ANNALS OF THE NEW YORK ACADEMY OF SCIENCES MARROW

The long QT syndrome (LQTS) is largely treated pharmacologically with β-blockers, despite the role of sympathetic activity in LQTS being poorly understood. Using the trigger–substrate model of cardiac arrhythmias in this review, we amalgamate current experimental and clinical data from both animal and human studies to explain the mechanism of adrenergic stimulation and blockade on LQT arrhythmic risk and hence assess the efficacy of β-adrenoceptor blockade in the management of LQTS. In LQTS1 and LQTS2, sympathetic stimulation increases arrhythmic risk by enhancing early afterdepolarizations and transmural dispersion of repolarization. β-Blockers successfully reduce cardiac events by reducing these triggers and substrates; however, these effects are less marked in LQTS2 compared with LQTS1. In LQTS3, clinical and experimental investigations of the effects of sympathetic stimulation and β-blocker use have produced contradictory findings, resulting in significant clinical uncertainty. We offer explanations for these contradicting results relating to study sample size, the dose of the β-blocker administered associated with its off-target Na+ channel effects, as well as the type of β-blocker used. We conclude that the antiarrhythmic efficacy of β-blockers is a genotype-specific phenomenon, and hence the use of β-blockers in clinical practice should be genotype dependent.


Introduction
Cardiac arrhythmias result from disruption in the orderly sequence of action potential (AP) activation and propagation through successive regions of the myocardium, causing failure of coordinated and effective cardiac contraction. 1 Of arrhythmic syndromes associated with congenital ion channel abnormality, long QT syndrome (LQTS) is characterized by prolonged electrocardiographic (ECG) QT intervals attributable to elongated ventricular AP duration (APD). 2 The LQTS has a prevalence of 1 in 2000 persons; 3 it is associated with predisposition to the normally self-terminating episodic polymorphic ventricular tachycardia (VT) torsades de pointes (TdP), with the potential to degenerate into ventricular fibrillation (VF) and/or sudden cardiac death. 2,4,5 LQTS patients also show atrial fibrillation (AF) more commonly than the remaining population, 6 with up to one-third of LQTS patients developing self-terminating atrial arrhythmias under daily life conditions. 7 doi: 10.1111/nyas.14425  LQTS includes many subtypes associated with abnormalities in up to 16 genes. 8 Table 1 summarizes the causative mutations and currents involved in LQTS1, LQTS2, and LQTS3. The three major LQTS genes KCNQ1, HERG, and SCN5A account for 49%, 39%, and 10% of LQTS cases, respectively. 9 All three subtypes cause prolongation of the APD (shown in Fig. 1A). LQTS1 is associated with a loss-of-function mutation in KCNQ1, which encodes the K V 7.1 channel; the mutation results in decreased I Ks current that is responsible for the slow activating component of the delayed rectifier potassium current during phase 3 repolarization of the cardiac AP (shown in Fig. 1B). LQTS2 is associated with a loss-of-function mutation in HERG encoding the K V 11.1 channel, and results in decreased I Kr current that is responsible for the rapidly activating component of the delayed rectifier potassium current during phase 3 repolarization of the cardiac AP (shown in Fig. 1C). LQTS3 is associated with a gain-of-function mutation in SCN5A, which encodes the Na V 1.5 channel; this results in increased late Na + current, I Na-L , as a consequence of impaired channel inactivation, with I Na being responsible for phase 0 depolarization of the cardiac AP (shown in Fig. 1D).

The trigger substrate model of cardiac arrhythmias
Reentry arrhythmia occurs when an AP wave fails to extinguish and results in excitation of previously excited, but now recovered, regions. The trigger-substrate model explains reentrant arrhythmias as occurring following the application of a single trigger to an arrhythmic substrate. 1,10 The trigger, typically an extrasystole, forms an area of refractoriness-a functional obstacle-around which the AP wave circulates. The substrate is formed of slowed conduction velocity, resulting in tissue ahead of the wave becoming excitable and transmural dispersion of repolarization (TDR) owing to heterogeneities in recovery from excitability; this results in a unidirectional conduction block preventing the wave from self-extinguishing. 11,12 Thus, arrhythmias arise when both a trigger and a substrate occur simultaneously (shown in Fig. 2). Sympathetic activity is associated with an increased risk of arrhythmias, as adrenergic stimulation increases the risk of both arrhythmic triggers and substrate. Unfortunately, despite the term arrhythmic trigger being used in cellular physiology to mean abnormal wave formation, which develops a functional obstacle, it is also used in clinical papers to refer to an environmental event that precedes and predisposes to arrhythmias, for example, exercise. Thus, to avoid confusion, we will use trigger (in italics) to refer to the physiological organ-level phenomenon of abnormal depolarization, and trigger (no italics) to refer to the clinical setting of the patient.

Sympathetic stimulation promotes arrhythmogenesis via enhancing triggered activity and developing arrhythmic substrates
Sympathetic input to the myocardium acts mainly on β-adrenergic receptors, primarily β1 adrenergic  receptors (β1-AR) and, to a lesser extent, β2-AR, but also the α1-adrenergic receptor (α1-AR), in order to promote chronotropic and inotropic effects. [13][14][15][16] AR are G protein-coupled receptors (GPCRs) that activate an intracellular cascade acting on a variety of surface membrane ion channels and intracellular Ca 2+ homeostasis proteins. β-AR are G αs GPCRs that activate adenylate cyclase to increase intracellular cAMP levels and trigger the protein kinase A (PKA) phosphorylation cascade. 13,14,[16][17][18] α1-AR is a G αq GPCR that activates PLC that then cleaves phosphatidylinositol, increasing IP 3 that binds to IP 3 receptor and diacylglycerol to initiate PKC phosphorylation changes. 19 Together, these phosphorylation changes can increase triggered activity and cause the development of an arrhythmic substrate, hence increasing the risk of cardiac arrhythmias. 16,17,[20][21][22][23][24] Figure 3 summarizes AR signaling and its effects on arrhythmic triggers.
β-Adrenergic receptors promote triggered activity Table 2 summarizes the effects of β1-AR stimulation on surface membrane currents and intracellular Ca 2+ homeostasis in relation to triggered activity. Briefly, the β-AR-PKA cascade initially phosphorylates Na V 1.5 that is responsible for phase 0 depolarization, giving rise to the rapid inward current (I Na ) and leading to faster inactivation kinetics and augmentation of I Na amplitude. 25  Reentrant arrhythmia occurs when an action potential (AP) wave fails to extinguish and results in reexcitation of previously excited, but now recovered, regions. Abnormal wave formation via triggered beats causes a functional obstacle, a region of refractoriness, around which the AP wave circulates. This is the initiating event-arrhythmic trigger. Arrhythmic substrates maintain reentrant arrhythmia and include slowed conduction velocity (CV), resulting in tissue ahead of the wave being excitable, and transmural dispersion of repolarization (TDR), resulting in unidirectional conduction block preventing the wave from self-extinguishing.
channel Ca V 1.2 that is responsible for the plateau phase 2, leading to an increase in the inward depolarizing Ca 2+ current (I Ca-L ) Ca 2+ entry. 18,23 In relation to pathology, reactivation of Ca 2+ channels and an increase in I Ca-L during phase 2 or 3, and made more likely by adrenergic modulation effects, will result in early afterdepolarizations (EADs) and hence an arrhythmic trigger. 22,23 Third, β-AR exerts complex effects on HERG channels, which gives rise to the rapidly activating component of the delayed rectifier potassium current (I Kr ) that contributes to AP plateau phase 2 and phase 3 repolarization, where adrenergic stimulation has been mainly shown to decrease (but also to increase) I Kr (depending on stimulation conditions). [33][34][35][36][37] Fourth, PKA phosphorylates K V 7.1, which gives rise to the slow activating component of the delayed rectifier potassium current (I Ks ) responsible for AP plateau phase 2 and phase 3 repolarization, leading to an increase in I Ks and faster repolarization. [38][39][40][41][42][43] Pathological disruption of the repolarization currents will promote triggered activity by reducing the repolarization current and offsetting delayed afterdepolarizations (DADs), hence increasing DAD amplitude and prolonging the APD associated with EADs. [44][45][46] Fifth, β-AR stimulation has also been shown to increase Na + -K + pump activity 47 associated with DADs through Na + -Ca 2+ exchanger (NCX) promotion. 48 Finally, β-AR-PKA phosphorylates a variety of intracellular targets relating to Ca 2+ homeostasis, including ryanodine receptor (RyR), thereby increasing sarcoplasmic reticulum (SR) Ca 2+ release 49 and phospholamban, which leads to disinhibition of sarco/endoplasmic reticulum Ca 2+ -ATPase (SERCA) and increased Ca 2+ recycling. 50 This ultimately increases the magnitude of intracellular Ca 2+ transients driving the depolarizing NCX activity, therefore promoting DADs. 51,52 α1-Adrenergic receptors promote triggered activity Table 2 summarizes the effects of α1-AR stimulation on surface membrane currents in relation to triggered activity. α1-AR activity contributes to the previously discussed generation of EADs and DADs. First, phosphorylation of the K V 4 channels, giving rise to the transient outward potassium current (I to ) responsible for the initial rapid repolarization phase 1 of the AP, reduces I to and allows the plateau phase to occur at higher voltages, permitting an increase in APD and I Ca-L , and thus is an important contributor to EADs. [53][54][55][56] Second, α1-AR inhibits the inwardly rectifying potassium channels, reducing background potassium current (I K1 ), hence importantly contributes to DADs by reducing the polarizing current, which normally offsets them. 57,58 Finally, α1-AR has also been shown to increase Na + -K + pump activity, 59 which accelerates the forward depolarizing mode of NCX and thereby contributes to the generation of DADs. 48 Overall, in healthy hearts, the increase in Ca 2+ transients and depolarizing currents due to sympathetic stimulation are balanced by an increase in the repolarizing potassium currents. Imbalance in these mechanisms in LQTS hearts, for example, due to failure to increase repolarizing potassium currents in LQTS1 and LQTS2 or to a persistent increase in I Na-L prolonging repolarization in LQTS3, will predispose them to triggered activity during adrenergic stimulation.

Sympathetic stimulation promoting TDR arrhythmic substrate
In addition, sympathetic activity promotes the development of an arrhythmic substrate (shown in Fig. 4) owing to regional heterogeneities in sympathetic input and response, including base-to-apex and epicardial-to-endocardial gradients of those  . Sympathetic input to the myocardium acts on α1and β1-adrenergic receptors (α1-AR and β1-AR). β1-AR is a G αs GPCR that activates adenylate cyclase (AC), increasing intracellular cAMP levels and triggering the protein kinase A (PKA) phosphorylation cascade that acts on a variety of surface membrane ion channels and intracellular Ca 2+ homeostasis proteins. α1-AR is a G αq GPCR that activates PLC that cleaves phosphatidylinositol increasing IP3, which then binds to the IP3 receptor and diacylglycerol (DAG), leading to a PKC phosphorylation cascade that acts on a variety of surface membrane ion channels. These changes increase the risk of early afterdepolarizations (EADs) and delayed afterdepolarizations (DADs) arrhythmic triggers.
parameters. 16 Regional differences in sympathetic input can arise as a consequence of differences in sympathetic nerve distribution as well as β-1AdR and α1-AR density. [60][61][62][63] Differences in response to sympathetic stimulation largely arise as a consequence of regional differences in expression levels and activity of membrane ion channels. 40,63-66 These regional differences consequently result in TDRs, 67 and hence an arrhythmic substrate. 66,68,69 As a consequence of its effect of increasing arrhythmic triggers and substrate, sympathetic activity is associated with an increased incidence of cardiac arrhythmias. 20,21 Thus, sympathetic βadrenergic antagonism in inherited arrhythmic syndromes, especially LQTS, has been an important area of research, as it elucidates the arrhythmogenic mechanism in those syndromes and offers a therapeutic target.

Implications of cardiac sympathetic stimulation in LQTS is a genotype-specific phenomenon
Briefly, after the identification of the LQTS genes, clinical reports demonstrated genotype-specific environmental triggers relating to sympathetic or parasympathetic (vagal) activity. In 1995, Schwartz and colleagues 70 provided an early account of the effect of increasing heart rate (HR), reflecting sympathetic activation, on the QT interval of seven LQTS3 and four LQTS2 patients. LQTS3 patients shortened their QT interval more than LQTS2 patients, and even more than healthy controls. Furthermore, Schwartz and colleagues reported cardiac events occurring more during rest or sleep for LQTS3 but during emotional or physical stress for LQTS2. These findings justified further clinical and experimental studies to investigate genotypespecific triggers.
These clinical triggers can be classified into three major types: 71 (1) physical stress, or exercise, characterized by progressively increasing HR via sympathetic stimulation from both neural and circulating catecholamines; (2) emotional stress and abrupt arousals characterized by sudden sympathetic activity via neural catecholamines, whereby HR does not significantly rise and without permitting time for QT shortening to faster rates; and (3)   parasympathetic vagal input, and hence the absence of significant catecholaminergic activity.
One study investigated the association of the aforementioned triggers with cardiac events in 371 LQTS1, 234 LQTS2, and 65 LQTS3 patients. 71 In LQTS1, patients experienced 62% of cardiac events during exercise and only 3% during rest/sleep. Furthermore, 68% of lethal events occurred during exercise, suggesting that exercise/physical stress, and hence sympathetic activity, appears to be the predominant trigger. In LQTS2, patients experienced 43% of cardiac events during emotional stress, 13% during exercise, and 29% during rest/sleep. Interestingly, 49% of lethal cardiac events occurred during rest/sleep. 71 However, sleep is not a homogenous state of vagal dominance. Periods of rapid eye movement (REM) sleep are associated with significant sympathetic activation that can trigger cardiac events. 72 Despite auditory triggers being rare in LQTS1 and LQTS3, 26% of LQTS2 patients experienced them, with 64% reported to have occurred during sleep. 71 Therefore, emotional stress and abrupt triggers during sleep (auditory stimuli) reflect sudden sympathetic activity that appears to be a predominant trigger in LQTS2 patients. The difference in triggers between LQTS1 and LQTS2 patients is recapitulated by other clinical reports, such as those finding that events in LQTS1 are associated with swimming (i.e., exercise), whereas events in LQTS2 are associated with auditory stimuli. 73,74 Together, these findings implicate sympathetic activity as a trigger in both LQTS1 and LQTS2 but operating by different mechanisms. In LQTS1, sympathetic activity promotes arrhythmia during sustained activity, whereas in LQTS2 it promotes arrhythmia during sudden activity.
In LQTS3, patients experience 39% of cardiac events during rest/sleep and only 13% and 19% during exercise and emotional stress, respectively. Furthermore, 64% of lethal events occur during rest/sleep. 71 This is consistent with previous reports, including those from a smaller study that found that symptomatic LQTS3 patients had cardiac events at rest or during sleep, with only one also during emotional stress. 70 In a family with a high incidence of nocturnal sudden death, patients were characterized with sinus bradycardia and bradycardia-dependent QT prolongation, indicating a bradycardic mode of cardiac death. 75 Therefore, in contrast to LQST1 and LQTS2,  This arises as a consequence of regional heterogeneities in sympathetic nerve distribution, adrenergic receptor density, and regional differences in expression levels and activity of membrane ion channels. Thus, sympathetic stimulation promotes transmural dispersion of repolarization (TDR), which allows the development of a unidirectional conduction block and, hence, the development of a substrate for reentry arrhythmia.
rest/sleep-reflecting a period of decreased cardiac sympathetic activity-appears to be the trigger in LQTS3 patients. Though this initially seems to suggest that sympathetic stimulation is protective, it is worth noting that cardiac events still occurred during periods of increased sympathetic activity, such as emotional stress and exercise. 70,71 Furthermore, as previously discussed, sleep does not necessarily mean an absence of sympathetic activity. 72

Molecular basis of the differential role of sympathetic stimulation in LQT subtypes
The differential response of LQTS subtypes to sympathetic stimulation can be understood by considering the role of the mutated channel. In the healthy myocardium, sympathetic stimulation increases both inward depolarizing currents, for example I Ca-L and I NCX , and outward repolarizing currents, for example I Ks , with the net outward current ensuring faster repolarization and shortened APD.

LQTS1
In LQTS1, decreased I Ks results in a scenario in which sympathetic stimulation still increases the depolarizing currents, but with a much weaker repolarizing current. Therefore, the APD will be paradoxically prolonged. This was demonstrated clinically both by exercise monitoring and epinephrine infusion. Clinical monitoring of LQTS1 patients during exercise demonstrated impaired HR response to exercise, reflected in prolonged QT due to extended ventricular repolarization time. 73,76,77 Additionally, infusing patients with epinephrine resulted in a sustained prolonged QT interval. 78 77,80 Thus, a sustained increase in the APD following adrenergic stimulation in LQTS1 will increase the risk of EADs, and therefore increase the risk of an arrhythmic trigger.
Furthermore, the previously discussed effect of adrenergic stimulation promoting TDR is potentiated in LQTS1 syndrome. Clinically, this was shown by prolonged ECG indicators of TDR, for example, Tpeak-end (Tcp-e) intervals, following exercise stress test 81 or epinephrine administration in LQTS1 patients. 77,82 Experimental investigation in the canine chromanol 293B pharmacological model of LQTS1 demonstrated that adrenergic stimulation prolongs APD in M cells, where I Ks is intrinsically weak but abbreviates APD in epicardial and endocardial cells, resulting in the persistent increase in TDR and possible induction of TdP. 77,80 Thus, the sustained increase in TDR following adrenergic stimulation in LQTS1 allows the development of an arrhythmic substrate. Therefore, in LQTS1 patients, sympathetic stimulation results in a persistent increase in both arrhythmic trigger and substrate and, consequently, increases the risk of cardiac arrhythmias. This accounts for the previously discussed reports that LQTS1 patients are at an increased risk of cardiac events during sympathetic stimulation, hence providing the support for the use of β-blockers in the treatment of LQTS1.

LQTS2
In LQTS2, decreased I Kr results in a scenario in which sympathetic stimulation increases the depolarizing currents, but with a much weaker rapidly activating repolarizing current I Kr . This results in the dominance of inward depolarizing currents, and hence prolonged APD. However, following a delay, the slower I Ks is increased and thus the APD will shorten. Therefore, in response to sympathetic stimulation, there will be a transient prolongation of the APD, which then reverses and shortens. This was demonstrated clinically both by exercise monitoring and epinephrine infusion, where the QT interval initially prolonged but later shortened to control values. 76,78,79 Experimental isoproterenol infusion in canine left ventricle wedge preparations, arterially perfused with I Kr blocker d-sotalol mod-eling LQTS2, initially prolonged then abbreviated the QT interval and the APD in M cells. 77 Similarly, the dofetilide pharmacological model of LQTS2 in guinea pig ventricular myocytes demonstrated that either exposure to isoproterenol or rapid pacing shortens the APD but only after an initial 3-min prolongation. 83 Thus, a transient increase in the APD following adrenergic stimulation in LQTS2 will increase the risk of EADs and, therefore, an arrhythmic trigger. 77,83 Furthermore, the previously discussed effect of adrenergic stimulation promoting TDR has been reported in LQTS2. In the d-sotalol canine pharmacological model, isoproterenol initially prolonged, then abbreviated, the APD of M cells but always shortened epicardial APD, transiently increasing TDR and the ability to induce TdP. 77 A transient increase in TDR following adrenergic stimulation in LQTS2 allows the reversible development of an arrhythmic substrate. Therefore, in LQTS2, patients demonstrate that sympathetic stimulation results in a brief increase in both arrhythmic trigger and substrate and, consequently, transiently increases the risk of cardiac arrhythmias. This accounts for the previously discussed reports that LQTS2 patients are at an increased risk of cardiac events during sudden sympathetic stimulation, hence providing the support for the use of β-blockers in the treatment of LQTS2.
Nonetheless, these effects of sympathetic stimulation are less pronounced in LQTS2 than in LQTS1. Clinically, the characteristic ECG features of TDR increase more prominently during exercise in LQTS1 than in LQTS2 patients. 81 Similarly, sympathetic stimulation with epinephrine produces a greater increase in TDR in LQTS1 than in LQTS2. 84 This may explain why cardiac events during exercise, and hence sympathetic activity, are less common in LQTS2 than in LQTS1. In turn, this predicts that β-blockers would be less effective in LQTS2 than in LQTS1 patients.

LQTS3
In contrast to LQTS1 and LQTS2, in which clinical and experimental findings have been consistent, LQTS3 reflects a more complex situation. In LQTS3, delayed channel inactivation results in excessive I Na-L , and hence prolonged APD. Increased I Na-L is associated with an increased risk of arrhythmias via a variety of mechanisms relating to increased EAD The effects of adrenergic stimulation have been reported to influence both the fast and late components of the Na + current. [25][26][27][28][29][30][31][32]86 Despite the important role of I Na-L in shaping the AP and contributing to arrhythmias, it remains a relatively poorly understood current, with the exact short-and long-term consequences of adrenergic stimulation remaining under investigation. 86,89 Because of this, reports investigating adrenergic effects on I Na-L have offered different results. Studies in rabbit ventricular myocytes, for example, reported that adrenergic stimulation increases I Na-L via both the PKA and CaMKII pathways. 89,90 In other experiments, adrenergic stimulation enhanced inactivation of the Na + channel, and hence reduced I Na-L . 25,28 Interestingly, however, in a number of different experiments, adrenergic stimulation has been shown to increase I Na-L , hence effectively exacerbating the inactivation defect of the Na V 1.5 channel. 25,91,92 The effect of increased I Na-L in LQTS3 on APD and arrhythmogenicity will depend on the interaction with the effects of adrenergic stimulation on other ion currents that contribute to repolarization and intracellular Ca 2+ homeostasis. Interestingly, clinical reports reveal that in response to an increase in HR or epinephrine administration, the QT interval and APD are shortened to normal control values, 78,79 or even abbreviated to values shorter than the controls. 70,78 Experimental investigation confirmed the clinical observations. In canine left ventricles wedge preparations, arterially perfused with ATX-II which augments I Na-L modeling LQTS3, isoproterenol constantly shortened the QT interval and the APD in M cells, even to values below the control. 77 Similarly, the anthopleurin pharmacological model of LQTS3 in guinea pig ventricular myocytes demonstrated that either exposure to isoproterenol or rapid pacing shortens the APD. 83 In a murine LQTS3 model, isoproterenol shortened rate-corrected APD and suppressed arrhythmias. 93 Together, these studies report a decrease in the APD following adrenergic stimulation in LQTS3, decreasing the risk of EADs and, therefore, of arrhythmic trigger. 77,83 Moreover, the previously discussed effect of adrenergic stimulation promoting TDR has not been reported in LQTS3. Interestingly, in experimental reports of canine left ventricle wedge preparations arterially perfused with I Na modifier ATX-II modeling LQTS3, isoproterenol infusion shortens the APD in endocardial, M, and epicardial cells, and hence persistently decreases TDR and suppresses TdP. 77 Similarly, in a Scn5a + / KPQ murine model of LQTS3, dobutamine sympathetic stimulation reduced the incidence of repolarization alternans. 94 Thus, decreased TDR following adrenergic stimulation in LQTS3 suppresses the development of an arrhythmic substrate. Therefore, clinical and experimental data appear to suggest that sympathetic stimulation results in a decrease in both arrhythmic trigger and substrate, and hence decreases the risk of cardiac arrhythmias.
However, some studies contradicting this conclusion have reported that sympathetic activity may be without effect, or may even be proarrhythmic. [94][95][96][97] These discrepancies arise because of differences in the protocol of sympathetic stimulation and dose of pharmacological agonists. These reports showed a protective effect of sympathetic activity in response to a progressive increase in adrenergic stimulation. However, under a protocol of sudden accelerations in HR in a murine model of LQTS3, the APD was prolonged and caused EADs and triggered arrhythmias. 95 Nonetheless, the same study found that isoproterenol infusion normalized the response to rate acceleration in vitro by preventing lengthening in APD, and suppressed arrhythmias upon premature stimulation in vivo 95 -hence, demonstrating that differences in protocol produce different effects on APD and, consequently, on the risk of arrhythmic trigger. These results were further complimented by in vivo LQTS3 murine reports: following an increase in ventricular rate produced by dobutamine sympathetic stimulation, a delay in ventricular repolarization adaptation was reported. 94 This may represent an increased arrhythmic risk by sympathetic stimulation following transient HR increase.
Additionally, the dose of sympathetic agonist administered may account for discrepancies between reports. In a canine experimental model of LQTS3 induced by anthopleurin-A, 0.5 μg/kg of epinephrine did not induce polymorphic ventricular tachyarrhythmia (PVA) and shortened the activation recovery interval at all sites and decreased TDR. By contrast, a dose of 1.0 μg/kg induced PVA, 97 which demonstrates a dosedependent opposite effect of sympathetic stimulation on arrhythmic tendency. 97   of sympathetic activity in determining arrhythmic risk, such that it is not merely whether sympathetic activity is increased but the intensity of the stimulation that determines if it is pro-or antiarrhythmic. A computational model of LQTS3 (in mutant guinea pig ventricular myocytes) replicated the previous results finding that the effects of isoproterenol on EAD and TDR were dose and pacing protocol dependent. 98 Therefore, both clinical and experimental evidence demonstrate that sympathetic activity plays a complex role in LQTS3.
It is important to treat experimental data with caution when considering generalizations to humans; animal models are limited by how similar their AP and ionic currents are to human counterparts. In this regard, many experiments use mouse models to study cardiac arrhythmias because of large structural and electrophysiological similarities to humans. [99][100][101][102][103] For example, I Na plays the same physiological role in both mouse and humans, being responsible for the rapid depolarization phase of the AP 102 and thus enabling the study of Na + channel abnormalities, such as LQTS3. [94][95][96]102 However, important differences exist, particularly regarding the less prominent Ca 2+ current and different expression and role of potassium channels in repolarization. 99,103 Considered together, this means mouse ventricular AP lacks the typical plateau phase present in human ventricular AP, as well as having a shorter APD. [102][103][104] This may influence whether changes in I Na-L are pro or antiarrhythmic, and hence affect the interpretation of findings. Nonetheless, transgenic Scn5a + / KPQ mice that model LQTS3 have been shown to reflect the LQTS3 human phenotype, including prolonged APD, arrhythmic tendency, and ECG features. [94][95][96]105 In summary, the effects of sympathetic stimulation are largely dependent on the clinical context and the experimental protocols used, which may influence the recommendation of use of β-blocker therapy in LQTS3 patients.

β-Blockers are the primary treatment for LQT syndrome
Currently, the predominant management strategy of LQTS is β-blocker use. [106][107][108] Justification for this arose from previously discussed early findings that associated sympathetic stimulation with a proarrhythmic phenotype, and the observation that the majority of cardiac events occurred during sympathetic activity, for example, exercise. However, evaluation of the efficacy of β-blockers in LQTS management is significantly limited to retrospective studies due to their apparent clinical efficacy, and hence ethical implications preventing the performance of prospective, placebo-controlled, randomized studies. 107 Furthermore, early studies lacked genotype profiling of patients, and hence examined β-blocker efficacy in LQTS without subtype specificity. Of those early studies, a large one was conducted on 233 LQTS patients who were symptomatic for syncope or cardiac arrest. For those patients not receiving antiadrenergic therapy, mortality 15 years after the first syncope was 60%, but for those on antiadrenergic therapy (β-blocker and/or left cardiac sympathetic denervation), mortality was 9%, 109 hence, supporting the use of β-blockers. However, as discussed previously, sympathetic activity influences arrhythmic risk differently in different LQT subtypes. Thus, the efficacy of β-blockers should be investigated in a genotypespecific manner.

β-Blocker therapy in LQTS1 and LQTS2
Genotype-specific clinical studies initially involved small sample sizes, such as a study of 69 LQTS1 and 42 LQTS2 patients in which β-blockers were found to have a significant effect in reducing cardiac event rate in both groups of patients. 110 Following these initial reports, larger clinical studies described the antiarrhythmic effects of β-blockers in both LQTS1 and LQTS2. For example, a study of 600 LQTS1 patients found that β-blockers were associated with a 79% reduction in the cumulative probability of first cardiac event; 111 in another study of LQTS2 patients, β-blockers were associated with a 63% reduction in the cumulative probability of first cardiac event. 112 Hence, the antiarrhythmic effect of β-blocker therapy was less in LQTS2 than in LQTS1 patients in the two studies. In yet another study, of 371 LQTS1 patients receiving β-blocker therapy 81% demonstrated recurrence-free survival, while 59% of 234 LQTS2 patients were recurrence free; nonetheless, the death rate was equally low in both patient groups (4%). 71 β-Blocker therapy was shown to reduce the incidence of cardiac events from 47% to 10% in LQTS1, and to 23% in LQTS2 patients; 113 and a recent meta-analysis found that β-blockers reduced the risk of cardiac events by 36 Ann 71% in LQTS1 and 52% in LQTS1. 114 Therefore, β-blocker therapy has been consistently associated with effective treatment of both LQTS1 and LQTS2 patients. [115][116][117] As previously discussed, sympathetic activity increases the risk of cardiac arrhythmias in LQTS1 and LQTS2 by increasing EADs and TDR. Thus, it is expected that the mechanism by which β-blockers exert their arrhythmic effects is by a reduction of incidence of EADs and TDRs. Indeed, clinical ECG monitoring of LQTS1 patients during exercise and recovery found that β-blockade reduced APD and hence EADs, measured as the QT interval; 118 this study also found that β-blockade reduced TDR, measured as the T peak-to-end interval (Tpe). 118 Additionally, propranolol suppressed the effect of epinephrine increasing TDR in LQTS1 and LQTS2 patients; 119 this effect was greater in LQTS1 than in LQTS2 patients. 119 Experimental pharmacological models of LQST1 and LQTS2 induced by chromanol 293B and d-sotalol, respectively, demonstrated that propranolol prevented the development of TDR induced by isoproterenol, and suppressed the development of spontaneous and stimulationinduced TdP. 77,80 Therefore, both clinical and experimental studies have demonstrated the effectiveness of β-blocker therapy in the management of LQTS1 and LQTS2 patients, with the therapy being more effective in LQTS1 than in LQTS2 patients. The evidence is consistent with a mechanism of action that involves a reduction in EADs and TDR, and hence a reduction in both arrhythmic trigger and substrate.

β-Blocker therapy in LQTS3
In contrast to LQTS1 and LQTS2, the use of βblocker therapy in LQTS3 patients has been more controversial. In a study of 28 LQTS3 patients, βblockers were found to have no significant effect on reducing cardiac event rate. 110 Consistent with this finding, later studies reported minimal or no protective effects of β-blocker treatment. For example, 65 LQTS3 patients receiving β-blocker therapy had a lower percentage of patients with recurrence-free survival (50%) and a much higher death rate (17%) compared with LQTS1 and LQTS2 patients. 71 Additionally, another study showed that β-blocker therapy reduced the incidence of cardiac event rate from 47% to just 32% in LQTS3 patients. 113 And in a murine LQTS3 model, acute β-adrenoceptor blockade by esmolol or propranolol, or chronic blockade by propranolol in vivo, did not suppress arrhythmias. 93 These studies are consistent with the conclusion that β-blocker therapy has no significant effect on reducing the risk of cardiac arrhythmias in LQTS3 patients.
On the other hand, other studies have described a proarrhythmic effect of β-blockers in LQTS3 patients. For example, in a family with a high incidence of nocturnal sudden death in combination with sinus bradycardia and bradycardia-dependent QT prolongation reflecting the LQTS3 phenotype, β-blocker therapy was contraindicated as it correlated with a bradycardic mode of cardiac death. 75 Similarly, in LQTS3 patients with the KPQ mutation, β-blockade correlated with slowed atrial, atrioventricular, and ventricular conduction. 120 In a pharmacological model of LQTS3 induced by ATX-II, the propranolol reversed the protective effects of isoproterenol; propranolol prevented APD abbreviation and promoted the development of TDR and TdP. 77 Similarly, in a murine LQTS3 model, β-blockade inhibited the beneficial effects of catecholamines, hence causing the prolongation of APD. 93 Studies in a murine LQTS3 model also demonstrated that 1 μM propranolol not only failed to suppress VT but also increased TDR, hence promoting an arrhythmic substrate. 96 Yet, more recent, larger scale studies have found a significant antiarrhythmic effect of β-blocker therapy in LQTS3 patients. In a study with 111 LQTS3 patients treated with β-blockers, females had an 83% reduction in cardiac events, 121 and yet no significant effect in males; this was explained by the low number of arrhythmic events in males, which prevented a conclusive finding. Nonetheless, there was no detrimental effect of β-blockers, as was previously reported. Similar results were reported by a later study involving 237 LQTS3 patients, which found a trend toward significant benefit of β-blockers in reducing cardiac events in LQTS3 females, but not in LQTS3 males. 122 And one of the largest studies conducted involving 4480 personyears demonstrated the efficacy of the β-blocker nadolol in preventing life-threatening events in all of the three major LQTS genotypes. 123 Additionally, rapid sympathetic activity has been implicated as a trigger in some LQTS3 patients, as discussed previously, 94,95 which supports the use of β-blockers, as they would prevent such triggers  95 Finally, in a canine LQTS3 pharmacological model, propranolol reversed the epinephrine-induced increase in TDR and prevented premature ventricular complexes and polymorphic VT induced by epinephrine; 97 and carbachol induced VT and VF in a murine LQTS3 model that were prevented by pretreatment with propranolol, demonstrating its antiarrhythmic effect. 124

Clarification of the contradictory findings in LQTS3
Potential explanations for the contradicting findings include study sample size, the dose of the β-blocker administered relating to its off-target Na + channel effects, and the type of β-blocker used.

Sample size
The relatively sample size of the early clinical studies, which reported no beneficial effect of propranolol in LQTS3, has been criticized. 114,121,124 These studies likely lack the statistical power necessary to detect a significant difference. More recent reports contain larger sample sizes and report an antiarrhythmic effect of β-blockers.

β-Blocker dose and Na + channel effects
In a murine LQTS3 model, VT triggering following programmed electrical stimulation was modulated in a dose-dependent manner by propranolol. 125 These effects were correlated with the effects of propranolol on TDR such that the concentration that suppressed VT triggering was associated with decreased TDR, and vice versa. 125 In a similar canine model, under moderately high concentrations, propranolol increased TDR; however, with even higher concentrations associated with further Na + channel inhibition, these effects were reversed, decreasing TDR. These findings are further supported by the computational model of LQTS3 mentioned above (in mutant guinea pig ventricular myocyted); 98 this study found that at low concentrations, propranolol reversed the beneficial effects of β-adrenoceptor activation and had proarrhythmic effects; at higher concentrations, however, propranolol had antiarrhythmic effects. Consistently, clinical data have shown that 40% of patients with ventricular arrhythmias responsive to propranolol receive doses significantly higher than those required for β-adrenoceptor blockade (>150 ng/mL). 126 A physiological explanation for the paradoxical effects of propranolol 125,127 is summarized in Figure 5. Independent of β-adrenoceptor blockade, propranolol has Na + channel blocking effects. 128,129 Thus, propranolol acting on the myocardium will block the Na V 1.5 channel causing a decrease in the amplitude of phase 0 depolarization of the AP, as well as a decrease in the Na + window/late Na + current. Therefore, in the endocardium, these effects will cause an abbreviation in the APD. 125,127 However, in the epicardium, while these effects do occur, the outcome is modulated by the presence of a dominant transient outward current (I to ), which will open at more negative potentials and efflux of K + will hyperpolarize the membrane. Though this would likely shift the membrane potential below the activation range of I Ca and hence decrease the number of open Ca 2+ channels, relevant Ca 2+ channels have already been activated during phase 0 depolarization. Furthermore, as a consequence of membrane hyperpolarization, the electrical gradient of the Ca 2+ ions across the membrane is increased. This means that open Ca 2+ channels will conduct more ions; I to will rapidly inactivate reducing the outward current giving rise to a delayed net inward current, I Ca ; and the delay in I Ca would, in turn, delay activation of outward repolarizing components, hence resulting in a prolongation of the epicardial APD. 74,76 This demonstrates the contrasting effects of propranolol on the endocardium and epicardium, resulting in increased TDR and thus a proarrhythmic phenotype via promoting spatial heterogeneity in the form of transmural gradients. 125,127 Nonetheless, at a higher critical threshold of propranolol concentration, the Na + channel block in the epicardium will be sufficient to reduce phase 0 depolarization to even greater extent. As a consequence, with the shift of potentials negative to the activation threshold of I Ca , the outward currents would overwhelm any activated inward current, preventing the previously discussed proarrhythmic sequence of electrical changes and, hence, exerting antiarrhythmic effects. 125,127 This demonstrates that the antiarrhythmic properties of propranolol are achieved at higher concentrations. However, the optimal dose of propranolol in LQTS3 has not been investigated clinically. Furthermore, the optimal antiarrhythmic dose in humans may be of such a high concentration that its use is not clinically feasible. 125 38 Ann    reported the superior efficacy of the nonspecific β-blockers nadolol and propranolol, over selective blockers, such as metoprolol and atenolol. [132][133][134] Similarly, a recent study with a population including LQTS3 patients found that nadolol was associated with a significant 62% reduction in the risk of life-threatening arrhythmic events. This was not the case for propranolol or selective β-blockers (i.e., metoprolol, atenolol, bisoprolol, carvedilol, and nebivolol). 123 Compliance. Regarding the reported differences in efficacy between the nonselective β-blockers nadolol and propranolol, they may be explained by differences in drug compliance. Noncompliance has been reported as an important factor underlying β-blocker therapy failures. 116,133 For example, in 216 LQTS1 patients, 67% of patients who suffered cardiac arrest were noncompliant. 116 And there have been no studies comparing the compliance of different β-blockers in LQTS patients. However, based on pharmacokinetic knowledge, previous reports predict a higher compliance for nadolol than propranolol. 135,136 For example, propranolol is highly lipophilic, allowing it to cross the blood-brain barrier (BBB) and reach concentrations within the cerebrospinal fluid similar to that in free plasma concentration. 135,137 By contrast, nadolol is significantly less lipophilic and hence does not cross the BBB. Therefore, propranolol is associated with more central nervous system side effects and potentially lower compliance than nadolol. This may account for previous reports of nadolol as more effective than propranolol in the treatment of LQT, or that propranolol is ineffective.
Long-term effects: receptor sensitization and electrical remodeling. The superior efficacy of nonselective β-blockers may be explained by the effects of selective blockers on receptor sensitization. The myocardium expresses β1, β2, and β3 adrenoceptors at proportions of 70-80%, 20-30%, and <2%, respectively. However, the functional expression of receptors has been reported to change both as a consequence of a variety of pathological conditions, such as heart failure, 138 and of pharmacological interventions, such as selective β1-receptor blockers. [139][140][141] Thus, selective β1adrenoceptor blockers will result in increased functional expression and activity of β2-adrenoceptors that, acting via the G as subunit, will result in down-stream positive inotropic and chronic effects similar to those generated by β1-adrenoceptor activation. Therefore, β2-adrenoceptors sensitization as a result of selective β1-adrenoceptor blocker use will compromise the ability of the β-blocker to prevent cardiac arrhythmias. 135,140,141 In addition to changes in receptor expression and sensitization, the long-term β-blocker use has been associated with electrical remodeling of the myocardium. These changes have been reported to be antiarrhythmic in some rhythm disturbances, such as AF. [142][143][144] Electrical remodeling arising from chronic β-blocker use primarily involves a decrease in the potassium repolarization currents, 144 including functional decreases in the transient outward (I to ) 143,145 and inward rectifier (I K1 ) 145 potassium currents. This causes prolonged repolarization and increases APD. [142][143][144]146 On the other hand, I Ca-L was not changed in amplitude, single channel kinetics, or expression. 143,144,147,148 However, chronic β1blockade in mice was associated with impaired intracellular Ca 2+ transients. 149 Chronic β-blocker use in LQTS, unlike AF, is likely to be proarrhythmic, as they contribute to further prolonging of the APD, and hence arrhythmic triggers. Nonetheless, the long-term consequences of β-blockade on ion channel remodeling in the abnormal LQTS heart, and how that will influence the arrhythmic phenotype, has not been investigated and offers an important area for future research. Na + channel block. Comparison between different β-blockers is further complicated by different abilities to block the Na + channel, which, as discussed previously, is an important aspect of their antiarrhythmic effects, especially in LQTS3. Of the β-blockers, propranolol exhibits the greatest Na V 1.5 blocking effect. 108,150,151 One study using whole-cell patch clamp recordings reported that propranolol, but not nadolol, had Na V 1.5-blocking effects. 150 Another study found that propranolol reduces both the peak and late Na + currents and causes a hyperpolarizing shift in both the activation and steady state curves of the Na V 1.5 channel. 151 On the other hand, nadolol reduces the peak but not the late Na + current, and causes a hyperpolarizing shift in the activation and steady state curves of the Na V 1.5 channel. 151 The difference in channel blocking effect may be due to the lipophilic structure of propranolol allowing it to move through the 40 Ann. N.Y. Acad. Sci. 1474 (2020) 27-46 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences membrane toward the blocking site. 151 Thus, despite propranolol having greater Na + channel inhibition, nadolol has been found to be more effective. This may be due to the previously discussed findings that Na + channel inhibition is not necessarily antiarrhythmic but, under certain concentrations, proarrhythmic by increasing TDR. Therefore, it could be that nadolol avoids these proarrhythmic changes and is, hence, more effective than propranolol at clinically relevant concentrations.

Conclusions
The antiarrhythmic efficacy of β-blockers is a genotype-specific phenomenon, and therefore the use of β-blockers in clinical practice should be genotype dependent. In LQTS1 and LQTS2, cardiac events most commonly occur during periods of increased sympathetic activity, which increases arrhythmic triggers, EADs, and substrate, TDR. Treatment with β-blockers is successful in reducing cardiac events via reducing arrhythmic triggers and substrates, thus supporting the use of β-blockers in the treatment of LQTS1 and LQTS2. LQTS3 represents a more complex situation both with respect to clinical reports and physiological mechanisms. In LQTS3, cardiac events most commonly occur during periods typically associated with decreased sympathetic activity suggesting a protective role, but some still occurred during periods of increased activity. Investigations into the effects of sympathetic stimulation on arrhythmic triggers and substrate and the use of β-blockers produced contradictory results. We offer insight to explain this by tying together disparate data. It becomes clear that understanding LQTS3 requires further research. Clinical studies with larger sample sizes and that control both for the type of β-blocker and concentration should clarify the optimal therapeutic protocol in LQTS3. If results are consistent with nadolol being the most effective at preventing arrhythmias in LQTS3, the mechanisms behind this could be investigated. We offer three testable hypotheses for nadolol being the drug of choice. First, nadolol's nonselectivity prevents the sensitization of other βadrenoceptors as occurs in selective β-blockers, and hence offers maximal antiadrenergic effect. Second, a study comparing the compliance of nadolol to propranolol in LQTS3 may show nadolol having higher compliance due to fewer side effects. Third, under clinically relevant concentrations, nadolol may shorten the APD of epicardial and endocardial LQTS3 cells, hence reducing TDRs in addition to suppressing EADs.