Human Induced Pluripotent Stem Cell-Derived Cardiomyocytes as Models for Cardiac Channelopathies: a Primer for Non-Electrophysiologists

Abstract

Life threatening ventricular arrhythmias leading to sudden cardiac death (SCD) are a major cause of morbidity and mortality. In the absence of structural heart disease, these arrhythmias, especially in the younger population are often an outcome of genetic defects in specialized membrane proteins called ‘ion channels’. In the heart, exceptionally well-orchestrated activity of a diversity of ion channels mediates the cardiac action potential. Alterations in either the function or expression of these channels can disrupt the configuration of the action potential, leading to abnormal electrical activity of the heart that can sometimes initiate an arrhythmia.

Understanding the pathophysiology of inherited arrhythmias can be challenging due to the complexity of the disorder and lack of appropriate cellular and in vivo models. Recent advances in human induced pluripotent stem cell (hiPSC) technology have provided remarkable progress in comprehending the underlying mechanisms of ion channel disorders or ‘channelopathies’ by modeling these complex arrhythmia syndromes in vitro in a dish. To fully realize the potential of iPSCs in elucidating the mechanistic basis and complex pathophysiology of channelopathies, it is crucial to have a basic knowledge of cardiac myocyte electrophysiology. In this review, we will discuss the role of the various ion channels in cardiac electrophysiology and the molecular and cellular mechanisms of arrhythmias, highlighting the promise of hiPSC-CMs as a model for investigating inherited arrhythmia syndromes and testing antiarrhythmic strategies. Overall, this review aims to provide a basic understanding of the electrical activity of the heart and related channelopathies, especially to clinicians or research scientists in the cardiovascular field with limited electrophysiology background.

Introduction

In a subset of individuals that are relatively young and otherwise healthy, life threatening ventricular arrhythmia can develop as a result of mutations in genes encoding ion channels, transporters, and accessory subunits that lead to altered cardiac electrophysiology^{1, 2}.

Over the past two decades, significant progress has been made in the identification of genetic defects underlying many recognized types of inherited arrhythmia syndromes (IAS). However, for most of these inherited disorders, clinical management is still hindered by insufficient knowledge of the functional consequences of the candidate mutation. This is largely due to lack of suitable model systems that appropriately mimic human cardiac electrophysiology. For example, widely used transgenic mice have a heart rate of about eight times faster than that of humans (500 vs. 60 beats/min), and the action potential duration (APD) in mouse CMs is relatively shorter secondary to differences in K⁺ currents³. Additionally, variable disease expressivity and sensitivity to therapeutic interventions between mutations and/or patients necessitate more personalized treatment. The recent emergence of induced pluripotent stem cell (iPSC) technology^{4, 5} has provided an unprecedented opportunity for generating and studying iPSC-derived cardiomyocytes from diseased patients. iPSC based studies allow investigation of patient- and/or mutation-specific disease mechanisms as well as identification and assessment of novel and/or custom tailored therapeutic strategies for arrhythmia syndromes.

This review provides a basic understanding of the fundamental electrophysiology of cardiac myocytes and the underlying mechanisms of IAS to help physicians and research scientists without expertise in this field to contribute towards the design of effective treatment strategies. The review will highlight the recent advances and opportunities from hiPSC-CMs models for investigating IAS and testing antiarrhythmic strategies.

Basis of Cardiomyocyte Excitability - Origin of the Membrane Potential

The heart is an electromechanical organ that pumps blood throughout the body undergoing roughly three billion cycles in an average lifetime⁶. To perform this function in the most efficient way, the heart contracts and relaxes in a highly coordinated way (both spatially and temporally) during each heartbeat referred to as a cardiac cycle. Each cardiac cycle starts with an electrical impulse, originating from a group of specialized pacemaker cells called the sinoatrial node (SAN), which subsequently propagates throughout the heart. This electrical excitation of the heart is made possible by an electrochemical gradient that exists across the membrane of each heart muscle cell or cardiomyocyte (CM)⁶.

The lipid bilayer membrane of a CM is embedded with integral proteins called ion channels that facilitate the passive diffusion of ions across the hydrophobic cell membrane. Ion channels are generally characterized by two important properties: 1) gating - ion channels open and close in response to a specific stimulus either binding of a ligand (e.g. neurotransmitters) or changes in the transmembrane (TM) potential; 2) selectivity - ion channels allow a high rate of selective passage of only a particular species of ions such as K⁺, Na⁺, Cl⁻ or Ca²⁺. The flux rate of channels is quite high, equivalent to 10⁷ ions for a single monovalent channel current magnitude of 1 pA. These properties of ion channels coupled with the presence of large negatively charged molecules (mainly proteins) inside the cell and energy-requiring activity of membrane pumps (primarily the Na⁺/K⁺-ATPase and Ca²⁺-ATPase) results in an unequal distribution of charged ions across the membrane (Figure 1A) generating an electrochemical gradient which sets the membrane potential (MP)⁷.

Figure 1. Electrophysiological Basis of the Electrical Activity in the Heart.

(A) Schematic for ion channel proteins embedded in a lipid bilayer and ionic basis for resting membrane potential in a cardiac myocyte. Based on the depicted concentration of cations, the calculated equilibrium potential (E = (RT/zF) * ln [X]_out/[X]_inside) for K⁺, Na⁺ and Ca²⁺ would be approx. −90 mV, +50 mV and +130 mV at 25^oC, respectively. (B) Schematic depicting different phases (0–4) of a typical ventricular action potential (top) with various depolarizing (arrows down) and repolarizing (arrows up) ionic currents. Bottom panel shows the relative amount of different currents between an adult ventricular myocyte (gray) and a ventricular-like hiPSC-CM (red). The associated ion channel abnormalities are also shown. Purple circles represent loss-of-function mutations and blue circles represent gain-of-function mutations. (C) Excitation-contraction coupling in a cardiac myocyte. Ca²⁺ entering via plasma membrane channels activates ryanodine receptors (ryanodine receptor 2 [RYR2]) and initiates Ca²⁺-induced Ca²⁺-release mechanism (CICR). CASQ2 is the cardiac isoform of the high-capacity Ca²⁺-sequestering protein, calsequestrin present inside SR. Phospholamban (PLN, brown oval) is shown on top of Sarco/endoplasmic reticulum Ca²⁺-ATPase type-2a (SERCA 2a). Mutations in proteins associated with SR Ca²⁺ release or uptake are associated with CPVT.

When ion channels open, they bias the MP of the cell towards the equilibrium potential of the permeant ion. The equilibrium potential for each ion, which can be calculated by the Nernst equation, is the MP where the chemical and the electrical forces are in balance such that there is no net flow of ions through the open channel. For example, under resting conditions, an atrial or ventricular myocyte is permeable to K⁺ because a specific type of K⁺ channel (Kir2.1) is constitutively in an open state, whereas other types of ion channels are closed unless activated by a change in MP. For this reason, the resting membrane potential (RMP) of a resting (i.e. between beats) CM is steered towards the K⁺ equilibrium potential of −90 mV. By convention, MP is measured as the voltage of the intracellular side of the membrane with respect to a ground electrode located outside the cell⁷.

The direction of ion currents conducted by an ion channel (into the cell [inward] or out of the cell [outward]) is determined by the electrochemical gradient of the permeant ions. For example, outward flow of K⁺ ions through K⁺ channels drive the myocyte towards −90 mV while inward flow of Na⁺ ions through opening of voltage gated Na⁺ channels forces the MP towards a positive value (up to +40 mV).

Action Potential of a Human Cardiac Myocyte

An action potential (AP) is generated by the orchestrated opening and closing of multiple ion channels present in the plasma membrane of an individual CM that conduct depolarizing, inward (Na⁺ and Ca²⁺) and repolarizing, outward (K⁺) currents⁸. The AP waveform in different regions of the heart is distinct owing to differences in the expression and/or the properties of the underlying ion channels. These differences and the unique architecture of the cardiac conduction system (AV node and His-Purkinje system) contribute to the normal unidirectional propagation of excitation through the atria to the ventricles and to the generation of normal cardiac rhythms in the heart^{9, 10}.

Figure 1B illustrates the five phases (0 – 4) of a ventricular AP. Phase 4 represents the resting state in a normal working CM wherein the RMP is ~ −90 mV. During phase 0, a depolarizing stimulus from a neighboring cell elevates the RMP to the threshold at which voltage-gated Na⁺ (Na_V) channels open, allowing Na⁺ ions to rapidly diffuse into the cell down their electrochemical gradient. The rapid phase 0 depolarization shifts the MP towards the Na⁺ equilibrium potential of +40 mV and is responsible for the rapid propagation of the cardiac impulse. Within a few ms after activation, Na⁺ channels rapidly inactivate and thus are unable to conduct Na⁺. The depolarized MP results in a brief activation of the transient outward potassium current (I_to), that in turn leads to a brief, and only partial repolarization (phase 1). Phase 2 is the characteristic prolonged “plateau phase” of the CM AP, where inward and outward currents are nearly equal (hence only a minor change in MP). During phase 2, activation of voltage-gated L-type Ca²⁺ channels (LTCC) leads to the entry of Ca²⁺ into the cell, which in turn activates nearby intracellular ryanodine receptors (RyR2) present in the membrane of the sarcoplasmic reticulum (SR) and release of SR-stored Ca²⁺ into the cytosol. During phase 2, the inward currents mediated by LTCC (I_CaL) and the sodium calcium exchanger (I_NCX) are electrically balanced by outward currents (I_Kr and I_Ks) mediated by two different voltage-gated delayed rectifier K⁺ channels. As LTCC inactivate late in the plateau phase, I_Ks and I_Kr return the MP towards the K⁺ equilibrium potential causing the “repolarization phase” or phase 3 of AP. Inwardly rectifying K⁺ channels (Kir2.1, that mediate I_K1) contribute to late phase 3 repolarization and to the maintenance of the resting MP in phase 4^11–13.

The inward and outward conductances during the plateau phase are delicately balanced and modest changes during this phase can lead to differing AP morphologies such as a spike-and-dome shape in the ventricular and a triangular shape in the atrial CMs. Atrial AP also displays a less negative RMP, attributable to the reduced expression of Kir2.1. I_to is more prominent in human atrial than in ventricular myocytes¹⁴. Moreover, atrial cells selectively express two additional K⁺ currents; the ultra-rapidly activating delayed outward rectifying current (I_Kur) and the acetylcholine-activated K⁺ current (I_KACh) that are absent in ventricular myocytes. I_KACh is small under normal conditions but it becomes prominent during increased vagal tone¹⁵. Activation of I_Kur and I_KACh during phase 3 explains the lower plateau potential and shorter APD of atria.

The AP morphology of pacemaker cells in the SAN and the atrioventricular node (AVN) is significantly different from that of ventricular and atrial cells. A hallmark feature of pacemaker cells is their ‘automaticity’, i.e. spontaneous beating in the absence of an external stimulus. The automaticity is driven by both voltage and Ca²⁺-dependent mechanisms and involves the funny current (I_f) carried by the hyperpolarization-activated cyclic nucleotide-gated (HCN) channels¹⁶. In pacemaker cells, the RMP at the onset of phase 4 is more depolarized (−50 to −65 mV) compared to ventricular cells. Due to this depolarized RMP, the Na⁺ channels are mostly inactivated. As a result, the phase 0 in pacemaker cells is mediated by the relatively slowly activating I_CaL, leading to very low upstroke velocity. Phase 1 is absent in pacemaker cells due to an absence of I_to. Phase 2 repolarization again represents the balance between I_CaL, I_Kr and I_Ks and the final phase 3 repolarization results from inactivation of I_CaL and the recovery from inactivation of I_Kr. I_f activates at the end of phase 3 repolarization and contributes to the initiation of phase 4 diastolic depolarization¹⁷.

Excitation Contraction Coupling

The synchronous beating of the heart depends on the coupling between the electrical activity (excitation) and the mechanical pumping (contraction), hence the name “excitation-contraction (EC) coupling” (Figure 1C)^{18, 19}. EC coupling involves finely controlled changes in the concentration of Ca²⁺ ions inside the myocyte, raising the cytoplasmic [Ca²⁺] from ~60–100 nM to 600–1000 nM and activation of the contractile apparatus²⁰. During phase 2 of the AP, Ca²⁺ ions enter the cell via LTCCs and activate nearby (within 15 nm) clusters of specialized SR Ca²⁺ release channels (RYR2, type 2 ryanodine receptors)²¹. Activation of RYR2 induces massive release of Ca²⁺ from the SR, the principal intracellular Ca²⁺ storage site, a process referred to as “calcium-induced calcium release” (CICR)²². The Ca²⁺ released into the cytoplasm from the SR then binds to and activates cardiac troponin C (TnC), the Ca²⁺-sensing protein of the contractile apparatus and begins myofilament contraction. During diastole, relaxation of the myocardium occurs by the pumping of cytosolic Ca²⁺ back into the SR by Ca²⁺-ATPase type-2a (SERCA2a) or into the extracellular space by Na⁺/Ca²⁺-exchanger type-1 (NCX). The NCX is electrogenic in nature and imports three Na⁺ ions in exchange for a single Ca²⁺ ion extruded, creating a depolarizing transient inward current (I_NCX). The sequential rapid release of Ca²⁺ from the SR into the cytosol followed by rapid reuptake by SR or extrusion from the cell creates a Ca²⁺ transient. The amplitude of these transients largely depends upon the amount of Ca²⁺ release from the SR by RYR2 and determines the strength of the systolic contraction²³.

Characteristics of Human iPSC-Derived Cardiomyocytes

Native human CMs are ideal to study the molecular mechanisms associated with cardiac ion channel abnormalities. However, it is not only highly invasive and cumbersome to obtain cardiac tissue from patients but also not currently possible to maintain these cells in long-term culture. Consequently, the majority of functional studies on specific mutations associated with IAS have relied on overexpressing the mutated channels in heterologous expression systems, including Xenopus oocytes, human embryonic kidney (HEK) cells, and Chinese Hamster Ovary (CHO) cells. An important shortcoming of these models is that they lack crucial constituents of cardiac ion channel macromolecular complexes that might be required to reproduce the exact molecular and electrophysiological phenotype linked with the mutation²⁴. Transgenic mouse models have been very useful in this regard (Table 1), but substantial differences in gene expression profile and physiology between species severely limit the validity of extrapolating data from rodents to humans. For example, the heart rate in mice is about eight times faster than that of humans and cardiac repolarization in mouse CMs relies primarily on I_to, I_K,slow1, I_K,slow2, and I_SS ion currents, whereas in humans, repolarization is mostly dependent on I_Ks and I_Kr. In addition, Ca²⁺ handling and myofilament proteins differ in expression and pathophysiology in mouse versus human hearts³.

Table 1.

Cardiac Ion Channels: Genetics, Properties and Role

Channel or current	Genes (encoding Major pore- forming α-subunit)	Minimal β-subunit/s*	Expression	Current properties and contribution to AP	Pathophysiological significance	Functional state in hiPSCs (compared to adult CMs)	Knockout mouse phenotype	Ref.
Voltage-gated Sodium Channels (INa): Characterized by a prominent transient (INaT) and a minor late component (INaL)
Nav1.5	SCN5A	SCN1B– SCN4B	All cardiac cell types (very low expression in nodal cells)	Tetrodotoxin (TTX) resistant (IC50 ≥ 1 µM), inward current responsible for the rapid upstroke of the AP (Phase 0) and for proper cardiac excitability and impulse propagation	BrS (LOF), sick sinus syndrome and LQTS3 (GOF leading to enhanced INaL)	Slightly lower functional expression**. A lower upstroke velocity in hiPSC-CMs does not necessarily denote low INa density because at the depolarized MDP typical for hiPSC-CMs, a fraction of Na+ channels remain inactivated and thus unable to open during the upstroke	SCN5A−/− mice are Embryonically lethal; SCN5A−/+ display conduction related abnormalities	24, 123
Nav1.8	SCN10A		Very low expression if at all in cardiac myocytes, predominantly in intracardiac neurons	TTX-resistant, affects cardiac conduction indirectly via neutrally mediated effect or modulation of Nav1.5 expression	BrS	No functional evidence	SCN10A−/− mice have shorter PR intervals	124, 125
Voltage-gated Calcium Channels (ICa)
L-type (long-lasting type) (ICaL) Cav1.2	CACNA1C	CACNB2, CACNG	All cardiac cell types (nodal and some atrial cells also express limited amounts of Cav1.3)	Inward current during the plateau (phase 2) of the AP, involved in EC coupling, conduction and automaticity	LQTS8 (Timothy syndrome, GOF) and BrS (LOF)	Robust expression of ICaL with similar characteristics	CACNA1C−/− leads to embryonically lethality. Heart- specific graded reduction in gene expression leads to hypertrophy while severe reduction causes early adulthood lethality	123, 126– 128
T-type (transient-type) (ICaT) Cav3.1	CACNA1G		Principally in pacemaker, atrial, and Purkinje cells	Activation at more Hyperpolarized potentials than LTCC and highly sensitive to Ni. Together with ICaL, plays a role in conduction and diastolic pacemaker current		ICaT expression has recently been reported in SAN-like pacemaker cells derived from hPSCs		31
Potassium Channels: Mediate AP repolarization. The human genome includes ~80 genes that encode for K+ channel subunits, but only a subset of these genes are highly expressed in the heart. Each K+ channel subunit is composed of either two (e.g. Kir channels) or six (e.g. voltage-gated K+ (Kv) channels) transmembrane (TM) segments
Transient Outward K+ Channels (Ito)
Ito,f (fast) Kv4.2, Kv4.3	KCND2, KCND3	KCNE2–3, KCNIP1–2, DPP6	Greater Ito,f density is present in the atria and Purkinje fibers than in the ventricular myocytes	Displays faster gating properties resulting in brief K+ efflux and a short-lasting repolarization (the phase 1 notch in the AP)	BrS (GOF), reduced current densities have been found in failing heart and chronic AFib	Slightly lower functional expression**. Slow recovery of Ito from inactivation and different subunit stoichiometry (i.e. relatively high Kv1.4 and low KCNIP2 mRNA expression) in hiPSC- CMs might lead to smaller contribution to AP. At depolarized MDP of hiPSC-CMs, a large proportion of the channels are unavailable	In Kv4.2 −/− mouse, Ito,s is upregulated and compensates for the loss	123, 129– 132
Ito,s (Slow) Kv1.4	KCNA4	KCNB1–4	Ventricular myocytes	Recovers slowly from inactivation in comparison to Ito,f. Contributes to early repolarization during phase 1
Delayed Outward Rectifier K+ Channels: So named because of delayed onset in their activation, which starts during the plateau phase of AP. These are voltage-gated K+ (Kv) channels. They play a prominent role in phase 3 repolarization. LOF or GOF mutations in these channels disrupt normal cardiac repolarization and result in various cardiac rhythm disorders, including congenital LQTS, SQTS, and familial AFib (Figures 1B and4).
IKur (ultra-rapid) Kv1.5	KCNA5	KCNAB1–2 (affect membrane trafficking and properties of activation and inactivation)	Atria-specific	Activates rapidly upon Membrane depolarization and undergoes very slow inactivation. Contributes to both early and late repolarization of the AP	Both LOF and GOF mutations in KCNA5 have been linked to familial AFib (likely explained by different underlying mechanisms of arrhythmia generation; i.e., AP prolongation leading to EADs vs AP shortening leading to reentry)	Lower proportion of atrial-like cells in hiPSC- CM cultures make an exact comparison difficult. RA-guided differentiation results in increased atrial-like phenotype and higher expression of Kv1.5 mRNA and IKur density in hPSC-atrial CMs	Precise role of Kv1.5 loss in transgenic mouse models is uncertain due to the compensatory mechanisms and effects of the inserted transgene	30, 133– 135
IKr (rapid) Kv11.1	KCNH2 (hERG1)		All cardiac cell types	IKr is characterized by fast inactivation, fast recovery and slow deactivation. It plays a major role in phase 3 repolarization	LQTS2 (LOF) and acquired LQTS due to high susceptibility of hERG to drug block, SQTS1 (GOF)	Comparable current densities as that of adult CMs	KCNH2−/− result in embryonic lethality while KCNH2 −/+ result in mild APD prolongation	59, 86, 123, 136
IKs (slow) Kv7.1	KCNQ1	KCNE1 (imparts unique slow activation- deactivation kinetics and physiological voltage range of activation to Kv7.1)	All cardiac cell types	Activates slowly during the depolarized plateau (phase 2 of the AP)	LQTS1 (LOF), SQTS2 (GOF)	Comparable current densities as that of adult CMs	KCNQ1−/− loss does not lead to a robust cardiac phenotype in mouse	58, 137
Inward Rectifier K+ Channel (IK1)
Kir2.1 (Kir2.2/2.3)	KCNJ2 (KCNJ12/ KCNJ4)		Kir2.1 is the Predominant isoform. Expression is higher in the ventricles than in the atria. Virtually absent in nodal tissue	Plays a major role in setting the RMP and terminal repolarization. Protects the ventricular cell from abnormal pacemaker activity	LQTS7 (Andersen-Tawil syndrome, LOF), SQTS and familial AFib (GOF)	IK1 density is considerably smaller in hiPSC-CMs, which (in combination with presence of If) contributes to the spontaneous activity and relatively depolarized RMP	Kir2.1−/− leads to perinatal lethality due to cleft palate. Kir2.1−/− myocytes showed broader APD and frequent spontaneous APs	32, 123, 138
Acetylcholine-activated K+ Channel (IKACh)
Kir3.1 (GIRK1) and Kir3.4 (GIRK4)	KCNJ3 and KCNJ5	GIRK1 Requires GIRK4 for Membrane targeting	Abundantly expressed in atria, SAN and AVN	Mediates Parasympathetic regulation of the heart rate. Activation of IKACh hyperpolarizes the MP and shortens the APD	Has been associated with AFib. Studies focusing on IKACh in hiPSC-CMs may be useful in modeling atrial arrhythmias	Atrial-like CMs were found to have substantially higher IKACh (activated by carbachol) density that also hyperpolarized RMP compared to ventricular- like CMs	KCNJ5−/− results in diminished vagal control of the heart rate	133, 139, 140
Hyperpolarization-activated, intracellular Cyclic Nucleotide-gated (HCN) Pacemaker Channel (If)
Pacemaker channels	HCN1, HCN4		HCN4 is the most abundant isoform in the nodal tissue. HCN2 has been shown to express in ventricles at very low levels	A non-selective inward current that is directly regulated by cAMP levels (affected by sympathetic stimulation). Contributes to diastolic depolarization in pacemaker cells (phase 4)	SA node dysfunction such as sick sinus syndrome	hiPSC-CMs express substantial If contributing to their spontaneous activity. In this regard, they resemble immature neonatal myocytes	Constitutive HCN4−/− results in Embryonic lethality while inducible cardiac- specific KO has bradycardia. Constitutive HCN2−/− results in sinus dysrhythmia	17, 32, 123

AFib, atrial fibrillation; AP, action potential; APD, action potential duration; AVN, atrio-ventricular node; EAD, early afterdepolarization ;EC, electro-mechanical coupling; GOF, gain-of-function; hPSC, human pluripotent stem cells; KO, knockout; LOF, loss-of-function; LQTS, long QT syndrome; MDP, maximum diastolic potential; MP, membrane potential; Ni, Nickel; RA, retinoic acid; RMP, resting membrane potential; SAN, sino-atrial node; SQTS, short QT syndrome.

^*An array of accessory subunits is required in most cases to reconstitute a native current phenotype in a heterologous expression system. Many of these accessory subunits have not been completely characterized, have regional or species variability¹⁰ and in some cases led to conflicting results depending on the expression system used. It is not possible to provide an exhaustive list of all the accessory subunits for each channel here; only the important ones are listed here.

^**An exact comparison of current density (e.g. for I_Na and I_to) is difficult due to different experimental conditions (e.g. temperature, different ionic concentrations and holding potential), regional variability in expression in an intact heart (e.g. epi- vs endocardium) and heterogeneous population of atrial-, ventricular- and nodal-like hiPSC-CMs in cultures.

The recent discovery of somatic cell reprogramming to generate iPSCs^{4, 5} has created great excitement because of the ability to produce for the first time, patient- and disease-specific human iPSC-derived cardiomyocytes (hiPSC-CMs). To generate hiPSCs, somatic cells can be obtained from patient’s hair, blood, skin, fat, urine, or oral mucosa^{24, 25}. These cells are then reprogrammed to a pluripotent state by introducing pluripotency-associated genes. The resulting iPSCs are then differentiated into CMs by using a number of strategies (for review, see [^26,27]). After approximately 8–12 days, clusters of beating cells typically appear which upon further maturation in culture can be enzymatically dissociated into single CMs for molecular and functional analysis.

The patch clamp technique is the gold standard to measure various AP parameters from hiPSC-CMs that correspond to specific ionic current properties. These parameters include RMP or maximal diastolic potential (MDP), cycle length (CL), AP amplitude (APA), maximum upstroke velocity (d_v/dt_max), and APD at various percentage levels of repolarization (i.e., APD20, APD50, and APD90). For example, MDP/RMP reflects activity of steady state K⁺ currents, primarily I_K1. The upstroke velocity reflects the magnitude of I_Na, while the AP durations (APDs) denote different phases of repolarization. APD20 is importantly regulated by I_CaL and I_to1, while APD90 is most dependent on I_Ks and I_Kr²⁵.

The common hiPSC differentiation protocols predominantly generate ventricular-like hiPSC-CMs^{28, 29}, although protocols aimed at specifically generating atrial-like³⁰ and nodal-like³¹ hiPSC-CMs have also been recently described. The distinction between different hiPSC-CMs is often made based on the AP morphology. For example, ventricular-like cells are characterized by a more negative MDP, a rapid AP upstroke, and a long plateau phase (see Figure 1B). The absence of a prominent plateau phase is defined as a characteristic of atrial-like cells, resulting in shorter APD compared to ventricular-like APs. Nodal-like APs display more depolarized MDP, a slower AP upstroke, and a prominent phase 4 depolarization²⁴. Recordings of subtype specific currents may also be used to distinguish atrial and nodal CMs within the hiPSC-CM population. For example, I_Kur and I_KACh are specific to atrial cells³⁰, while I_f and T-type Ca²⁺ current (I_CaT) are specific for nodal cells³¹ (Table 1).

In general, hiPSC-CMs are electrophysiologically immature compared to isolated adult ventricular CMs (AVCMs) as reflected by the lower upstroke velocity (secondary to less negative MDP), and less prominent notch (phase 1)^{24, 32}. One of the most remarkable differences between hiPSC-CMs and native adult human primary CMs relates to the fact that hiPSC-CMs (including ventricular-like and atrial-like cells) display innate automaticity resulting in spontaneous beating, like in fetal human CMs²⁴. This is due to substantially reduced density of I_K1 and expression of pacemaker current (I_f) in hiPSC-CMs that is virtually absent in AVCMs. Recent studies have aimed to overcome these limitations by artificially enhancing the I_K1 density through either viral over-expression of Kir2.1^{33, 34}, or by in-silico injection of I_K1 with kinetics of Kir2.1^{35, 36}. Over-expression of Kir2.1 in human embryonic stem cell-derived cardiomyocytes (hESC-CMs) abolished cell automaticity resulting in similar AP characteristics to those of adult CMs but the Ca²⁺ handling properties remained immature³³. However, in a more recent study, enhanced Kir2.1 expression in hiPSC-CMs improved both AP and Ca²⁺ handling properties³⁴. Similarly, enhanced I_K1 by in-silico injection in hiPSC-CMs, induced a more physiological and stable MDP, a ventricular-like AP morphology, and increased AP upstroke velocity comparable to that from freshly isolated human myocytes^{35, 36}. Additional approaches including three-dimensional (3D) tissue engineering, mechanical loading, modulation of substrate stiffness, phasic electrical stimulation, hormonal treatment (e.g., thyroid hormone), and long-term culture have been utilized to enhance the maturity of hiPSC-CMs^37–44.

Mechanisms of Arrhythmia

Although most hearts beat with remarkable fidelity and resilience, under certain circumstances the rhythm of the heart can fail. Accurate identification of specific mechanisms underlying arrhythmia may at times be challenging for the clinician and may require invasive EP study. Differentiating and understanding these complex mechanisms is critical to the development of appropriate diagnosis and treatment strategy. The fundamental mechanisms of arrhythmia can be classified into two major categories: 1) enhanced or abnormal impulse generation (focal activity) and 2) improper impulse conduction (reentry). Focal activity includes automaticity and triggered activity^{13, 45}.

Automaticity refers to the ability of CMs to initiate spontaneous AP. Normally, the SAN has the highest intrinsic rate and sets the heart rate. Abnormal patterns of automaticity of the SAN can result in either sinus tachycardia or bradycardia. If the SAN is unable to generate impulses or if these impulses fail to propagate, latent pacemakers in other regions of the heart can take over initiating excitation. Atrial and ventricular myocytes, which in the normal heart typically do not exhibit spontaneous activity, may exhibit automaticity properties (e.g. under β-adrenergic stimulation or reduced extracellular K⁺). The spontaneity in SAN arises as a result of net inward current during the phase 4 of AP causing diastolic depolarization⁴⁶ (Figure 2A).

Figure 2. Key Electrophysiological Mechanisms of Cardiac Arrhythmia.

Ectopic impulse generation by (A) Enhanced automaticity. (B) APD prolongation (in blue) leading to the development of Phase 2 and Phase 3 EADs (red). (C) Development of delayed after depolarizations (DADs) that occur due to Ca²⁺ overload. (D) Reentry requires a vulnerable substrate, which can be caused by APD shortening or dispersion of refractoriness. Schematic shows a depolarizing wavefront around an anatomical obstacle. Under normal conditions, the depolarizing waves around the obstacle cancel each other out. Certain conditions e.g. ischemia might generate areas of unidirectional block, or sufficiently slow conduction to enable recovery of excitability in time for re-excitation by the depolarizing wavefront.

Triggered activity arises from premature activation of cardiac tissues by two types of afterdepolarizations⁴⁷. Early afterdepolarizations or EADs occur during phase 2 or 3 of AP repolarization, delayed afterdepolarizations or DADs occur after repolarization is complete. When the amplitude of an afterdepolarization is large enough, a spontaneous AP can be triggered⁴⁸. EADs are often but not always associated with prolonged APD which occur when the depolarizing currents (e.g., I_NaL, I_CaL or I_NCX) predominate the repolarizing currents (e.g., I_Kr, I_Ks and I_K1). Two mechanisms have been proposed for EADs. First, the prolonged repolarization of the APD allows recovery and reactivation of I_CaL⁴⁹ resulting in enhanced inward I_CaL that further depolarizes the membrane. This sets up a positive feedback loop, triggering a premature AP¹³. Secondly, at MPs negative to the threshold of I_CaL activation (but before full repolarization), spontaneous Ca²⁺ release from the SR can activate I_NCX resulting in membrane depolarization¹³ (Figure 2B). DADs arise as a result of abnormal spontaneous Ca²⁺ release from the SR into the cytoplasm. The resulting diastolic increase in [Ca²⁺]_i induces a transient increase in inward I_NCX, generating DADs⁴⁶ (Figure 2C). When the depolarization induced by the DADs is sufficiently large, I_Na can be activated resulting in triggered activity that also underlies arrhythmogenicity observed in Catecholaminergic Polymorphic Ventricular Tachycardia (CPVT)¹³.

Under normal conditions, the myocardium has a unidirectional propagation of depolarization wavefront that cancel each other out when the two meet from the opposite sides. Abnormal AP propagation can occur when a depolarization wavefront encounters a region that lags in recovery (i.e. longer effective refractory period (ERP)) from a previous depolarization, and thus the wavefront is blocked unidirectionally. However, the same region might be outside its ERP when the wave reaches to the other end of this region (e.g. via propagating around the periphery of the region). This will lead to retrograde excitation of this region and the wave front loops back and due to slow conduction might re-excite the adjacent regions that have recovered excitability. This circulating wave results in a circuitous and repetitive mode of depolarization, termed re-entry⁵⁰ (Figure 2D).

The pathogenesis of arrhythmias in the whole heart is a complex phenomenon^{51, 52} involving different layers of the myocardium, 3-D cellular coupling and different regions with unique structural and electrical features that can remodel with time. Although abnormal electrical behavior can be modeled in single cells, some events (e.g., re-entry) can be demonstrated only in multicellular preparations such as monolayers of hiPSC-CMs^{53, 54}. Caution must be exercised in extrapolating the findings from cellular studies to the whole heart.

Platforms for Functional Analysis of Human iPSC-CMs

iPSC technology has great potential to advance precision medicine. However, to be able to better utilize this human platform, appropriate methods for functional analysis of patient-specific iPSC-CMs are essential for understanding the pathogenic mechanisms of cardiac disorders as well as performing preclinical safety studies for novel drug candidates. A concise overview of the commonly used methodology for iPSC-CM functional analysis is described below.

Patch Clamp

High fidelity intracellular electrical recordings from CMs provide the most detailed and precise evaluation of the electrophysiological properties of CMs. Sharp electrodes penetrating the cell membrane can accurately record MP and thus characterize APs in detail. The patch clamp technique⁵⁵, provides the opportunity to measure both membrane potential (or AP) and underlying ionic currents. The technique involves gently pressing a fine-tip glass pipette, containing a solution with an ionic composition similar to the cytosol, against the cell membrane using a precision micromanipulator under guidance from an inverted microscope, followed by application of a mild suction to obtain a high resistance seal between the glass pipette and the cell membrane. Once a giga-ohm resistance seal is formed, the small patch of membrane plugging the pipette tip can be ruptured to provide electrical access to the inside of the cell, a recording configuration known as the whole cell patch clamp technique (Figure 3A) that allows the measurement of APs in current-clamp mode or membrane currents in voltage-clamp mode⁵⁶.

Figure 3. Platforms for Functional Analysis of Human iPSC-CMs.

Graphical illustration summarizing different methods currently used for hiPSC-CM functionality analyses. (A) Patch Clamp (B) MEA (C) Fluorescence Imaging (D) Impedance. Representative MEA^{64, 121} and Fluorescence Imaging^{70, 122} traces reprinted with permission of the publisher.

Many seminal studies of channelopathies have employed the patch clamp method for electrophysiological analysis of hiPSC-CMs^57–60. The biggest disadvantage of the patch clamp method remains its low throughput. The recordings are laborious, technically challenging and requires skilled personnel. Moreover, the invasive and terminal nature of the technique allows recordings to be performed only for a short period of time from single cells. However, it provides very high time (<1ms) and voltage- or current-amplitude resolution (unmatched by any other technique) while allowing control of the intracellular and extracellular ion composition^{7, 55}. Further, using specific blockers and recording protocols that can control the MP, activity of an individual ion channel (or whole cell current) can be separated from other ion currents with high fidelity.

Multielectrode Array (MEA)

MEA is a non-invasive and non-terminal method that allows long-term measurements of extracellular field potential signals from clusters and monolayers of hiPSC-CMs⁶¹. CMs are plated on the wells of the MEA plate containing electrodes on the bottom, and field potential is measured. The field potential duration (FPD) correlates roughly to the QT interval in the in vivo EKG and the APD in vitro. The field potential signal is composed of an initial rapid spike which corresponds to the AP upstroke, a slow wave/plateau phase, followed by a repolarizing wave that corresponds to repolarization^{56, 62, 63} (Figure 3B). Over the past few years MEAs have been increasingly used for predicting cardiotoxicity (manifested as prolonged FPDs and development of triggered activity) of hiPSC-CMs and can be scaled up for moderate throughput. The MEA platform is widely accepted and a relatively easy to understand assay^64–66. MEA allows generation of detailed activation maps and conduction velocity measurements; however, the conduction velocity in hiPSC-CMs monolayers is relatively slow (10–20 cm/s compared to 60 cm/s in adult human left ventricle) likely due to the immaturity of hiPSC-CMs⁶⁷. The disadvantages of the MEA method are its low resolution and inability to analyze ionic currents.

Fluorescence Imaging

Fluorescence microscopy methods that utilize Ca²⁺- or voltage-sensitive dyes provide a non-invasive method for measuring intracellular ion fluctuations and voltage changes. However, such indicator dyes can bind to intracellular molecules likely interfering with the normal functioning of the cell and most Ca²⁺ indicator dyes, including Fura-2 and Fluo-4, are cytotoxic. Despite novel developments in software tools used for data analysis, Ca²⁺ imaging remains a rather slow method, both for performing the assay itself and for data analysis⁶⁸. Voltage-sensitive dyes, such as di-4-ANEPPS, respond to changes in MP by changing their fluorescence emission. They are commonly used for studying whole hearts but have their limitations in single cell and monolayer imaging due to significant cytotoxicity⁶⁹. To overcome these challenges, genetically encoded fluorescent indicators (GECI for intracellular calcium ions and GEVI for voltage) were recently introduced to improve the efficiency of cellular phenotyping in hiPSC-CMs. These genetically encoded indicators consist of a sensing element (e.g. CaM for Ca²⁺), often fused to an autofluorescent protein (like circularly permuted enhanced GFP; cpEGFP) that alters its fluorescent intensity based on conformational changes in the sensing element⁷⁰ (Figure 3C).

In a recent study, Shinnawi et al.⁷⁰ demonstrated successful expression of genetically encoded voltage (ArcLight) and calcium (GCaMP5G) fluorescent indicators in hiPSC-CMs derived from patients with different IAS⁷⁰. The use of these reporters in comparison to traditional voltage- and Ca²⁺-sensitive dyes offers several advantages including significant photostability, superior signal-to-noise ratio and minimal cytotoxicity. Most fluorescent dyes (except Fura-2, which is ratiometric) and all genetic indicators allow monitoring of only relative changes in membrane voltage and intracellular Ca²⁺ levels and do not provide absolute values⁷⁰.

Impedance

Functional assays based on cellular impedance (an indirect measure of CM contractility) offer a non-invasive, label free and high throughput analysis method⁷¹. Cells are seeded onto a multi-well cell culture plate that contains gold film electrodes embedded onto the bottom of the wells. The mechanical displacement of the cells during CM contraction is measured as variations in impedance (equivalent to resistance in a DC circuit) that directly correlates with the beating frequency (Figure 3D). These measurements utilize weak alternating current (AC) between the electrodes with tissue culture medium as the electrolyte. The electronic hardware monitors the voltage across the electrodes, and the impedance is calculated using the AC version of Ohm’s law where impedance (Z) rather than resistance (R) is calculated as Z = V/I. Monitoring of the impedance signal does not alter cellular physiology either in an excitatory, suppressive, or cytotoxic fashion⁷¹. Clearly, this method has great potential in ascertaining the beat rate/arrhythmogenicity of hiPSC-CMs, however caution should be taken to avoid over-interpreting the data. It is important to realize here that the changes in the beating pattern of the cells are recognized only due to a measurement of a minute change in resistance and not due to their direct electrical activity⁷². Impedance measurements have been validated for mouse ESCs⁷³, hiPSC-CMs⁷³, rat neonatal primary CMs^{74, 75} and further cell lines, e.g. cardiac muscle cells (HL-1) and 3D cell clusters (hESC-CM™)⁷⁶. The recently introduced Cardio ECR platform (ACEA Biosciences) combines MEA with contractility measurements for a simultaneous and more comprehensive analysis of the excitation-contraction coupling⁷⁷.

Inherited Arrhythmia Syndromes (IAS) and hiPSC Modeling

Molecular and biophysical studies over the past two decades have linked genetic mutations in ion channels or ion channel regulatory proteins (genotype), to IAS (phenotype). These disorders include LQTS, SQTS, Brugada syndrome, (BrS), CPVT, and familial atrial fibrillation (AFib)⁴⁶. Table 2 summarizes the genotype-phenotype correlation of various channelopathies. As illustrated in Figure 4, mutations in different ion channels can lead to a very similar pathological phenotype and different mutations in the same ion channel gene can cause different channelopathies. For example, mutations in genes encoding K⁺, Na⁺, and Ca²⁺ channels can all lead to AP prolongation and manifest as LQTS, while different mutations in K⁺ can cause both LQTS and SQTS.

Table 2.

hiPSC-CM Models of Inherited Arrhythmia Syndromes

Syndro me	Causal gene	Effect on ion channel	Experimental approach	Cellular phenotype	Ref
LQTS1	KCNQ1 (R190Q)	Trafficking defect, altered channel activation and deactivation properties	Patch clamp	Reduced IKs, APD prolongation, increased susceptibility to catecholamine-induced tachyarrhythmia, attenuation of this phenotype with beta blockade	58
	KCNQ1 exon7 deletion	Possible Haploinsufficiency and trafficking defect of KCNQ1	Patch clamp	Reduced IKs, APD prolongation, reduced wild type KCNQ1 mRNA and protein, small molecule ML277 partially restored APD and reversed the decreased IKs.	141
	KCNQ1 (R594Q R190Q)	Trafficking defect	Patch clamp, MEA,	Prolonged APD, reduced IKs activation that was reversed by hERG allosteric modulator LUF7346	142
LQTS2	KCNH2 (A614V)	Trafficking defect	Patch clamp, MEA	Reduced IKr, APD prolongation, induction of EADs and triggered activity, potential improvement with pinacidil	86
	KCNH2 (G1681A)	Trafficking defect	Patch clamp	APD prolongation and EADs	57
	KCNH2 (R176W)	Trafficking defect	Patch clamp, MEA	Reduced IKr, APD prolongation, mild incidence of EADs	59
	KCNH2 (N006I)	Trafficking defect	Patch clamp, MEA	Prolonged APD, reduced IKr activation that was reversed by hERG allosteric modulator LUF7346	142
LQTS3	SCN5A (V1763M)	Gating defect- inactivation of sodium channel	Patch-clamp	Enhanced INaL, APD prolongation, phenotype reversed by Nav1.5 blockade by mexiletine	91
	SCN5A (F1473C)	Gating defect- deficiency in sodium channel inactivation	Patch clamp	Delayed repolarization, arrhythmia, prolonged QT interval with increase in pacing improving the phenotype	92
	SCN5A (V240M R535Q)	Gating defect- deficiency in sodium channel inactivation	Patch clamp	Insignificant increase in APD, delayed time to peak INa inactivation	90
	SCN5A (R1644H)	Gating defect- deficiency in sodium channel inactivation	Patch clamp, MEA	Prolonged APD, high EADs, and accelerated recovery from inactivation of Na+ currents. Rescue of abnormal phenotypeby mexiletine and ranolazine	143
LQTS7	KCNJ2 (R218W), (R67W), (R218Q)	Trafficking defect	MEA, calcium imaging	Strong arrhythmic events, higher incidence of irregular Ca2+ release. Flecainide, but not pilsicainide, suppressed irregular Ca2+ release and arrhythmic events	97
LQTS8	CACNA1C (G1216A)	Gating defect- loss of voltage- dependent channel inactivation	Patch clamp, calcium imaging	APD prolongation and DADs, abnormal calcium handling, irregular and slow contraction. Roscovitine rescued abnormal cellular phenotype	60
LQTS14	CALM1 (F142L)	Gating defect- impaired Ca2+- dependent inactivation	Patch clamp, MEA, calcium imaging	Prolonged APD, defective ICaL inactivation, altered rate-dependency and response to isoproterenol. Repolarization abnormalities reversed by verapamil	144
LQTS15	CALM2 (N98S)	Gating defect- suppression of L- type Ca2+ channel (LTCC) inactivation	Patch clamp	Lower beating rate, prolonged APD, and impaired ICaL inactivation, correction of the mutant allele rescued abnormal phenotype	145
	CALM2 (D130G)	Gating defect- disruption of Ca2+/CaM- Dependent inactivation of LTCC	Patch clamp, Fluorescence imaging	Prolonged APD, disrupted Ca2+ cycling properties, and diminished Ca2+/CaM- dependent inactivation of ICaL. Suppressing the mutant gene rescued abnormal phenotype	146
BrS	SCN5A (R620H), (R811H	Trafficking defect	Patch clamp, calcium imaging	Reduced INa and maximal upstroke velocity of AP, abnormal Ca2+ transients, and variation in beating interval	81
	PKP2 (c.2484C>T)	Trafficking defect	Patch clamp, MEA, calcium imaging	Reduced INa, deficit restored by transfection of WT gene	147
CPVT1	RYR2 (M4109R)		Patch clamp, calcium imaging, MEA	DADs, isoproterenol enhanced DADs and developed triggered activity. Flecainide and thapsigargin eliminated DADs. Ca2+ transient irregularities that worsened with adrenergic stimulation and Ca2+ overload and improved with β-blockers	106
	RYR2 (F2483I)		Patch clamp, calcium imaging	Longer duration of Ca2+ release from SR long after repolarization, arrhythmias and DADs with adrenergic stimulation	107
	RYR2 (S406L)		Patch clamp, Calcium imaging	Elevated diastolic Ca2+ concentrations, reduced SR Ca2+ content, and arrhythmias. Dantrolene ameliorated arrhythmia and restored normal Ca2+	108
	RYR2 (P2328S)		Patch clamp, Calcium imaging	Increased non-alternating variability of Ca2+ transients in response to isoproterenol and β-agonists decreased AP upslope velocity	108
	RYR2 (R420Q)		Calcium imaging	Less developed ultrastructure, isoproterenol either ineffective, caused arrhythmias, or markedly increased diastolic Ca2	148
	RYR2 (L3741P)		Calcium imaging, MEA	Altered intracellular Ca2+ homeostasis, β- adrenergic stimulation potentiated spontaneous Ca2+ waves and prolonged Ca2+ sparks. Flecainide ameliorated disease phenotype	149
	RYR2 (I4587V)		Patch clamp, Calcium imaging	Increased diastolic Ca2+ waves and DADs with pacing, while S107 suppressed the DADs	150
CPVT2	CASQ2 (D307H)		Patch clamp, Calcium imaging	β-adrenergic agonist caused DADs, oscillatory arrhythmic pre-potentials, and diastolic [Ca2+]i rise	108, 148, 151
	CASQ2 (D307H)		Patch clamp, Calcium imaging	Ca2+-transient irregularities, EADs, and reduced threshold for store overload- induced Ca2+-release, β-blockers prevented arrhythmia	152

APD, action potential duration; EAD, early after depolarization; DAD, delayed after depolarization; CALM1, Calmodulin 1; CALM2, Calmodulin 2; PKP2, Plakophilin-2.

Figure 4.

Diagram showing overlap between the genes associated with Brugada Syndrome (BrS), Short QT Syndrome (SQTS), Long QT syndrome (LQTS), Catecholaminergic Polymorphic Ventricular Tachycardia (CPVT) and Atrial Fibrillation (AFib).

Brugada syndrome

BrS is an inherited disorder characterized by ventricular arrhythmias and SCD occurring in otherwise healthy individuals at a relatively young age (< ~40 years), typically during conditions of high vagal tone (i.e., during sleep). On EKG analysis, a typical pattern is observed comprising ST-segment elevation in the right-precordial leads V1–V3^{78, 79}. In high-risk patients, with (recurrent) ventricular tachyarrhythmias, implantable cardioverter defibrillator (ICD) is considered, as pharmacological options are limited.

While several genetic mutations have been associated with BrS, all lead to a general imbalance of currents during the early phases of the AP tipping in favor of repolarization. The most common and best studied of these is the LOF mutation of the SCN5A gene (encoding Na_v1.5)⁵⁰ leading to reduced Na⁺ channel availability, either through decreased surface channel expression or through altered channel gating properties^{78, 79}. Although much of our current understanding regarding the molecular mechanisms of BrS was gleamed from heterologous expression studies, very recently, patient-specific iPSC-CMs harboring LOF genetic mutations in SCN5A were shown to recapitulate the pro-arrhythmic phenotype associated with either BrS^{80, 81} or overlapped LQT3/BrS⁸².

Long QT syndrome

LQTS is a potentially lethal cardiac disorder, characterized by QT interval prolongation and increased risk of ventricular arrhythmias, particularly torsade de pointes (TdP) that can lead to SCD. Increased arrhythmic risk associated with a prolonged QT interval can be congenital or acquired. Congenital LQTS is associated with mutations in ion channels and/or accessory proteins, whereas acquired QT prolongation is most commonly associated with block of the hERG K⁺ channel by some antiarrhythmic agents or as an unwanted side effect of many other medications. Additionally it can be associated with electrolyte imbalance (hypokalemia, hypocalcaemia, and hypomagnesaemia). The mutations associated with LQTS result in a favoring of depolarizing currents over repolarizing ones during phase 2 and 3 of the AP⁵⁰. At the single cell level, LQTS results from prolongation of the APD that can initiate EADs and TdP arrhythmia that sometimes degenerates into potentially lethal ventricular fibrillation.

LQTS1

LQTS1 is the most common type of LQT disorders that accounts for ~40% of total cases, often triggered by adrenergic drives, e.g., emotional stress, physical exertion, diving and swimming, etc. LQTS1 results from LOF mutations in the KCNQ1 gene that encodes the α-subunit of the channel conducting I_Ks. LQTS1 patients respond well to β-blockers while some are less responsive or even resistant to this medication⁸³.

Moretti et al⁵⁸ were the first to model LQTS1 using hiPSC-CMs derived from a patient carrying the R190Q KCNQ1 mutation. This missense mutation caused a trafficking defect that reduced KCNQ1 protein surface expression, leading to a 70% reduction in I_Ks. Consequently, the diseased cells recapitulated the hallmark LQTS phenotype characterized as APD prolongation and an increase in arrhythmia with β‑adrenergic agonists. Propranolol pre-treatment reduced isoproterenol-induced arrhythmias in spontaneously beating LQTS1 hiPSC-CMs⁵⁸. In a later study, Ma et al⁸⁴ derived iPSC-CMs from a LQTS1 patient harboring a different KCNQ1 mutation (exon 7 del) and demonstrated that patient-specific iPSC-CMs faithfully recapitulated the clinical phenotype, and responded to a small molecule ML277, an investigational agent under development to treat arrhythmia. Egashira et al⁸⁵ generated iPSC-CMs from a sporadic LQTS patient with 1893delC mutation in the KCNQ1 gene with clinical suspicion of LQTS1. MEA recordings revealed markedly prolonged FPD, a characteristic of LQTS phenotype. Further analysis identified a trafficking defect in KCNQ1 as the cause of the patient’s clinical phenotype.

Taken together, these findings clearly demonstrate the potential of patient-specific hiPSC-CMs as robust models for IAS despite their relatively immature electrophysiological profile.

LQTS2

LQTS2 accounts for 30% of all LQTS cases and is caused by LOF mutations in the KCNH2 (hERG1) gene that encodes the α-subunit of the channels that conduct I_Kr. An auditory stimulus is often a trigger for arrhythmic events in LQTS2 patients. The first reports of iPSC-based models of LQTS2 were reported by Itzhaki et al⁸⁶ and Matsa et al⁸⁷. It was demonstrated⁸⁶ that hiPSC-CMs derived from a patient with missense mutation A614V in KCNH2 exhibited significant prolongation of the AP, smaller I_Kr, EADs, and triggered arrhythmias. Importantly, the authors evaluated the potency of several pharmacological agents that worsened or mitigated the cellular LQTS2 phenotype. In a later study⁸⁷, hiPSC-CMs were derived from a patient with the A561T mutation in KCNH2. The authors⁸⁷ used allele-specific RNA interference (RNAi) to selectively degrade mutant mRNA while leaving wild-type mRNA intact. This approach normalized the APD, increased I_Kr, and decreased the frequency of EADs. In a recent study, Mehta et al.⁸⁸ provide solid evidence in favor of hiPSCs having promising drug repurposing applications with regard to the development of precision medicine treatment of LQTS. The authors demonstrated that Lumacaftor, a drug already in clinical use for cystic fibrosis due to its ability to foster trafficking of mutant CFTR channels to the cell surface, can also rescue the pathological phenotype of class 2 (trafficking deficient) LQTS2 hiPSC-CMs derived from patients not protected by β-blockers.

LQTS3

LQTS3 is caused by GOF mutations in the SCN5A (Nav1.5) gene that accounts for ~10% of all LQTS cases. Most of these mutations result in enhanced late Na⁺ current (I_NaL) during the plateau and repolarization phase of the AP due to defective inactivation of the Nav1.5 channel. The net increase in the depolarizing current ultimately causes APD prolongation and proarrhythmia. In contrast to the triggers of LQTS1 and LQTS2, which are often increased heart rate and stressful auditory stimuli, respectively, LQTS3 patients exhibit QT prolongation at lower heart rates and, consequently, have an increased risk for cardiac events during rest or sleep⁸⁹. LQTS3 models of patient-specific hiPSC-CMs have been shown to recapitulate key electrophysiological disease features such as increased I_NaL and prolonged APD at the single-cell level^{64, 90}. Ma et al⁹¹ modeled LQTS3 in hiPSC-CMs derived from a patient with the V1763M mutation and demonstrated enhanced I_NaL and prolonged APD. Interestingly, mexiletine, a Na⁺ channel blocker, preferentially inhibited I_NaL and rescued the abnormal electrophysiology in diseased cells.

In an elegant study, Terrenoire et al⁹² investigated the underlying cause of pathogenicity in a patient with variants in both SCN5A (F1473C) and KCNH2 (K897T). The authors demonstrated by patch clamp recordings that the observed clinical phenotype of severe QT prolongation and arrhythmias was primarily due to late Na⁺ channel defect and not influenced by the KCNH2 polymorphism. Interestingly, specific to the patient’s hiPSC-CMs, I_NaL displayed very pronounced rate-dependence such that enhancing the pacing rate markedly reduced I_NaL and increased its inhibition by mexiletine. This type of analysis clearly demonstrates the promise of patient-specific hiPSC-CM platform in improving management of IAS and for developing patient-specific clinical regimens.

LQTS7

LQTS7, also known as Andersen-Tawil syndrome (ATS), is a rare autosomal dominant genetic disorder typified by mild QT interval prolongation and appearance of significant U waves, frequent ventricular ectopy, bidirectional ventricular tachycardia (VT), and more rarely syncope, recurrent polymorphic VT, and cardiac arrest. Exercise is an important trigger of ventricular tachyarrhythmia and syncope in some patients with ATS, therefore patients are commonly treated with β-blockers. ATS type-1 (ATS1) has been linked to mutations in KCNJ2⁹³ and accounts for about 70% of all ATS cases. The genetic cause of the remaining 30% of cases (ATS2) remains unknown⁹⁴. KCNJ2 encodes for the Kir2.1 potassium channel that contributes to I_K1. Most KCNJ2 mutations in ATS1 cause LOF and dominant-negative suppression of Kir2.1 channel function, leading to reduced I_K1^{95, 96}. Recently Kuroda et al⁹⁷, modeled ATS using the hiPSC platform. A significant challenge in modeling LQTS7 is that since I_K1 density is already compromised in hiPSC-CMs, it is difficult to obtain a clear difference between healthy control and diseased hiPSC-CMs using voltage clamp techniques. But interestingly, in this study, the authors employed MEA and Ca²⁺ imaging platforms to uncover the pathogenesis of ATS in patient-derived hiPSC-CMs. hiPSC-CMs from three ATS patients carrying either one of the KCNJ2 mutations (R218W, R218Q or R67W) clearly exhibited arrhythmic events and abnormal Ca²⁺ release that was significantly suppressed by flecainide and a reverse-mode Na⁺/Ca²⁺ exchanger (NCX) inhibitor.

More recently, a new role of hiPSC-CMs has emerged in deciphering the pathogenicity of variants of uncertain significance (VUS). Recent advances in sequencing technologies have made it feasible to sequence the entire exome of patients suspected of having LQTS to identify causal genes. Such testing can identify hundreds of novel nonsynonymous coding variants in any given individual⁹⁸ and determining which one of them is pathogenic or benign is currently the biggest challenge in the management of IAS, especially when the variant is of uncertain significance. Using the iPSC platform, Gélinas et al⁹⁹ demonstrated the pathogenicity of a novel variant Gly52Val in the KCNJ2 gene identified by whole-exome sequencing in an LQTS8 patient. Patient-specific hiPSC-CMs displayed significant APD prolongation and incidence of arrhythmia. Immunohistochemistry and whole-cell current recordings in HEK cells heterologously expressing the WT or the mutant Kir2.1 channel further confirmed a surface trafficking defect in the mutant channels.

In future studies, the creation of isogenic hiPSC lines with single variant changes using clustered regularly interspaced short palindromic repeats (CRISPR) genome editing will allow direct comparison of phenotypes at a cellular level. A unique advantage of genome editing technology in disease modeling is the ability to study variants in an isogenic background to rule out phenotypic variability from epigenetic differences or unknown genetic modifiers¹⁰⁰. This novel approach of combining iPSCs and genome editing technology to decipher VUS pathogenicity in LQTS could aid in tailoring drug treatment and ICD therapy, and could potentially be applied to other cardiac disorders to accelerate progress toward realizing the promise of precision medicine in IAS.

We recently validated this approach in our lab¹⁰¹ by generating hiPSC-CMs from a VUS carrier of LQTS2 (T983I KCNH2) and demonstrated their aberrant electrophysiological phenotype (reduced I_Kr and prolonged APD). We further delineated the VUS pathogenicity using CRISPR genome-editing to selectively correct the variant which normalized the abnormality and introduced the homozygous KCNH2 variant in an otherwise healthy individual which exacerbated the disease phenotype. To our knowledge, this is the first report of combining human iPSC and genome editing platforms to decipher pathogenicity of uncertain variant.

LQTS8

LQTS8 (Timothy Syndrome), results from rare single amino acid substitutions in CACNA1C gene that encodes the LTCC. It is the most severe form of LQTS, characterized by multi-organ dysfunction and high mortality. Besides marked QT interval prolongation and severe ventricular arrhythmia, LQTS8 patients can present with congenital heart defects, AV block, syndactyly, autism, malignant hypoglycemia, and an abnormal immune system^{102, 103}. The G406R and G402S mutations in CACNA1C reduce the rate of channel inactivation and thereby prolong the plateau phase of the cardiac AP^{102, 103}. In 2011, Yazawa and colleagues⁶⁰ modeled LQTS8 using patient-specific hiPSC-CMs. The diseased cells recapitulated electrical abnormalities such as prolonged APD, DADs, and altered Ca²⁺ transients. Roscovitine accelerated I_CaL inactivation, reduced APD, restored the irregular Ca²⁺ transient, and decreased the frequency of abnormal depolarizations in LQTS8 hiPSC-CMs⁶⁰. Although roscovitine rescued the electrophysiological abnormality in LQTS8 hiPSC-CMs, the exact mechanism by which it restored cardiac function was not clear. In a separate study¹⁰⁴, it was shown that roscovitine exhibits its therapeutic effects in part by inhibiting CDK5, a key mediator involved in the regulation of Ca_V1.2 channels in CMs. These studies provide mechanistic insights into the regulation of Ca_V1.2 and the development of future therapeutics for Timothy syndrome patients¹⁰⁴.

Catecholaminergic Polymorphic Ventricular Tachycardia

CPVT is an inherited disease characterized by abnormal intracellular Ca²⁺ handling and signaling in CMs. The underlying mechanism of CPVT involves abnormal Ca²⁺ release from the SR, which is often due either to GOF mutations in RYR2 (CPVT1)¹⁰⁵ or LOF mutations in calsequestrin (CPVT2), a SR calcium buffering protein. Such abnormal Ca²⁺ release may cause DADs, similar to digitalis toxicity, leading to triggered activity and ventricular tachyarrhythmias. Clinically, CPVT is precipitated by increased levels of catecholamines (e.g. epinephrine) under conditions of emotional stress and physical exertion. β-blockers are the drugs of choice for CPVT patients, but they often fail to prevent fatal arrhythmias.

To date, several groups have successfully modeled CPVT using the hiPSC platform^106–108. CPVT1 patient-specific hiPSC-CMs carrying the F2483I mutation in the RYR2¹⁰⁷ recapitulated DADs after catecholaminergic stimulation while Ca²⁺ imaging studies revealed higher amplitudes and longer durations of spontaneous Ca²⁺ release at basal state compared to control hiPSC-CMs. In addition, the CICR events continued after repolarization that were abolished by forskolin. In yet another CPVT1 model of hiPSC-CMs harboring the RYR2-M4109R mutation¹⁰⁶, hiPSC-CMs showed increased DAD incidence and Ca²⁺ transient irregularities as compared to control hiPSC-CMs, which were further exacerbated by β-adrenergic stimulation¹⁰⁶. Flecainide eliminated DADs in this CPVT1 hiPSC-CM model¹⁰⁶. These data indicate that hiPSC-CMs can recapitulate the cardinal features of CPVT, making them useful models to investigate IAS mechanisms and to expedite personalized therapy.

Short QT Syndrome

SQTS is a rare inheritable, autosomal dominant disorder characterized by accelerated cardiac repolarization, abnormally short QT intervals and an increased propensity to develop atrial and ventricular tachyarrhythmia. Cardiac arrest seems to be the most frequent symptom (up to 40%)¹⁰⁹. SQTS has been associated with mutations in six genes (KCNQ1, KCNH2, KCNJ2, CACNA1C, CACNB2, and CACNA2D1) with GOF mutations in KCNH2 (SQTS1) being the most prevalent subtype^{110, 111}. Since it is a repolarization abnormality, the causative ion channel genes overlap with those of BrS and LQTS (Figure 4). Additionally, Afib or atrial flutter are also observed in SQTS patients (particularly SQT1)¹¹². Besides ICD, several pharmacological approaches have been applied to treat SQTS, all aiming to restore the duration of ventricular repolarization to physiological values. Gaita et al.¹¹³ tested several antiarrhythmic drugs (flecainide, sotalol, ibutilide, and hydroquinidine) on two SQTS families carrying the N588K mutation in KCNH2¹¹³. Hydroquinidine was found to be the only agent that effectively prolonged the QT interval and abolished the inducibility of ventricular arrhythmias during programmed electrical stimulation.

To date, there has been only one report of modeling SQTS using the hiPSC platform. El-Battrawy and colleagues¹¹⁴ demonstrated that hiPSC-CMs derived from a patient carrying a N588K mutation in the KCNH2 recapitulated single-cell phenotype of SQTS. In contrast to the single cell phenotype of LQTS2, SQTS patient-specific hiPSC-CMs displayed markedly enhanced I_Kr density and shortened APD compared to healthy control. These cells further demonstrated abnormal Ca²⁺ transients and arrhythmic activities. Carbachol increased the arrhythmic events in SQTS while quinidine, but not sotalol or metoprolol, prolonged the APD and abolished arrhythmic activity induced by carbachol. These studies provide novel opportunities to further elucidate the pathophysiological mechanisms of SQTS and development of effective treatment strategies.

Atrial Fibrillation

AFib represents the most common type of supraventricular arrhythmia especially in the elderly population, oftentimes associated with underlying ischemic or structural heart disease, hypertension, diabetes, autonomic imbalance, as well as electrical remodeling of the atria. An underlying genetic component (Figure 4) has been associated with AFib only in a small number of cases of ‘lone AFib’ (AFib in the absence of predisposing factors), however inheritance is complex and mostly ‘non-Mendelian’¹¹⁵. The primary arrhythmia mechanisms contributing to AFib are focal ectopic firing and reentrant activity (Figure 2D). Ablation techniques have improved outcomes over the last two decades, but are associated with a risk of complications, and are only successful in a limited number of patients with permanent AFib. Pharmacologic therapy remains as the first-line treatment, and an important adjuvant to ablation therapy, however it is limited by a lack of efficacy and serious side effects such as ventricular proarrhythmia.

One of the major barriers to an improved mechanistic understanding of AFib, and thus drug development, has been a lack of appropriate models⁵³. Until recently, hiPSC-based models of AFib were lacking because most hiPSC-CM differentiation protocols tend to generate a predominant ventricular-like phenotype. This methodological hurdle has been recently solved by manipulation of retinoic acid (RA) signaling, a physiological pathway that also drives atrial specification in vivo^{30, 116}. Recently, Laksman et al.⁵³ used human ESCs to generate an atrial-specific tissue model of AFib for pharmacologic testing. These atrial-like CM sheets showed uniform AP propagation, and rapid reentrant rotor patterns as seen in AFib could be induced. It was demonstrated⁵³ that anti-arrhythmic drugs such as flecainide and dofetilide modulated reentrant arrhythmogenic rotor activation patterns in a manner that helps explain their efficacy in treating and preventing AFib.

Limitations of hiPSC-CMs

hiPSC-CMs have emerged as a promising experimental tool for translational cardiovascular research and drug development. However, their applicability as a human adult CM model is still hindered by several limitations. We have already mentioned on several occasions in this review, one of the major limitations of hiPSC technology — the immaturity of hiPSC-CMs. Compared to adult CMs, immature hiPSC-CMs have several differences, including a lack of T-tubules, abnormal morphology, altered gene expression, and reduced sarcomere organization^{117, 118}. Genetic modifications during reprogramming, incomplete reprogramming resulting in residual epigenetic memory, induced alterations during passaging in culture, and differentiation-dependent variability and heterogeneity of iPSC-CMs are additional problems¹¹⁹. With regard to the challenges specific to disease modeling, it is not yet possible to model all human diseases using the hiPSC platform, for example diseases based on complex interactions of multiple genetic and environmental factors or on long incubation time, such as atherosclerosis and congenital heart diseases¹²⁰. Moreover, differences in the individual genetic background present a particularly vexing problem to the ‘disease-in-a-dish’ approach due to the uncontrolled impact of genetic modifier loci. Thanks to the recent advances in genome editing tools such as zinc finger nucleases (ZFN), transcription activator–like effector nuclease (TALEN), and CRISPR genome editing, this limitation can now be addressed using genetically manipulated isogenic controls¹¹⁹. The robust combination of hiPSCs and genome editing will provide a unique opportunity to systematically and faithfully recapitulate human diseases in vitro, revealing new insights into the pathophysiology of monogenic and complex disorders¹¹⁹. However, in order to use hiPSCs for precision medicine, the development of improved differentiation protocols and 3D bioengineering approaches to create mature, adult-like CMs will be of paramount importance.

Conclusions and Future Perspectives

IAS represents a group of potentially lethal disorders that remain a diagnostic and therapeutic challenge due to complexity of the disorder and variable disease penetrance. Studies based on in vitro expression of mutations and transgenic animal models have been instrumental in offering the practitioner tools for deciphering these rare disorders but their potential remains limited. At present, there are still many unanswered questions such as, why mutations in the same gene can cause different types of arrhythmias, or why patients carrying the same mutation may exhibit completely different phenotypes ranging from being completely asymptomatic throughout life to experiencing SCD at an early age. In this regard, the role of the genetic modifiers that modulate the outcome in a mutation-and patient-specific manner remains unknown. Recent advances in the derivation of hiPSCs directly from diseased patients now provide a great opportunity to unravel these diverse mechanisms unique to a mutation/patient. Combined with the rapidly advancing genome editing approaches, iPSC technology has already emerged as a powerful platform for studying genotype–phenotype association and for predicting patient response to individualized therapy⁵⁰. Cleary, iPSCs hold enormous potential in advancing the practice of precision medicine. In the future, a more thorough knowledge of the disease processes of these syndromes may enable us to use targeted therapy alone and minimize the need for ICDs.

Nonstandard Abbreviations and Acronyms

AFib

Atrial fibrillation

AP

Action potential

APD

Action potential duration

AVCM

Adult ventricular cardiomyocytes

BrS

Brugada syndrome

CALM

Calmodulin

CICR

Calcium-induced calcium release

CM

Cardiomyocytes

CPVT

Catecholaminergic polymorphic ventricular tachycardia

DAD

Delayed afterdepolarization

DPP6

Dipeptidyl Peptidase Like 6

EAD

Early afterdepolarization

EKG

Electrocardiogram

GOF

Gain-of-function

HCN

Hyperpolarization-activated cyclic nucleotide-gated

hERG

Human ether-a-go-go related gene

hESC

Human embryonic stem cells

ICD

Implantable cardioverter defibrillator

iPSC

Induced pluripotent stem cells

iPSC-CM

Induced pluripotent stem cell-derived cardiomyocytes

KCHIP2

Kv channel-interacting protein 2

Kir

Inward rectifier K⁺ channels

Kv

Voltage-gated K⁺ channels

LOF

Loss-of-function

LQTS

Long QT syndrome

LTCC

L-type Ca²⁺ channels

MEA

Multielectrode array

MiRP1

MinK-related peptide 1

MP

Membrane potential

NCX

Na⁺/Ca²⁺- exchanger type-1

PKP2

Plakophilin-2

RA

Retinoic acid

RMP

Resting membrane potential

RYR2

Type 2 ryanodine receptors

SAN

Sinoatrial node

SCD

Sudden cardiac death

SERCA2a

Sarco/endoplasmic reticulum Ca²⁺-ATPase type-2a

SQTS

Short QT syndrome

TM

Transmembrane