Ventricular voltage‐gated ion channels: Detection, characteristics, mechanisms, and drug safety evaluation

Abstract Cardiac voltage‐gated ion channels (VGICs) play critical roles in mediating cardiac electrophysiological signals, such as action potentials, to maintain normal heart excitability and contraction. Inherited or acquired alterations in the structure, expression, or function of VGICs, as well as VGIC‐related side effects of pharmaceutical drug delivery can result in abnormal cellular electrophysiological processes that induce life‐threatening cardiac arrhythmias or even sudden cardiac death. Hence, to reduce possible heart‐related risks, VGICs must be acknowledged as important targets in drug discovery and safety studies related to cardiac disease. In this review, we first summarize the development and application of electrophysiological techniques that are employed in cardiac VGIC studies alone or in combination with other techniques such as cryoelectron microscopy, optical imaging and optogenetics. Subsequently, we describe the characteristics, structure, mechanisms, and functions of various well‐studied VGICs in ventricular myocytes and analyze their roles in and contributions to both physiological cardiac excitability and inherited cardiac diseases. Finally, we address the implications of the structure and function of ventricular VGICs for drug safety evaluation. In summary, multidisciplinary studies on VGICs help researchers discover potential targets of VGICs and novel VGICs in heart, enrich their knowledge of the properties and functions, determine the operation mechanisms of pathological VGICs, and introduce groundbreaking trends in drug therapy strategies, and drug safety evaluation.


INTRODUCTION
The cardiac cycle begins when an action potential (APs) is spontaneously generated in the sinoatrial node (SAN), the primary pacemaker in the heart. The coordinated propagation of synchronized electrical impulses relies on effective cooperation among various components in the heart system to maintain cardiac rhythm. 1 Specifically, the AP from the SAN passes sequentially through the atria, the atrioventricular node (AVN), and His-Purkinje conducting tissue before ultimately reaching the ventricles. 1 APs, which are generated and modulated by the opening and closing of ion channels in the plasma membrane, are the fundamental electrical excitation signals responsible for the beating of cardiomyocytes and are distinct among various components in the heart due to various expression of ion channels. 1 Among the voltage-gated ion channels (VGICs) involved in ventricular APs, sodium (Na + ), potassium (K + ), and calcium (Ca 2+ ) channels (Nav, Kv, and Cav channels, respectively) predominate. 2 The functions of VGICs, the mechanisms underlying cardiac physiology and pathology, and appropriate diagnostic and treatment strategies have been explored for decades through electrophysiological techniques. Moreover, these techniques have been developed and expanded over time; from laborious, low-throughput methods limited to whole-cell experiments, they have been refined into automated, highthroughput methods. These developments have dramatically augmented the ability of researchers to further explore VGICs. In the real world, aside from genetic mutations affecting VGICs, many drugs can bind to these channels, block ion flow and disrupt the regulation of APs, potentially leading to drug-induced arrhythmia, or "proarrhythmia." 3,4 It is necessary to evaluate the risks of potential drug candidates by using the different techniques mentioned above, 5,6 according to the US Food and Drug Administration (FDA) guidelines. In this respect, enhancing the quality of preclinical safety screening is particularly important for validating the safety of therapies to avoid potential adverse effects on ion channels and prevent billions of dollars in losses because of late-stage premarket drug withdrawals in the development process of drug development before marketing.

Electrophysiological techniques
The manual patch clamp (MPC) technique (Table S1) is the gold standard for analyzing electrophysiological characteristics (APs and specific ion channel currents) in cardiac myocyte research studies under physiological/pathological conditions or in response to drug application. Three main cell models are used: 1) freshly isolated ventricular myocytes from wild-type (WT), diseased or genetically modified animal models; 2) heterologous expression systems specifically expressing the human ion channels of interest; and 3) human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) from healthy individuals and patients. 2,7 In addition, cardiac ion channels can be examined by single-channel MPC recording (Table S1); for example, this technique can be applied to a potential new channel with a putative pore-containing structure, 8 or channels that cannot be expressed or trafficked on the cell membrane in heterologous expression systems, 9 or channels that are potentially altered in the diseased heart. 9 The main limitation of MPC is its low throughput. Therefore, the automated patch clamp (APC) (Table S1) enables much higher-throughput experiments while nevertheless achieving high-quality seals, thereby facilitating the use of the MPC and is now routinely used in cardiac drug discovery and safety testing. [10][11][12] In addition, microelectrode arrays (MEAs) (Table S1) offer an alternative noninvasive that enables noninvasive, high-throughput assays evaluating extracellular field potential (EFP) of excitable cells 13 ; MEAs have also been increasingly used in cardiology to test the safety of drug candidates. 13,14 Furthermore, impedance techniques (Table S1) have recently been combined with EFP recording on the same platform to provide a noninvasive, high-throughput and long-term measurement strategy for assessing the synchronous beating of monolayer cardiomyocytes, the duration of EFPs, and the proarrhythmogenic capacity of drug candidates without altering cellular physiology 15 ; this combined approach offers a more comprehensive analysis of excitationcontraction (EC) coupling than either component alone.
Generally, low-throughput MPC is a critical tool for examining the electrophysiological characteristics of cardiac cells and the biophysical properties and functions of ion channels. APC and MEAs have developed into an indispensable platform for pharmaceutical companies and academic laboratories to conduct potential drug target discovery, drug screening, and cardiac safety with high efficiency and accuracy. hiPSC-CM-based APC, 16 MEA, 17 and impedance 15 screening assays are increasingly used to evaluate antiarrhythmic effects, adverse effects or interindividual variations in patients or healthy individuals and to acquire more comprehensive validation data.

Joint techniques
Cryoelectron microscopy (cryo-EM) (Table S2), which can resolve the structure of macromolecular complexes at the atomic conformation level, has provided researchers with a more in-depth molecular picture of ion selectivity, voltage gating, and intersubunit interactions in channel complexes and thereby provides insights into important biological phenomena, such as electrophysiological feature variations among different VGIC isoforms, 18,19 feature changes after the application of various compounds, 20 and the mechanisms of mutationrelated arrhythmia. 18 Hundreds of disease-associated missense mutations have been mapped onto all major domains in the structure of many VGICs. 18,21 Cryo-EM structure analysis could provide novel insights into both VGIC-drug interactions and the mechanisms of action of such drugs. 20,22 Moreover, electrophysiological techniques can help evaluate whether the functional properties of truncated or mutated VGICs purified for cryo-EM analysis are similar to those of WT full-length VGICs. 18,19 Optical imaging methods (Table S2) using voltageor Ca 2+ -sensitive dyes are less invasive than MPC and are able to measure changes in the MPs, intracellular calcium concentrations, electrical activity and EC coupling of cardiac cells. 23 However, some sensitive dyes are limited by cytotoxicity and short half-lives. Genetically encoded fluorescent Ca 2+ indicators, such as ArcLight and GCaMP, were developed and applied to cardiac research to monitor functional changes in hiPSC-CMs in a long-term, noninvasive, high-throughput manner. 24,25 The combination of optical imaging and electrophysiological techniques allows simultaneous recording of optical AP signals and calcium transient signals and permits both high spatial resolution and accurate functional evaluation.
Optogenetics approaches (Table S2), using light to control the perturbation of membrane voltage through the opening of optogenetic channels have been used to modulate cardiomyocyte excitability and heart rate with high precision and to explore the mechanisms of arrhythmia generation. 26,27,28 Optogenetic channels can also be used to study the relationship between cardiac myocytes and nonmyocyte cells and provide a feasible way to explore direct evidence of electrical coupling between these cells in normal or injured regions of the heart. 29 Automated frequency-dependent cardiotoxicity screening can be conducted by applying optogenetic stimulation similar to physiological heart rates in hiPSC-derived cardiomyocytes. 30

Normal electrophysiology of AP and EC coupling
A typical ventricular AP consists of five phases (P0-P4) that are mediated by different depolarizing and repolarizing ionic currents ( Figure 1A). 2 The initial phase (Phase 0) of a cardiac AP occurs after the resting state (Phase IV) of the previous AP and arises from a very large inward I Na current mediated by Nav channels. Then, Kv channels are activated to mediate transient outward potassium currents I to , leading to partial repolarization in Phase I. During Phase II, L-type Cav channels (LTCCs) are activated, generating an inward I CaL current. In addition, the Na + /Ca 2+ -exchanger (NCX) opens in forward mode and mediates an inward I NCX current by exchanging an influx of 3Na + for an efflux of 1Ca 2+ . On the other hand, the voltage-gated delayed rectifier potassium channels open and mediate outward rectifier currents (I Kr and I Ks ). Membrane potential (MP) changes extraordinarily little due to the nearly equal inward and outward currents during this phase, which is also known as the plateau phase. In the late plateau phase, LTCCs are inactivated, and the dominant outward currents, I Kr and I Ks , result in repolarization in Phase III. Toward the end of Phase III, I Kr and I Ks decline, and the inwardly rectifying potassium channels Kir2.x mediate the I K1 current to drive repolarization and maintain a resting MP (Phase IV).
The beating of the heart relies on EC coupling ( Figure  1B). During AP generation, the LTCC-mediated increase in the cytosolic Ca 2+ concentration instantaneously triggers the opening of the ryanodine receptor 2 (RyR2) channel, a Ca 2+ channel in the sarcoplasmic reticulum (SR), which causes Ca 2+ release from the SR, and thereby further increases the cytosolic Ca 2+ concentration. This Ca 2+induced Ca 2+ -release (CICR) prompts Ca 2+ -sensing protein troponin C to initiate contraction (systole). Cytosolic calcium levels are reduced via the Ca 2+ -ATPase type-2a (SERCA2)-mediated influx of Ca 2+ back into the SR and the NCX-mediated efflux of Ca 2+ back to the extracellular space, resulting in the dissociation of calcium and troponin and then muscle relaxation (diastole). 2,31,32

Abnormal electrophysiology as a trigger of arrhythmias
Disruption of the normal generation and duration of Aps is associated with arrhythmias in the heart. 1,31 Two types F I G U R E 1 Normal AP generation and EC coupling of cardiomyocytes; abnormal electrophysiology as a trigger of arrhythmias. (A) A typical ventricular action potential (AP) and the depolarizing and repolarizing ionic currents underlying its different phases (P0-P4). (B) Voltage-gated ion channel (VGIC) distribution and contribution to AP excitation-contraction coupling in cardiomyocytes. Na + , Ca 2+ , and K + of afterdepolarizations, early afterdepolarizations (EADs) and delayed afterdepolarizations (DADs), could induce premature APs and contribute to arrhythmias. EADs occurs during Phase II or III ( Figure 1C). Prolongation of action potential duration (APD) due to the reduction of repolarization currents (eg, I Kr and I Ks ) or the increase in I Na,L current could give rise to abnormal recovery from the inactivation of LTCC channels and further depolarize the membrane due to the reactivation of inward currents I CaL . 1,2 DADs can result from depolarization after the end of AP repolarization ( Figure 1C), potentially due to Ca 2+ overload caused by enhanced SR Ca 2+ release and the inappropriate activation of the reverse mode of NCX, which mediates outward I NCX current by exchanging an influx of 1Ca 2+ for an efflux of 3Na + . 1,2 Under physiological conditions, Nav channel activity is regulated by cytosolic Ca 2+ levels, such as elevation of cytosolic Ca 2+ levels resulting in destabilization of inactivation and increase of the amount of available channels to open for the next AP. [33][34][35] And Na + influx, in turn, affects the modulation of cytosolic Ca 2+ levels. 32 However, under pathological conditions, an abnormal increase in Na + during diastole can result in inappropriate timing of reverse flow through the NCX channel (3Na + efflux and 1Ca 2+ influx), further increasing the cytosolic Ca 2+ concentration and altering normal EC coupling ( Figure 1D). 32

4
VENTRICULAR AP-RELATED ION  CHANNELS: CLASSIFICATION,  STRUCTURE, FUNCTION, REGULATION,  AND DISEASE RELEVANCE   4.1 Nav channels Cardiac voltage-gated Nav channels initiate AP in electrically excitable cells. The specificities among isoforms (Table 1) are attributed to the distinct α-subunit encoded by the corresponding gene and the different combinations of β subunits. 36 β subunits regulate channel surface expression, voltage dependence and gating kinetics. 36 SCN5A-encoded Nav1.5 is the most abundantly expressed Nav channel in ventricle and atrium (Table 1) and is responsible for the generation of APs and the conduction of cardiac impulses in cardiomyocytes. 18,34,37,38 Additional evidence has shown that other isoforms are also expressed in the ventricular myocytes (Table 1). 32,[39][40][41][42][43][44][45][46][47][48] 4.1.1 Nav1.5 In ventricle cardiomyocytes, Nav1.5 channels are known to be located in lateral membrane, transverse tubules (Ttubules), and intercalated discs, ensuring propagation of are represented by red, yellow, and blue dots, respectively. The cardiac VGICs Nav1.5, Kv7.1, and Kv11.1 are primarily localized in intercalated discs (IDs), T-tubules and lateral membranes. TTX-sensitive Nav channels are primarily localized in T-tubules and colocalized with the Na/Ca exchanger (NCX). The sarcoplasmic reticulum (SR) channel ryanodine receptor 2 (RyR2) is located near most L-type Ca 2+ channels, and Cav1.2 is located in T-tubules. RyR2 is regulated by type 2 calsequestrin (CASQ2). A novel TRIC-A channel can also directly interact with RyR2 and act as a counterion channel to modulate Ca 2+ release from the SR. The potassium channel Kir2.1 interacts with Nav1.5 in T-tubules. Kir2.1-mediated outward K + drives repolarization, while the rapid increase in membrane potential (MP) depolarization and the MP overshoots during Phase I are driven by the influx of Na + , which is mediated by Nav1. 5 electrical impulse in longitudinal, transverse directions of cardiomyocytes, and between adjacent ones, respectively ( Figure 1B). 49 Nav1.5 channels are closed at the resting MP (Phase IV). In response to membrane depolarization, Nav1.5 could be activated. Within 200-300 μs, a large inward peak I Na (I Na,P ) is formed to trigger overshooting of AP in Phase 0. At the end of this phase, most Nav1.5 channels are rapidly inactivated within 2-5 ms, rendering the channel refractory until repolarization is completed in Phase III. During Phase IV, after recovering from inactivation, the channels are closed and can again be reopened by membrane depolarization (Figure 2A). In Phase II or III, a small population of total Nav1.5 channels could be reactivated before complete inactivation and then generate a relatively small, persistent sodium cardiac inward current called late I Na (I Na,L ). 50,51 Under physiological conditions, I Na,P but not I Na,L plays a central role in ventricular AP generation, while under pathological conditions, I Na,L can play an important role. 50,51 Abnormal increases in I Na,L prolong the duration of the AP plateau, triggering EADs or further elevating intracellular Ca 2+ levels by driving the NCX exchanger to function in reverse mode, thereby inducing DADs and contributing to arrhythmogenesis. 51 To date, some high-resolution structures of Nav channels, 21,52,53 including rNav1.5C, 18 have been elucidated (Table 1). In general, key structural features of Nav1.5, the structural basis for its physiological function and its dysfunction in cardiac arrhythmias could be learned from the revealed Nav1.5 structures. Unlike other Nav α subunits, in Nav1.5, the regulatory interface with auxiliary β1 and β2 subunits, encoded by SCN1B and SCN2B respectively, is not as strong due to the substitution F I G U R E 2 Cardiac voltage-gated sodium channel (Nav1.5) structure, accessory proteins and signaling. (A) The contribution of I Na,P (upper) mediated by Nav1.5 to action potential (AP) Phase 0 (lower). (B) The cardiac voltage-gated sodium channel Nav1.5 comprises one α subunit and one or two auxiliary β subunits. The α subunit consists of four homologous but nonidentical repeats (DI-DIV) connected by cytoplasmic linkers (ICL I-II , ICL II-III , and ICL III-IV ) and is responsible for voltage dependence, pore formation, and surface expression. Each domain contains S1-S6 subunits that are connected by loops located intra-and extracellularly. The carboxyl-terminal domain (CTD) of S1 and the amino-terminal domain (NTD) of S6 are located in DI and DIV, respectively. The S4 subunit of each domain contains the voltage sensor. S5 and S6 of each repeat form the pore domain, and the connecting P-loops between the S5 and S6 regions curve back into the pore to form the extracellular selectivity filter (SF), which is responsible for Na + selectivity. ICL III-IV serves as an inactivation gate (IG), which closes the pore within 1-2 ms after opening. The β subunit, consisting of an extracellular domain, an intracellular domain, and a single transmembrane helix, plays an important role in modulating the surface expression, kinetics, and functions of Nav channels. The β1 and β2 subunits do not stably associate with the Nav1.5 α subunit. Y304 in the Nav1.7 α subunit, which is connected to E48 in the β1subunit by a hydrogen bond, is substitute at L316 at the equivalent position in Nav1.5. Similarly, C895, which forms a disulfide bond with the β2 subunit, is substituted at L869. Several accessory proteins interact with Nav1.5 channels to form macromolecular complexes that regulate ion trafficking, posttranslational modifications and gating. Nav1.5 activity is driven by Ca 2+ levels via its interactions with CaM and CaMKII. Nav1.5 can be activated in response to β adrenergic stimulation via the activation of CaMKII. CaM binds not only to an IQ motif in the CTD of Nav1.5 but also to its IG. The CaMKII inhibitor KN93 was recently reported to alter the kinetics of Nav1.5 inactivation by interrupting the CaM-IG interaction but did not suppress CaMKII function. 64 of residues for β subunit interactions. 18 Recent studies suggest β1 subunit may differentially control expression and function of α-subunit via acute and chronic feedback mechanisms. 54 β2 is pivotal for the correct localization of NaV1.5. 55 Nav1.5 is insensitive to the inhibition of tetrodotoxin (TTX), a selective sodium channel blocker nonprotein toxin, due to the substitution of binding residues at the outer mouth of the Nav1.5 selectivity filter (SF). 18 Nav1.5 is blocked by the class Ic antiarrhythmic drug flecainide with comparatively high affinity and slow binding kinetics compared to class IA and IB antiarrhythmic drugs due to the larger hydrophobic ring structure of flecainide interacting with the central cavity of Nav1.5. 18 In addition, the structural template of Nav1.5 for arrhythmia mutations provides a better understanding of the mechanism of various SCN5A variants in different positions. 18 The traditional view is that a single α-subunit of Nav1.5 functions as a monomer, while increasing evidence has shown that the α-subunit of Nav1.5 could be oligomerized into dimers within the first intracellular loop and results in coupled gating properties with the accessory protein14-3-3 interaction. 56 Inhibition of 14-3-3 could abolish the dominant negative (DN) effect and the biophysical coupling between α-subunits. 56 In addition, several accessory proteins have been demonstrated to interact directly with the α subunit of Nav1.5 channels ( Figure 2B) to form macromolecular complexes with Nav1.5 and modulate the expression, trafficking and biophysical function of Nav1.5 (Table 2). 49 Calmodulin (CaM), a ubiquitously expressed calcium-binding TA B L E 2 Accessory proteins reported to interact with and regulate Nav1.5 Accessory proteins 239 protein, and CaM-dependent protein kinase II (CaMKII), an adrenergically activated kinase, serve as important components affecting channel function. 34 CaM binds with IQ motif of Nav1.5 carboxyl-terminal domain (CTD) in Ca 2+ -free forms and Ca 2+ -bound forms at the basal levels of intracellular Ca 2+ concentration. [33][34][35] While this CaM-Nav1.5 interaction is altered when the elevation of intracellular Ca 2+ concentration, therefore changing the rate of Nav1.5 inactivation. [33][34][35] Increasing evidence has shown that at the high level of intracellular Ca 2+ concentration, CaM also directly binds to the inactivation gate (IG) of Nav1.5 to destabilize the IG and promote faster recovery from inactivation. [57][58][59][60][61] CaMKII not only phosphorylates Nav1.5 at key site S571 62 but also directly interacts with Nav1.5 to regulate the expression and function of Nav1.5. 63 On the other hand, protein phosphatase 2A (PP2A) was recently found to interact with the Nav1.5/ankyrin-G/CaMKII/Βiv-spectrin macromolecular complex and balance CaMKII-dependent phosphorylation. 62 The CaMKII inhibitor KN93 but not autocamtide-2-related inhibitory peptide (AIP) could interrupt the CaM-IG interaction by forming the ternary complex CaM-IG-KN93 and then inhibit Nav1.5 recovery from inactivation without altering the kinetics of inactivation. 64 Therefore, determining the effects of accessory proteins and signaling pathways on modulating Nav1.5 provides us with a more comprehensive understanding of Nav1.5 roles in cardiac tissues in both health and disease states and is beneficial for the discovery of potential drug targets. Future investigations of the kinetics of CaM-Nav complexes and the effects of structure-guided mutations on the roles of Nav1.5 in the absence/presence of Ca 2+ transients will provide us with a more comprehensive understanding of the mechanisms and significance of Ca 2+ -dependent Nav roles in cardiac tissues in both healthy and disease states. Mutations in SCN5A ( Table 3) are associated with inherited life-threatening arrhythmias, such as long QT syndrome type 3 (LQTS3), and Brugada's syndrome (BrS). 18,65 Slower channel inactivation and thus conducting an increase in I Na,L are responsible for the gain-of-function (GOF)-associated LQTS3. 66,67 On the other side, reduction of membrane expression of functional channel due to synthesis deficiency 68 or trafficking defects, 69 impairment of gating (such as slower activation or faster inactivation) 56,59,[70][71][72] or permeation disruption 73,74 cause SCN5A loss-of-function (LOF)-associated BrS. 65 The αsubunit of Nav1.5 oligomerization also explains the existence of several BrS variants displaying DN effects, providing new therapeutic targets for BrS caused by SCN5A LOF variants. 56 Moreover, LOF mutations in SCN1B and SCN2B are also implicated in BrS. 54,55,75,76 In addition to most isolated GOF or LOF variants of SCN5A which are typically associated with a distinct clinical and electrocar-diographic phenotype, variants could also lead to overlapping syndromes 77,78 or inherited arrhythmia syndrome different from BrS and LQT3. 79 There is also a category of benign (atypical) SCN5A mutations which shows normal function alone but leads to a reduction in sodium currents when coexpressed with WT in vitro as typical SCN5A BrS mutations do. 80 Moreover, missense variants in CAV3-encoded caveolin-3, which forms macromolecules with Nav1.5 and serves as a negative regulator for I Na,L , could result in I Na,L increase and thus cause LQTS9, providing new therapeutic strategies to correct I Na,L . 81 Drugs that inhibit I Na,L 50 could shorten the AP duration or QT interval and could therefore be considered a potential treatments for I Na,L -associated diseases. 82 Thus, abnormal changes in I Na,L could be considered as a target for drug development and safety evaluation.
High-throughput assays of cardiac Nav1.5 I Na,P have been widely used in cardiac safety screening, but screening studies do not routinely measure I Na,L . 83 However, it is important for potential therapeutic candidates that could minimize I Na,L without affecting I Na,P to be selected. 51 The variety of different protocols and measurement strategies applied in the use of these drugs have contributed to remarkable variations in the reported data on I Na,L and screening results for inhibitory compounds. 84,85 I Na,L is small, and studies have had difficulty generating reproducible data; thus, the best choice for an I Na,L enhancer should increase I Na,L with no obvious effect on I Na,P . 85 In addition, it is necessary to double check the median inhibitory concentration (IC 50 ) of potential drugs in the absence of enhancers, eliminating the modification effect of enhancers on the activity of compounds, 85 and to evaluate the IC 50 of drugs in different stimulation states with regard to variations in the effects on I Na,P and I Na,L in different stimulation states. 18

4.1.2
Other Nav channels in the heart TTX-sensitive Nav channels including neuronal Nav (eg, SCN1A-encoding Nav1.1, SCN2A-encoding Nav1.2, SCN3A-encoding Nav1.3, SCN8A-encoding Nav1.6), which were first identified in neurons, and skeletal muscle Nav (eg, SCN4A-encoding Nav1.4), which was first identified in skeletal muscle, have been unexpectedly found in Ttubules of ventricle myocytes (Table 1), contributing a small portion to the total sodium current under physiological conditions due to their much lower expression level than Nav1.5. [39][40][41][45][46][47] While, in inherited forms of cardiac arrhythmia, augmentation of TTX-sensitive Nav channels phosphorylated by β-AR stimulation/CaMKII stimulation, contributes to abnormal increases in I Na,L and arrhyth- mogenic Ca 2+ release. 32,41 Compared with other TTXsensitive Nav channels, the location of Nav1.6 ( Table 1) is the closest channel to RyR2 (< 100 nm) 41,86 indicating that Nav1.6 is capable of impacting Ca 2+ cycling proteins and Ca 2+ dynamics in both health and disease. 86 GOF variants of SCN8A-encoding Nav1.6 ( Table 3) potentially leads to sudden unexpected death in epilepsy (SUDEP) due to arrhythmia of the brain and the heart. 87 The GOF variant of SCN8A (N1768D) causes hyperexcitability of ventricle myocytes by increasing calcium transient duration, prolonging the early phase of APD, and increasing the incidence of DADs but not by compensatory changes in Nav1.5 expression. 87 Selective pharmacological blockade of Nav1.6 and silencing of Nav1.6 indicate that Nav1.6 can potentially contribute to β-AR stimulation-induced I Na,L and arrhythmias. 41 This explains why catecholaminergic polymorphic ventricular tachycardia (CPVT) models respond to treatment with some Na + channel blockers. 41 Besides SCN8A, the possible roles of SCN1A, or SCN4A mutations in pathophysiology of cardiac congenital syndrome were also investigated ( Table 3).

TA B L E 3 Mutations in cardiac voltage-gated Na
In addition, TTX-insensitive SCN10A-encoding Nav1.8 (Table 1) channels, which are mainly expressed in the peripheral nervous system, are also found in the heart [42][43][44]48 at a higher level in the atrial myocardium than in the ventricular myocardium, 43 exhibiting a more depolarized voltage dependence of inactivation and slower inactivation kinetics than other faster sodium channels like Nav1.5. 44 Nav1.8 contributes to abnormal increases in I Na,L and consequently prolongs the APD and elevates proarrhythmogenic diastolic SR Ca 2+ in cardiac disease. 42 Genetic deletion of Nav1.8 produces a smaller I Na,L increase than in WT cardiomyocytes during β-AR stimulation. 42 LOF and GOF variants in SCN10A (Table 3) are associated with BrS 88 and SUDEP, 89 respectively. Gating dysfunction with enhanced of inactivation results in LOF of Nav1.8. 88 In contrast, dysfunction with slower inactivation could result in GOF of Nav1.8 and then allow more Na current entry. 89 Thus, Nav1.8 also plays a significant role in the initiation of proarrhythmic triggers via I Na,Linduced SR Ca 2+ leakage.

Ca channels
In response to membrane depolarization, voltage-gated calcium (Cav) channels activate and mediate extracellular Ca 2+ influx into the cytosol, which serves as the second messenger of electrical signaling, initiating many physiological processes, such as excitability, contraction and cell death. 90 The Cav1 and Cav3 groups mediate Ltype and T-type currents, respectively, and are involved in cardiac function. Cav1 is more highly expressed than Cav3 in ventricular myocytes, while Cav3 is mainly expressed in SAN cells (Table 4). 90 Ryanodine receptors (RyRs), intracellular Ca 2+ channels in the sarcoplasmic/endoplasmic reticulum (SR/ER), control the rapid release of Ca 2+ from SR/ER into the cytoplasm to initiate CICR, a key event that triggers skeletal and cardiac muscle contraction. 91,92 Among three mammalian isoforms (RyR1, RyR2, and RyR3), RyR2 is primarily expressed in cardiac muscles, [91][92][93] and higher expressed in the ventricle (Table 4). 37,38 Cav1.2 channels, located in T-tubules of ventricular myocytes ( Figure 1A), are assumed to be the major subtype of Cav1 channels that mediate the entry of Ca 2+ , which is required for the AP plateau ( Figure 3A), and EC coupling, triggering activation of RyR2 and initiating CICR ( Figure  1A). 90 The Cav1.2 complex consists of one pore-forming subunit α1c and the auxilary subunits α2δ and intracellular β ( Figure 3B). 94,95,[96][97][98] The full-length cryo-EM structure of cardiac Cav1.2 has not been revealed, while skeletal Cav1.1 was the first Cav channel to have its full-length cryo-EM structure reported with an overall resolution of 4.2 and 3.6 Å. 99,100 The structure of Cav1.1 provides a structural template for the Cav1 family and comparisons for molecular interpretations of the functions and disease mechanisms between eukaryotic Cav and Nav channels. 99,100 Because the α1 subunits from Cav1.1 and Cav1.2 are highly homologous, Cav1.2 modeling could be based on the Cav1.1 structure for analyzing the molecular determinants of opening and closure of Cav1.2. 94 Voltage-independent upward movement or voltage-dependent movement of S4 segments maintain the opening or closure of the gating, respectively. 94 The voltage sensor S4-S5 are coupled with pore S6 segments by directly interacting with a ring of small residues, which are regarded as interesting sites for studying electromechanical coupling. 94 In addition, in complex with channel-selective agonists/antagonists, structural analysis helps to elucidate their specific binding sites and reveal the structural reasons why similar types of molecules (such as nifedipine and Bay K8644) exert opposite antagonist and agonist effects on Cav1.1 channels. 20 The auxiliary subunits α2δ and β generally modulate the surface expression and biophysical kinetics of α1c. 97  In addition, several accessory proteins interact with the α1c subunit of Cav1.2 and regulate Cav1.2 expression and function (Table 6). Cav1.2 is involved in the β-AR/cAMP/PKA signaling pathway 95 ( Figure 3B). β1-AR/cAMP signaling is diffusive and global, while β2-AR/cAMP is relatively localized. 95 The CTD of β2-AR not only binds to A-kinase anchoring protein (AKAP) but also directly binds to Cav1.2 to mediate local signaling via the cAMP-dependent PKA pathway and facilitate localized cAMP signaling. 95 The PKA-dependent phosphorylation of amino acids in the CTD of the α1c subunit has been demonstrated to be decisive for the β-AR-mediated upregulation of cardiac I CaL . 95,101 In addition, some amino acids in the CTD of the α1c subunit are targets of the β1-AR/CaMKII signaling pathway. 95 In parallel with the roles of the β-AR system, BDNF-TrkB binding regulates myocardial Ca 2+ cycling and EC coupling by triggering CaMKIIdependent signaling. 102 GOF variants in CACAN1C (Table 5) cause timothy syndrome (TS), which is a multisystemic disorder including LQTS8, autism, and dysmorphic features. 103 Complete loss of inactivation kinetics leading to a prolonged calcium influx during action potentials, 103 or left shift in the activation curve leading to increase in window currents 104 could result in GOF of the Cav1.2 channel. Variant E1496K slowed inactivation, causing isolated LQTS8 without TS. 105 On the other side, LOF variants ( Table 5) which disruption of protein trafficking, 106,107 gating, 108 or Ca 2+ permeation 108,109 account for genotyped BrS cases. These results implicated the importance of the Cav1.2mediated calcium signaling in human physiology and heart disease.

Ryr2
The near-atomic-resolution cryo-EM structure of RyR2 from porcine hearts has been recently revealed in both the open and closed states, 110 or with key modulators, 111,112 offering the opportunity to characterize the roles of the structural elements and modulators during gating shifts. Each subunit of the homotetrameric RyR2 consists of a large cytosolic domain, which is responsible for interaction with protein modulators, linking gap between the SR and transverse tubule (T-tubule) membranes, and carboxylterminal transmembrane domain (CTTD), four identical of which are responsible for forming the central ionconducting pore ( Figure 3C). 110 The cryo-EM structures of the RyR2 complex and the abovementioned Cav1.1 establish a solid foundation for future revealing the Cav1.2 complex, the complex formation between Cav1.2 and RyR2, and excitation-contraction coupling. Several proteins (Table 6) interact with the cytosolic region of Ryr2 to regulate its open probability. For example, CaM 91 inactivates Ryr2 during diastolic cytosolic calcium elevation, thus playing an important role in Ca 2+ alternans. 113 The CaM binding sites on cytosolic sites of Ryr2 will be shifted and dependent on Ca 2+ concentration binding to CaM. 112 Enhancement of CaM function promotes, whereas impairment of CaM function suppresses Ca 2+ alternans. 113 Several enzymes, such as PKA, CaMKII, PP1, and PP2A, interact with Ryr2 and exert phosphorylation/dephosphorylation effects on Ryr2. 114 The hyperphosphorylation of RyR2 by PKA 114 and/or by CaMKII 115 causes abnormal Ca 2+ leakage from the SR. RyR2 is also coupled to proteins at the luminal SR surface, such as type 2 calsequestrin (CASQ2), 116,117 which increases the open probability and facilitates high rates of Ca 2+ efflux during systole. 116 Moreover, RyR2 also interact with other channels. RyR2 is located beneath most Cav1.2 (within ∼12 nm) in Ttubules and is connected with the α1 subunit of Cav1.2 by sorcin, which is a Ca 2+ -binding protein reducing the open probability of Ryr2, bridging the gap between SR and the sarcolemma for interchannel cross-talk. 118 In addition, trimeric intracellular cation (TRIC) channels represent a recently discovered class of cation channels that were first identified in rabbit skeletal muscle in 2007. 119 TRIC-A is a subtype that is abundantly expressed in excitable cells, having slightly higher permeability for K + than Na + and mediating counterion movements by releasing Ca 2+ from the SR. 120 The cryo-EM structure of the symmetrical trimer TRIC-A has been reported. 8 Moreover, TRIC-A also directly interacts with the cytosolic region of RyR2 via its carboxyl-terminal tail domains ( Figure 3C) to modulate intracellular Ca 2+ homeostasis and thereby facilitates Ca 2+ release from the SR. 121 The open probability and current amplitude of TRIC-A are increased by a positive shift in the MP 8 but are blocked by exposure to a highconcentration Ca 2+ bath on the luminal side during the resting state. 8,121 TRIC-A gene deletion decreases the sensitivity of individual RyR channels to β-AR/PKA stimulation, eventually resulting in Ca 2+ release impairment 122 and irregular ECG. 121 These studies indicate that TRIC-A promotes the release of Ca 2+ from the SR via RyR2 and maintains RyR2 function at low Ca 2+ to neutralize the (PP1 and PP2A) exert phosphorylation/dephosphorylation effects on Ryr2. Type 2 calsequestrin (CASQ2) interacts with the luminal surface of Ryr2 to increase the open probability. In addition, class A of trimeric intracellular cation (TRIC-A) channels on the SR membrane directly interact with the cytosolic region of RyR2 ( Figure 1B) (Table 5) are implicated in ventricular tachyarrhythmias, including type 1 of CPVT type (CPVT1), which is characterized by stress-induced ventricular tachycardia in the absence of a structurally abnormal heart. 123 GOF variants could induce channel instability by disrupting the interdomain interactions after channel activation, 124 or increase the open probability of RyR2 and pathological SR Ca 2+ release, 115,125 and thus causing susceptibility to CPVT. On the other hand, RyR2 LOF variants have been identified among survivors of cardiac arrest without exhibiting the CPVT phenotype and further regarded as RyR2 Ca 2+ release deficiency syndrome (CRDS) via an EAD-mediated mechanism. 126 I to , I CaL , and I NCX were alternatively increased, although catecholamine-induced SR Ca 2+ release was suppressed in LOF variant D4646A, thus causing AP waveform alteration and finally enhancing the propensity for arrhythmogenic EADs. 126 In CPVT cardiomyocytes with the RyR2 variant R176Q, a viral vector containing a CaMKII inhibitor (autocamtide-2-related inhibitory peptide, AAV9-GFP-AIP) completely suppressed the abnormal increase in spontaneous Ca 2+ transients, suggesting that CaMKII suppression represents a potential therapy for CPVT. 127 A KN93-mediated increase in RyR2 Ca 2+ release in cardiomyocytes was found to be due to disruption of the CaM-RyR2 interaction rather than inhibition of CaMKII. 64 Gene transfer of CaM, exhibiting a slower Ca 2+ dissociation rate and longer RyR2 refractoriness, alleviated arrhythmias in a CASQ2-associated CPVT mouse model. 128 Previous studies have illustrated that flecainide prevents ventricular tachyarrhythmia in patients with CPVT by blocking of the TTX-sensitive Nav channel. 41 Recent research has shown that the antiarrhythmic effect of flecainide mainly relies on blocking RyR2 channels but not TTX-sensitive Nav channels. 129 The secondary amine on the piperidine ring in flecainide is necessary for its activity in RyR2 channels. 129 In general, the regulation of RyR2 modulators (RyR2-CaM interaction) represents an important therapeutic target for regulating cardiac alternans in cardiac ventricular arrhythmia.

Kv channels
Cardiac Kv channels play prominent roles in resting potential maintenance, AP repolarization, and the AP plateau phase. 130 The rapidly activated and inactivated transient outward potassium current (I to ) contributes to early ventricular AP repolarization and underlies the initial "notch" before the AP plateau phase in humans and other larger mammals ( Figure 4A). 130 1 Unlike in human and mammalian models, ventricular AP in rodent models exhibits fast repolarization without a plateau phase due to I to rather than I Kr playing the major role in repolarizing currents. 1,130,133 The significant prolongation of repolarization duration, which is affected more by a reduction in I to than a reduction in I Kr , underlies the mechanism for heart failure with preserved ejection fraction (HFpEF, typical heart failure symptoms with a normal ejection fraction)-related ventricular arrhythmias and sudden cardiac death in rodent models. 134 A rabbit ventricular APD could be shortened, and its plateau could be lost when mouse I to currents were integrated. 1 Thus, differences in I to densities in different species contribute to variations in the waveforms of action potentials. 1,130 Kv4.3 is composed of one pore-forming α subunit and K + channel interacting protein 2 (KChIP2) β subunit 130,133 and is regulated by several accessory protein interactions ( Figure 4B) (Table 8). A reduction in Kv4.3 expression and I to, fast in heart disease, is associated with β-AR/CaMKII-mediated activation 133 and β-AR/NF-kBmediated activation. 135 Moreover, an increasing number of studies have speculated that Kv4.3 and Nav1.5 not only regulate each other's functions, but also have the ability to interact with each other. 136,137 Nav1.5 and Kv4.3 reside is visualized in close proximity (<40 nm) at the membrane. 136 Overexpression of Kv4.3 protein significantly decreased AP upstroke and Nav1.5 current density without affecting Nav1.5 total protein expression and its kinetic properties. 137 In addition to auxiliary subunit of KchIP2, Navβ1 subunit also associated with Kv4.3 138,139 and regulated the I Na /I to balance by yielding an increase in I Na and a decrease in I to 136 . Mutations in KCND3-encoded Kv4.3 or SCN5A-encoded Nav1.5 further showed the functional relationship between Kv4.3 and Nav1.5. 136 GOF and LOF mutations in KCND3 (Table 9) respectively decreased and increased the Nav1.5 current, respectively. 136 On the other hand, SCN5A LOF mutations increased I to by facilitating Kv4.3 cell surface expression or by slowing its steady-state inactivation. 136 Thus, during the early phase of ventricular AP repolarization, a fine balance may exist between I Na and I to . GOF mutations in the KCND3 contributes to increase of peak I to via efficient protein trafficking and gating, resulting in the imbalance of those two currents, the accentuation of the AP notch, and the development of BrS and/or early repolarization syndrome. 136,[140][141][142]

TA B L E 7
The slow delayed rectifier current, I Ks , is mediated by KCNQ1-encoding Kv7.1 and plays an important role in regulating the repolarization phase that terminates cardiac APs and thereby ends contraction ( Figure 4C). In the heart, the KCNE1, encoding the auxiliary β-subunit KchIP2, interacts with Kv7.1 α chains and affects both voltage-sensing S4 movement and the gate, 143 making the activation of the complex much slower than that of Kv7.1 alone 144 ( Figure 4D). Cryo-EM analysis revealed a unique feature of Kv7.1: pore opening requires lipid phosphatidylinositol 4,5-bisphosphate (PIP2) binding during membrane depolarization, thereby increasing current and slowing inactivation. 144 In addition, Kv7.1 is regulated by accessory protein interactions (Table 8), β-AR/PKA-mediated phosphorylation 145 and PP1-mediated dephosphorylation 146 ( Figure 4D).
Mutations in KCNQ1 (Table 9) are the leading cause of several congenital cardiac diseases, including LQTS and SQTS. 147 LQT1, the most common genotype-positive LQTS, is associated with LOF mutations in the KCNQ1encoded Kv7.1 α subunit and is often triggered by β-AR stimulation. 148 Trafficking defects, 149 gating defects, [150][151][152] or permeation disruption 153 have been postulated to be the mechanism of decreasing I Ks or hindering I Ks currents at physiologically relevant membrane potentials but limiting the upregulation of I Ks by PKA activation. 145 Because CaM regulates channel gating by interacting with voltage sensor domains, mutations impair CaM binding (located near the IQ motif of KCNQ1 C-terminus) and alter both channel assembly and gating, thus decreasing I Ks current density and contributing to LQT1. 152 Thus, dysfunction of Kv7.1 caused by KCNQ or related accessory protein mutations decreases I Ks or limits the upregulation of I Ks by PKA activation and then contributes to LQT1. I Ks are more sensitive to β-AR stimulation than I Kr . 132 Enhancement of I Ks by increasing Kv7.1 phosphorylation to shorten the APD during rapid heart rates might represent an effective antiarrhythmic strategy. LOF mutations in KCNE1 are associated with LQTS5. 154,155 On the other hand, SQTS2 is associated with GOF mutations in KCNQ, which could enhance I Ks via acceleration of the activation kinetics or prolongation of deactivation time constant 156  Permeation disruption V302M 184,185 : disrupts the channel to conduct potassium without altering subunit assembly or suppressing cell surface expression.
LQTS, long QT syndrome; SQTS, short QT syndrome; BrS, Brugada syndrome; ERS, early repolarization syndrome; SUDS, sudden unexplained death syndrome; ATS, Andersen-Tawil syndrome; GOF, gain-of-function; LOF, loss-of-function. depolarization in Phase 0-II. Then, it quickly recovers from inactivation and reopens during the initial Phase III repolarization, followed by slow deactivation that permits sustained Phase III and early Phase IV of the AP ( Figure  4E). 158 Kv11.1 channels exhibit longer-lasting and higheramplitude tail currents than have been found for other outward current channels that contribute to cardiac AP repolarization and duration. 159 Kv11.1 is composed of one pore-forming α subunit and two β subunits (MinK and MiRP1 encoded by KCNE1 and KCNE2) ( Figure 4F). The structure of the hERG channel with depolarized voltage sensors and open pores was revealed using cryo-EM. 19 A small central cavity includes extended pockets, which is specific to Kv11.1, explaining the notable susceptibility of this channel to a wide range of drugs. 19 This high-resolution structure of the hERG channel in the open state also provided the opportunity to investigate the potential mechanisms for the state-dependent blockade of hERG by drugs. 22 Kv11.1 is regulated by accessory protein interactions and signaling pathways ( Figure 4F) ( Table 8). Phosphorylation of Kv11.1 could be induced by the stimulation of β-AR/cAMP/PKA or G protein-coupled receptors (such as angiotensin II receptor AT1 and the α-adrenoceptors)/PKC signaling pathway, resulting in a decrease in I Kr. 158 LOF mutations in Kv11.1 (Table 9) are characterized by reduced I Kr and are associated with LQTS2, perhaps due to the disruption of the α subunits responsible for channel synthesis/translation, a reduction in intracellular transport or the accessory protein interactions required for channel trafficking on the membrane, or the impairment of channel gating structure as well as permeation. [160][161][162] Among those mechanisms, trafficking defects is the dominant one, responsible for approximately 80-90% of LQT2 by decreasing the folding efficiency of Kv11.1 proteins and increasing their retention in the endoplasmic reticulum (ER). [160][161][162] Comprehensive analysis of hundreds LQT2-linked mutations in four Kv11.1 structural domains and found that deficient protein trafficking is the dominant mechanism for all domains except for the distal Cterminus. Comprehensive and accurate analysis of mutations between normal and abnormal trafficking across multiple structural domains would aid in understanding the deleterious nature of these mutations. 162,163 Increasing high-throughput assays are developing and as alternative to traditional western blot assay to collect functional data. 163 In addition, LOF mutation in KCNE2 is associated with LQTS6, a rare type of LQTS. 155,164 The allosteric modulation (Table 10) of Kv11.1 was investigated to explore methods of alleviating channel dysfunction and increasing I kKr current and may represent a useful new approach for treating inherited and drug-induced LQTS2. 165 SQTS1 is caused by GOF mutations (Table 9) in the Kv11.1 channel and is the most prevalent SQTS subtype. Mutations that impair the inactivation of Kv11.1 166 might explain the lack of efficacy of many class III antiarrhythmic drugs (such as sotalol and ibutilide 167 ) in some patients. Interestingly, hydroquinidine, aclass I antiarrhythmic drug inhibiting the Nav1.5 channel, could also block Kv11.1, significantly intervene with ventricular tachyarrhythmia induction 167 and prolong the QT interval in SQTS patients with Kv11.1 mutations. 168 In addition, ivabradine, as a class 0 antiarrhythmic drug inhibiting hyperpolarizationactivated cyclic nucleotide-gated (HCN) channels, could also block I Kr currents by binding in the vicinity of the lipid-facing surface M651 residue, which is directly coupled to the conformational dynamics of residues in the pore helices, 22,169 and exert antiarrhythmic effects in SQTS1 hiPSC-CMs with the N588K mutation. 170 This represents one important method by which the efficacy of drugs used for SQTS treatment can be evaluated in hiPSC-CMs with mutant Kv11.1 or multiple ion channels to predict effects in SQTS patients. 171 It would be worthwhile to further examine the effects of traditional inhibitors and to develop novel specific inhibitors to expand the clinical options available for these patients.

4.3.4
Kir2.  37,172 In contrast to adult ventricular CMs, a substantial lack of I K1 in hiPSC ventricular CM is regarded as one mechanistic contributor to the immature electrophysiological properties of spontaneous AP. Artificial expression of Kir2.1 might overcome this limitation, render the electrophysiological phenotype to be mature, and ablate proarrhythmic AP traits. 173 The structure of K ir channels is relatively simple in comparison with Nav, Cav and the members of the voltagegated Kv channels. Each subunit of the Kir2.1 tetramer has only two membrane-spanning helices (S1-S2) but without the four membrane helices that form the voltage sensor in Kv channels. 174  sory protein interactions and signaling pathways ( Figure  4F) (Table 8). Newly synthesized Kir2.1 could be sent to specific membrane subdomains for functional expression by Golgi according to a recognition site formed by the residues in the CTD and amino-terminal domain (NTD) and interaction with adaptor protein complex 1 (AP1). 174 AKAP79 directly interacts with Kir2.1 through the intracellular N and C domains to promote anchoring other kinases (PKA, PKC) and close to Kir2.1 phosphorylation sites. 175 PIP 2 is an essential cofactor for activating Kir2.1 channel function. 176,177 In addition, Kir2.1 closely interacts with Nav1.5 (Figure 1) 178,179 and shares a coupled forward trafficking process with Nav1.5. 179 Normal trafficking of Kir2.1 could enhance the functional expression of Nav1.5 compared to Nav1.5 alone, while trafficking-deficient variants disrupt Kir2.1 functional expression at the membrane and also exert a DN effect on Nav1.5 functional membrane expression. Thus, in addition to controlling resting MP, I K1 could also modify Nav1.5 function and cell excitability. In turn, suppression of Nav1.5 by the CaMKII inhibitor KN93 178 or by trafficking-defective Nav1.5 variants could trap Kir2.1 channels, 180 thus decreasing I K1 in addition to I Na .
Most KCNJ2 LOF mutations (Table 9) are associated with type 1 Andersen-Tawil syndrome (ATS), in which LQTS7 is the primary cardiac manifestation. 181,182 I K1 reduction could prolong the terminal phase of the cardiac AP and contribute to the development of DAD and ventricular arrhythmias in ATS. 183 LOF mutations could suppress I K1 via impairment of PIP2 gating, 176,182 membrane trafficking, 174,179 or potassium conduction. 184,185 On the other hand, GOF mutations in KCNJ2 (Table 9) cause SQT3. 186,187

Drug-induced cardiovascular arrhythmias
In addition to gene mutation-induced congenital arrhythmias, drug therapy could exert side effects on cardiac VGIC and increase the risk of life-threatening arrhythmias, such as drug-induced LQTS (diLQTS) and torsades de pointes (TdPs) that is morphologically distinctive polymorphic ventricular tachycardias with short-long-short cycles patten. 3,4,188 LQT on the surface electrocardiogram correlates with ventricular AP repolarization prolongation at the cellular level. 189 Drugs can induce AP repolarization prolongation by inhibiting I Ks , or, more frequently, I Kr . Due to the robustness of I Kr , defective I Ks by blockade of Kv7.1 might produce little AP prolongation in humans and other large mammals 190 but might further prolong AP and induce LQT1 when challenged with β-AR stimulation 190 or reduce repolarizing currents by drugs, especially I Kr . 191,192 Kv11.1 is recognized as a predominant target for diLQTS and TdPs due to its intrinsic arrhythmogenic activity, although it is one of the interests of the development of antitachyarrhythmia drugs. The list of drugs that inhibit Kv11.1 includes not only includes antiarrhythmics (such as dofetilide) but also antipsychotics (such as Pimozide), gastroprokinetic agents (such as cisapride), antihistamines (such as astemizole), and other drug classes. 19 Among all potassium channels, Kv11.1 is unique in having a small central cavity with extended pockets so that it is susceptible to direct blockade by a wide range of drugs. 19 In addition, some drugs could exert inhibitory effects on Kv11.1 trafficking 193 or coexisting effects of channel blocking and trafficking defects, thus causing diLQTS and TdPs. 194 Due to the increasing attention that diLQTS has attracted from clinics, drug developers, and pharmaceutical regulators, 3 cardiovascular safety concerns are the most common reasons for the withdrawal of approved drugs from the market or the termination of potential drugs during preclinical or clinical trials. 3 For example, the noncardiovascular drug cisapride has been withdrawn from the US market because it produces a modest increase in the QT interval in children, causing TdP; 195 the drug exerts this effect by inhibiting Kv11. 1. 196 Since the outbreak of the coronavirus disease 2019 (COVID-19), many repurposed drugs are proposed as potential therapies for this disease; their risks, causing LQTS or TdPs is being evaluated. 197

Development of drug safety evaluation guidelines
Since the guidelines, including the International Council for Harmonisation (ICH) S7B (nonclinical) and E14 (clinical), 198 were announced in 2005, Kv11.1 channel safety screening data of new drug candidates before beginning clinical trials has become a great need in the pharmaceutical industry. 199 However, promising drug candidates might be eliminated by the guidelines because variations in the potency of Kv11.1 blocking could result from varying patch clamp protocols and a poor ability to statistically quantify experimental variability. 200 Moreover, promising drugs might be Kv11.1 blockers but exceptions in terms of causing TdPs or arrhythmia. Some also block other cardiac currents 201 (Table 11) necessary for TdPs development but do not obviously prolong AP repolarization. [202][203][204]  proposed to be more effective than Kv11.1 assays in predicting TdPs. 205,206 Although the current paradigm has largely kept potential torsadogenic drugs off the markets, but a new cardiac safety paradigm with comprehensive modelinformed approach rather than exclusively by potency of Kv11.1 block and by QT prolongation is urgent to adopted to improve the deficiencies of current paradigm, more specifically discern a real proarrhythmic risk of promising drugs, and enhance the development of effective and safe products or therapeutics. 6 In 2013, several organizations formed a team to develop the Comprehensive In Vitro Proarrhythmia Assay (CiPA) initiative, 6 a new paradigm developed with the goal of presenting a deeper understanding of the mechanism of TdPs and improving the assessment of the proarrhythmic effects of potential drugs. It is driven by mechanistically based in vitro assays of drug effects on multiple cardiac channels coupled in silico reconstruction of cardiac AP, and comparison of predicted and observed responses in human-derived cardiac myocytes. Twenty-eight drugs with well-characterized three torsadogenic risk groups (Table S3) have been selected and considered as test cases to build/calibrate model for testing and validation of in silico and stem cell CIPA models. 207 Several working groups are involved in developing the CiPA: 1. The ion channel group is developing voltage-clamp protocols by MPC or APC for several key cardiac ion channels. It is believed that at least six ion channels are involved in cardiac APs: Nav1.5, Kv4.3, Cav1.2, Kv11.1, Kv7.1, and Kir2.1. 208 A study evaluated the predictive ability of these six ion channels using APC and showed that four ion channels provided good predictions, whereas the analysis of three channels wrongfully predicted one high-risk drug to be safe. 209 Improved systematic approaches for accurately estimating the potency and safety margins are required. 200 Increasing APC-based assays have been explored in Kv11.  212 to improve the evaluation strategies. 2. The in silico group is building computer models to reconstruct electrophysiological activities and drug effects on multiple human cardiac currents by integrating experimental data within a heart cell and subsequently outputting the net impact on the cellular APD and QT interval for predicting drug-induced proarrhythmic risks. 209 For example, by using an in silico model, several proposed drugs against COVID-19 are estimated to have a significant risk for LQTS; thus, mandatory monitoring of the QT interval should be performed among patients in use of drugs. 197 In silico models are keeping updated to expand the index for discriminating TdPs compounds 213 and to satisfy a series of general principles for the validation of proarrhythmia risk prediction. 214 Those principles will help shape the future important directions of more accurate prediction models. 214 For example, development of better simulating models to capture the drug response not only in normal humans but also in specific patient populations. 214 With the application of in silico modeling, machine learning could identify cellular electrophysiological phenotypes associated with patients who has certain cardiac diseases and further predict which patients face an elevated risk of ventricular arrhythmias and sudden death. 215 However, information such as comparisons among drugs with similar chemical or affinity profiles is not yet possible incorporated into in an silico model. 216 Thus, newer proarrhythmia risk prediction models could be developed to aid in decision making. 216 For example, a computational pipeline was recently developed to predict Kv11.1 blocker proarrhythmic risk from drug chemistry and distinguish drugs that have similar chemistry and effects on the AP and QT interval but different proarrhythmic risk levels. 216 3. The myocyte group used iPSC-CM assays to evaluate the in vitro and in silico assay results. 217 Native human cardiomyocytes are ideal but with difficulties to obtain, maintain in long-term culture. 2 Native cardiomyocytes from different species have variations in the waveforms of APs and drug responses due to differences of potassium currents densities. 1,130 Rodent is not an appropriate specie for modelling human repolarization due to dominant I to ; dogs and rabbits are relative closely to human due to the major role of I Kr in repolarization. 218 Thus, the need for proarrhythmia evaluation in preclinical studies based on human models is emphasized. Currently, hiPSC-CMs have provided a perfect platform for proarrhythmia evaluation and safety evaluation of human cardiomyocytes in preclinical studies, and various AP parameters could be measured using highthroughput systems. [15][16][17]30,219 4. The clinical translation group will use clinical Phase I ECGs to evaluate potential unanticipated effects.
In addition, evaluating the effect of compounds on the overall APs rather than a single ion channel current has been proposed to be a more appropriate approach. 5,6 However, the more depolarized resting MP of hiPSC-CMs than that of primary cardiomyocytes is the limitation and challenge of their use in safety evaluation. This is due to the distinct expression level of ion channels expression compared to primary cardiomyocytes, especially the low expression of the I K1 channel protein Kir2.1. 173 Exogenous overexpression of the Kir2.1 173,220 or electronic injection of an I K1 -like current by dynamic clamp into hiPSC-CMs 221 to compensate and thus achieve more stable AP facilitates clinical applications, drug discovery, and cardiotoxicity screening. Although it is unclear when the CiPA project will lead to new guidelines (as organizations are generally conservative when considering changes to effective standard protocols), the CiPA initiative and other similar projects worldwide are promoting the development of questions and answers (Q&As) to facilitate the application of the ICH S7B and E14 guidelines. 200 With the development of these techniques, other cardiac safety liabilities, such as dysfunction of EC coupling and contractile and structural cardiotoxicity, may also be added to electrophysiological tests in the same platform to complement CiPA for regulatory use. 222,223 In general, with the development of medium-or highthroughput test systems to produce efficient, reliable result output and of basic knowledge of VGICs to update the detection assay designs and analysis methods, drug safety evaluation will receive more attention in preclinical research. Evaluation will be conducted as early as possible to avoid further unnecessary investments in unusable compounds during later stages of drug development.

CONCLUSION AND PERSPECTIVE
This review provides detailed descriptions of major ion channels in ventricular myocytes, including their expression, structures, regulators, and contributions to normal excitability and congenital pathology. It has been discussed that the application basic and newly discovered knowledge of cardiac ion channels and the continuous development of techniques employed in studies of cardiac ion channels can lead to more attentions to comprehensive proarrhythmic risk assessment in human cardiomyocytes platform in preclinical studies and promote development of cardiovascular safety evaluation guidelines. Recent research on potential targets of interest in cardiomyocytes, such as TTX sensitive or TTX-insensitive Nav and Ryr2 regulatory TRICA channels, has opened new avenues for improving our understanding of the molecular mechanisms of Ca 2+ homoeostasis, EC coupling, and associated cardiac disease pathogenesis. The development of a Nav-selective inhibitor or a heart-specific Nav channel-KO mouse model will be beneficial for further confirming the pathological mechanism of specific Nav channels. 41,43 The selective inhibition of Nav channels may offer a potential therapeutic target to alleviate arrhythmias during states of Ca 2+ overload. 41,43 The development of hIPSC-CM, highthroughput techniques for cellular phenotype detection (such as Aps and contraction), computational simulation models facilitate integration of multiple channels, achieving a comprehensive view of channelopathies as a global phenomenon in human myocytes. Modeling of patient-specific iPSC-CMs 149,224 provides great benefit for the precision medicine treatment of congenital cardiac arrhythmia and for the screening of promising or already approved drugs to test for mutation-specific antiarrhythmic effects.
Over time, technological developments will certainly further promote the study of an increasing number of scientific questions related to cardiac physiology and pathology and reveal additional ion channels with potential involvement. Based on cryo-EM structures of many VGICs in basic science, a large body of experimental and clinical observations concerning VGICs has been interpreted and summarized by the structural template. 99,100 The development of clinical and translational medicine could be advanced by the discovery of the potential drug targets within many VGICs, as well as drugs characteristics, targets-drugs interaction, and computational models for integrating and predicting information. Recently, a novel multiscale approach has been developed to predict drug-induced arrhythmia directly based on structural models of drug-channel interactions and kinetics by using integrative experimental and computational modeling and machine-learning from the atom to the rhythm in the heart. 216 For potential targets, cryo-EM structures map and classify hundreds of clinical arrhythmia variants onto all major domains in the structure of many VGICs, 18,21 reveal the common or distinct clusters of arrythmia mutations among different types of VGICs 99,100 or different isoforms of the same VGIC, 21 provide the molecular basis for understanding disease mechanisms, and thus allow the development of structure-based diagnosis and drug discovery for arrhythmias in the future. 18 For clinical or potential drugs, the cryo-EM structure of VGIC-drug interactions provide structural insights into the binding affinity and mechanism of drugs, 20,22 which is beneficial for modifying the structure of drugs, screening alternatives or synthesizing new compounds. For example, cryo-EM structure of Kv11.1 channel in the open state 19 promotes the investigation of the state-dependent blockade of Kv11.1 by the heart-ratelowering agent ivabradine, 22 which could also exert antiarrhythmic effects in SQTS1 hiPSC-CMs with the N588K mutation. 170 The development of novel additional pharmacological approaches (eg, activators/allosteric modulators of potassium Kv11.1 and Kv7.1 channels) are needed to counteract both congenital LQTSs, although currently available therapies (implantable cardioverter defibrillators) have yielded good clinical responses. 225 For example, Lumacaftor, a drug already in clinical use for cystic fibrosis, has been demonstrated to interact with a site distinct from where classic Kv11.1 blockers bind, thereby restoring Kv11.1 trafficking defects and alleviating LQTS2. 165 Polyunsaturated fatty acids (PUFAs) and their analogs Narachidonoyl taurine have been found to speed up Kv7.1 channel opening and restore channel gating of many different mutant channels 226 PUFAs and their analogs are effective in shortening the cardiac action potential in pharmacologically prolonged ventricular action potential and QT interval in isolated guinea pig hearts 227 and in hIPSC-CM. 228 Therefore, activators of Kv7.1 are also worth developing to treat LQT1 based on structure-function studies on diverse IKs channel mutations. However, PUFAs analogs vary in selectivity and different effects for Kv7.1, Nav1.5, and Cav1.2 through nonidentical mechanisms. It is necessary to determine the specific binding sites of PUFAs analogs among normal VGICs and to further identify the most therapeutically relevant PUFAs and PUFA analogs in the treatment of different LQTS subtypes. Moreover, if negative allosteric modulators are used in combination with patient-specific hIPSC-CM, drugs that have been withdrawn from market or excluded from clinical application due to diLQT effects may be reconsidered or even rescued to clinical use 229 after safety validation by electrophysiological approaches.
In addition, site-specific and target-oriented approaches using nanomaterials (NMs) have been increasingly applied but might exert potential toxicity on ion channels and cardiac electrophysiology. 230,231 Maybe revealing these NMinduced structural changes in ion channels could facilitate the modification of bioactive NPs to optimize NM-based drug delivery and safety. 230,232 In general, in-depth studies that combine electrophysiological approaches with other technologies are being used to explore the expression, function, mechanism, and structure, and activity modulation of WT VGICs and a broad variety of mutated VGICs, providing critical contributions to our knowledge of the roles of VGICs in both normal and diseased cardiac functions, thus facilitating to the discovery of potential structurally and functionally guided drug targets for the modification of channel function and for the treatment of inherited or drug induced cardiac diseases, providing a basis for structure-and mechanism-based personalized clinical management, prompting safety control committees to establish more integrated strategies for drug screening, and enabling improved prediction of cardiac risks to provide safer and more effective drugs for clinical use.

A C K N O W L E D G M E N T S
We thank Dr Qinqin hu, Dr Peng Li, Dongsheng Xu, Na Wu, Jingjing Ma, Yiqing Hu, Zhenyang Guo, and Zheyan Fang for comments on the manuscript. This work was supported by the National Natural Science Foundation of China (81521001, 81870182) and the National Key Basic Research Programme (2016YFC1301204, 2020YFC1316700).

A U T H O R C O N T R I B U T I O N S
Hua Li, Xiangdong Wang, and Junbo Ge proposed the conception, study design, and had the final approval of the manuscript submitted. Lulan Chen and Yue He participated in the data collections and analysis, the drafting of the manuscript, and the submission.

C O M P E T I N G I N T E R E S T S
There is no conflict of interest involved in this review.