1014X, W1001X, Y652X	50, 150, 356, 409
Multiple mechanisms (e.g., gating and trafficking)	F29L, M124R	23, 206, 390
KCNE2 LQT6 mutations
Gating	T10M, V65M	151, 182

I) Decreased hERG synthesis. Defective biogenesis of hERG may occur via mRNA processing abnormalities or mRNA instability (150, 504). The hERG R1014X mutant mRNA transcript is degraded by “nonsense-mediated mRNA decay,” a cellular damage-control mechanism that destroys mRNA transcripts harboring nonsense mutations (premature stop codons), which prevents translation of a shortened hERG peptide (150, 356). Other nonsense-mediated decay mutants include W1001X (150) and Y652X (409). These mutations would result in haplotype insufficiency, as there would be 50% loss of function while the remaining WT hERG channels function normally. Nevertheless, a 50% reduction in I_Kr can lead to clinically significant LQTS.

II) hERG trafficking defects. Defective hERG trafficking is the most common mechanism of loss of hERG function (22, 23, 436), and most KCNH2 missense mutations cause trafficking defects in hERG (22). Trafficking mutants may result in either dominant negative loss of hERG function (22, 201) or haploinsufficiency (131). If WT hERG associates with trafficking-defective mutants to form heteromeric channels in the endoplasmic reticulum (ER) or Golgi apparatus, dominant negative suppression of I_Kr results: 93% reduction in I_Kr was observed by this mechanism (123).

Defects in trafficking may be further subdivided by the underlying mechanism, protein trafficking, or protein misfolding (22, 124, 512); hERG protein becomes core glycosylated in the ER, with modifications made in the Golgi apparatus. When the mature hERG protein traffics to the plasma membrane, it weighs 155 kDa, versus 135 kDa for the core-glycosylated-only hERG protein. This difference in molecular weight aids in determining whether hERG possesses a trafficking and/or protein folding defect (512). The hERG Y611H and hERG V822M LQT2 associated mutations were found to have a molecular weight of 135 kDa, and thus had decreased trafficking to the plasma membrane due to protein misfolding. These mutant hERG channels were retained in the ER, and subsequently ubiquitinated and degraded in proteasomes (512).

III) hERG gating defects. Alteration in activation, deactivation, and/or inactivation kinetics can result in loss of function of hERG and a decrease in I_Kr at physiologically relevant membrane potentials. Any hERG mutation that, for instance, enhances the speed of inactivation, or causes a depolarizing shift in channel activation, ultimately leads to loss of hERG function at voltages ranging from the resting potential to the plateau phase of the action potential, depending on the nature of the gating defect (48, 77, 290, 357, 363, 509). Several hERG gating defect mutations have been described in mammalian cell lines (48, 363, 509).

Both the I711V and R835W mutations represent examples of c-NBD region LQT2 mutations in the COOH terminus of the channel, with altered channel gating. These hERG mutants deactivate faster at −50 mV compared with WT hERG. In addition, the R835W mutation confers a rightward, depolarizing shift (+16 mV) in the V_0.5 of activation, while inactivation measured at 0 mV remains unchanged (23). Within the NH₂ terminus of hERG, the LQT2 hERG D16A mutation exhibits a +13 mV rightward shift in the V_0.5 of activation, with a minor slowing of inactivation at 0 mV and a slowing of deactivation at −50 mV (23, 302).

The hERG G584S pore mutation increases the rate of channel inactivation (509). G584S represents a pore mutation with a milder loss-of-function phenotype directly and solely attributed to its effect on inactivation. This finding adds further to the complexity of hERG mutant phenotypes as most other pore mutations cause increased arrhythmic events compared with non-pore hERG mutations (283). Recently, the hERG T613A mutation in the outer region of the pore helix, a regulatory site for C-type inactivation, was found to cause greater than 80% inhibition of maximal hERG current via a hyperpolarizing shift in channel inactivation when expressed in Xenopus oocytes alone; coexpression of WT and mutant channel resulted in an intermediate loss of channel function, without dominant-negative suppression of current (329).

Gating defects can arise from mutations across a wide range of locations in hERG, and predispose individuals to LQT2 of varying severity dependent on the functional impact of the mutation.

IV) hERG K⁺ permeation defect. The potassium selectivity filter in the pore of the hERG channel allows for potassium permeation with great specificity. Conserved amino acid residues comprise the selectivity filter, and mutation of any of these residues or nearby amino acids may greatly alter potassium permeation, even if only one of four hERG α-subunits within the mature channel bears the mutation (i.e., dominant negative suppression of I_Kr). The hERG G628S mutation, for instance, traffics to the cell surface and gates normally, which was confirmed by voltage-clamp fluorometry experimentation, but the mutant hERG channels do not conduct potassium under physiological conditions due to blockade of potassium permeation by intracellular potassium (116). The G628S mutation was also found to be cause lethal arrhythmia in transgenic rabbits with the mutation, as these rabbits possessed greater than 50% incidence of sudden cardiac death after one year (59).

B) COMBINED MECHANISMS OF LOSS-OF-FUNCTION IN LQT2.

Many mutations leading to LQT2 confer a loss-of-function hERG phenotype by multiple molecular mechanisms. Kanters et al. (206) recently reported the mechanistic basis of channel dysfunction of the hERG F29L mutation in a founder population of Danish families, which results in a malignant form of LQT2. The mutation causes clinically significant prolonged QT, with a penetrance of 73%. The F29L mutation was found to 1) reduce trafficking of hERG to the cell surface, posited to occur based on decreased glycosylation of the hERG F29L channel; and 2) reduce steady-state inactivation current density, positively shift the voltage dependence of inactivation, and increase the rate of deactivation. Similarly, the M124R mutation in the NH₂ terminus residing close to the hydrophobic surface of hERG, results in faster channel deactivation, in addition to reduced trafficking to the cell surface (23, 390). Furthermore, the exact amino acid substitution can impact the mechanism of loss of function of the hERG channel: the LQT2-associated G626A mutation in the conserved K⁺ selectivity filter in the channel pore traffics to the cell surface normally, while the LQT2-associated G626D, G626V, and G626S mutations had trafficking defects. The exact amino acid substitution has potential therapeutic implications: of the three G626 trafficking mutants, only hERG G626S had its trafficking defect corrected by E4031, an experimental pharmacological agent that increases forward trafficking of hERG channels in addition to its pore blocking effects (23).

2. LQT6

KCNE2 (or MiRP1) protein, encoded for by the KCNE2 gene, modulates hERG activity as an accessory β-subunit (Figure 5A). Mutations in KCNE2 cause LQT6 due to a decrease in I_Kr (see Figure 7 and Table 4) (2). The V65M mutation in KCNE2, for example, induces faster inactivation of colocalized hERG/mutant KCNE2 channel complexes (182). The KCNE2 T10M was found linked to patients with cardiac arrhythmia and LQTS, with particular susceptibility to arrhythmia onset in the setting of auditory stimuli and hypokalemia, both common triggers of arrhythmia in LQT2 patients as well. Coexpression of KCNE2 T10M with hERG decreases tail currents and causes a hyperpolarizing shift in steady-state inactivation, along with slowed recovery from inactivation compared with WT channels in CHO cells. The impact of the KCNE2 T10M mutation in the setting of known LQT2 arrhythmia triggers supports the hypothesis that KCNE2 associates with hERG in the human heart to regulate I_Kr (151). As KCNE2 is known to interact with other K⁺ α-subunits in addition to hERG, defective KCNE2 may lead to varying phenotypes. Indeed, the KCNE2 M54T mutation leads clinically to LQT6 in addition to sinus bradycardia, the latter clinical finding mediated by reduction of the HCN4 pacemaker channel activity (297). The complexity of interactions of KCNE2 with a variety of potassium channels may inform differential phenotypes observed clinically in those harboring a KCNE2 mutation.

3. Autoimmune-associated LQT syndrome

Recently, Yue et al. (494) discovered that anti-Ro antibodies block the hERG channel, resulting in an autoimmune-associated LQTS. The anti-Ro antibodies, present in the serum of patients with a variety of common autoimmune illnesses, were found to inhibit I_Kr in HEK293 cells, and a guinea pig autoimmune model with circulating anti-Ro antibodies had a prolonged QTc interval, caused by inhibition of native I_Kr. Anti-Ro likely blocks hERG by binding within the pore region of the channel. Patients with anti-Ro-positive autoimmune illness may be well advised to take particular caution to avoid medications that prolong the QT interval (494).

E. Molecular Pharmacology

hERG channel modulation by pharmacological agents has become an increasingly studied research area because it is now well established that a wide variety of clinically used drugs block hERG and thus inhibit I_Kr, leading to drug-induced LQTS. The section below details mechanisms and the biomedical relevance of drug-induced LQTS, followed by a discussion of various pharmacological tools that hold promise as novel agents for the treatment of LQT2 (and LQT6) in the future.

1. Drug-induced LQTS

Most drugs that block I_Kr conduction bind to the open state of the hERG channel, binding the inner pore thereby preventing permeation of potassium (67, 215, 267, 400, 483, 513). The anti-arrhythmic pharmacological agent MK-499 interacts mostly with polar and/or aromatic residues within the channel pore, including Thr623, Tyr652, and Phe656, to induce hERG block (266). Different hERG-blocking drugs may interact with different combinations of these four residues (266, 319). For instance, the anti-arrhythmic agent dofetilide interacts with Phe656 to mediate hERG binding (233). While the aromatic and polar residues above are integral to hERG's supreme sensitivity to block by pharmacological agents, C-type inactivation may increase affinity of drugs to hERG's pore residues and enhance drug block (125, 171).

Some drugs, including pentamidine (90, 228) and probucol (156), cause a reduction in I_Kr by reducing trafficking of hERG to the plasma membrane, leading to prolongation of the QT interval. The widely clinically employed drugs fluoxetine and ketoconazole block hERG directly and decrease hERG channel density at the cell surface (337, 411). Several other pharmacological agents precipitate drug-induced LQTS in susceptible patients, via hERG blockade or modulation of other important ionic currents (e.g., I_Na, I_Ks) (439, 482).

KCR1, a 12-pass transmembrane segment regulatory protein, interacts with hERG and mitigates the proarrhythmic effects of hERG-blocking agents (178, 223). In heterologous conditions, KCR1 decreases the sensitivity of hERG to a variety of well-established hERG channel blockers, such as sotalol, quinidine, and dofetilide (223). Loss of function of KCR1 is associated with greater risk of arrhythmia: the KCR1 E33D mutant found in a patient with ventricular fibrillation and QT-interval prolongation conferred reduced protection against drug-induced hERG blockade (168). The KCR1 I447V variant is associated with lower frequency of drug-induced arrhythmia, and reduced sensitivity hERG block by dofetilide when coexpressed heterologously (322). These results suggest that functional KCR1 provides protection against a drug-induced decrease in I_Kr, with a mechanism of action thought to entail modulation of KCR1 α-1,2-glucosyltransferase activity (291).

2. Drug screening for hERG toxicity

Given the propensity for clinically relevant pharmacological agents to block hERG, leading to drug-induced arrhythmia, it is now a required practice for drugs early in development to undergo hERG toxicity screens. High-throughput screening techniques, including radioligand binding, ion flux and fluorescence assays, and hERG-Lite which measures cell surface density of hERG, detect candidate drugs with strong hERG-blocking and/or traffic-altering properties (436). The development of automated patch-clamp electrophysiology enhances the throughput of data acquisition tremendously, allowing for numerous hERG recordings from different cells, applying different drugs and at different concentrations, to be collected simultaneously. Automated patch-clamp technology is the best initial high-throughput hERG toxicity drug screening method to date (163). A focus on hERG toxicity screening early in the drug development process may save time, and prevent or minimize adverse clinical outcomes due to inadvertent I_Kr block.

3. Pharmacological treatment in LQT2

β-Blocker treatment represents first-line pharmacological therapy in LQT2 patients. For a more detailed discussion of β-blocker treatment in congenital LQTS, refer to section VI of this review.

A growing interest in hERG activators has developed. Such agents may be especially useful in the setting of congenital LQT2 and/or drug-induced LQT syndromes. hERG activators may be subdivided into type 1 activators type 2 activators, based on mechanism of hERG activation (436).

Type 1 hERG activators include RPR260243, an experimental small molecule compound that slows hERG deactivation in a temperature- and voltage-dependent fashion as its primary means of channel activation (204, 321). RPR260243 may slow deactivation via binding to the S4–S5 linker and the cytoplasmic regions of the S5 and S6 hERG domains (321).

Type 2 hERG activators act primarily by slowing channel inactivation, and/or shifting the voltage dependence of inactivation to more positive membrane voltages (69, 138, 153, 164). Type 2 activators may also have other relatively minor effects on deactivation kinetics and voltage dependence of activation, as well as single-channel open probability (69, 138, 164, 165, 407, 511). Several type 2 hERG activators have been described, including experimental compounds PD118057 (320), ICA105574 (28, 134, 263), and NS1643 (153, 155, 318), and key insights into their mechanism of action alongside their hERG binding location have been found, each having specific binding residues that mediate their mechanism of hERG activation within the “type 2” paradigm.

Other compounds have predominating mechanisms of hERG activation that do not fall into the “type 1” or “type 2” classes. For instance, mallotoxin and KB130015 primarily enhance hERG activation while shifting the voltage dependence of activation to more negative membrane voltages, likely by binding within the channel pore (140, 499). Some compounds, such as A935142 (407), have both “type 1” and “type 2” characteristics. As activators bind hERG at various locations within the channel, efforts to correlate binding location and channel stoichiometry with mechanism of hERG activation by small molecule compounds will aid the design of more selective and potent hERG channel activators by a variety of mechanisms (476).

IV. I_NA DYSFUNCTION IN CONGENITAL LQTS

A. Introduction

LQT3, LQT9, LQT10, and LQT12 account for 5–10% of congenital LQTS (7) and are associated with gain-of-function mutations in the cardiac voltage-gated Na⁺ channel complex of proteins. As such, patients diagnosed with Na⁺ channel-associated LQT often, though not exclusively, have a unique risk profile for arrhythmia triggers and require different pharmacological approaches to disease management than patients with LQT1 or LQT2 (370). Electrophysiological experiments have elucidated the connection between aberrant function of single ion channels and the unique clinical presentation of this set of diseases. In each case, congenital mutations result in a failure of the cardiac Na⁺ current (I_Na) to properly inactivate. However, even within the subset of Na⁺ channel-associated LQT, differences in mechanisms of pathological I_Na generation and risk stratification in some patients makes LQT3 an important case study in the value of patient-specific medicine. Below we will discuss the role of I_Na in the conduction of electrical impulses in the heart, mechanisms by which congenital mutations may alter these currents, and advances in the discovery of potent and selective pharmacological agents to treat disease.

B. Physiology

In the heart, the majority of voltage-gated Na⁺ current is carried by the cardiac isoform of the voltage-gated Na⁺ channel, Na_v1.5 (137). Upon membrane depolarization to a threshold around −50 mV, channels activate and allow the influx of Na⁺ down its electrochemical gradient. The rapid influx of positive charge across the cell membrane drives the membrane potential from its resting voltage to a peak of +50 mV, near the Na⁺ equilibrium potential (Figure 1). The rapid activation of Na_v1.5 is therefore critical to maintaining the speed of cardiac conduction. This rapid depolarization coordinates the stimulation of voltage-gated Ca²⁺ and K⁺ channels necessary for excitation-contraction coupling and, eventually, repolarization (299).

In a healthy human heart, the Na⁺ current is rapidly attenuated in milliseconds as channels inactivate, a process that was first observed and defined in Hodgkin and Huxley's giant squid axon experiments (173). Inactivation is responsible for the refractory period in neurons and myocytes during which the firing of new impulses is not possible. Similar properties are observed in Na_v1.5, with inactivation progressing with a time to half peak current on the order of 1 ms and a mean open channel duration of roughly 0.5 ms (49) (Figure 8, A AND B).

FIGURE 8.

Physiology and molecular biology of I_Na. A: consecutive recordings of single Na⁺ channels recorded under cell-attached conditions in HEK293 cells transfected with Na_v1.5 cDNA (unpublished data). B: WT Na⁺ currents recorded under whole cell patch-clamp conditions and obtained by depolarizations from −120 to −30 mV (unpublished data). At high gain (inset), whole cell currents return nearly to 0 nA. C: voltage dependence of activation (red-filled square) and inactivation (blue-filled circle) (unpublished data). D: topology of Na_v1.5 in the plasma membrane with accessory proteins implicated in LQTS.

Na⁺ channel inactivation is a complex multistep process that involves multiple channel components and is itself highly voltage dependent. The voltage dependence of inactivation is independent of the voltage dependence of channel activation, as inactivation may occur from either closed or opened states (66). Closed-state inactivation can be measured with subactivation threshold depolarizing pulses. When these pulses are sufficiently long to approach equilibrium, these experiments are referred to as “steady-state inactivation” experiments. As the holding potential becomes more depolarized, channels enter a closed-inactivated state and are unavailable to open upon supra-activation threshold depolarization with a midpoint near −70 mV (Figure 8C). Because this midpoint is in the range of the resting membrane potential in a ventricular myocyte (−85 mV), small changes in the V_0.5 of steady-state inactivation caused by mutations or drugs, or small changes in resting membrane potential, can have a large impact on channel availability, Na⁺ current density, and conduction velocity (86).

The voltage dependence of inactivation from the open state has been more difficult to dissect, as channel fast-inactivation is a nonequilibrium process. The rate of decay of Na⁺ currents increases with increasing voltage over the physiologically relevant range, reaching half-inactivation in 1.5 ms at −40 mV and 0.5 ms at +20 mV (34). Recovery from inactivation has the opposite voltage dependence of the onset of inactivation: channels recover more rapidly at hyperpolarized potentials, and very slowly at potentials more positive than activation threshold (approximately −45 mV) (508). The presence of multiple inactivation states is evident in the time course of recovery from inactivation. Following an initial depolarizing pulse to inactivate channels, channel recovery during a period at resting membrane potential follows a biexponential time course, suggesting the existence of two distinct populations of channels (508): one in the fast-inactivated state and one in the slow-inactivated state.

C. Molecular Biology

Na_v1.5 is encoded by the gene SCN5A (452). Unlike the canonical voltage-gated K⁺ channels, Na_v1.5 is composed of a single polypeptide chain (137). The protein folds into four homologous asymmetric domains, each containing six transmembrane helices (S1–S6) (70) (Figure 8D). S1 through S4 form a voltage-sensing unit in each domain. S4 spans the membrane and contains several charged arginine residues capable of “sensing” changes in membrane potential. Depolarization drives a conformational change in S4 that is transmitted to the pore-forming helices, S5 and S6. Along with the extracellular reentrant loop that connects them, S5 and S6 from each domain form the central ion conduction pathway (70).

The intracellular segments of Na_v1.5 have been shown to be critical to proper channel function. The linker between domains III and IV contains the three amino acid isoleucine, phenylalanine, methionine (“IFM”) motif, shown to be a key molecular determinant of fast inactivation (313, 465). Upon channel opening, this motif has been proposed to bind to a hydrophobic cluster in the intracellular linkers between S4 and S5 of domains III and IV (268, 415). Consistent with these findings, disruption of the structure of the DIII/DIV linker by proteolytic cleavage (434) or mutation (47) can impair inactivation, and coexpression of a cytosolic peptide containing the IFM motif can potentiate inactivation (111).

The intracellular COOH-terminal domain has been shown to play an important role in channel function. The COOH terminus has a structured proximal domain and unstructured distal region (91). It has been proposed that interactions between the COOH terminus and DIII/DIV linker are critical in the stabilization of the inactivated state (285), and furthermore that this interaction may be facilitated by calmodulin (435). In fact, several LQT3-causing mutations have been identified in the COOH terminus (34, 343).

Multiple techniques have been used to study the role of voltage-sensing units in channel function. The asymmetric nature of eukaryotic voltage-gated Na⁺ channels allows the voltage sensors to modulate different aspects of channel gating. Mutagenesis (66, 384), cysteine accessibility (380, 381, 383), and voltage-clamp fluorometry (71, 72) have revealed distinct properties of each voltage sensor. Movements of the DIV voltage sensor have been shown to correlate with the voltage dependence and kinetics of inactivation (71, 147), and DIV-S4 activation is necessary and sufficient for inactivation gating (66). The putative docking site for the inactivation gate resides in the intracellular NH₂-terminal loops of S4 in DIII and DIV, potentially bridging our understanding of voltage-sensing domains and voltage-dependent gating transitions (268, 415).

1. Accessory proteins

In the heart, Na_v1.5 is associated with several accessory proteins that modulate its trafficking and biophysical properties. Four genes that encode β-subunits of the voltage-gated Na⁺ channel have been identified, SCN1B through 4B, and all are expressed in the heart (253, 342). These β-subunits have an Ig (immunoglobulin)-like NH₂-terminal domain, a single transmembrane helix, and a short cytoplasmic COOH-terminal domain. While they are similar in structure, the β-subunits interact with the α-subunit through different mechanisms. β1 and β3 interact noncovalently with the α-subunit, and β2 and β4 interact via covalent disulfide bond (490).

Immunohistochemical studies suggest that β-subunits play an impo rtant role in subcellular localization of different channel isoforms (253). Na_v1.5 predominantly colocalizes with the β2 and β4 subunits, encoded by the gene SCN2B and SCN4B, in the intercalated disks. β4 has no significant effect on Na_v1.5 activation or inactivation gating (490). Small amounts of β1 subunit were also found to colocalize with Na_v1.5 (101, 211, 253). Unlike β4, β1 displays a dramatic functional effect on inactivation of Na_v1.5 currents expressed in tsA cells. β1, but not β2, caused a significant positive shift in the V_0.5 of steady-state inactivation (101, 211, 342).

Na_v1.5 is also associated with several other structural and signaling complexes in the cardiomyocyte. Caveolin and syntrophin are two such proteins that have been implicated in functional perturbation of I_Na and Na⁺ channel-associated LQT. The six known caveolin proteins are encoded by three caveolin genes: CAV1, CAV2, and CAV3 (30). Cav-3 is the most abundant isoform in cardiac tissue where it is associated with cytoplasmic signaling factors (396), is responsible for the formation of lipid raft microdomains and caveolae (218), and potentially plays a role in t-tubule organization (133). Cav-3 null mice lacked caveolae and had abnormal t-tubule development in skeletal muscle. Biochemical and microscopic techniques have revealed that Na_v1.5 colocalizes with Cav-3 in lipid rafts, which facilitate the increase in peak I_Na density in response to β-adrenergic stimulation (486).

Caveolin and Na_v1.5 also both associate with the cytoskeletal dystrophin complex (106). Syntrophin-α₁ is a member of the dystrophin-associated protein complex and is the most abundantly expressed syntrophin in the heart. Syntrophins mediate the interaction of dystrophin with the PDZ-binding domain on the COOH terminus of Na_v1.5 (135). Isolated cardiomyocytes from dystrophin knockout mice showed markedly decreased I_Na density, and ECGs showed significant QRS prolongation, consistent with an interaction between the dystrophin-associated protein complex and ion channels in the heart.

D. Molecular Pathophysiology

1. LQT3

Like other forms of LQT, the hallmark of LQT3 is a prolongation of the QT interval on the ECG arising from a prolongation of the ventricular action potential and delayed ventricular repolarization. The discovery that mutations in multiple genes can give rise to LQT has allowed for an improved understanding of the risk stratification for cardiac events in this set of patients (Figure 9, A–C). While patients with K⁺ channel-associated LQT are generally at higher risk or cardiac events during exercise, patients with LQT3 have elevated risk at rest (370, 371). However, even within the umbrella of LQT3 there is a range of triggers for arrhythmia that is not easily explained by a “one size fits all” model, suggesting patient-specific differences in disease mechanism (244).

FIGURE 9.

I_Na dysfunction leading to congenital LQTS. A: topology of Na_v1.5 in the plasma membrane. Several representative LQT3-associated mutations were selected because there is strong evidence that they induce I_NaL by channel bursting (red-filled triangle) or by late reopening (red-filled circle), or prolong the action potential without I_NaL (red-filled square). B: ECG from a LQT3 patient (unpublished data). C: simulated action potential (top) and I_Na (bottom) in WT (green) and heterozygous LQT3 (purple) conditions. I_Na peak current was truncated to enhance view of I_NaL. D: representative current recordings from WT (purple) and F1473C (orange) Na_v1.5 expressed in HEK293 cells at low and high gain (inset) elicited by depolarization to −10 mV (35). E: steady-state channel availability for WT (green triangle) and F1473C (purple square). [D and E from Bankston et al. (35).]

LQT3 is a disease of impaired Na_v1.5 inactivation. Failure of the channel to inactivate or remain inactivated during the plateau or repolarization phases of the ventricular action potential allows for a depolarizing current sufficient to prolong the action potential, leaving patients susceptible to asynchronous EADs or DADs (184, 240). In most cases, impaired inactivation is manifest as a sustained noninactivating inward current, or late current (I_NaL) (Figure 9D). Rather than inactivating completely, many disease-associated mutations cause a persistent I_NaL during prolonged depolarizations. It should be noted that I_NaL of smaller magnitude has also been detected in tissue isolated from normal hearts (432); thus in pathology some I_NaL may result from the failure to control or modulate normal activity. Nevertheless, while currents through these channels still provide the rapid influx of Na⁺ necessary for cardiac conduction, at high gain one can observe I_NaL that amounts to as little as 1% of the magnitude of the peak. Because the membrane impedance during the plateau phase of the action potential is high, this small depolarizing current has a substantial effect on AP duration.

The mechanisms by which gain-of-function Na⁺ currents may arise have been studied in depth, and recent computational work suggests that these divergent mechanisms may underlie tissue-specific phenotypes (185), as well as arrhythmia risk stratification (32). Two mechanisms have been established for the generation of persistent I_NaL, and an additional set of mechanisms has been proposed for aberrant Na⁺ conduction in the absence of I_NaL. Table 5 lists a set of Na_v1.5 mutations known to perturb channel function by specific mechanisms.

Table 5.

Representative LQT3-associated mutations classified by mechanism of gain of function

Mechanism	Mutations	Reference Nos.
INaL: bursting	ΔKPQ (1505–1507)	47
	Y1795C	243
INaL: late reopening	R1644H	109
	N1325S	109
	G1631D	447
	S1904L	34
	R1623Q	202
INaL: cAMP-dependent	D1790G	416
Window current	E1295K	6
	G1631D	447
	F1483del	397
Nonequilibrium gating	I1768V	87

A) CHANNEL BURSTING.

Channel bursting is best exemplified by the ΔKPQ mutation, a deletion of amino acids 1505–1507 in the DIII/DIV linker near the IFM motif. ΔKPQ was first identified in 1995 (453) and characterized that same year (47). Like WT currents, currents passed by single ΔKPQ channels typically activate and inactivate quickly. However, during a small fraction of depolarizing pulses, channels enter a slow gating mode, or “burst” mode, characterized by a persistent noninactivating current.

Modal transitions have been shown to be highly dependent on inter-pulse duration and resting membrane potential. I_NaL from ΔKPQ channels in heterologous expression systems (289) is highly rate dependent, decreasing dramatically with increased pulse frequency. In isolated transgenic ΔKPQ mouse cardiomyocytes, changes in pacing rate caused pause-induced spontaneous depolarizations (127). This is consistent with a decrease in EAD frequency and polymorphic VT during increased pacing in intact mouse hearts (120), and explains the elevated risk of arrhythmia and sudden cardiac death at rest in human patients with LQT3 (370, 371).

The inverse dependence on interpulse duration suggests that transitions between background and burst gating modes occur predominantly at rest. At higher resting potentials, a larger fraction of channels exist in closed-inactivated states, leaving fewer channels available to transition into bursting states. Indeed, I_NaL is smaller during test pulses from more depolarized holding potentials (185). Taken together, Clancy and Rudy (85) were able to develop a computational model of ΔKPQ function by allowing transitions between background and burst modes only when channels exist in the closed state. This model was able to reproduce single-channel and whole cell gating properties, as well as prolongation of the ventricular action potential and the occurrence of EADs.

B) LATE REOPENING.

Following the identification of modal transitions in ΔKPQ channels, other SCN5a mutations were identified in patients with LQT3 that cause I_NaL by a different mechanism. At the whole cell level, the N1325S and R1644H mutations had markedly different inactivation (109) and recovery from inactivation time courses (448). Single-channel records revealed that these gating defects were in fact not due to bursting channels, but to dispersed channel reopening that occurred during prolonged depolarizations (109). Additional studies revealed that these late reopenings contribute to a fraction of the I_NaL from ΔKPQ channels, as well (73).

An additional mutation, S1904L, was later shown to also induce late reopenings and I_NaL (34). Interestingly, this mutation does not reside in the DIII/DIV linker, but rather exists in the distal COOH-terminal domain. Transition metal ion FRET showed that this mutation destabilized a critical intramolecular interaction within the COOH-terminal domain (144), contributing to our understanding of the inactivation gate as a macromolecular complex of different channel components including the DIII/DIV linker and COOH terminus (285).

A computational model has been constructed to simulate late reopening currents at the single-channel and whole cell level (34). Function of the S1904L mutation could be simulated by reducing the rate of transition into the slow inactivated states, allowing channels to reside for a prolonged time in fast inactivated states from which they may reopen. An interesting prediction of this model is that the amplitude of I_NaL will not display the inverse rate dependence present in bursting channels, and in fact the patient experienced arrhythmic events that were not limited to periods of rest.

C) WINDOW CURRENT.

In WT channels, the voltage dependence of channel availability (V_0.5 = −71 mV) and channel activation (V_0.5 = −25 mV) overlap only slightly. Therefore, there is a very small range, or “window,” of voltages in which channels may activate but not completely inactivate. This range may provide a small noninactivating current known as “window current.” Typically, this window current is very small and only exists at voltages at which repolarizing currents through delayed rectifier K⁺ channels are strong, and therefore contribute little to net membrane current.

Changes in steady-state properties of the cardiac sodium current have been proposed as mechanisms for AP prolongation and LQT, even in the absence of sustained I_NaL during prolonged depolarizing steps (Figure 9E) (6, 344, 395, 461). These mutations cause depolarizing shifts in channel availability such that the window current is either larger, or exists at more depolarized voltages where rectification of K⁺ currents increases the relative contribution of I_Na to total membrane current. These patients are therefore more susceptible to drug-induced LQT, most commonly arising from block of hERG channels, less commonly from block of other K⁺ channels and drugs that potentiate inward currents, including I_Na (439, 482). Additionally, several disease-causing mutations have been shown to induce depolarizing shifts in channel availability in addition to canonical I_NaL during prolonged depolarization (35, 448, 463).

D) NONEQUILIBRIUM GATING DEFECTS.

In addition to changes in steady-state parameters of channel activity, time-dependent changes in gating kinetics have also been posited as a mechanism for AP prolongation. As membrane potential declines into the repolarization phase of the AP, WT channels recover slowly from inactivation. As membrane repolarization is more rapid than recovery from inactivation, WT channels never recover at potentials that facilitate channel activation. Some LQT3 mutant channels, such as the I1768V mutation, have enhanced recovery from inactivation and are able to recover at voltages above the channel activation threshold (87). Importantly, this nonequilibrium reopening is different than window current because it occurs at potentials at which activation and steady-state channel availability do not overlap.

E) PHOSPHORYLATION-DEPENDENT I_NAL.

There exists one known mutation, D1790G, for which in vitro studies found I_NaL only arises after PKA-dependent phosphorylation (416). D1790G channels showed marked I_NaL in response to cAMP-dependent PKA stimulation and lacked I_NaL under basal conditions. Furthermore, the D1790E mutation showed no response to cAMP application, suggesting that the negative charge at residue 1790 is important in modulating PKA phosphorylation. Interestingly, alanine mutagenesis of consensus PKA phosphorylation sites showed that S38 and S525 are important in mediating I_NaL in D1790G channels, despite residing in disparate parts of the channel. These effects were found despite the fact that most spontaneous arrhythmic events in LQT3 patients occur during rest or sleep, suggesting perhaps a role of basal phosphorylation of the Nav1.5 channel in D1790G patients (371).

2. LQT associated with mutations to accessory proteins

In addition to mutations in SCN5A, mutations in the accessory proteins important in cardiac I_Na have been associated with LQT (11), although there are few documented cases of such diseases.

3. LQT10

Mutations in the β4 subunit give rise to LQT10. The L179F mutation was identified in a 21-mo-old girl with bradycardia and severe QT prolongation (261). When coexpressed with Na_v1.5 in HEK cells, this mutation caused both a dramatic I_NaL and a depolarizing shift in the V_0.5 of steady-state inactivation that increased the window current. Interestingly, this β-subunit mutation induced a larger I_NaL than the ΔKPQ deletion in the α-subunit. Mutations in SCN4B have also been associated with Sudden Infant Death Syndrome (SIDS) (413). The SIDS-associated S206L caused a depolarizing shift in stead-state inactivation and a dramatic increase in I_NaL. Recently, a mutation was reported in the β1b subunit, a splice variant of β1, that caused a speeding of recovery from inactivation, significant I_NaL, APD prolongation in HL-1 cells, and QT prolongation and stress-induced syncope in the patient (342). This was the first report of a mutation in β1b associated with LQT.

4. LQT9 and LQT12

LQT9 is a Na⁺ channel-associated arrhythmia arising from mutations in Cav3, the gene encoding Cav-3, a cardiac isoform of caveolin (30). Four Cav-3 mutations were identified in a sample of 905 LQTS patients (438). Sixty-seven percent of patients identified with mutations in Cav3 exhibited syncope or cardiac arrest during periods of rest, consistent with the risk profile for patients with mutations directly in the α-subunit, Na_v1.5. Two mutations, F97C and S141R, were studied further by coexpression in HEK-293 cells with Na_v1.5. The S141R mutation sped recovery from slow inactivation, and both mutations caused a significant I_NaL. Consistent with the functional importance of the Cav-3/Na_v1.5 interaction, three additional mutations in Cav3 have been shown to induce I_NaL and are associated with SIDS (92).

In mouse skeletal muscle, knockout of Cav-3 caused a disruption of the dystrophin complex in lipid rafts (133). Perhaps not surprisingly, mutations in syntrophin-α, a member of the dystrophin protein complex, are also associated with LQTS and give rise to variant 12 of the disease (473). Two mutations in SNTA1, the gene encoding syntrophin-α₁ have been identified. The A257G mutation nearly doubled the peak current density, increased window current, and sped recovery from fast inactivation, although it did not induce I_NaL relative to peak current density in isolated neonatal rat cardiomyocytes (473). It is likely, however, that absolute I_NaL was greater in A257G myocytes because of the increase in channel expression. The A390V mutation does cause a persistent I_NaL and a depolarizing shift in steady-state inactivation in a nitric oxide synthase-dependent manner (431).

E. Molecular Pharmacology

The pharmacology of Na⁺ channel-associated LQT has been difficult to navigate because the diversity in patient phenotypes (392, 397) and risk of on- and off-target effects of standard antiarrhythmic therapies (276) has limited their use clinically. As a result, genetically identified LQT3 patients are the most likely to receive implantable cardioverter-defibrillators (ICDs) as a precautionary measure despite previously having been asymptomatic (372). There is therefore a large unmet medical need for effective pharmacological intervention, and our understanding of Na⁺ channel blockade and β-adrenergic receptor blockade in LQT3 continues to evolve.

1. Local anesthetics

The utility of local anesthetics in the treatment of LQT3 is rooted in their ability to inhibit the voltage-gated Na⁺ channel with preference for I_NaL (Figure 10). Soon after the discovery of the ΔKPQ mutation and its role in LQT3, it was reported that the commonly prescribed local anesthetic and class 1b antiarrhythmic agent lidocaine inhibited I_NaL more effectively than peak current (19). This phenomenon has been observed in block of several mutations by other clinically relevant local anesthetic-like drugs including mexiletine (35, 258), flecainide (288), and ranolazine (127), as well as recent experimental drugs GS967 (43) and eleclazine (128, 200), making them useful in normalizing the QT-interval (130, 281, 284, 370) and preventing arrhythmic events in some patients (347). This approach has recently been supported by a study of patients with a wide range of LQT3 mutations in which a robust improvement in QT interval and a reduction in arrhythmic events was induced by oral mexiletine (258). Because of the efficacy of these drugs in treating LQT3 patients and because of a common underlying mechanism, preferential inhibition of late versus peak Na⁺ channel current, development of more selective and effective blockers of I_NaL is underway by the pharmaceutical community. Nevertheless, it is important to recognize and understand the mechanisms underlying inhibition of this current.

FIGURE 10.

Representative recordings at low and high (insets) gain of F1473C inhibition by 50 μM mexiletine (A), 50 μM ranolazine (B), and 10 μM flecainide (C). D: mexiletine, ranolazine, and flecainide inhibit I_Na with preference for I_NaL. [From Bankston et al. (35).]

Local anesthetic-like drugs bind to a conserved phenylalanine (127, 336) in the central cavity of the channel in a highly state-specific manner. This specificity is imparted both by state-dependent changes in drug access to the binding site (243, 336) and by allosteric modulation of channel gating (162). Clinically relevant channel blockers have been shown to bind preferentially either to inactivated states, as is the case for lidocaine, flecainide, and mexiletine, and, in the case of ranolazine, to open channels (275).

The state dependence of drug binding has been assayed by electrophysiological as well as biochemical experiments. Lidocaine induces a hyperpolarizing shift in steady-state inactivation(202), slows recovery from inactivation, and stabilizes voltage-sensing units shown to provide the rate-limiting step in fast inactivation (27, 66, 382). Lidocaine has also been studied at the single-channel level. In cell-attached patches, lidocaine reduced the propensity of ΔKPQ channels to burst, increased the number of depolarizing sweeps with no channel openings, and had no effect on the mean open time of background or burst openings (31, 108). Taken together, these findings are consistent with stabilization or promotion of inactivated channels. Ranolazine, on the other hand, has no effect on steady-state inactivation and has therefore been interpreted as an open-channel blocker (275).

It is therefore not surprising that mutations that alter channel gating by different mechanisms will produce unique allosteric effects on drug binding. For example, flecainide (243) and mexiletine (347) have been shown to more potently inhibit some mutations than others when expressed in heterologous systems. Mutation-specific variability is therefore a major hurdle in disease management (88), as heightened sensitivity to Na⁺ channel blocking drugs can be pro-arrhythmic in certain settings (112, 276). In fact, flecainide is a useful diagnostic agent in patients suspected to have loss-of-function Na⁺ channel mutations in Brugada Syndrome because of its ability to reduce peak I_Na and unmask latent disease phenotypes in a clinical setting (58). It has therefore been proposed that an understanding of channel pathology and drug properties may help guide a pharmacogenomic approach to disease management (88).

2. β-Blockers

There remains some debate over the clinical utility of β-blockers in treatment of LQT3. A mainstay therapy for patients with QT prolongation, clinical data suggest that prophylactic β-blocker therapy in patients with LQT3 may be less effective at preventing cardiac events as it is in K⁺ channel-associated LQT (282, 331). Interestingly, however, β-blocker therapy did not show evidence of arrhythmia induction, as might be expected from the heart rate-slowing effects of β-blockade in LQT3 patients with elevated risk of arrhythmia at rest. This has prompted further investigation into the mechanism of β-blocker activity in models of LQT3.

Experiments in animal models of LQT3 (64, 333) showed a protective effect of sympathetic stimulation against APD prolongation and torsades de pointes (387), suggesting a deleterious effect of β-blockade. However, experiments in heterologous expression systems have shown that β-blockers have a direct effect on Na⁺ channels (33). β-Blockers exhibit preferential inhibition of I_NaL and bind to the local anesthetic binding site, and at high doses have a similar pharmaceutical profile as local anesthetics.

Ahrens-Nicklas et al. (12) constructed a computational model of ΔKPQ-expressing ventricular myocytes stimulated by the β-agonist isoproterenol and inhibited by the β-blocker propranolol. Isoproterenol shortened APD and EADs, consistent with the inverse-rate dependence of I_NaL and animal models of LQT3. At low doses, propranolol mitigated this effect and prolonged APD, but at higher doses where propranolol inhibits Na⁺ currents, it provided a beneficial reduction in APD. Taken together, the experimental and computational data suggest a potential benefit of β-blockade in treatment of LQT3, but more robust blinded studies must be performed to better understand the clinical relevance.

V. OTHER SUBTYPES OF CONGENITAL LQTS

A. LQT4: Ankyrin-B Syndrome

The correct targeting of ion channels and transporters to their respective subcellular domains is critical for normal physiological function. It was discovered that autosomal dominant loss-of-function mutations in ANK2, which encodes ankyrin-B, can cause LQTS (273, 366). Ankyrin-B is an adaptor protein that functions in the cytoskeletal network to ensure proper localization and stabilization of ion channels and transporters. Although ankyrin-B was originally characterized as a 220-kDa protein, multiple splice variants exist (474). The protein is composed of a membrane-binding domain, a spectrin-binding domain, and a regulatory domain (271). It has been demonstrated to play important roles in the normal localization of the Na⁺/Ca²⁺ exchanger, the Na⁺-K⁺-ATPase, as well as the IP₃ receptor (IP₃R) which mediates intracellular calcium release (95, 269, 270, 273, 429). In addition, ankyrin-B has been shown to regulate voltage-gated Na⁺ channels (74), ATP-gated K⁺ channels (236), as well as ryanodine receptors (65) in the heart.

While a loss-of-function ankyrin-B mutation was initially described as LQT4, it was later discovered that mutations in ankyrin-B are not always associated with QT prolongation and may present with a variety of cardiac phenotypes including bradycardia, sinus arrhythmia, as well as atrial fibrillation (96, 232, 271). Most of these mutations, such as E1425G, are located in or near the regulatory domain of ankyrin-B (271). These mutations in general lead to disruption of normal targeting and regulation of transporters as well as ion channels, leading to disrupted calcium handling that contributes to arrhythmia (272, 273). However, variation in severity of cardiac phenotypes exists between different mutations and can be correlated to the degree of loss of function observed in vitro (271).

B. LQT7: Andersen-Tawil Syndrome Type 1

Andersen-Tawil syndrome type 1 arises from LQT7 mutations, which are loss-of-function mutations in the KCNJ2 gene (327). This gene encodes K_ir2.1, an inward rectifier potassium channel that underlies the cardiac I_K1 current (104, 248, 303, 496), which contributes to setting the myocardial resting potential and aids terminal repolarization of the action potential (264). Because the membrane potential of cardiomyocytes is more positive than the potassium equilibrium potential, the channel actually conducts an outward current in physiological conductions. K_ir2.1 expression has been demonstrated in both human atria and ventricles (458). The protein consists of two transmembrane helices, a pore-loop, as well as an NH₂- and COOH-terminal region. Four K_ir2.1 subunits may come together to form a functional channel. It has been shown that PIP₂ binding to K_ir2.1 plays an important role in channel activation (179, 410).

Andersen-Tawil syndrome is an autosomal dominant, multisystemic disorder characterized by periodic paralysis, ventricular arrhythmias, as well as dysmorphic features including low-set ears, hypertelorism, and hypoplastic mandible (362, 417). Early reports of patients with Andersen-Tawil syndrome described QT prolongation on ECG; however, most ECGs of patients with this disorder are characterized by prominent U waves that merge with the T wave, and a characteristic prolonged “Q-T-U” interval. The physiological basis for the U wave is not well understood, but the large U waves in Andersen-Tawil syndrome may be the result of delayed after depolarization or tissue-dependent differences in repolarization including the mid myocardium and His-Purkinje system (317, 328, 502). Andersen-Tawil syndrome type 1 results from KCNJ2 mutations, almost all of which are dominant negative (424, 425). Many of these mutations, such as R218W and G300D, located in the COOH-terminal region, either decrease PIP₂ affinity or affect residues involved in PIP₂ binding (103, 249, 398, 501). Some disease mutations, such as Δ312–314, lead to diminished channel trafficking to the cell surface (44). Disease-associated mutations can also be found in the NH₂-terminal region as well as the channel pore (424).

C. LQT8: Timothy Syndrome

Timothy syndrome is caused by gain-of-function mutations in the CACNA1C gene, which encodes Ca_v1.2, the α1_C (pore-forming) subunit of a cardiac voltage-gated calcium channel. The channel generates the L-type calcium current, which allows calcium influx during the plateau phase of the action potential, sustaining membrane depolarization and leading to calcium release from intracellular stores which triggers excitation-contraction coupling in cardiomyocytes. The channel undergoes both calcium-dependent and voltage-dependent inactivation. Inactivation of calcium channels during the action potential causes an imbalance between the inward calcium and outward potassium currents, resulting in repolarization. The α1_C subunit consists of four homologous domains of six transmembrane helices each (S1–S6), including a voltage-sensing domain (S1–S4) and a pore domain (S5–S6), similar to voltage-gated sodium channels. Different splice variants of the α1_C subunit may be expressed in the heart (445).

Timothy syndrome is an extremely rare genetic multisystemic disorder that includes QT prolongation, syndactyly, craniofacial abnormalities, congenital heart defects, as well as autism. All known Timothy syndrome mutations to date are located at the COOH-terminal end of S6 helices in different domains of α1_C (54, 142, 403, 404). Most commonly, Timothy syndrome arises from the de novo missense mutation G406R on exon 8a, which is present in a minor cardiac splice variant of CACNA1C (404). This mutation is located at the COOH-terminal end of the first S6 helix of the α1_C subunit and dramatically disrupts voltage-dependent inactivation of Ca_v1.2, leading to increased calcium influx that can result in APD prolongation. Timothy syndrome resulting from this mutation came to be known as the “classic” variant. Two patients were later found to have the mutations G406R and G402S, respectively, on exon 8, the major splice variant in the heart (403). Similar to the classic Timothy syndrome mutation, these two “atypical” Timothy syndrome mutations also disrupt channel inactivation. Interestingly, neither of the two atypical Timothy syndrome patients had syndactyly, as opposed to 100% of patients with classic Timothy syndrome, suggesting differential expression patterns of the α1_C splice variants. One Timothy syndrome mutation, I1166T, shows minimal effect on Ca_v1.2 inactivation, but is thought to cause disease by increasing window current, which occurs in voltage ranges where channels activate but do not completely inactivate (54). Efforts are underway to develop mechanism-based therapies. In a landmark study, it was found that roscovitine, a cyclin-dependent kinase inhibitor, partially recovers inactivation of mutant Ca_v1.2 channels (487, 488) and reduces APD in ventricular-like cardiomyocytes derived from iPSC reprogrammed from the skin cells of a Timothy syndrome patient (489). The study provides a model for future development of therapies using human iPSC-derived cardiomyocytes.

D. LQT13

Mutations in the KCNJ5 gene can also cause LQTS. KCNJ5 encodes a G protein-coupled inwardly rectifying potassium channel, K_ir3.4 (or GIRK4). K_ir3.4 and another GIRK isoform K_ir3.1(GIRK1) may form homo- and heterotetramers to generate the I_KACh current in the heart (99, 219, 222), a potassium current that is stimulated by acetylcholine through a GPCR-mediated pathway which has been demonstrated to slow heart rate during parasympathetic stimulation (136, 467). Similar to other inward rectifier channels, K_ir3.4 contains two transmembrane helices with a pore-loop in-between as well as an NH₂- and COOH-terminal region. Activation of GPCR receptors by acetylcholine causes the G_βγ subunit to activate K_ir3.4, resulting in an outward K⁺ current and hyperpolarization of the membrane potential, which in pacemaker cells results in slowing of heart rate (214, 341, 468).

The presence of I_KACh has been best demonstrated in sinoatrial node, atrioventricular node, and atrial cells of the heart (225). While it has also been shown to be expressed in the ventricles (42, 183, 217, 466, 485), its role there is not well understood. Identification of LQTS mutations on KCNJ5, however, suggests a role for this current in controlling ventricular action potential (484). For example, the disease mutation G387R, a conserved residue located at the COOH-terminal region of K_ir3.4, has been shown to cause autosomal dominant LQTS. The mutation has been shown to reduce trafficking of both K_ir3.4 as well as K_ir3.1 to the plasma membrane and decrease current amplitude. Interestingly, a number of these LQTS patients also have atrial fibrillation, suggesting a role for K_ir3.4 in electrical conduction in both atria and ventricles.

E. LQT14 and LQT15

Calmodulin is a ubiquitous calcium-binding protein that critically mediates cellular calcium signaling. Calmodulin is important not only for the calcium-dependent inactivation of L-type calcium channels and the inactivation of Na_v1.5 channels (435, 514), but also in the trafficking and assembly of KCNQ1 (379). Three genes, CALM1-3, all encode the identical calmodulin protein. For example, a recent study suggests that de novo mutations in CALM1 or CALM2 are associated with severely prolonged QT interval, ventricular arrhythmias, as well as atrioventricular block in infants (93). Located in calmodulin EF-motifs that directly bind calcium, these mutations result in a reduction in calcium affinity that may lead to abnormalities in sodium currents, potassium currents, as well as calcium dynamics in the heart. In addition to exhibiting prolonged QT, these patients show neurodevelopmental delay, consistent with widespread calmodulin expression in different tissues (316).

VI. GENOTYPE-DRIVEN CLINICAL MANAGEMENT OF CONGENITAL LQTS

A. Introduction

Clinical management of congenital LQTS aims to minimize symptomatic arrhythmia and prevent life-threatening cardiac events. As described in detail in the preceding sections, the three most common congenital Long QT syndromes (LQT1 due to mutation in KCNQ1; LQT2 due to mutation in hERG; and LQT3 due to mutation in Na_v1.5) result in lengthening of the APD in cardiomyocytes, leading to a prolonged QT interval on the ECG.

By consensus guidelines, management strategies in congenital LQTS today are mostly uniform, regardless of genotype. However, given the remarkable strides made in understanding the molecular identity of LQTS subtypes, as well as the specific impact of a wide range of LQT-associated mutations, the ultimate hope is that understanding a patient's genotype and specific LQT mutation can lead to rational treatment considerations. This section reviews the features of molecular physiology that may influence genotype-specific strategies for clinical management.

B. Risk Factors for Arrhythmogenesis

1. Mutation location, type, and cardiac risk

In LQT1, mutations in the transmembrane regions including cytoplasmic loop (C-loop) domains of KCNQ1 confers a twofold greater risk of cardiac events [defined as syncope, aborted cardiac arrest (ACA), or sudden cardiac death (SCD)] compared with patients harboring mutations in the COOH-terminal domain of the KCNQ1 protein (280). The C-loops are involved in sympathetic regulation of KCNQ1 (256), and mutations in this region may underlie an increased risk of cardiac events in these patients compared with mutations in other regions of the channel due to the substantial regulation of KCNQ1, and therefore I_Ks activity, by sympathetic stimulation (39).

In LQT2, Shimizu et al. (389) reported that pore region missense mutations, located in the S5-loop-S6 portion of the hERG channel, confer greater risk of cardiac events than non-pore mutations in hERG (389). Specific triggers of cardiac events were further subcategorized in LQT2 based on mutation location and age at the time of the cardiac event. A study of 634 LQT2 patients revealed that while patients with pore mutations in hERG confer a >2-fold risk of cardiac events compared with individuals with non-pore mutations overall, non-pore transmembrane domain mutations specifically led to a 6.8-fold increase in exercise-induced cardiac events (216), suggesting that adrenergic pathways interact with hERG non-pore transmembrane domains either directly or indirectly. Thus the location of the hERG mutation variably impacts susceptibility to common triggers in LQT2 patients.

In LQT3, rates of cardiac events are high, and risk is likely related to the degree of channel dysfunction or late current. The canonical SCN5A ΔKPQ mutation results in 2.4-fold greater risk of experiencing cardiac events through age 40 compared with the SCN5A D1790G mutation (244). Furthermore, while preliminary data suggest similar risk for mutations in the COOH-terminal versus transmembrane region of SCN5A, a 2.5-fold risk of ACA and SCD is conferred by mutations that cause both sodium late current and window current, compared with mutations causing only one such gain-of-function defect (38).

2. Genotype-specific triggers of arrhythmia

The first recognition of the relevance of genotype in management of LQTS occurred in the context of a seminal study linking molecular identity in LQTS with clinical features, establishing genotype-specific triggers of arrhythmic events (371). Six hundred seventy symptomatic patients with LQT1, LQT2, and LQT3 were followed, and strong differences arose in the trigger of arrhythmic events based on patient genotype (371). LQT1 patients experienced 62% of their cardiac events during exercise, and swimming or diving was found to be a particularly potent trigger of arrhythmia (10, 279). In contrast, LQT2 patients experienced the highest percentage of arrhythmic events in the setting of emotional arousal (43%), including sudden loud noises and startle (e.g., alarm clock awaking the patient from sleep), or other triggers of emotional stress and fear; 13% of LQT2 patient triggers occurred in the setting of exercise. LQT3 patients experienced the highest percentage of arrhythmic events, 39%, at rest or during sleep (371). LQT1 patients, by extension, experience more cardiac events in the setting of sympathetic nervous system stimulation induced by exercise than LQT2 or LQT3 patients. More generally, these findings highlight the varying pathophysiology of the LQT subtypes, and the specific life-style modifications and treatment considerations dependent on genotype (described later in this section).

3. Role of heart rate in LQT variant susceptibility to arrhythmia

The variable triggers of arrhythmia in congenital LQTS can be explained in part by the role of heart rate in arrhythmogenesis. Heart rate contributes greatly to action potential and QT duration (298), and therefore impacts the most common forms of congenital LQT syndrome in a genotype-specific, and even mutation-specific manner.

In LQT1, a higher basal heart rate increases risk of cardiac events, providing some prognostic value, unlike in LQT2 patients (40, 57, 373). Functional I_Ks current is pivotal to normal QT adaptation, or shortening with heart rate. At increased heart rates (and/or during β-adrenergic stimulation of the heart), I_Ks facilitates cardiac repolarization and shortening of the QT interval: I_Ks has slow deactivation kinetics, rendering the channel open for a longer period of time. Residual I_Ks channel activation from previous cardiac stimulation can enhance repolarization in the setting of short diastolic intervals (119). Indeed, a hallmark of LQT1 is the failure of the QT interval to adapt to increases in heart rate, as often seen during diagnostic exercise stress testing (29, 230, 470). This feature of I_Ks molecular physiology may also explain the predilection toward arrhythmias during swimming and diving in LQT1, which can produce a complex autonomically-driven “diving reflex” with abrupt changes in heart rate.

In contrast to LQT1 and LQT2, LQT3 patients experience greater risk of arrhythmic events at slower heart rates (332, 370). Bradycardia can produce prime conditions for SCN5A channel dysfunction, in a mutation-dependent manner. For example, the gain-of-function late sodium current generated by channel “burst” gating in the canonical ΔKPQ deletion mutation has been shown to be strongly attenuated with increases in pacing rate (289). The fraction of channels available to enter bursting channel states is larger at slower pacing frequency (371). The observation of strong association of arrhythmic events with bradycardia has led to concerns about the use of β-blockers, which slow heart rate, in the LQT3 subtype (33). However, it is important to note that β-blockers have multiple intracellular targets that may be protective against arrhythmia, and direct effects of β-blockers on sodium channel function have been reported (33). Additionally, some LQT3 mutations can exert their effect via mechanisms other than channel bursting, and therefore can have different features of rate dependence (34). For example, the LQT3 SCN5A S1904L mutation triggers arrhythmia more often at faster heart rates (185). Biophysical characterization alongside paired computational studies demonstrates increased late channel reopenings at faster pacing frequencies with the S1904L mutation, in direct contrast to ΔKPQ and F1473C mutations (35, 185).

4. Genotype-guided exercise testing

Since changes in heart rate can affect propensity for development of arrhythmia, exercise testing in congenital LQTS has evolved into an essential tool in the diagnostic workup of affected patients. Almost 40% of LQTS patients do not have a prolonged QTc interval on an ECG at rest. Exercise stress testing can aid in diagnosis and risk stratification of congenital LQTS, with particular utility in uncovering cases of concealed LQTS that have otherwise gone undiagnosed (177, 339). Exercise ECG during stress and recovery is a particularly helpful tool in uncovering concealed LQT1 syndrome (177).

Given the diagnostic utility of stress testing, attempts at using this test to guide genotype and mutation-specific risk stratification have been made. Laksman et al. (230) tested the hypothesis that exercise testing can predict QT interval response and effect of β-blockers in LQT1 patients with mutations in the C-loops (which, as described above, represent the site of final sympathetic regulation of KCNQ1). In this retrospective analysis, C-loop KCNQ1 missense mutations were not associated with an increased QTc interval during exercise stress testing or response to β-blocker therapy (230). Despite the known limitations of retrospective analysis, this study highlights the need for caution in the use of exercise stress tests for genotype-specific risk stratification.

5. Female gender and pregnancy as risk factors for arrhythmogenesis

Female gender is a risk factor for developing torsades de pointes in LQTS (186, 234), and the baseline QTc interval in healthy females is prolonged compared with males. Furthermore, women are at increased risk of developing drug-induced torsades de pointes as a consequence of hERG-blocking medications (254). In men, testosterone levels correlate with QTc interval duration (507). In a large study of 1,166 LQT2 syndrome patients, followed through age 40, ACA and SCD occurred significantly more frequently in females (26%) than in males (14%). With regard to location of the mutation, males with pore mutations had a greater than twofold risk of a first ACA or SCD compared with males with non-pore mutations. Pore mutations did not increase risk of these life-threatening cardiac events in females (265).

Women with LQTS are particularly prone to developing arrhythmias in the peripartum period. Seth et al. (377) demonstrated that the postpartum state (defined as 9 mo postdelivery) confers a 2.7-fold increased risk of a cardiac event and a 4.1-fold increased risk of a life-threatening cardiac event. The majority of these events did not occur during labor and delivery. This risk further appears to be genotype specific; the LQT2 genotype, in particular, conferred greater risk of cardiac events postpartum, compared with females with LQT1 and LQT3 genotypes (377). Changes in sex hormones associated with menopause may also affect risk of arrhythmia. Within five years leading to the onset of menopause, the odds ratio for a syncopal event for LQT2 females is 3.4 compared with LQT1 females, and after menopause, this increases to 8.1 (60). In LQT2, females older than 15 years of age have a 3.7-fold increased risk of suffering cardiac events compared with males, while there is no significant difference in risk of cardiac events between males and females younger than 15 years of age (495). A study of 634 LQT2 patients revealed that females >13 years of age were nine times more susceptible specifically to “arousal” triggers compared with males >13 years of age (216).

These clinical observations suggest that sex hormones may impact propensity for arrhythmia formation in LQTS in a genotype-specific manner. Sex hormones are known to play a role in cardiac ion channel regulation and arrhythmia. Estrogen has been shown to decrease I_Kr activity and prolong the QTc interval (227), and β-estradiol inhibits hERG channel expression in cultured HEK293 cells (24). Progesterone modulates both I_Ks and L-type calcium channels, leading to a shortened QTc interval (294).

Animal models, including canine, mouse, rat, rabbit, and guinea pig, have been extensively employed to better understand gender differences in cardiac electrophysiology (186), and strongly support the role for male sex hormones shortening the QT interval via enhanced I_Kr activity (129, 245, 477). In a LQT2 rabbit model, estradiol enhances risk of sudden cardiac death, while progesterone is protective (307). In canines, females have prolonged APD50, APD70, and APD90 values compared with males. Dofetilide, a hERG channel blocker, prolonged the APD90 of Purkinje fibers to a greater extent in female canines compared with males with rate dependence, and the canine studies demonstrate consistency of sex-related differences in cardiac repolarization between humans and canines (4). Further research is needed to elucidate the mechanisms underlying sex hormone and arrhythmia formation, and to possibly uncover therapeutic avenues through sex hormone sensitive pathways.

C. Genotype-Specific Management of Congenital LQTS

Congenital LQTS represents a potentially lethal condition, yet advances in device therapy, medical options, and surgical techniques have aided in the clinical management of the disorder (421). Today, most clinical management of LQTS is uniform across subtypes and consists of targeting of the sympathetic nervous system, through β-blocker therapy and left cervicothoracic sympathetic denervation (LCSD) in selected refractory cases. Primary and secondary prevention of cardiac arrest with an implantable cardioverter-defibrillator (ICD), alongside antiarrhythmic therapy, is considered after weighing risks and benefits for an individual patient.

As more is uncovered regarding the molecular physiology of these conditions, there is growing potential for tailored therapy to a given patient's genotype and specific LQT-causing mutation.

1. Lifestyle modification

Given the striking association of cardiac events with genotype-specific triggers, counseling should be given regarding lifestyle modification. As several examples, LQT1 patients are counseled about exercise limitations, most specifically swimming and possibly competitive athletics. LQT2 patients would be well advised to limit loud startling noises in their surroundings, including alarm clocks and telephones (38).

2. β-Blockers and targeting of the sympathetic nervous system

According to consensus guidelines (335), β-blocker treatment carries a class I recommendation in the setting of a clinical diagnosis of LQTS. β-Blockers reduce overall risk of cardiac events in children and adults, and treatment represents a class IIa indication in all individuals confirmed as carriers of LQTS mutations (118).

However, the efficacy of β-blockers is strongly genotype-specific. β-Blockers produce greatest symptom relief and cardiac event reduction in LQT1 patients, followed by LQT2 patients, and are least effective in LQT3 patients (334, 440). High-risk LQTS patient populations, including LQT1 males and LQT2 females, experienced 67 and 71% risk reduction with β-blocker treatment, respectively (145). Along with genotypic considerations, mutation-specific benefits exist based on location of the mutation in the ion channel. For instance, β-blocker treatment resulted in 88% risk reduction of life-threatening cardiac events in LQT1 patients harboring cytoplasmic-loop missense mutations compared with LQT1 patients harboring missense mutations in regions outside of the cytoplasmic loop (39). β-Blockers confer trigger-specific benefits as well, reducing the risk of exercise-induced cardiac events by 78 and 71% in LQT1 and LQT2 patients, respectively. However, β-blockers were not protective against events triggered by arousal or rest/sleep in LQT1 and LQT2 (146, 216). These findings are in line with the molecular physiology of these mutations. In nondiseased conditions, upregulation of repolarizing currents by β-adrenergic stimulation can enhance repolarization and favorably adapt QT interval for the concomitant increase in heart rate. However, in the setting of increased heart rates and LQT1 (and in some cases LQT2) disease, delayed rectifier current is not available for cardiac repolarization. One possible mechanism for β-blocker effect could be prevention of sudden and sustained increase in heart rates, which in itself is associated with arrhythmia formation in these subtypes.

In LQT2 events associated with sudden noises and startle, both adrenergic and other neural pathways may be triggered. As such, β-adrenergic blockade in LQT2 may not fully mitigate the risks. Although β-blockers remain first-line medical treatment in LQT2, a life-threatening cardiac event rate of 6–7% while on therapy remains a serious concern (334).

It is important to note that β-adrenergic stimulation affects multiple intracellular targets, including the pathways for intracellular calcium cycling that regulate L-type calcium current, the ryanodine receptor, and sarcoplasmic reticulum calcium uptake (18, 506). In LQTS, with impaired repolarizing currents (or enhanced depolarizing currents), β-adrenergic stimulation can be pro-arrhythmic, regardless of the genotype producing the impaired repolarization. β-Blockers therefore are commonly prescribed to all LQTS patients regardless of genotype.

In selected cases, left cervicothoracic sympathetic denervation (LCSD), which surgically interrupts adrenergic input to the heart, can be considered. In a study of 147 high-risk LQTS patients, cardiac event rate dropped by 91% per patient post-LCSD (369). When carried out at experienced centers, this treatment modality can be very effective in refractory cases.

3. Other pharmacological treatments in LQT2

Treatment of LQT2 syndrome by increasing serum K⁺ concentration has been explored, as hERG conductance is modulated by changes in extracellular K⁺, and LQT2 patients can develop low body and/or serum potassium levels (436). I_Kr activity decreases in the setting of hypokalemia (359) and increases in the setting of hyperkalemia, due mainly to extracellular K⁺ effects on inactivation gating (446, 454, 455).

As such, potassium supplementation and therapy with spironolactone, a potassium-sparing diuretic, were tested as treatment options to enhance I_Kr activity, and each treatment significantly shortened the QTc interval (89, 117). However, there was no evidence of a decrease in cardiac events. Safety dilemmas have been raised, including hyperkalemia, with the use of spironolactone in LQT2 treatment (38). Potential adverse effects of spironolactone may be explained by the fact that its metabolism into canrenoic acid was actually shown to induce hERG blockade and reduce channel activity (63).

4. Other pharmacological treatments in LQT3

The efficacy of local anesthetic drugs in LQT3 treatment is due to preferential inhibition of late sodium current (more than peak current). Limitation of the use of late sodium current blockers is due, in part, to off-target drug effects as well as to the relative selectivity of block of late versus peak sodium channel currents. The introduction of newer compounds with significantly higher selectivity for I_NaL block offers great potential for more effective clinical management of LQT3. The need for improved pharmacotherapy is supported by a high breakthrough cardiac event rate for LQT3 patients on β-blocker treatment, affecting as many as 15% of patients, generating a growing interest in I_Na-specific blocking agents (143). The use of the local anesthetic class of sodium channel blockers lidocaine (during acute settings and arrhythmic storm) and mexiletine (for suppressive therapy) are commonly clinically employed. Mexiletine administration shortens the QT interval in LQT3 patients, and not in LQT2 patients (370). Mexiletine blocks mutant Na_v1.5 channel function, diminishing the late sodium current more than the peak sodium current by stabilizing the inactivated state of the Nav1.5 channel (449). Ranolazine, on the other hand, is thought to block Nav1.5 current via open channel block (275). The utility of local anesthetic-like drugs is weighed against their off-target undesirable side effects. For example, flecainide inhibits peak I_Na and has been reported to unmask Brugada syndrome-type patterns on surface ECG (indeed, this class of agents is clinically utilized as a provocative challenge in the diagnosis of Brugada syndrome) (58). Various mechanisms of drug block of Nav1.5 underlie, in part, the mutation-specific efficacy of LQT3 pharmacological agents (see sect. IV).

Pharmacological sodium channel blockade appears on the surface to be a specific and targeted strategy against gain-of-function mutations in the sodium channel, and in individual cases does provide remarkable reduction in arrhythmia burden. However, a mortality benefit for these agents in LQT3 has yet to be established. Anti-arrhythmic therapy in LQT3, as with all LQT subtypes, must be considered carefully after thorough review of all patient-specific factors, including mutation or genotype-specific risk, clinical variables (e.g., family history, age, gender), and overall treatment plan (including role of ICD placement and/or LCSD) (143). Given the complexities of drug effect, these agents are often employed empirically in LQTS, with close attention to clinical response.

D. Emerging Translational Strategies in Congenital LQTS

1. Cellular models of LQTS

The physiological significance of cardiac channelopathies can be difficult to model at the cellular level because of the complex interplay between several ionic currents in the cardiomyocyte. While heterologous expression systems and animal models are invaluable tools in our understanding of disease phenotypes, they lack the genetic background and molecular biology of human cells. Induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) represent a useful platform for studying human cardiac ion channel function in health and disease in a cell system with human molecular and physiological components, acquired from unaffected patients and/or those carrying mutations.

In 2010, iPSC-CMs were generated from skin cells taken from two LQT1 patients with the autosomal dominant KCNQ1 R190Q mutation (277). These cardiomyocytes were spontaneously beating and classified as ventricular-, atrial-, or nodal-like based on cellular electrophysiology and morphology. Compared with WT controls, LQT1 cardiomyocytes demonstrate decreased I_Ks amplitude and prolonged APD, consistent with LQT1 pathogenesis. More recently, Jouni et al. (199) derived iPSC-CMs from a LQT2 patient harboring a hERG A561P mutation, to study the electrophysiological consequences of hERG loss-of-function in a patient-specific genetic background. Ventricular-, atrial-, and nodal-like action potentials were recorded, and a trafficking defect was characterized as the mechanism of loss of I_Kr, leading to a prolonged APD. Furthermore, a study by Jones et al. (195) reported the use of mRNA silencing in human iPSC-CMs to show that the hERG1b isoform is a significant component of cardiac repolarization, and that loss of hERG1b function provides a substrate for arrhythmogenesis. Several additional studies have investigated LQT3 in the context of iPSC-CMs (121, 252, 420).

In the context of drug development, iPSC-derived cardiomyocytes provide a means for testing compounds in patient-specific genetic backgrounds, allowing for a targeted approach in optimizing therapy. Terrenoire et al. (420) studied human iPSC-CMs from a patient harboring a LQTS mutation in SCN5A and a polymorphism in hERG. Upon electrophysiological characterization of ionic currents, it was determined that the SCN5A F1473C mutation was responsible for the LQTS clinical phenotype via increased I_NaL, while the hERG K897T variant was a benign polymorphism. Experiments in the iPSC-CMs revealed key rate-dependent properties of I_NaL and its inhibition by mexiletine; these results provided insight into control of arrhythmias in the proband. This study offers an example of how iPSC-CMs can elucidate mechanisms of LQTS pathogenesis to inform patient-specific therapy.

Use of iPSC-CMs is not without limitations. Discrepancies may still exist in the expression profile of cardiac ion channels between iPSC-derived cardiomyocytes and native human cardiac cells due in part to differences in cell maturity and growth environment. Cardiomyocyte differentiation can produce a mix of atrial-like and ventricular-like cells of varying morphology. Furthermore, recent publications using iPSC-CMs report a resting membrane potential depolarized to near −60 mV (252, 420), which can alter properties of important cardiac channels such as Na_v1.5 and affect state-dependent pharmacological inhibition. Efforts are underway to improve iPSC-CMs by using physiological systems to promote cellular maturity such that they more closely resemble native mature cardiomyocytes. As efficiency and specificity of cell differentiation improve, iPSC-CMs hold great promise as a model cell system for the study of LQTS disease and patient-specific pharmacology (115).

2. Pharmacological and medical approach to LQTS

More specific targeting of the mechanisms underlying ion channel defects that predispose to arrhythmia guides the advancement and rational drug design of pharmacological agents in the treatment of congenital LQTS. The experimental drug GS-967 preferentially inhibits late I_Na over peak I_Na with considerable selectivity, and does so more potently and efficaciously than flecainide and ranolazine, while also reducing arrhythmia burden in experimental rabbit cardiomyocytes (43). Computational modeling of GS-967 suggests therapeutic benefit of late I_Na block by decreasing arrhythmogenesis in the setting of LQTS (481). Another experimental compound, F15845, selectively blocks late I_Na potently, preventing cardiac angina and arrhythmia in animal models (324, 433). The benefits of any selective late I_Na blocker must be weighed against the relative effects on peak I_Na, in addition to off-target effects that include block of peak I_Na and I_Kr. Studies to date suggest that selective late I_Na blockers are not proarrhythmic; clinical trials of these and related drugs are currently underway, and it is expected that, in the near future, these studies will inform whether this drug class represents a viable treatment option in congenital LQTS (25).

Along with more selective targeting of LQT-associated mutant ion channels, growing interest in both mutation-specific and tissue-specific treatment strategies may become fruitful. For instance, ectopic rhythms leading to ventricular fibrillation in LQT patients has been shown to originate 50% of the time from the Purkinje fiber cells in the cardiac specialized conduction system (158, 306, 414). The Purkinje system is especially prone to arrhythmia formation, owing to differences in membrane current density, APD and morphology, and calcium cycling (102, 132, 158). In a computational model of the His-Purkinje system, LQT3-associated mutations were shown to confer a more severe phenotype in Purkinje fibers compared with LQT1 and LQT2 mutations (185), predictions that were confirmed in isolated murine Purkinje fiber cells expressing ΔKPQ mutant Na⁺ channels (175). Continued research into the functional implications of mutations in different cell types may uncover further therapies directed at tissues most prone to arrhythmia formation. For example, targeting the insertion of these Purkinje fibers in the ventricular endocardium via catheter ablation may prove to be an effective strategy in reducing arrhythmia burden (471).

VII. CONCLUSIONS

The growing body of mechanistic insights into the pathophysiology of channelopathies and other arrhythmia-causing mutations has helped elucidate the molecular underpinnings of congenital LQTS and genotype-phenotype correlation in disease. Although insights from benchside to bedside have facilitated progress toward better therapeutic strategies, there remains a need for tailoring management toward individuals in a genotype-driven and mechanism-specific manner to optimize care. Techniques such as iPSC-derived cardiomyocytes and high-throughput patch clamping that allow for screening of compounds in patient-specific contexts may help pave the way for future development of therapeutics. In addition, continued progress toward fundamental understanding of mechanisms of ion channel function and drug-channel interaction will guide the development of more effective, mechanism-based molecular agents in the treatment of LQTS.

GRANTS

This work was supported by the American Heart Association and National Heart, Lung, and Blood Institute Grants F30HL129656, 5K08HL116790, and 5R01HL123483.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.