Murine Electrophysiological Models of Cardiac Arrhythmogenesis

Abstract

Cardiac arrhythmias can follow disruption of the normal cellular electrophysiological processes underlying excitable activity and their tissue propagation as coherent wavefronts from the primary sinoatrial node pacemaker, through the atria, conducting structures and ventricular myocardium. These physiological events are driven by interacting, voltage-dependent, processes of activation, inactivation, and recovery in the ion channels present in cardiomyocyte membranes. Generation and conduction of these events are further modulated by intracellular Ca²⁺ homeostasis, and metabolic and structural change. This review describes experimental studies on murine models for known clinical arrhythmic conditions in which these mechanisms were modified by genetic, physiological, or pharmacological manipulation. These exemplars yielded molecular, physiological, and structural phenotypes often directly translatable to their corresponding clinical conditions, which could be investigated at the molecular, cellular, tissue, organ, and whole animal levels. Arrhythmogenesis could be explored during normal pacing activity, regular stimulation, following imposed extra-stimuli, or during progressively incremented steady pacing frequencies. Arrhythmic substrate was identified with temporal and spatial functional heterogeneities predisposing to reentrant excitation phenomena. These could arise from abnormalities in cardiac pacing function, tissue electrical connectivity, and cellular excitation and recovery. Triggering events during or following recovery from action potential excitation could thereby lead to sustained arrhythmia. These surface membrane processes were modified by alterations in cellular Ca²⁺ homeostasis and energetics, as well as cellular and tissue structural change. Study of murine systems thus offers major insights into both our understanding of normal cardiac activity and its propagation, and their relationship to mechanisms generating clinical arrhythmias.

I. INTRODUCTION

A. Scope of Review

Cardiac arrhythmias result from disruption of the orderly physiological sequence of electrical excitation processes that initiates coordinated and effective cardiac contraction. Of the wide clinical variety of arrhythmic variants, ventricular arrhythmias, particularly ventricular fibrillation (VF), often preceded by ventricular tachycardia (VT), potentially lead to sudden cardiac death (SCD), defined as unexpected death from cardiac causes occurring <1 h after onset of symptoms (971, 1152). This major worldwide source of morbidity and mortality causes >300,000 and ∼70,000 deaths/year in the United States (USA) (535) and United Kingdom (UK) (215), respectively. Cardiac causes likely account for 56.4% of nontraumatic, sudden death in autopsies of patients aged 5–35 years. Among these, arrhythmic causes likely account for ∼30% of cases. Although most of the latter cases result from ischemic heart disease (87), autopsy fails to reveal a cause in ∼4% of SCD cases and 14% of all resuscitation attempts performed on patients aged <40 years (206, 738–740, 1149). Furthermore, such deaths often occur in the absence of structural abnormalities. This suggests the possibility of underlying channelopathy (116, 1122). Of deaths in infants <1 yr, 60–80% are autopsy-negative. They are accordingly defined as sudden infant death syndrome (SIDS) (47), and 10–20% of these cases may result from channelopathy (575). Finally, arrhythmogenesis as a possible adverse effect of pharmacotherapeutic agents constitutes an important clinical problem with significant implications for pharmaceutical drug discovery (564).

Atrial arrhythmias are similarly becoming increasingly clinically and demographically important. Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, with an overall prevalence of ∼1–2% of the general population (30, 683, 1093). AF is often associated with advancing age. It affects 4.7 and 9% of individuals of age >65 years and between 80 and 89 years, respectively (285, 1212). It predisposes to further, major, cardiac and cerebrovascular morbidity and mortality (1093). Thus it increases risks of stroke fivefold (1268).

Sinus node disorder (SND) causes sinus bradycardia, sinus pause/arrest, chronotropic incompetence, and sinoatrial node (SAN) exit block (271). Its incidence increases exponentially with age to 1 in ∼600 cardiac patients aged >65 years. It is responsible for ∼50% of the million permanent pacemaker implants per year worldwide often in otherwise healthy individuals (272, 731).

Cardiac arrhythmogenesis poses significant clinical challenges in terms of both risk stratification and management (986). The latter are limited by our current inadequate understanding of the physiological mechanisms underlying initiation, maintenance, or propagation of cardiac arrhythmias, whether in the atria, ventricles, or the conducting tissues within or between them. Much valuable data have derived from human studies. However, much of this is inevitably observational. Physiological animal models of arrhythmic disease, whether involving pharmacological interventions or targeted genetic changes, permit more analytical studies of mechanisms and their consequences. Of these systems, transgenic mouse models are a relatively recent addition to other animal, rabbit, and canine systems. They offer a means of genetic and physiological manipulation that can be effectively directed at particular molecular targets strategic to cardiac electrophysiological function.

This review describes studies using some of these approaches, relating arrhythmic phenomena to cardiac electrophysiological properties. The latter in turn bear upon the generation of atrial or ventricular action potentials (APs), any abnormal, triggered activity accompanying such events, and associated metabolic and structural changes. The analysis is made in the light of the features of different exemplars modifying the activity of specific ion channels, cellular properties, or tissue or chamber structure. This involves first summarizing the roles of the major ion channels that underlie cardiac electrophysiological excitation, and the consequent excitation-contraction coupling processes involving the release and reuptake of sarcoplasmic reticular (SR) store Ca²⁺. Alterations in these processes are next related to the properties of genetic murine exemplars of ion channel dysfunction. These include models for altered gap junction function compromising the spread of excitation, losses, and gains of function in the Na⁺, K⁺, and Ca²⁺ channels, and their subunits and associated proteins, affecting cell excitability, and ryanodine receptor (RyR2) modifications affecting cellular Ca²⁺ homeostasis. Finally, further upstream, metabolic, energetic, and structural changes are considered in relation to human arrhythmic conditions.

B. Normal Physiology of Cardiac Excitation

1. Ion current contributions to cardiac action potentials

Effective cardiac function depends on repetitive cycles of AP excitation followed by recovery and the propagation of these events through the myocardial or conducting tissue as coherent electrical waves. This conduction takes place successively through the SAN, atria, atrioventricular bundles, Purkinje conducting tissue, and ventricular endocardial and epicardial myocardium. Repetitive atrial and ventricular excitation cycles normally depend on SAN automaticity driven by pacemaker cells (see sect. IIIA). The typical human ventricular AP waveform begins with a rapid (∼400 V/s) initial, phase 0, depolarization. This is driven by a rapidly increasing voltage-gated Na⁺ current (I_Na) (∼400 μA/μF). This drives the myocardial membrane potential from its normal negative (approximately −90 mV) resting potential to a positive (+40 to +60 mV) voltage close to the Na⁺ Nernst potential (Figure 1A and Table 1). This is followed by a, phase 1, initial rapid repolarization from this positive value that results from both inactivation of I_Na and the activation of fast and slow transient outward (I_to), K⁺, I_to,f, and I_to,s currents (Figure 1, B and C) (reviews in Refs. 71, 827, 863, 1020, 1255).

Figure 1.

Basic features of cardiac electrophysiological excitation. Inward (A) and outward (B) ionic current contributions attributable to surface membrane ion channels to human (C) and mouse (D) ventricular action potential (AP) waveforms.

Table 1.

Human and murine ventricular and atrial expression of cardiac ionic currents mediating excitable activity

			Human		Mouse
Current/Symbol	Protein	Gene	Ventricle	Atrium	Ventricle	Atrium	Action Potential Contribution
Voltage-gated inward currents
Fast Na+ current, INa	Nav1.5	SCN5A	+++	+++	+++	+++	[0]
L-type Ca2+ current, ICaL (dihydropyridine receptor: DHPR)	Cav1.2	CACNA1C	+++	++	++	++	[2]
Voltage-gated outward currents
Fast transient outward K+ current, Ito,f	Kv4.2	KCND2	++	+++	++	+++	[1]
	Kv4.3	KCND3
Slow transient outward K+ current, Ito,s,	Kv1.4	KCNA4	++	+++	++	+++	[1]
Delayed rectifier K+ current, IKr	Kv11.1	KCNH2 (HERG)	+++	+	+	+	[3]
Delayed rectifier K+ current, IKs	Kv7.1	KCNQ1	+++	+	+	+	[3]
4-Aminopyridine-sensitive, rapidly activating, slowly inactivating K+ current, IKslow1	Kv1.5	KCNA5	−	−	++	+
4-Aminopyridine-insensitive, rapidly activating, slowly inactivating K+ current, IKslow2	Kv2.1	KCNB1	−	−	++	+
Sustained 4-aminopyridine-sensitive delayed rectifier K+ current, Iss	Kv1.5	KCNA5	−	−	++	+
Atrial-specific 4-aminopyridine-sensitive ultrarapid delayed rectifier K+ current, IKur	Kv1.5	KCNA5	−	++	−	++
Inward rectifiers
Inwardly rectifying current, IK1	Kir2.1	KCNJ2	+++	++	+++	++	[3], [4]
	Kir2.2	KCNJ12
	Kir2.3	KCNJ4
Acetylcholine-activated, K+ current, IKACh	Kir3.1	KCNJ3	−	+++	−	+++
	Kir3.4,	KCNJ5
ATP-sensitive potassium channel, IKATP	Kir6.2	KCNJ11	++	++	++	++
Leak currents
Two-pore domain K+ leak current, IK2p	K2p3.1	KCNK3	+++	++	++	+++
Ca2+-activated K+ currents, IKCa	KCa2.x-	KCNNx	−	+++	−	+++
Exchange currents
Transient, inward, Na+-Ca2+ exchange current, Iti	NCX	SLC8A1	++	++	++	++

Action potential contributions (human ventricle): [0], phase 0 rapid depolarization; [1], phase 1 initial rapid repolarization; [2], phase 2 plateau; [3], phase 3 repolarization; [4], phase 4 electrical diastole.

This partial recovery is then followed by a phase 2, plateau, phase brought about by the activation of Ca²⁺ current (I_CaL) through voltage-gated L-type Ca²⁺ channels often termed dihydropyridine receptors (DHPRs), reflecting their pharmacological sensitivities (107, 131, 556, 1319).

The action potential is then terminated by a phase 3 repolarization. This returns the membrane to its resting potential. This is driven by outward currents mediated by a species-dependent variety of K⁺ channels. In human ventricles, these include the delayed rectifier I_Kr (1169) and I_Ks (1097), inwardly rectifying I_K1 (693), and two-pore domain K⁺ leak currents (I_K2p) (1020).

This culminates in full repolarization to electrical diastole (phase 4), in which the inward rectifying current I_K1 plays a major part in maintaining a fully polarized (approximately −90 mV) resting potential. With repolarization, Nav1.5 channels recover their capacity for reexcitation. This recovery takes place over both absolute and relative relative (effective) refractory periods (ERPs). Following this, Nav1.5 channels become capable of beginning the next excitation cycle.

Atrial APs begin from more depolarized resting potentials. The latter mainly reflects their smaller I_K1. They show triangular waveforms with a more prominent phase 1 recovery reflecting a larger I_to. Atrial myocytes also specifically express ultrarapid delayed rectifier K⁺ (I_Kur) (1256), acetylcholine-activated K⁺ currents (I_KACh), and Ca²⁺-activated K⁺ currents (I_KCa) (1020, 1257). They show a less prominent phase 2 plateau phases than observed in ventricular APs. This is the consequence of smaller I_Kr, I_Ks, and I_K1 currents but a more prolonged phase 3 repolarization (822).

Finally, ATP-sensitive K⁺ current (I_KATP) occurs throughout the heart but generally accounts for relatively little current due to its inhibition by intracellular ATP (320, 324). However, I_KATP may be activated under conditions of energetic stress (324, 419, 486, 1024).

2. Excitation-contraction coupling

Excitation-contraction coupling in ventricular myocytes is initiated by the surface membrane depolarization described above and the transmission of this electrical change into the transverse tubules. This results in an opening of voltage-gated L-type Ca²⁺ channels within the membranes of the transverse tubules, which then mediate the inward Ca²⁺ currents (I_CaL) responsible for the phase 2 plateau phase of the AP_. The accompanying influx of extracellular Ca²⁺ produces a local elevation of cytosolic Ca²⁺ concentration ([Ca²⁺]_i) in regions of the sarcolemmal-SR junctions (107). This triggers an opening of cardiac SR RyR2 Ca²⁺ channels by a process of Ca²⁺-induced Ca²⁺ release (CICR) (298). The resulting release of intracellularly stored SR Ca²⁺ produces the elevation of [Ca²⁺]_i that drives the Ca²⁺-troponin binding which triggers mechanical activation. The cytosolic Ca²⁺-mediated regulation of cardiac, RyR2-Ca²⁺ release channel activity is facilitated by SR luminal Ca²⁺ (395, 622, 1069, 1285) and cytosolic ATP. It is inhibited by cytosolic Mg²⁺ (see sect. VIIA) (1313). RyR2 is also sensitive to thiol-oxidation and reactive oxygen species (ROS) which may disrupt its interdomain stability (see sect. VIIIA) (785). The CICR process in cardiac muscle contrasts with the control of RyR1 opening by the more rapid direct allosteric control by voltage-dependent configurational changes in the DHPR in skeletal muscle. It explains the contrasting dependence and relative independence of these respective excitation-contraction coupling processes upon extracellular [Ca²⁺] (458, 462). It also contrasts with RyR2 function in inexcitable osteoclasts in which the RyR2 assumes a surface rather than an endoplasmic membrane site (460, 1314–1316).

The smaller atrial myocytes possess less prominent transverse tubular networks, particularly in hearts of small mammals such as the mouse. Their SR membranes are differentiated into 1) corbular regions that form junctional elements close to the cell periphery. These regions are flanked by clusters of surface membrane L-type Ca²⁺ channels and SR membrane RyR2-Ca²⁺ release channels. 2) Noncorbular SR in the cell interior also contains membrane regions that express RyR2 but do not show this proximity to the cell surface. Depolarization of the cell surface membrane triggers entry of extracellular Ca²⁺ through the activation of I_CaL, resulting in a local increase in [Ca²⁺]_i. This induces RyR2-mediated Ca²⁺-induced Ca²⁺ release at corbular SR. The resulting local elevation of [Ca²⁺]_i then initiates Ca²⁺-induced Ca²⁺ release and its propagation through adjacent, deeper regions of noncorbular SR. This results in an inward, centripetal, propagated Ca²⁺-induced Ca²⁺ release wave by noncorbular cytoplasmic SR within the cell interior (128, 419, 521, 710, 1335, 1338).

3. Recovery from excitation

Contractile relaxation follows reduction in the elevated [Ca²⁺]_i back to baseline levels. This permits Ca²⁺-troponin dissociation. Cytosolic [Ca²⁺] is reduced by Ca²⁺ transport activity by SR Ca²⁺-ATPase (SERCA2), sarcolemmal Na⁺/Ca²⁺ exchange through the Na⁺/Ca²⁺ exchanger (NCX), and slow systems represented by sarcolemmal Ca²⁺-ATPase (PMCA) and mitochondrial Ca²⁺ uniport. The relative contributions of these different Ca²⁺ transport mechanisms to the restoration of resting levels of [Ca²⁺]_i varies with species (110). In rabbit ventricular myocytes, SERCA2a activity eliminates 70%, NCX removes 28%, and the slow systems ∼1% of released Ca²⁺ with similar levels for ferret, dog, cat, guinea pig, and human ventricle (450). Corresponding contributions in rat and mouse ventricle are 92, 7, and 1%, respectively (78, 140, 653).

Of these processes, the NCX exerts potentially important effects on membrane potential through its electrogenic effects. These can contribute to arrhythmic tendency. Each cycle of NCX activity translocates 1Ca²⁺ from the intracellular to the extracellular space in return for a transfer of 3Na⁺ in the opposite direction. This results in a net current (I_NCX) whose reversal potential depends on both the Na⁺ and Ca²⁺ Nernst potentials, E_Na and E_Ca, giving E_NCX = 3E_Na − 2E_Ca. E_NCX therefore falls within the range of voltages traversed by normal physiological activity. Consequently, whether I_NCX takes an inward, depolarizing or outward, hyperpolarizing direction varies with the changes in both [Ca²⁺]_i and membrane potential that take place through the cardiac cycle. Thus I_NCX takes an inward direction and exerts a depolarizing effect on membrane potential under conditions when [Ca²⁺]_i is elevated and thereby drives Ca²⁺ efflux. Conversely, I_NCX takes an outward direction exerting a hyperpolarizing effect on membrane potential when [Ca²⁺]_i is relatively low, thereby driving Ca²⁺ influx. The pattern of I_NCX activity also varies with species-related differences in AP waveform. The long plateau phase in rabbit ventricular APs results in a sustained I_NCX-mediated Ca²⁺ influx through the AP plateau phase (165). Following repolarization, a large outward electrochemical gradient then drives Ca²⁺ efflux. In contrast, the short AP in rat ventricles results in a rapid initial I_NCX-mediated Ca²⁺ influx, but this is followed by a more marked Ca²⁺ efflux. This results in a more rapid removal of the added Ca²⁺ load arising from electrical excitation and a lower subsequent diastolic NCX activity (1042). Either of these effects potentially modify the action potential duration (APD) (Figure 1,A and B) (265, 1025).

Of the overall energetic cost of contractile and excitable cardiac activity, ∼60–70% of this cellular ATP is consumed by cardiac muscle contraction. The remainder maintains Ca²⁺ homeostasis and the transmembrane ion gradients.

C. Propagation of Excitation

1. Electrical current flow between cardiac cells

The subsequent propagation of cellular-level events through conducting atrial or ventricular tissue first involves a local spread of electrotonic excitation currents. These are driven typically in the more rapidly conducting cardiac tissues by I_Na (137, 496, 1043). Cable theory classically describes this current flow along a constant intracellular, axial resistance r_a from one depolarized myocyte to its quiescent neighbor (492, 577). In cardiac tissue, r_a reflects the electrical resistances formed, respectively, by the cytosol and intercellular gap junctions connecting successive adjacent cells. An axial current, i_a, having traversed the resistance r_a, then discharges the membrane capacitance, c_m, of neighboring quiescent cells. Where this resulting depolarization exceeds the activation threshold of its voltage-sensitive Na⁺ channels, this results in a regenerative production of further transmembrane depolarizing currents in that cell. This continues the AP propagation process. Thus, in cardiac cells, resistance to conduction through intercellular gap junction channels, the membrane capacitance, and the magnitude of I_Na are critical to AP propagation (958).

2. Determinants of conduction velocity

The velocity θ of AP conduction along a simple one-dimensional uniform cylindrical fiber of excitable tissue can be approximated by a nonlinear cable equation (550, 902). This has been applied to biological membrane with circuit elements that each incorporate a capacitance of unit fiber length, c_m (typically expressed in units of μF/cm) in parallel with a linear membrane resistance of unit fiber length r_m (kΩ·cm). The membrane additionally contains further, nonlinear, time- and voltage-dependent, membrane conductances representing the properties of its contained individual ion channels. These together generate the time (t)-dependent total membrane ionic current i_i (A/cm) in unit fiber length, x (cm). Successive circuit elements are connected by terms arising from cytoplasmic resistances and the gap junction resistances between cells (see sect. IV) (262, 1181). Any one of these factors could be modified by changes including tissue fibrosis or inflammatory processes (see sects. V, E and F, and IXA).

The membrane potential V at any given membrane site then depends on the charging of its unit length by currents traversing the membrane, i_i, as well as the axial current flow, i_a, coming from neighboring regions along the length x of the membrane area in question through the equation

$$\frac{1}{r_{\text{a}}}\left(\frac{{\text{d}}^{2}V}{\text{d}{x}^2}\right)=c_{\text{m}}\left(\frac{\text{d}V}{\text{d}t}\right)+i_i$$ (1)

This equation reduces at constant conduction velocity, θ = dx/dt to

$$\frac{1}{{\theta }^2{r}_{\text{a}}}\left(\frac{{\text{d}}^{2}V}{\text{d}{t}^2}\right)=c_{\text{m}}\left(\frac{\text{d}V}{\text{d}t}\right)+i_i$$ (2)

This simplified interpretation identifies r_a, c_m, and i_i as key determinants of θ, although interdependences between some of the terms involved preclude analytic solution (471). Explicit prediction instead requires numerical solution of a stiff equation involving iterative estimates of θ (492). This is particularly given potential further contributions from other membrane components including transverse tubular membrane resistances and capacitances, tubular luminal geometry and its resistance, and nonlinear capacitances (9, 457, 1047). A full cable analysis of AP propagation would also be required to incorporate the three-dimensional nature of cardiac geometry.

Nevertheless, a number of useful, simple relationships between θ and its determinants arise from computational studies of one-dimensional electric current flow in skeletal muscle fibers whose APs upstrokes are similarly dominated by fast I_Na (330, 567, 880). These confirm the importance of i_Na during the AP upstroke, and that the maximum Na⁺ current

$$i_{\text{Na}\left(\max \right)}\alpha \hspace{0.17em}\log \left[P_{\text{Na}\left(\max \right)}\right]\left(R^2=0.9965\right)$$ (3)

P_Na(max) is the maximum permeability produced by the fast Na⁺ channels. It is accordingly dependent on Na⁺ channel density. Of the key determinants of θ, r_a does not influence the AP waveform. It thus does not influence either its rate of upstroke voltage change, as given by the first derivative (dV/dt), or its second derivative (d²V/dt²), though the cable Equation 2 predicts that θ² α 1/r_a. In contrast, increased c_m does alter AP waveform. It reduces both dV/dt and d²V/dt² as well as reducing θ, giving the simple approximations

$${\theta }^2\hspace{0.17em}\alpha \hspace{0.17em}1/c_{\text{m}}$$ (4)

$${\left(\text{d}V/\text{d}t\right)}_{\max }\hspace{0.17em}\alpha \hspace{0.17em}\log \left(c_{\text{m}}\right)$$ (5)

$${\left({\text{d}}^{2}V/\text{d}{t}^2\right)}_{\max }\hspace{0.17em}\alpha 1/c_{\text{m}}$$ (6)

The subscript max designates the maximum value of the parameter in question. Finally, it is difficult to obtain analytic relationships between i_Na(max) and θ. Nevertheless, it is possible to demonstrate the straightforward empirical relationship

$$\theta \alpha \hspace{0.17em}{i}_{\text{Na}\left(\max \right)}$$ (7)

and the following effects of i_Na(max) upon AP waveform

$${\left(\text{d}V/\text{d}t\right)}_{\max }\hspace{0.17em}\alpha \sqrt{i_{\text{Na}\left(\max \right)}{{\displaystyle }}^3}$$ (8)

$${\left({\text{d}}^{2}V/\text{d}{t}^2\right)}_{\max }\hspace{0.17em}{{\alpha \hspace{0.17em}i}_{\text{Na}\left(\max \right)}}^3$$ (9)

This cable analysis, confined to a simple cylindrical, geometrically one-dimensional, structure, can be generalized to a continuous electrically coupled myocyte network. This provides cable equations extended from one to three dimensions analyzing the conduction velocity resulting from a matching of current and load (577, 599). Such an approach has been used to characterize the passive cable properties of cardiac muscle including its relationships between dV/dt and macroscopic (>1 mm) propagation and altered cell to cell coupling (576, 1246).

3. Repolarization gradients and action potential wavelength

Finally, normal atrial and ventricular myocardium shows a highly ordered sequence of AP repolarization and return of the membrane to resting conditions. In the ventricular myocardium, this typically proceeds transmurally from epicardium to endocardium and from apex to base. These features reflect regional differences in K⁺ channel density and kinetics. This results in the normal spatial repolarization gradients that may normally protect the orderly electrical and mechanical activation sequence. This thereby ensures correctly timed and coordinated mechanical activation and relaxation of the chamber concerned. In contrast to the ventricles, the thinner walled atria do not show a marked transmural differentiation into epicardial and endocardial tissues. The geometrical distribution of excitable events then occurs only within the plane of the atrial wall.

Together, the velocity θ of AP propagation and the recovery parameters of APD or the effective refractory period (ERP) define the AP wavelength (λ = θ·APD or θ·ERP). The APD reflects the period during which the membrane deviates from its normal resting value. It may closely correlate with the ERP in normally functioning canine and human hearts (243, 329, 624). The latter provides an indication of the time during which there is reduced likelihood of ectopic or reentrant action potential activation in the membrane behind the propagating excitation wavefront. However, ERPs can be selectively affected by factors that need not similarly affect APD. These include alterations in the maximum amplitude and activation or inactivation kinetics of I_Na and I_K, the amplitudes and durations of applied stimuli and myocyte injury and ischemia (975). Subsequent sections will consider VERP-ERP differences and their significance in proarrhythmic situations including Brugada syndrome (sect. VC5) (50, 734) and hypokalemia (sect. VE)(978).

However, measurement of ERP, typically from the shortest S1S2 interval following the last (S1) pacing stimulus at which a subsequent extrasystolic S2 stimulus elicits an AP, itself poses problems under particular experimental circumstances. Thus 1) some protocols are not amenable to such direct determinations of ERPs, yet 2) ERPs themselves vary with pacing protocol. 3) The observed ERPs critically depend on the relationship between stimulus and recording sites. A detectable AP requires its successful propagation through the entire tissue pathway from stimulating to recording electrode. The resulting ERP consequently actually reflects recovery from refractoriness over the entire line of tissue between stimulus and recording sites. 4) This condition poses further problems for determinations of spatial ERP heterogeneities between recording sites. The difficulties are compounded if stimulus and recording sites have differing ERPs. 5) ERP determinations are further affected by conduction velocities in the paths intervening between stimulus and recording sites. This was demonstrated in comparisons of ERP measurements at the stimulus site itself, ERP_s, and the ERP_r, given by the time interval separating AP upstrokes, at the recording site. ERP_r then did not equal ERP_s when the conduction velocity of the AP produced by the S1 stimulus differed from the conduction velocity of the AP produced by the S2 stimulus. These discrepancies were accentuated at high pacing rates close to refractoriness, and with increasing distance between stimulation and recording sites (734). 6) Whereas absolute refractory period may be regarded as an invariant value, ERP depends on the value of the effective stimulus intensity at the site of membrane excitation. Thus an adaptive S1S2 protocol demonstrated differing ERPs corresponding to differing magnitudes of S2 extrastimulus. Representations involving ERP, rather than APD, are thus strongly dependent on applied stimulus voltage (281).

The wavelength parameter nevertheless provides a useful description of the spatial extent of excitation by the traveling wave. Larger values of λ would result in a reduction of the likelihood that areas of depolarization and repolarization meet on encountering tissue heterogeneity. The resulting safety factor would then ensure that the traveling wave completely passes over the heterogeneity without disruption (1251). Conversely, a decreased value of λ would increase the likelihood of a wave breakup into into multiple wavelets, formation of scroll-waves (244, 866, 1317), and a positive feedback formation of further wavebreaks giving wavelets along chaotic conduction pathways corresponding to VF (1084). For example, where alterations in heart rates decrease APD, ERP, or θ, they could thereby alter AP wavelength λ and thereby potentially exert arrhythmic effects (595, 753).

Figure 2 illustrates such situations for a sequence of murine AP waveforms (753). The relevant parameters can be quantified, in terms of their basic cycle lengths (BCL), APDs at 90% repolarisation (APD₉₀), latencies, and corresponding diastolic intervals (DIs) at 90% repolarization (DI₉₀) separating the current and preceding action potentials. Together these yield active and resting wavelengths λ' and λ₀' (Figure 2A). The latter sum together to give a basic cycle distance, BCD' (Figure 2B). When an AP with long λ' passes over a heterogeneity that can potentially cause conduction block, the back of the propagating wave blocks retrograde propagation. This leaves only an orthograde excitation wave (Figure 2C). In contrast, when λ' is short, the back of the wave passes the heterogeneity before retrograde excitation has passed through the unidirectional block. This initiates a new propagating retrograde wave which potentially sets up a sustained reentrant circuit (Figure 2D).

Figure 2.

Extension of cable analysis to action potential wavelength, wave-break, and re-entry. A: typical murine monophasic right ventricular action potential (AP) waveform, indicating basic cycle length (BCL), action potential duration at 90% recovery (APD₉₀), latency and diastolic interval (DI) of the current (n^th) and preceding [(n−1)^th]AP. B: these variables yield the active and resting wavelengths λ' and λ₀' for which the basic cycle distance, BCD' = λ' + λ₀'. C: orthograde propagation of an AP with long λ' over a heterogeneity results in the back of the propagating wave blocking retrograde propagation. D: propagation of an AP with a short λ' results in the back of the wave passing the heterogeneity before retrograde excitation has crossed the unidirectional block. This results in initiation of a new propagating retrograde wave and a re-entrant circuit. [From Matthews et al. (753).]

II. ARRHYTHMOGENESIS: DISRUPTION OF ORDERED EXCITATION AND CONDUCTION

A. Electrophysiological Conditions Initiating and Perpetuating Arrhythmia

1. Triggering events at the cellular level

Arrhythmias result from an inappropriate generation, or a breakdown in the orderly sequencing, of cardiac electrical activity. They often follow triggering events, and then take place in tissue with intrinsic instabilities resulting in arrhythmic substrate (284, 529, 931, 1111). Events reflecting such phenomena could occur at the single-cell level, during propagation of excitation at the tissue level, or at the level of entire cardiac chambers (560, 981).

At the cellular level, triggered activity results from extrasystolic membrane depolarization that could potentially generate premature APs, and therefore, triggered beats, following an otherwise normal AP, if the voltage changes that they produce are sufficiently large. Of these, early afterdepolarizations (EADs) intercept the repolarization time course of a prolonged AP. They thereby permit time for L-type Ca²⁺ channel recovery from inactivation in a still depolarized membrane. The reactivated inward I_CaL then produces a further depolarization. This initiates a positive feedback process resulting in the afterdepolarization potentially triggering AP firing (498). In contrast, delayed afterdepolarizations (DADs) follow full AP repolarization. They can result from an enhanced SR Ca²⁺ release. This in turn increases activation of electrogenic NCX or Ca²⁺ activated Cl⁻ transient inward (I_ti) currents. DADs are associated with conditions of Ca²⁺ overload as occurs in digitalis toxicity and catecholaminergic polymorphic ventricular tachycardia (CPVT) (509). In the atria, triggering can also arise from the pulmonary or the superior caval veins and may thereby trigger episodes of AF (190) (see sects. VIA and VIIA).

Rarer causes of arrhythmias initiated by abnormal AP triggering at the cellular level include the enhanced automaticity resulting from accelerations in depolarization of pacemaker tissue. This might follow increased sympathetic activity, hypokalemia, or pharmacological intervention. In addition, parasystole could result from a parallel activation of two or more pacemaker regions (32).

2. Spatial electrophysiological heterogeneities at the tissue level

At the tissue level, failure of the AP wave to completely extinguish after normal activation, leading to reexcitation of regions that had hitherto recovered excitability can result in a reentrant excitation. This can occur in the presence of spatial electrophysiological heterogeneities that result in 1) an obstacle around which the AP can circulate provided 2) this occurs with slowed conduction velocities that would permit each region to recover excitability before the wave returns, in the presence of 3) a unidirectional conduction block. The latter might occur with spatial gradients in the latency separating stimulation and depolarization or the time between depolarization and the end of the ERP (see sect. VC) (740). Either would prevent the wave from self-extinguishing (778).

Figure 3 reconstructs the generation of arrhythmic substrate through such a combination of conditions. It illustrates the consequences of introduction of a slow conducting myocardial pathway passing through nonconducting myocardium (path 1; dark gray). This is bordered by a second pathway of normal myocardium (path 2; white). A normal AP (blue arrow) would propagate along path 2 following excitation (Figure 3Ai). The myocardium then becomes refractory. As indicated above (see sect. IC2), the resulting normal action potential traveling along path 2 possesses an excitation wavelength λ (yellow region). Consequently, the impulse conducting along path 1 cannot reenter the circuit as it would collide with refractory tissue in path 2 (Figure 3Aii). Similarly, when an abnormal impulse from an ectopic focus is triggered immediately following the normal AP, it cannot enter path 1 as this remains refractory (Figure 3Bi). It therefore splits at the end of path 2 to conduct retrogradely along path 1 and orthogradely along path 2 (Figure 3Bii). In contrast, a self-perpetuating reentrant excitation can result when 1) an action potential conducting retrogradely along path 1 enters the beginning of path 2 (Figure 3Ci) under conditions of 2) reduced conduction velocity (θ) and/or reduced effective refractory period (ERP). The latter result together in a reduction in excitation wavelength (λ = θ × ERP), to values smaller than the dimensions of the available circuits (Figure 3Cii). This results in persistent reentrant excitation (567).

Figure 3.

Conditions underlying generation of re-entrant arrhythmia. A: basic features of arrhythmic substrate, consisting of slow conducting myocardial pathway (path 1; dark gray), nonconducting myocardium, and second normally conducting pathway (path 2; white) (i). Normal action potential (blue arrow) propagates with velocity θ and effective refractory period (ERP) resulting in propagation wavelength (λ = θ × ERP) (yellow region) along path 2. It initiates a slow conducting impulse traveling along path 1 (i). In normal activity, the latter impulse cannot re-enter the circuit as it collides with refractory tissue in path 2 (ii). B: an abnormal triggered impulse immediately following the normal action potential cannot enter path 1 as this remains refractory. C: self-perpetuating re-entrant excitation occurs when a retrogradely conducting AP along path 1 (i) enters the beginning of path 2 with reduced conduction velocity and effective refractory period and therefore reduced excitation wavelength smaller than the dimensions of the propagation pathways. [From King et al. (567).]

Data from both canine right ventricular (RV) wedge preparations (801, 802) and the Scn5a+/ΔKPQ mouse model (see sect. VIB) further implicate reentry arising from repolarization abnormalities as exemplified by their epicardial dispersions of repolarization in triggering ventricular arrhythmia. The simplest example of this might arise from substrate for reentrant excitation arising from relative changes in two key parameters describing the recovery from excitation. Thus windows of reexcitation have been suggested in situations where critical intervals result from positive time differences between full action potential repolarization and the refractory period. These parameters are often quantified by APD₉₀ and VERP, respectively. This could take place within, or between, adjoining areas of myocardium, and is exemplified in section VE4 (978).

At the level of individual cardiac chambers, disruption of the normal sequence of repolarization following the depolarization wave accentuates arrhythmic tendency (1012). Mammalian ventricular myocardium normally shows a consistent and regular sequence of repolarization. This proceeds from epicardium to endocardium resulting in a transmura l repolarization gradient that optimizes the normal sequence and coordination of electromechanical activation and relaxation (see sect. VIA4). These spatial differences are mainly determined by regional differences in repolarizing K⁺ channel densities. Disturbances in these gradients may permit regions of depolarization to reentrantly reexcite already recovered areas (see sect. VIA2).

Transmural gradients may be particularly important in producing repolarization differences across relatively short distances. They have been implicated in a number of both canine models subject to pharmacological manipulation and murine models genetically modified to reproduce Brugada syndrome (BrS) (see sect. V) and long QT syndrome (LQTS) (see sect. VI). In addition, a number of animal models show ventricular, base-to-apex, heterogeneities in AP characteristics (see sect. VI). In human clinical situations, increases in this dispersion have been associated with arrhythmogenesis in cardiomyopathies and have been related to increased incidences of T wave alternans and VT (181). Finally, left-right interventricular differences in APD have been implicated in arrhythmogenesis in BrS. BrS patients show characteristic right precordial electrocardiographic ST elevation, right bundle branch block, and changes specific to RV epicardial AP waveforms (605) (see sect. VD).

3. Temporal electrophysiological heterogeneities at the tissue level

Temporal electrophysiological heterogeneities may appear as beat-to-beat variations in AP amplitude or duration. These instabilities have been clinically associated with the appearance of alternans in electrical properties between beats. T-wave alternans reflecting alternating time courses in successive ventricular APs classically precedes breakdown of regular electrophysiological activity and an onset of major arrhythmias. Both T-wave alternans and dispersion of the QT interval thus constitute risk stratification markers in both patients susceptible to sustained ventricular arrhythmias (46, 823, 966) and experimental situations (872). At the cellular level, several hypotheses have suggested possible, potentially coexistent, mechanisms. Voltage-driven alternans can arise from instabilities in membrane voltage resulting from steep APD restitution properties (413, 838). These can reflect the properties of depolarizing, I_Na, or repolarizing currents including I_to (701), Kir3.x in the case of atrial alternans (126), and currents giving rise to EADs (1005). Membrane voltage is also influenced by the Ca²⁺-sensitive I_CaL, I_NCX, I_Ks, and Ca²⁺-activated SK channels (1004, 1050, 1309, 1333). Nonlinearities can occur in intracellular Ca²⁺ cycling itself (894, 920, 1218). The latter can become unstable with SR Ca²⁺ overload, RyR2 sensitization, or at high pacing rates. These can result in steep SR Ca²⁺ release versus SR Ca load relationships giving Ca²⁺ restitution properties predisposing to [Ca²⁺]_i-driven alternans (204, 1050). At the tissue level, the resulting alternans may be spatially concordant in which the alternations in adjacent regions of a tissue are in phase, or discordant, when these are out of phase.

APD alternans may be an important mechanism for the generation of arrhythmic substrate. Spatially concordant alternans is not itself arrhythmogenic. However, it may represent a stage that precedes the development of discordant alternans. Discordant alternans in adjacent tissue areas produces APD gradients across their intervening regions which are separated by a nodal line region not showing alternans (872, 1236). Discordant alternans greatly amplifies the dispersion of refractoriness, generates regions of conduction block, and predisposes to figure-of-8 reentry phenomena. Triggered activity within the area with the short APD that propagates directly to the nodal line then likely collides with the electrical activity in the area with the long APD while this is still depolarized and refractory. It will then become extinguished. However, where its propagation is less direct over a longer distance, it would reach the nodal line at a later time when the area with the long APD has recovered (Figure 3 in Ref. 1250). It would then induce reexcitation and a reentrant circuit leading to VT, wavebreak, and evolution into VF (928, 1236).

4. Heterogeneities arising from restitution phenomena

APD restitution phenomena were first reported in classic cardiac electrophysiological papers describing alternations in excitation duration, later measurable by intracellular recording as APDs, with increasing heart rate (436, 778). The subsequent classical analysis described below arose from observations that APDs recorded from canine papillary muscle decreased with increasing steady-state pacing rate. The latter findings had led to the development of a restitution theory seeking to integrate such observations (838). This theory has been recently successfully applied to murine hearts (979, 982) (see sects. VC6 and VIA5).

The restitution analysis seeks to graphically predict the occurrence and magnitude of alternans with alterations in heart rate (Figure 4A). It does so through the expected oscillatory properties of a negative feedback system. The output of such a system (O) is considered to be the result of an amplification G of an input I. The latter is itself controlled by an independent variable, X

$$O=G\left(I\right)$$ (10)

Figure 4.

Temporal heterogeneity in the generation of re-entrant substrate. A: classical restitution curves in which action potential duration (APD_n) of the n^th AP decreases with the decreasing, preceding, (n−1)^th, diastolic interval (DI) observed at successively shortened basic cycle lengths (BCL). The accompanying progressively increasing slope requires successively greater number of cycles of alternans to intervene before the system reaches a new steady-state APD (points 1 and 2). When unity slope is reached, alternans become sustained (point 3). Slopes exceeding unity result in waxing oscillations (point 4) in APD. This culminates in conduction block and/or tachyarrhythmia resulting from wave-break. B: fuller analysis of generic restitution function relating APD₉₀ corresponding to the APD at 90% AP recovery to the corresponding DI₉₀. In addition to conventional measures of critical diastolic interval (DI_crit) and maximum gradient (m_max), this maps the maximum APD (APD_max) at low heart rates, DI₉₀ at the effective refractory period (DI_ERP), and the horizontal axis intercept of the restitution function (DI_limit), corresponding to absolute refractoriness. This permits definition of conditions for stability (unshaded), instability (gray), as well as relative (dark shading) and complete loss of capture (left shaded area). C and D: typical records reflecting arrhythmic phenotypes in monophasic action potential recordings from regularly paced (triangular markers) murine Scn5a+/− right ventricular (RV) epicardia showing nonsustained VT (BCL 134 ms) (C) and the initiation of sustained polymorphic VT (BCL 124 ms) (D). [From Martin et al. (740) and Matthews et al. (752).]

The output O in turn influences the input, to an extent which depends on a fraction of the output (F) and the independent variable (X)

$$I=X-F\left(O\right)$$ (11)

The solution of these simultaneous Equations 10 and 11 corresponds to the points of their graphical intersection. These thereby yield the set points giving the input I and output O at any value of X. The original analysis adopted as input variable the diastolic interval (DI) over which the membrane is restored to the resting potential. Over this period, the membrane is recovering following the AP. The output variable is the APD. This is itself dependent on the preceding DI. Thus, for any given, n^th, beat

$${\text{APD}}_n=f\left({\text{DI}}_n\right)$$ (12)

The independent variable controlling DI is the BCL. The DI in the subsequent, (n+1)^th beat, DI_n+1, depends on both the BCL and the APD in the previous, n^th beat, APD_n

$${\text{DI}}_{n+1}=f\left({\text{APD}}_n\right)=\text{BCL}-{\text{APD}}_n$$ (13)

An A curve was obtained by plotting the output variable of APD against the input variable of DI. A family of D lines each taking the form

$$\text{APD}=-\text{DI}+\text{BCL}$$ (14)

with a consistent negative unity gradient but variable ordinate intercepts set by the BCL was then plotted between the same axes. The intersection between the D line and the A curve would give the solution for the steady-state APD and DI.

A perturbation in heart rate would result in an immediate transition between two separate D-lines representing the respective BCLs. A horizontal line from the steady-state point on the A curve to the new D line would then give the magnitude of the first subsequent DI. A vertical line drawn from there to the corresponding A curve would give the corresponding subsequent APD. Graphical representations for the different cases are shown numbered 1–4 in Figure 4A. Each case would yield different predictions for the outcomes generated by continuation of this process. These would depend on the gradient of the A curve at its intersection with the D line, and its variation with DI in this region.

Where this intersection occurs in a region of the A curve where its slope is zero (Figure 4A, point 1), a final steady-state APD is immediately reached without oscillations. Where the slope at the intersection falls between zero and unity (Figure 4A, point 2), the successive projection lines converge towards and ultimately attain the set point. Each corner on the A curve then represents an individual oscillation, producing a transient alternans. With an intersection at a critical DI, DI_crit, where the the A curve assumes a unity slope, the projection lines form a square. The oscillation then does not converge (Figure 4A, point 3). This results in a sustained alternans whose magnitude is determined by the size of the square. Finally, where the intersection takes place in a region on the A curve where its slope is greater than unity, the progressive projections take a centrifugal trajectory. They thus veer away from the left-hand limit of the A curve, producing conduction block (Figure 4A, point 4). This state, particularly when heterogeneous across the myocardium, may cause reentry (754).

These different conditions can then be related to the remaining parameters describing cellular excitability in a fuller generic analysis of the restitution function. Figure 4B illustrates this development using the relationship relating APD₉₀ to DI₉₀ through different BCLs. In addition to the conventional measures of critical diastolic interval (DI_crit) and maximum gradient (m_max) this maps the maximum APD, APD_max, at low heart rates, DI₉₀ at the effective refractory period (DI_ERP), and the horizontal axis intercept of the restitution function (DI_limit) corresponding to absolute refractoriness. These additional limits permit definition of different stability conditions within the plane of the restitution function. Thus stability would correspond to the condition when the gradient of the restitution function, m ≤ 1 as outlined above (unshaded areas). Instability would be expected under conditions when DI_crit > DI₉₀ > DI_ERP (filled areas). This would manifest in occurrence of either nonsustained (Figure 4C) or sustained arrhythmia (Figure 4D). These are exemplified in the recordings from murine Scn5a+/− RV epicardia at two respective BCL values (134 and 124 ms, respectively). Relative loss of capture would be expected in the interval DI_ERP > DI₉₀ > DI_lim, (dotted areas) and complete loss of capture when pacing takes place at a BCL shorter than the absolute refractory period when DI₉₀ < DI_lim (hatched areas) (752).

The first restitution (A-) curves were deduced from transmembrane APs in isolated frog (Rana catesbiana) ventricles. These were paced at successively higher rates until refractoriness was reached. These gave a wide range of DIs. APD then varied minimally at low pacing rates. However, with increases in rate, the A-curve gradients became progressively but reproducibly steeper. Recordings obtained immediately following rate changes were variable. They showed a hysteresis above and below the steady-state line for accelerating and decelerating rates, respectively.

Parameters other than APD, such as voltage, may also show alternans and could be similarly graphically analyzed. However, although often occurring together, the existence or extent of voltage amplitude alternans did not parallel the magnitude of duration alternans. Furthermore, wide interspecies variations exist in the time course and the final steady states of adaptations to change in BCL. Imposed increases in heart rate most frequently result in initial reductions in APD with reduced DI, then increasing to nevertheless still reduced steady-state APD values, in human, guinea pig, mouse, and frog ventricular APs. However, APD initially lengthens then rapidly decreases to shorter values expected at shorter DI in rabbit, dog, and cat ventricles. This likely reflects incomplete I_to decay in rabbit and transient L-type Ca²⁺ current facilitation in dog and cat. Finally, APD actually lengthens and remains prolonged with decreased DI in rat heart. Ca²⁺ homeostasis has been implicated in restitution changes invoking actions of [Ca²⁺]_i upon the activity of numerous other channels and carriers within the cell (166). Ionic mechanisms for APD accommodation to rate changes remain unclear. They almost certainly involve inactivation processes in depolarizing I_Na and I_Ca currents or enhancement of repolarizing K⁺ currents, with both processes accumulating with successive APs. However, there are numerous differing reports on their relative importance. This has hampered attempts at in silico modeling of APD restitution. Recent clinical reports similarly suggest complexities in the use of such plots in predicting human arrhythmia (818).

Conduction velocity restitution plots of θ against DI similarly reflect the changes in θ with DI in consecutive AP waves. Conduction velocity alternans results in a compression and rarefaction of APs as they travel through the tissue. Slowed θ also reduces the distance between nodes. This results in a higher number of nodes thus predisposing to reentry. θ restitution thus depends primarily on Na⁺ channel and intercellular coupling properties, whereas APD restitution is also affected by Ca²⁺ (212, 530), and K⁺ channel properties and is thus engaged at lower pacing rates (928). In silico modeling studies suggested that θ alternans initiated at higher pacing rates may be responsible for the breakdown of concordant to discordant alternans (1236). However, θ restitution, as represented by plots of θ against DI, are not amenable to the systems analysis of the kind illustrated in Figure 4 and Equations 10–14. Thus the variable θ does not directly feed into that of DI. Furthermore, θ restitution is primarily concerned with the wavefront, whereas APD restitution describes the recovery from excitation that follows (753). Nevertheless, a recent analysis has unified restitution analysis involving APD and θ, respectively, into a single λ restitution analysis. This may offer a more general approach relating restitution conditions to arrhythmic tendency (see sect. VC6) (753).

B. Experimental Models in Studies of Arrhythmic Phenomena

1. Animal models for arrhythmic disease

Genetic exemplars useful for clarification of physiological and arrhythmic effects of particular molecular modifications are available from a range of species. Rabbit ventricular cardiac APs share similar ionic currents and positive plateau phases whose durations resemble the corresponding features in human hearts. Of rabbit ion channel variants, 1) transgenic rabbits with LQTS contain pore mutations in KCNQ1 and KCNH2. They showed LQTS1 and LQTS2 phenotypes corresponding to their absence of I_Ks and I_Kr, respectively (152). LQTS2 rabbits showed high incidences of spontaneous SCD (>50% at 1 yr) due to polymorphic VT. Optical mapping studies revealed increased spatial dispersions of repolarization in LQTS2 rabbits. In both, elimination of one repolarizing current (e.g., I_Kr) was associated with downregulation of the other (e.g., I_Ks). This contrasts with the upregulation found in mouse LQTS models (383, 384). 2) Rabbits with chronic atrioventricular block showed biventricular hypertrophy, QT interval prolongation following I_Ks and I_Kr downregulation, spontaneous torsades de pointes, and shortened lifespan (1147). 3) Hypercholesterolemic rabbits showed an electrophysiological and neural cardiac remodeling including cardiac hypertrophy, QT prolongation, and a vulnerability to VF (682). Of hypertrophic variants, there are 4) hypertrophic cardiomyopathic (HCM) rabbit models affecting the β-myosin heavy chain through the β-myosin heavy chain (MyHC)-R400Q mutation (725) and a cardiac-restricted expression of the mutant β-MyHC-Q403 known to cause clinical HCM (950). These exhibited cardiac hypertrophy, myocyte disarray, interstitial fibrosis, and SCD. However, they often did not exhibit electrophysiological properties normally accompanying proarrhythmia (see sect. IXB).

The earliest described spontaneous mutation causing nonhuman HCM involved a feline G to C variant in cardiac myosin binding protein C (MYBPC3) producing a alanine to proline (A31P) substitution (771). This was followed by similar reports with a feline C820T mutation (770). Both variants resulted in a computationally predicted alteration in protein conformation.

Canine hearts show ion current and AP waveforms characteristic resembling those in human myocardium. 1) Dog hearts with chronic atrioventricular block showed complex structural hypertrophic and electrophysiological remodeling. The latter was associated with I_Ks and I_Kr downregulation (1205), NCX upregulation (1068), and predispositions to torsades de pointes and SCD (375, 858, 1068, 1206, 1207). 2) Boxer dogs provide a spontaneous model of arrhythmogenic right ventricular cardiomyopathy (ARVC) and SCD. Their clinical and pathological features closely resembled the human condition. These included VT arising from an enlarged RV showing myocyte loss with fatty or fibrofatty replacement, myocarditis, and apoptosis, in some cases accompanying altered RyR2 expression (81, 769). 3) Finally, dogs affected with X-linked Duchenne muscular dystrophy also showed calcified myocardia and surrounding dense connective tissue associated with ventricular arrhythmias (791).

2. Transgenic mouse models as disease systems

In addition to spontaneous mutations producing phenotypes recapitulating human disease, genetic modifications can be produced from insertion into the animals' genome a mutated human gene (transgene), or introduction of a mutation at a particular locus by gene targeting (160). Among mammalian models, the mouse remains the most amenable to such genetic modifications. The resulting animals often reproduced the corresponding congenital human arrhythmic disorders, particularly when they involved established monogenic mutations (200, 246, 563, 831, 995). Many of these latter genetic variants are clinically rare. Nevertheless, they may model better defined pathological mechanisms than variants representing more common and often more complex conditions. They thus potentially yield important and widely applicable pathophysiological insights. Use of genetic exemplars additionally avoids the need to replicate clinical phenotypes through the use of potentially nonspecific pharmacological manipulations (319, 594, 774, 1054, 1055).

Conventional gene targeting involves homologous recombination of an engineered exogenous DNA fragment containing a targeting vector and selection marker conferring resistance to a cytotoxic drug with the genome of an embryonic stem (ES) cell. The modified ES cell is then screened for presence of the fragment using the cytotoxic agent. It is then injected into a blastocyst that develops into a chimeric animal where the desired gene is present in the gametes. Targeted ES cells injected into wild-type (WT) mouse blastocysts can then contribute to the germline of chimeric mice. These can then be used to generate progeny containing the targeted gene. Selective breeding then creates heterozygous, then possibly homozygous, offspring. Knock-in models contain alterations in the genetic code resulting in a modified mRNA producing a protein with loss or gain of, or modified, function. In knockout models, no mRNA is produced from the altered gene (246).

Alternative methods expediting genome modification generate DNA double-strand breaks by directly injecting DNA or mRNA coding for site-specific nucleases into the one-cell embryo (660). The latter utilize zinc-finger nucleases (ZFNs) (163, 350, 1160), transcription activator-like effector nucleases (TALENs) (1101), or the RNA-guided clustered regularly interspaced short palindromic repeat (CRISPR)/CRISPR-associated (Cas) nuclease system (648, 1290). The double-stranded breaks occur at a specified genomic locus. They are then repaired by error-prone nonhomologous end joining to resulting mutant alleles (163, 350, 1101, 1216). The Cre-lox system has proven a useful technique to restricting expression of a mutated gene to the heart. It is particularly applicable where the mutation is embryonically lethal when involving the entire organism. It also makes it possible to achieve temporal control over gene expression (414, 601, 813). It ensures that recombination events only occur only in specific cell types expressing the site-specific DNA cyclization recombinase (CRE).

3. Murine hearts as models for human arrhythmic disease

Murine hearts are similar in overall anatomy to human hearts. However, there was initial uncertainty as to whether their considerably smaller tissue volumes could sustain the polymorphic arrhythmias shown by larger hearts (343, 526). In the latter event, they would not accommodate the multiple drifting rotors then thought necessary to produce the scroll waves required to generate polymorphic arrhythmia (1264). However, later reports from rabbit heart demonstrated that a single drifting rotor generated scroll waves leading to such polymorphic arrhythmia could thus exist in relatively small volumes of tissue (371). Polymorphic arrhythmia was subsequently observed in mouse ventricles (172, 301, 505, 736, 752–754, 1162). Some of these also demonstrated transitions from a monomorphic to the polymorphic character pattern (982) as observed in larger hearts (1259).

In common with findings in human hearts, murine ventricular APs show rapid depolarization phases driven by inward I_Na (384). Furthermore, human and murine hearts have similar transmural AP conduction velocities (428, 668). It is generally thought that the effective refractory period ends before the completion of the repolarization process (301, 582, 977, 980, 995; but see Ref. 59). These features expedited their use in studying physiological effects of SCN5a mutations that can underlie BrS and LQTS3, and of effects of class I, Na⁺ channel blocking drugs (see sect. V and Table 5).

Table 5.

Phenotypic similarities between arrhythmogenic properties in murine Scn5a+/− and Scn5a+/Δkpq hearts and human BrS and LQTS3

Model	Phenotype	Mouse	Human	Reference Nos.
Scn5a+/− compared with BrS	Arrhythmic phenotype unmasked by flecainide	+	+	150, 344, 914, 1270
	Arrhythmic phenotype relieved by quinidine	+	+	92, 421, 792
	Positive ΔAPD90 persistent with flecainide and quinidine	+	−	743, 1095
	RVOT/right ventricular initiation of arrhythmia	+	+	734–736, 741, 754, 1336
	Fibrotic change with age	+	+	500–503
	Presence of male/female phenotypic differences	+	+	500–503
Scn5a+/Δkpq compared with LQTS3	Arrhythmic phenotype reduced by flecainide	+	+	94, 914, 1096, 1263
	Arrhythmic phenotype exacerbated by quinidine	+	+	207
	Negative ΔAPD90 rescued by flecainide	+	+	94, 1096
	Absence of male/female phenotypic differences	+	+	684, 1096, 1321

However, AP waveforms in mouse ventricles show recovery phases distinct in character from those found in the ventricles of larger animals and humans. The latter show clear-cut plateau phases. These are attributable to activation, followed by inactivation, of I_CaL in combination with voltage-dependent changes in the specific K⁺ currents. The K⁺ current contributions are dominated by rapid, I_Kr, and slow, I_Ks, delayed rectifier currents. Of these, I_Kr rapidly activates with phase 0 AP depolarization. However, its inactivation, that rapidly follows, makes the channel nonconducting during phases 0–2 of the AP. The initial phase 3 repolarization then releases the inactivation and reopens the channel. Its subsequent, slow, deactivation then permits a sustained phase 3 and early phase 4 I_Kr. In contrast, I_Ks activates slowly with depolarization to a relatively positive greater than −20 mV potential, and barely inactivates. It gradually increases over phase 2 to become a major phase 3 K⁺ conductance (696, 697, 1169, 1170).

In contrast, AP repolarization in mouse ventricles results in shorter, triangulated, APDs (∼30–80 ms) compared with those of humans (∼150–400 ms, respectively) (Figure 1D) (236). K⁺ currents are similarly central to this AP repolarization, although there are smaller I_CaL contributions. However, repolarization is driven mainly by fast, Kv4.3 and Kv4.2-mediated, I_to,f, as well as the more slowly inactivating Kv1.4-mediated I_to,s, components of the rapidly activating and inactivating transient outward current I_to (Figure 1, B and D). Both I_to,f and I_to,s become activated at potentials greater than −30 mV (829). They show distinct time constants for inactivation and recovery from inactivation. For I_to,f, these fell in the range of 20–100 ms for both processes. In the case of I_to,s, these processes took hundreds of milliseconds and seconds, respectively (833). Murine hearts also showed rapidly activating but slowly inactivating 4-aminopyridine-sensitive and -insensitive, Kv1.5 (KCNA5)-mediated I_K,slow1 and Kv2.1 (KCNB1)-mediated I_K,slow2, as well as a steady-state I_ss (829, 1284).

Recent evidence suggests murine hearts do express both I_Kr and I_Ks, but their roles are unclear (53, 64, 276). I_KATP and I_K1 continue to be important in repolarization and electrical diastole (324, 828, 830). Mouse ventricles also show additional repolarizing rapidly activating and sustained delayed rectifier K⁺ current, I_ss, and slow K⁺, I_Kslow, currents (669). Murine hearts additionally exhibit inward rectification properties attributed to I_K1. These strongly reduce K⁺ conductance at voltages greater than −20 mV in phases 0–2, permit outward currents with repolarization to greater than −40 mV late in phase 3, and stabilize the phase 4 diastolic resting potential (357, 1020).

Atrial myocytes additionally show an ultra-rapid I_Kur in phase 1. There is also the I_KACh, which is activated through its interaction with the βγ part of acetylcholine muscarinic receptor-associated G_i proteins particularly in the SAN but also in the atria and ventricles (1020). Studies exploring effects of parasympathetic challenge in Girk4−/− mice implicate I_KACh in AF (589). Finally, I_KATP is normally a small current. However, its activation by reduced intracellular ATP levels associated with energetic stress exerts triangulating effects on AP waveform (320, 486).

Finally, recent evidence suggests that induced pluripotent stem cells derived from carriers of clinical conditions might replicate some of the expected cellular properties. This may have potential utility in the development of novel treatment strategies in the LQTS2 (488), LQTS3 (718), CPVT (307, 525), and some overlap syndromes (248), given the appropriate correlations between basic molecular physiology and our understanding of arrhythmia in structurally intact hearts.

C. Experimental Studies on Murine Systems

1. Studies at the organism or tissue levels

Arrhythmogenesis itself ultimately takes place at the whole heart level and likely depends on electrophysiological properties through populations of coupled cells in addition to events within single cells. These may include myocardial properties, such as reentry mechanisms and gradients of excitation or recovery from excitation operating at the tissue and whole chamber level. Culture methods that might reproduce such cell populations show initial promise (158). Studies in mouse systems bred on electrophysiologically stable, typically 129/sv or C57BL/6, genetic backgrounds, permit examination of such phenomena whether related to mechanisms of SND, and both atrial and ventricular arrhythmia. These could carry well-defined genetic modifications strategically selected to reflect genotypes associated with specific disease conditions. Such modifications could also be used to expedite experimental clarification of the role of alterations in the function of particular ion channels in arrhythmic mechanisms. Use of WT hearts also permits investigations of acute, reversible, often pharmacological, manipulations, replicating clinical situations and potentially modifying arrhythmogenicity. Investigations of either case typically begin by verifying reproducible arrhythmic phenotypes. This would relate the experimental system to the corresponding clinical condition, providing a translational background for studies proceeding to clarify their underlying mechanisms. The latter investigations are directed at successive, interacting, organism, organ, tissue, in addition to the cellular and molecular levels.

Of these, in vivo electrocardiographic (ECG) studies in ambulatory or anesthetized intact animals (e.g., Refs. 184, 1040) can demonstrate spontaneous rhythm disruptions, electrocardiographic abnormalities (742, 1340), or chronic changes in electrophysiological properties. The latter have been demonstrated in particular genetic conditions with age (e.g., sect. VE) (503). They have also been instrumental in the analysis of the effects of genetic connexin modifications on ventricular conduction and excitation reflected in their PR, QRS, and QT intervals (378, 803, 1132). They could be extended to localizations of atrioventricular conduction abnormalities to supra-Hisian and particularly, infra-Hisian, conduction tissues, in common with human clinical findings in a murine myotonic dystrophic model (976). These studies additionally established strategic conditions of temperature, anesthesia, and selection of electrocardiographic parameters as important in such determinations (111). Recent studies have successfully correlated QT intervals with simultaneously measured APDs in Langendorff-perfused murine hearts. This further makes ECG recording a useful quantitative tool (236, 782, 1219, 1341).

Ex vivo isolated Langendorff-perfused preparations permit closer investigations for triggering arrhythmic events occurring spontaneously during intrinsic activity. In addition, arrhythmic substrate may be detected through an appearance of spontaneous arrhythmia during spontaneous activity, regular epicardial pacing, or septal pacing. Arrhythmia could also be provoked by programmed electrical stimulation (PES). This typically involves imposition of extrasystolic (S2) beats following successively decrementing S1S2 time intervals following trains of regular S1 pacing beats (240, 392, 561, 1095). Alternatively, dynamic pacing protocols could apply sequences of successively incremented steady-state pacing rates (e.g., Refs. 735, 752–754, 979, 982). Some atrial studies also include burst pacing protocols involving delivery of successive high-frequency stimulus trains (245, 402, 659).

Different recording techniques can then explore for underlying pathophysiological and pharmacological processes through closer studies of AP initiation, propagation, repolarization, and refractoriness (392, 978). Of these, intracellular recordings using sharp glass microelectrodes provide definitive indications of absolute resting potentials and AP waveforms including their durations and their maximum rates of rise (dV/dt)_max (e.g., sect. VIID) (569, 1340).

Extracellular recording methods offer stable and prolonged in situ estimates of such parameters and can be applied to specific cardiac regions in intact preparations. Unipolar recordings made against remotely positioned reference electrodes represent tissue electrical activity in the form of single large negative intrinsic deflections reflecting arrival of the excitation process immediately beneath the electrode contact with tissue. These are superimposed upon the extrinsic deflections produced by far-field effects. The latter cause positive deflections as the excitation wave travels from remote regions of tissue towards, and negative deflections as the wave subsequently moves away, from the recording electrode. The result is a biphasic waveform (646, 1271). The steep negative intrinsic deflection in the unipolar electrogram coincides with the transmembrane AP upstroke. This makes it possible to determine wave velocity from the time at which it assumes maximum negative slope (275).

Bipolar extracellular recordings comparing voltages at two recording electrodes positioned close to the conducting path for tissue excitation similarly yield biphasic waveforms. The latter reflect a summation of two unipolar recordings. The two terminals are affected to almost the same degree by the extrinsic potentials and far field effects. The latter are therefore minimized by taking the difference between recordings from the positive and negative poles of the amplifier (493). Comparison between biphasic waves obtained under conditions of bipolar recording with biphasic waves obtained under conditions of unipolar recording correlate the fast part of the intrinsic deflection in the unipolar recording with the top of the differential spike obtained from the bipolar recording. Thus, in bipolar, in contrast to unipolar, electrograms, the intrinsic deflection coincides with the top of the spike.

Bipolar electrogram (BEG) recordings thus yield electrogram latencies and durations (EGDs). They have been used in conjunction with PES procedures. The resulting extracellular waveforms have been analyzed by programmed electrogram fractionation analysis (PEFA). This involved plotting conduction latencies in the recorded BEG deflection components at progressively shortened S1S2 intervals in clinical studies. This analysis had first successfully assessed clinical arrhythmogenic risk in HCM and cardiac channelopathy (461, 1009, 1011, 1012). It thus associated arrhythmogenic reentrant substrates with increased EGD. The latter reflects the distribution or spread (“fractionation”) of ventricular myocardial conduction velocities at reduced S1S2 intervals. This approach proved applicable to murine systems (see sects. V, D2 and G4, and VIB3) (64, 416, 1095, 1096).

Monophasic AP (MAP) recordings additionally reproduce many features of AP waveforms without requiring cell impalement, particularly the AP repolarization phases, and their quantification as APDs and ERPs (559, 560, 562, 563, 1129). They also are potentially translatable from studies in mouse to human hearts (796). These in turn yield APD₉₀/ERP ratios. Increases in such ratios are associated with increased arrhythmogenicity (978).

MAPs were initially recorded by suction electrodes as injury potentials (443, 512). Subsequent contact MAP recordings measured the voltage drop between a positive contact electrode firmly pressed against, and an adjacent, closely placed (∼5 mm), negative electrode in light contact with, the myocardial surface. The volume conductor hypothesis suggested that the pressure exerted by the positive electrode depolarizes the underlying cells to approximately −20 to −30 mV by opening membrane stretch-sensitive ion channels. This locally inactivates their contained voltage-gated Na⁺ channels. This makes the local membrane unexcitable and assuming a fixed reference potential, but leaves the adjacent cells capable of AP generation. During electrical diastole, the inactivated region acts as a current source as it is relatively depolarized compared with the surrounding myocardium. The resulting current flow between source and sink regions originates from a large number of cells, given the ∼1–2 mm tip diameter of the positive electrode. It is determined by the potential gradient and the number of cells contributing to the source-sink interface, and reverses during electrical systole. Alterations in the number of cells generating the MAP voltage by altering the contact pressure, studying the atria rather than the ventricles, or differently sized hearts, alter the amplitude of the MAP recording.

Comparisons with intracellular recordings confirm that MAP recordings faithfully represent the AP time course, particularly its repolarization phase, which largely comprises low-frequency electrical signals, where they show smooth upstrokes over <5 ms rise times to overall amplitudes >10 mV uncontaminated by intrinsic or QRS deflections and fall to smooth stable diastolic baselines (217, 328, 582). The latter condition can be affected by motion artifacts resulting from cardiac contraction, but such artifacts may be minimized by spring mounting of the electrode. MAP recordings reproduce alterations in, but not absolute values of, resting and plateau voltage. They yield significantly smaller (∼7 V/s) values of (dV/dt)_max than intracellular recording (200–300 V/s). This likely reflects the lower seal resistances made between tissue and MAP compared with those made by intracellular electrodes. In addition, MAP recordings reflect the activity of large numbers of potentially sequentially activated, rather than single cells.

Electrical contact or optical mapping methods provide patterns and timings of wavefront propagation of either electrical excitation or spectrofluometrically measured Ca²⁺ release (292, 296, 427, 632, 736, 803, 1109). Unipolar, multi-electrode array (MEA) recording from the left or right atrial or ventricular epicardial surfaces can follow AP propagation from either stimulated or spontaneously beating hearts. They thereby yield isochronal maps displaying activation times (AT) to the point of maximal negative slope (dV/dt)_max of each recorded electrogram. Recovery times (RT) can be measured as the (dV/dt)_max for a negative T wave or the (dV/dt)_min of a positive T wave. They thereby provide activation recovery intervals (ARI) separating their respective AT and RT. ARI proved comparable to APD values obtained from MAP electrode measurements. It is then also possible to determine activation and repolarization time differences (ATD and RTD) between the first and last ATs and RTs and the ARI differences (ARID) between the shortest and longest ARI (see sect. VC7) (736).

MEA recording also permits determination of conduction velocity parameters. It can provide both effective conduction velocities representing AP conduction through recording points in the array as a whole, or from local vector analyses giving both the magnitude and direction of AP conduction (see sects. VB2 and VIIC6) (996, 1336). Fluorescence imaging permits a spatiotemporal characterization not only of voltage, but also changes in calcium cycling in tissue labeled with voltage-sensitive (e.g., di-4-ANEPPS, RH237) or Ca²⁺ indicators (such as rhod-2-AM). This involves appropriate excitation of the signals, and separating and collecting emissions of the appropriate wavelengths. These give results that clarify the pattern of the activation and calcium release, and presence of diastolic calcium leak (59, 1218).

2. Studies in single cells and at the molecular level

Single-cell studies in intact or perforated whole cell patch-clamped myocardial cells enzymically isolated from Langendorff-perfused hearts have studied Ca²⁺ (1128) and K⁺ current activation (561) and activation, inactivation and recovery from inactivation of Na⁺ channels under both acute, including hypokalemic, conditions and in genetically modified, including Scn5a+/− and Scn5a+/ΔKPQ, myocytes (see sect. VC2) (416, 741, 867). In addition, a recent, loose patch-clamp, technique (22, 23, 1304) has been introduced to examine differences in current-voltage relationships of peak I_Na in intact Scn5a+/−, RyR2-P2328S, and WT atria. These have included studies of the effects of high extracellular Ca²⁺, caffeine, or cyclopiazonic acid (CPA). All these agents are expected to acutely increase [Ca²⁺]_i (see sect. VIIC6) (568). This technique shows significant future promise for studies of biophysical events in intact hearts (935). Transfection methods whether involving α- or β-subunits in expression systems such as those offered by Chinese hamster ovary (CHO) or human embryonic kidney (HEK) cells have been used to study and both activation and recovery properties for subsequent modelling (696, 697, 887, 1170) as well as reconstruct channel behavior in response to simulated APs (273, 406, 1060).

Isolated atrial and ventricular fluophore-loaded myocardial cells were also used to examine Ca²⁺ signaling properties by themselves or following acute pharmacological interventions. These have studied measures of SR Ca²⁺ release or reuptake, or cellular Ca²⁺ entry or expulsion, and their effects upon arrhythmic properties. These were first explored in WT murine hearts (see sect. VIIB) (62, 63, 352, 446, 1335, 1338) prior to their introduction to genetically modified systems (see sect. VIIC) (e.g., Refs. 364, 387, 1334).

At the molecular level, assessments of longer term factors contributing to arrhythmia have involved examination of gene expression characteristics, by quantitative reverse transcriptase polymerase chain reaction (qPCR) and western blotting, of a wide range of ion channels and regulatory genes. These have included genes for Na⁺ channel subunits, Ca²⁺ channel subunits and Ca²⁺-handling proteins, K⁺ channel subunits, hyperpolarization-activated cyclic nucleotide-gated (HCN) channels, gap junction proteins, transcription factors, and other selected genes, including transforming growth factor-β1 (TGF-β₁) and vimentin gene expression (see sects. VIF and IXA) (245, 392, 408).

3. Studies of contributions from morphological change

Finally, explorations can be made of morphological changes extending from gross organ and tissue architecture, through hypertrophic and fibrotic tissue changes, to electronmicroscopic analysis, associated with particular arrhythmic conditions. These studies have assessed contributions of age and/or genotype to the observed arrhythmic changes. This has included studies of gross tissue architecture, as well as quantitative digital microscopy studies of the results of picrosirius red staining and vimentin fibroblast immunostaining, in different cardiac chambers of the right and left sides of the heart (see sects. V, E and F, and IX) (408, 501, 502, 968, 1180).

4. Computational modeling studies

Quantitative computational reconstructions have explored the consequences of the individual changes in properties obtained by the studies above for properties at the systems level. These could optimally link findings in murine hearts (133) to properties of human hearts under circumstances of both normal (154, 369, 489, 1153) and abnormal function (551, 552, 1154, 1155). Recent modeling approaches now potentially permit further incorporation of electroneutral solute and osmotic fluxes, and interacting alterations in the resulting ion compositions and cell volumes (331, 332). At higher levels of organization, they extend to whole chamber reconstructions of the spread and recovery of the electrophysiological change (1140), with significant promise of clinical translatability (936).

Studies analyzing experimental findings in murine hearts have initially examined situations involving relatively simple geometry as exemplified by the spread of excitation from the SAN into and through the atria. These provided models that implicated I_Na in AP propagation through the SAN and from SAN to atria, its effects upon heart rate, and their reconstruction of features of clinical SND. This clarified possible roles for Na⁺ channels in both normal SAN function and SND following targeted disruption of the murine cardiac sodium channel gene (see sect. VF3) (408, 456, 632, 634, 1276). They extended to investigations of the effects of fibrotic, resulting in morphological change (see sect. IXA3) (245). Computational reconstructions have also been used to guide interpretation of ion channel changes (e.g., Ref. 1276). They have also provided theoretical predictions of the effects of particular ion channel variants for which experimental murine models are not available as in some short QT syndrome (SQTS) variants (see sect. VIH, e.g., 6–8). Computational studies thus represent an approach that will become increasingly valuable as physiological, anatomical, and molecular data become progressively available.

5. Murine exemplars for the study of cardiac arrhythmias

The remainder of this article reviews experimental and theoretical studies of electrophysiological mechanisms underlying cardiac arrhythm ias. They make systematic examinations of specific murine exemplars variously representing ventricular, and sinoatrial arrhythmic clinical disease. In each case, these could be used to assess the relevance of the biophysical and physiological principles underlying stable and unstable atrial or ventricular excitation as outlined above. Such instabilities could disrupt the normally orderly sequence of cardiac electrophysiological activation and recovery and its propagation.

Of these, primary electrophysiological disorders could arise from reversible pharmacological manipulations or loss- or gain-of-function genetic modifications that involve proteins directly concerned with cardiac excitable activity or its modulation. They thereby influence electrophysiological stability even in the presence of normal cardiac anatomy. Such changes can occur against a background of changes in heart rate. Section III discusses exemplars bearing on sinoatrial pacing disorders. The resulting propagated electrophysiological activity can then be characterized in terms of a wavelength (λ) of excitation as well as events falling within the resting wavelength following recovery from activity (λ₀) that follows (see sects. IC and IIA). By defining the spatial extent of excitation terms, λ would clarify the tendency for wave breakdown leading to arrhythmic substrate. A first term determining λ is the conduction velocity θ. This in turn is first dependent (see sect. IC, 1 and 2) on the intracellular resistance r_a, which determines passive spread of local circuit currents through Cx molecules between successive myocytes. Murine hearts with expected alterations in these are explored in section IV. Local circuit current activation resulting in AP propagation relevant to cellular excitation in turn depends on I_Na (see sect. IB1). The consequences of compromised I_Na are accordingly examined using examples discussed in section V. Section VI then discusses the consequences of abnormalities in the remaining, recovery, term determining λ, viz. the APD or ERP. Sections VII and VIII then explore inputs to excitable properties from intracellular events. These often bear upon Ca²⁺ homeostasis and cardiac energetics. These may alter both excitation as well as triggering events following AP repolarization. Finally, development of anatomical abnormalities could potentially result in arrhythmic substrate through alterations in the relationship between the spatial indicators of electrophysiological activity, particularly λ, and the effective length of the conducting excitation paths. These are considered in section IX.

III. MODELS FOR SINOATRIAL PACING DISORDER

A. Murine Models for Sinoatrial Pacing Function

Successive cycles of cardiac activity are normally driven by SAN automaticity, effectively constituting an oscillator generated by a combination of time-dependent outward, and constant or voltage-dependent inward currents activated during AP repolarization (723). The high background SAN pacing rate in murine compared with human hearts, and the wide variation of heart rates through different species may suggest similarly varying roles of the different ion mechanisms that have been implicated in this function. Similarly, isolation of murine SAN cells is challenging owing to the small size of murine pacemaker regions (136, 137, 197, 210, 440, 632, 637, 722). Nevertheless, mouse strains carrying specific genetic modifications in particular ion channels have offered useful models for pacemaker dysfunction (Table 2) (700).

Table 2.

Murine models involving sinoatrial pacemaker function

Gene	Genotype	Phenotype	Reference Nos.
Rapid K+ current IKr
Erg1b	Erg1b−/−	Loss of IKr; bradycardic episodes	210, 626
Hyperpolarization-activated cyclic nucleotide-gated (HCN) current If
Hcn1	Hcn1−/− (global)	Bradycardic, reduced If	310, 837
Hcn2	Hcn2−/− (global)	Sinus dysrhythmia and reduced If	699
	Hcn2−/− (cardiac-specific)	Sinus dysrhythmia	311
Hcn3	Hcn3−/− (global)	Sinus rhythm, normal If	311
Hcn4	Hcn4−/− (global)	Embryologically lethal	1094
	Hcn4−/− (cardiac-specific)	Embryologically lethal; bradycardia; reduced If	1094
	Hcn4-R669Q/R669Q (global)	Embryologically lethal; bradycardia; slowed If	410
	Hcn4−/− (global; inducible)	Sinus pauses; reduced If	426
	KiT-Hcn4−/− (inducible)	Sinus pauses; reduced If	435
	Ci-Hcn4−/− (inducible, cardiac-specific)	Bradycardia; reduced If	74
	Hcn4–573X KI (inducible, cardiac-specific)	Bradycardia; normal If	20, 425
Voltage-gated Ca2+ current ICa
Cacna1d	Cacna1d−/−	Bradycardia; 70% reduced ICa,L; residual ICa,L due to Cav1.2	720, 751, 1344
Cacna1c	Cacna1c-DHP−/−	Dihydropyridine-insensitive Cav1.2, continued DHP effects implicating Cav1.3	1066
Cacna1g	Cacna1g−/−	ICaT due to Cav3.1 entirely lacking; slowed atrioventricular conduction; moderate bradycardia	724
Cacna1h	Cacna1h−/−	Normal ECG characteristics	186

Note: mutations involving Scn5a+/− (Scn5a) are covered in Table 4.

B. Role of the Rapid K⁺ Current, I_Kr

In rabbit SAN, I_Kr and I_Ks are activated during the AP upstroke and then deactivate relatively slowly during the end of the repolarization and diastolic depolarization phases. I_Kr likely controls AP repolarization and sets the maximum diastolic potential. Nanomolar E-4031 concentrations partially blocking I_Kr thus positively shift the maximum diastolic potential, decrease the (dV/dt)_max and amplitude, and prolong the repolarization duration of the AP (855, 1189). The resulting reduction in recruitment of other depolarizing, diastolic currents (210) then slows pacing rate. Complete I_Kr block by micromolar E-4031 terminates automaticity and leaves a depolarized (−30 to −40 mV) resting potential (855). Mouse pacemaker cells showed similar effects of E-4031-mediated I_Kr block (210). Erg1b−/− hearts lack fast I_Kr and demonstrate bradycardic episodes (see sect. VIF2; Ref. 626).

C. Role of the Hyperpolarization-Activated Cyclic Nucleotide-Gated Current, I_f

The initial phase of the diastolic depolarization that follows likely includes “membrane clock” contributions from inward, depolarizing, cAMP-sensitive I_f carried by hyperpolarization-activated cyclic nucleotide-gated (HCN) channels (122). I_f is activated within diastolic depolarization voltage ranges (−40 to −60 mV) (72, 73, 153). Diastolic depolarization may include additional contributions from time-independent background current including I_Cl (114) or TRPM4-mediated I_b (252) and decay of outward I_Kr (400). Nevertheless, studies of the I_f blockers Cs⁺ (256), UL-FS 49 (1127), ZD-7228 (733), and ivabradine on pacemaker activity demonstrated a definite role for I_f in controlling diastolic depolarization rates not only in rabbits (1126) but also mice (280, 643).

D. Murine Models with Altered HCN4 Channels

Genetic alterations involving I_f, mediated by HCN channels, are associated with both mild and severe clinical arrhythmic conditions (72). Of the four HCN isoforms, HCN4 is likely to be the major contributor to HCN-mediated pacemaker currents in adult SAN of most species, including mouse and human. Hearts with homozygous Hcn4 deletions showed marked bradycardia and decreases (75–90%) in I_f, and a residual current, possibly carried by Hcn1 and Hcn2, before lethality at embryonic day (ED) 9.5–11.5 (1094). This is consistent with HCN4 function becoming important only after this embryonic stage. In common with both global and cardiac-specific constitutive Hcn4−/− mice, Hcn4-R669Q/R669Q mice with altered cyclic nucleotide binding that would result in an inability to respond to adrenergic stimulation (410) were similarly bradycardic before dying at ED 11–12 (179, 906). Similarly, constitutive deletion of the transcription factor implicated in SAN, Shox2, permitted normal development until an onset of severe bradycardia and decreased Hcn4 expression at approximately ED 10.5 and lethality between ED 11.5 and 12.5. Similar findings occurred with Hcn4-KO mice (297).

However, hearts from adult mice with inducible Hcn4 knockouts, whether globally with a tamoxifen-inducible Cre construct (426) or specific to Hcn4-expressing cells (435), showed limited changes in cardiac pacemaker generation and modulation. Hearts showed normal mean basal heart rates, albeit interrupted by ∼8–16 sinus pauses/min with duration ∼320 ms. Yet patch-clamped SAN cells showed reduced (60–80%) I_f that nevertheless retained normal activation-voltage curves. AP activity was absent in a significant proportion of cells. However, where such activity was present, spontaneous firing rates were normal. Sympathetic challenge by isoproterenol or treadmill activity in ambulant mice produced alterations in heart rates similar to those found in control mice albeit with increased frequencies of post-exercise sinus pauses (426). However, muscarinic stimulation produced more marked bradycardic effects in Hcn4-KO than WT (426, 435).

Tamoxiphen-inducible models have been produced by crossing floxed Hcn4 mice (74) with mice carrying the Cre-recombinase controlled by the cardiac-specific α-myosin heavy chain (αMHC) promoter (1075), sparing neuronal Hcn4 channel expression. These showed a similar ∼70% reduction in I_f. But this now accompanied progressive, up to 50%, reductions in heart rates without sinus pauses, with reduced rate responses to isoproterenol challenge during telemetric recording. There was also an unexpected prolonged PQ interval and atrioventricular block, eventually proving lethal. Mouse models in which it was possible to induce an elimination of Hcn4-expressing cardiomyocytes and their substitution with collagen fibers were similarly bradycardic with a similarly reduced tachycardic response to isoproterenol adminstration on ambulant ECG recording (424). However, the elimination of HCN4-expressing cells additionally resulted in sinoatrial pauses, and atrioventricular node conduction changes increasing PR interval and producing complete heart block and supraventricular tachycardia or VT.

The hHCN4-573X mutation has also been correlated with clinical SAN dysfunction. It produces a dominant negative abolition of cAMP-mediated HCN4 modulation (20, 1028). Ambulant transgenic mice with an α-MHC promoter and a Tet-Off system-controlled cardiac-specific overexpression of Hcn4-573X (20) showed ∼20% reductions in resting heart rates but no changes in PQ and QTc intervals. They retained albeit reduced tachycardic effects of exercise. Their I_f showed slower activation kinetics and a negative, −20 mV, shift in voltage dependence. Their isolated SAN cells were either quiescent or, in a small fraction of cells, showed subthreshold membrane potential oscillations followed by regular firing at reduced rates. Isoproterenol challenge restored regular pacemaker activity but to lower rates than when applied to WT (425).

E. Murine Models With Modifications in the Remaining HCN1-3 Channels

The occurrences of HCN1 and HCN2 are species dependent, and their levels of expression are lower than that shown by HCN4. There are nevertheless reports of Hcn1 signal in the SAN (425). Recent immunohistochemical studies demonstrated Hcn1 in WT mouse SAN (310). However, a generalized knockout Hcn1−/− mouse developed normally (837). Nevertheless, Hcn1−/− mice were bradycardic with low cardiac output, sinus dysrhythmia, and recurrent sinus pauses on telemetric in vivo ECG and echocardiographic study. Single Hcn1−/− spindle and elongated SAN cells showed significantly reduced I_f amplitudes. I_f activation kinetics were slowed, consequently resembling that of cloned HCN4 channels. Isolated Hcn1−/− SAN cells accordingly showed reduced pacing frequencies (310). These findings suggest a role for Hcn1 in stabilizing SAN pacemaker function. Hcn1 was also expressed in mouse cardiac conducting tissue (425, 726). Hcn1 has also been demonstrated in rabbit SAN (804). Together with Hcn4 it could there contribute to I_f (25). Finally, Hcn1 has been implicated in rhythmic and resting neuronal activity in mouse brain (122).

Hcn2 occurs mainly in adult mouse ventricles. It is hardly detectable in atria and SAN. Nevertheless, global and cardiac-specific constitutive Hcn2-deficient mouse models showed sinus dysrhythmic features. Their resting heart rates were normal, but resting ambulatory ECGs showed increased RR variability (699). Isoproterenol and exercise challenge accomplished similar maximum rates as those in control mice and reduced the sinus dysrhythmia. Their isolated SAN cells showed slowed activation kinetics and ∼30% reductions in I_f and a −5 mV hyperpolarization in maximum diastolic potential.

Hcn3 subunits have long been regarded as little relevant to cardiac activity. Hcn3 was not observed in SAN (145, 178, 311, 425). There have been no reports of Hcn3 mRNA or protein expression in rodent SAN cells (311, 425). There are varying in situ hybridization, real-time PCR or protein expression reports in mouse and rat hearts (425). Global and constitutive Hcn3−/− knockout mice were born with normal Mendelian ratios. Ambulant ECG recording showed normal sinus rhythm. However, under bradycardic conditions, T-wave amplitudes and duration as well as QT intervals all increased. Epicardial cells showed shortened APDs and ∼30% reduced I_f. These findings suggest a role as a ventricular background current that opposes ventricular AP repolarization (311).

F. Role of Ca²⁺ Currents, I_Ca

Late diastolic depolarization may further involve contributions from depolarizing, I_CaL, and T-type Ca²⁺ channel current, I_CaT (998). It is compromised in mice homozygously lacking either L-type, Cav1.3 (Cacna1d), or T-type, Cav3.1 (Cacna1g), channels (768). Both genetic and pharmacological data implicated I_Ca,L, with its more negative threshold in SA than ventricular myocytes, in both diastolic depolarization and the AP upstroke through the actions of Ca_v1.3 and Ca_v1.2 (Cacna1c), respectively (720, 721). In vivo ECG studies in anesthetized mice demonstrated that dihydropyridine Ca²⁺ channel blockers induced bradycardia (620). Studies in genetically modified murine platforms suggested a role for Ca_v1.3 in automaticity and Ca_v1.2 in excitation contraction coupling.

Ca_v1.3 channels activated at the more negative potentials (approximately −50 mV) within the range corresponding to diastolic depolarization (720). Both intact Ca_v1.3−/− mice and Cav1.3−/− atria were bradycardic, the former even following autonomic block (901). This altered function was reflected in corresponding properties of isolated SAN (1344) and pacemaker cells, the latter showing erratic pacemaker function (720) The Ca_v1.3 inactivation produced a 70% reduction in I_Ca,L in pacemaker cells with the persisting I_Ca,L then likely arising from Ca_v1.2. Thus I_Ca in Ca_v1.3−/− pacemaker cells showed a greater dihydropyridine sensitivity (720) and faster inactivation kinetics (1344) than in WT. Maximal pacing rates in Ca_v1.3−/− hearts following isoproterenol challenge were slightly lower than those shown by WT hearts (751). This could reflect the effect of β-adrenergic activation negatively shifting thresholds for Ca_v1.3-mediated I_Ca,L, activation to approximately −55 mV (720). In contrast, in Ca_v1.2-DHP−/− mice with dihydropyridine-insensitive Ca_v1.2, dihydropyridines continued to exert in vivo bradycardic effects implicating Ca_v1.3 in this effect (1066).

The I_Ca,T inhibitors Ni²⁺ and tetrametrine slowed SAN cell pacemaker activity (399, 1008). Yet there is relatively little I_Ca,T activity through diastolic membrane voltages. Mouse SAN expresses both Ca_v3.1 and Ca_v3.2 mRNA. However, although their SAN ionic currents and automaticity have not been studied, Cav3.2−/− mice had normal ECG characteristics (186). In contrast, Cav3.1−/− SAN and AVN cells entirely lacked I_Ca,T with no evidence for a residual I_Ca,T that would be expected from the remaining Ca_v3.2. In contrast, only Ca_v3.1- and not Cav3.2-mediated I_Ca,T occurred in postnatal rat atria (315, 834). These findings suggest that Ca_v3.2 is expressed in developing heart, whereas Ca_v3.1 expression dominates the phenotype in adult heart. Spontaneous activity in isolated SAN pacemaker cells was slowed by ∼30%. Cav3.1−/− mice showed slowed atrioventricular conduction and a moderate bradycardia persistent even following pharmacological autonomic block reflecting the slowing of intrinsic SAN automaticity (724).

Finally, murine models for N-type (Ca_v2.2: Cacna1b) and R-type (Ca_v2.3: Cacna1e) channels have been investigated for changes in autonomic rather than intrinsic pacing mechanisms (479, 698, 1247).

G. Role of Na⁺ Currents, I_Na

The upstroke components of pacemaker cell APs are primarily driven by Ca²⁺_. Nevertheless, the SAN expresses both TTX-sensitive neuronal Nav1.1, likely important in pacemaking (75, 77, 635, 714), and TTX-resistant Nav1.5 that may mediate intranodal conduction (632, 635). In neonatal rabbit SAN, TTX (3 μM) slowed pacemaker activity by 63% through slowing late phase diastolic depolarization, decreasing AP thresholds and reducing overshoot. I_Na contributed a window current that declined with age (75). I_Na was activated by ramp depolarizations to extents that varied with ramp slopes. This suggested an inward current contribution resulting from incomplete inactivation of the I_Na that was initiated during the AP upstroke during diastolic depolarization (76). Na_v1.1 becomes downregulated in the adult consistent with roles in increasing basal heart rate primarily in neonates (75, 76). However, adult mouse pacemaker cells expressed both TTX-sensitive and -resistant I_Na. Nav1.1 may assume a pacemaker role in adult mice (197, 635, 722). I_Na inhibition by lidocaine reduced heart rate (620), and low, 50–100 nM, TTX concentrations reduced pacing rates in Langendorff-perfused mouse hearts (714). The roles for the effects of Nav1.5 haploinsufficiency in intranodal and sinoatrial conduction is discussed in further detail in section VF2.

H. Contributions of Ca²⁺ Homeostatic Processes

Further contributions to SAN pacing may arise from RyR2-mediated Ca²⁺ release in conjunction with the electrogenic effects of a resulting increase in I_NCX and the decay of delayed rectifier currents. Whereas ryanodine-mediated block of Ca²⁺ release reduced the rate of late diastolic depolarization, Cs⁺-mediated block of I_f reduced initial diastolic depolarization (652, 972). Distinct cytosolic, early, delayed, and late Ca²⁺ transients have been observed in SAN cells. These have been attributed to events reflecting action potential activation, nuclear Ca²⁺ change, and subsarcolemmal local Ca²⁺-induced Ca²⁺ release, respectively (523). Of these, the last has been implicated in a distinct intracellular “Ca²⁺ clock” driven by spontaneous SR Ca²⁺ release that could contribute to the pacemaker process (161, 459, 616, 719, 722, 946, 1121). Increases in RyR2-mediated Ca²⁺ spark activity that would drive this pacing mechanism might occur when a critical SR Ca²⁺ level is regained during diastole (193), or arise from particular kinetic properties of SR Ca²⁺ uptake or release proteins (132, 615, 1121, 1199).

The properties of SAN cells are consistent with such a mechanism. SAN cells both contain the requisite SR Ca²⁺ stores and express RyR2 and RyR3 (748). They also have high basal cAMP levels. This could facilitate a protein kinase A (PKA)-dependent phosphorylation of RyR2 that would enhance RyR2-mediated SR Ca²⁺ release. The consequent, appropriately timed elevations of local [Ca²⁺]_i would then enhance I_NCX whose electrogenic effects would then contribute to the observed depolarizing potentials (1198). Increases in [Ca²⁺]_i could also activate physiologically important enzymes such as calcium/calmodulin-dependent protein kinase II (CaMKII) (132, 616). This scheme would be consistent with observations in which the rapid Ca²⁺ chelator 1,2-bis(o-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA) slowed and ultimately abolished AP firing in isolated guinea pig SAN myocytes (161). Such suggestions are potentially testable through the use of genetic paradigms for RyR2 or SERCA outlined in section VII, D and E.

I. Roles of Other Charge Carriers

Of the remaining charge carriers, both murine genetic exemplars and pharmacological probes are lacking for the sustained current I_st (136, 783). Its function in SAN automaticity has been studied by modeling studies (783, 1327). Further contributions to pacing may arise from Na⁺-K⁺-ATPase (I_NaK), I_NCX, and its modification by intracellular Ca²⁺ handling (132, 136). Preliminary reports additionally implicated transient receptor potential cation channel, subfamily C, member 3 (TrpC3), the endoplasmic reticular Ca²⁺ sensor stromal interacting molecule (STIM), and the surface membrane channel Ca²⁺ release-activated Ca²⁺ channel protein 1 (Orai1), involved in store-operated Ca²⁺ entry (SOCE) in inexcitable cells in this complex regulatory mechanism (524, 670). Finally, studies in Girk4−/− mice deficient in Kir3.4 implicate I_KACh in regulation of pacemaker activity and heart rate following sympathetic stimulation (767).

IV. TISSUE CONNECTIVITY AND ARRHYTHMIC DISORDER

A. Gap Junction Proteins and Tissue Electrical Connectivity

Following initiation at the SAN, a wave of regenerative electrophysiological activity is conducted through successive cardiac structures initiating their activation (see sect. IC3). A first determinant of the conduction (θ) term in the expression for the wavelength λ of this conducted AP arises from the longitudinal resistance, r_a. The latter depends on gap junction-mediated passage of the local circuit current between cells which spreads the excitation. Each gap junction comprises two apposed and connected connexon hemichannels. The latter are protein hexamers of one of a variety of genetically distinct connexon protein isoforms. The six connexin (Cx) subunits surround an aqueous, conducting, pore electrically coupling successive cells (283, 495, 578, 606, 1192).

Of the major murine cardiac Cx isoforms (Table 3), Cx40 occurs at intercalated disks in atrial myocytes, atrioventricular node, and ventricular conduction system (247). Cx43 occurs in both atrial and ventricular myocytes and distal conduction system (495, 942, 1179). Cx45 mainly occurs in SAN, atrioventricular node, and conducting bundles (1179). Altered gap junction expression slows conduction enhancing arrhythmogenecity. It can occur by itself or can accompany alterations in other AP conduction determinants such as fibrotic change or other remodeling accompanied by excitability change (496, 803, 948, 1038, 1091). Thus Cx43 distribution and expression both alter with most human ventricular remodeling following cardiac overload. This was first demonstrated with the fiber disarray in infarct border zones (888). Early compensated hypertrophic remodeling following aortic stenosis was accompanied by increased Cx43 expression (587). HCM is accompanied by either Cx43 downregulation (888) or lateralization (1036). Dilated cardiomyopathies (DCMs) also correlate with reduced and lateralized Cx43 expression.

Table 3.

Selected loss-of-function murine cardiac connexin models

Gene	Genotype	Phenotypes	Reference Nos.
Cx30.2	Cx30.2−/−	Reduced PQ interval	590–592
Cx40	Cx40−/−	Prolonged ECG P-wave; prolonged PQ intervals; supraventricular arrhythmias; normal ventricular conduction	57, 176, 396, 573, 1026, 1065, 1172, 1191
Cx43	Cx43+/−	Normal P wave durations; reduced ventricular conduction velocities; homozygote lethal	296, 937, 1132
	Cx43- (inducible)		286
	Cx43- (cardiac specific)		389
Cx45	Cx45−/−	Normal AV conduction, i.e., normal PQ and QRS duration	596, 602
	Cx45−/− (cardiac specific)		326

B. Connexin-Deficient Murine Hearts

1. Sinus node function

Mice deficient in Cx43 (286, 389, 803), Cx40 (57, 1109, 1172, 1191), Cx30.2 (1026), combined Cx40 (596) and Cx45, or Cx30.2 (1026) showed normal sinus node function. However, Cx40−/− hearts demonstrated shifts in leading pacemaker activity from SAN to secondary sites including the sulcus terminalis, RA free wall, and right superior vena cava. Cx40−/− mice also showed prolonged ECG P-wave durations (57), PQ intervals, and QRS durations (1191). This could possibly reflect local conduction blocks prolonging intra-atrial activation path length and therefore conduction time. However, their resting membrane potentials, (dV/dt)_max, AP amplitude, or APD were normal compared with WT Cx40+/+ (57).

2. Atrial arrhythmia

Both animal and human AF occur with altered atrial Cx40 expression and distribution (820). Cx40−/− atria showed reduced RA epicardial conduction velocities (57), with prolonged ECG P-wave durations in some (57, 396, 1191) but not all reports (1109, 1172). They also showed increased incidences of supraventricular arrhythmias (176, 396, 573, 1191). In contrast, Cx43-haploinsufficient or conditional Cx43 knockouts showed normal P wave durations (286, 1132) and conduction velocities (1132). Increased proportions of Cx40 decreased, while increased proportions of Cx43 increased conduction velocity in cultured atrial myocyte strands from Cx40- or Cx43-deficient mice (85).

3. Atrioventricular conduction

Mouse atrioventricular conduction systems contain Cx40, Cx30.2, and Cx45 (496, 592, 948, 1026), as well as the murine-specific Cx30.2 (590). Cx40 knockout prolonged PQ intervals (57, 1026, 1065, 1172, 1191). This was attributable to prolonged His-ventricle interval (1026) with or without increased atrial-His intervals (1172). They also showed QRS widening and fractionation (57, 596, 1026) during anterograde but not apical ventricular pacing (1191) despite normal ventricular conduction velocities. These findings are consistent with delayed bundle branch conduction (1109, 1191). This delayed conduction was directly observable in both right (947, 1109) and left bundle branches (947).

In contrast, Cx45 haploinsufficiency did not affect ECG measures of atrioventricular conduction. PQ and QRS duration then were similar to control. However, it accentuated the effects of Cx40 insufficiency on PQ and QRS duration (596). Finally, Cx30.2 knockout surprisingly was associated with reduced PQ durations. This suggested increased atrioventricular conduction velocities attributable to decreased atrial-His as opposed to His-ventricle intervals (590, 592). In contrast, combined Cx40 and Cx30.2 deficiency resulted in normal atrial-His and His-ventricular intervals consistent with opposing actions in, respectively, increasing or decreasing conduction velocity (1026). Loss or redistribution of Cx40 accompanies other murine models of AF including overexpression of Ras homolog gene family member A of RHOA, angiotensin converting enzyme, tumor necrosis factor (TNF), and cAMP response element modulator (CREM) (539, 572, 990, 1013).

4. Ventricular arrhythmia

Complete knockout of the main ventricular connexin Cx43−/− causes perinatally lethal pulmonary outflow tract malformation (937). Cx43+/− mice, 50% Cx43-haploinsufficient, were viable but showed increased ventricular activation delays reflecting reduced conduction velocities in some (296, 1132) but not all studies (803, 948). Cultured strands of perinatally obtained Cx43+/− and Cx43−/− myocytes accordingly showed 0 and 96% reductions in conduction velocities, respectively (84). Several conditional Cre/LoxP Cx43 knockout mouse models provided expression systems with >50% Cx43 reduction (237, 389). Even very low Cx43 levels then permitted conduction velocities that were ∼50% of normal. Nevertheless, a >80% reduction in Cx43 did slow ventricular conduction, particularly in the transverse as opposed to longitudinal myocardial fiber direction. These effects were greater in the RV than the left ventricle (LV) (389, 948). They accompanied arrhythmias initiating in the RV (389, 948) and lethal ventricular tachyarrhythmias during telemetric recording (286, 389).

Computational predictions correspondingly demonstrated nonlinear dependences of conduction velocity upon junctional conductance. Gap junction uncoupling enhanced both safety factor (957) and (dV/dt)_max (1083). Consequently, large gap junction conductance reductions were required to slow conduction (520, 1043).

V. ELECTROPHYSIOLOGICAL EXCITATION AND ARRHYTHMIC DISORDER

The Na⁺ Channel, Nav1.5 (Scn5a), α-Subunit

Cardiac Na⁺ channel function is essential for myocardial excitation. In addition to voltage-dependent activation properties, it demonstrates transitions into and from fast and slow inactivated states, resulting in entry into and recovery from refractoriness to reexcitation (171, 970, 1184, 1200). These properties in turn are altered by Nav1.5 modification through a wide range of agents.

The main, Nav1.5, Na⁺ channel, α-subunit 260-kDa protein, encoded by SCN5A, comprises four homologous domains (I–IV). These are assembled around a central ion-selective pore. Each domain contains six transmembrane segments (S1–S6). Of these, S5 and S6 and their connecting loop regions form the pore module. The charged amphipathic S4 segments move towards the membrane extracellular face on depolarization. This gating transition initiates conformational changes opening the pore (878). Conformational changes in domains I–III show relatively rapid kinetics and result in ion channel opening (177). In contrast, the slower changes in domain IV permit an inactivation component connecting S6 of domain III to S1 of domain IV. This occludes the pore to cause channel inactivation (171). The α, Nav1.5, subunit thus provides the primary and often sufficient requirement for functional, voltage-gated, channel opening, ion selectivity, and channel activation and inactivation (11).

SCN5A itself shows evidence of alternative promoter usage and splicing resulting in multiple transcripts of Nav1.5 protein resulting in differing functional properties. Nav1.5 may additionally undergo posttranslational glycosylation and phosphorylation by protein kinases A and C, and modification by tyrosine kinases and phosphatases. Nav1.5 may also be a regulatory target for intracellular Ca²⁺, calmodulin (CaM), and CaM-dependent protein kinase II (CaMKII) (963), as well as ROS and the cell NAD⁺/NADH ratio (see sect. VIII, A and C). I_Na is also modifiable by either indirect, PKA-dependent, and direct, PKA-independent mechanisms. The PKA-independent response may involve stimulation through G protein subunit-α (G_sα)-caveolin-3 binding resulting in access to caveolar-associated Na⁺ channels (963).

In ventricular myocytes, Na⁺ channels may exist in three subpopulations. These have distinct properties, localizations, and associated molecules (889). Nav1.5 channels at intercalated disks were associated with β2 and β4-subunits, ankyrin-G, synapse-associated protein 97 (SAP97), and junctional proteins (889, 1006). Genetic variants reducing ankyrin-G/Nav1.5 affected intercalated disk but not sarcolemmal Nav1.5 channels. They reduced I_Na and produced human or mouse BrS (716, 787). Nav1.5 demonstrated by recordings at the lateral and tubular membranes were associated with the α1-syntrophin PDZ (postsynaptic density protein/Drosophila disc large tumor suppressor/zonula occludens-1 protein) domain of syntrophin/dystrophin through the COOH-terminal, S-I-V motif of Nav1.5 β1 and β3-subunits (1063). Mutations deleting the COOH-terminal motif of α1-syntrophin affected lateral membrane but not intercalated disk Nav1.5. However, they also reduced I_Na (1063). There have been recent suggestions of differing kinetic properties of I_Na arising from Nav1.5 at the two latter sites. There is also a tetrodotoxin (TTX)-sensitive, Na⁺ channel subpopulation unlikely to reflect Nav1.5. However, this likely makes only minimal I_Na contributions (662).

Genetic alterations involving Nav1.5 and its associated molecules are correspondingly associated with a wide variety of cardiac arrhythmic disorders. These include BrS (see sect. VB; Refs. 32, 95, 189), progressive cardiac conduction disease (see sect. VB3; PCCD; Lev-Lenegre syndrome) (737, 1021), congenital LQTS3 (see sect. VIB; Ref. 1224), SND (see sect. VF; Ref. 100), and SIDS (575, 862; editorials in Refs, 456, 633, 985). The comparative studies of these using murine models implicated abnormalities in some or all the fundamental electrophysiological processes of conduction, depolarization, and repolarization in arrhythmic substrate. Single SCN5A mutations can also accompany combinations of as opposed to single clinical symptoms. These could include BrS with cardiac conduction disorder (609, 1027, 1073) SND and/or atrial standstill (1073) and LQTS3 (see sect. VIB7; Refs. 115, 412).

B. The Brugada Syndrome

1. Occurrence and inheritance

BrS is characterized by increased risks of arrhythmogenic SCD occurring particularly in middle aged (∼40–45 yr) males (147, 289, 912). Although relatively recently identified, BrS is implicated in 4–12% of unexpected sudden deaths. It accounts for up to 20% of deaths in patients with structurally normal hearts worldwide. This suggests a population incidence of such SCD of ∼0.05% (31). There is a greater male prevalence. There are also significant worldwide variations in incidence. Thus BrS may be the most common cause of sudden death in anatomically normal hearts below age 50 yr in South Asia (749).

BrS is inherited as an autosomal dominant trait with incomplete penetrance, so genetic testing may aid diagnosis. However, the known causative mutations are demonstrable only in ∼30% of cases. The most common mutation concerns SCN5A encoding Nav1.5, in 15–30% of BrS cases. There are links to almost 300 different mutations in the gene (537). Mutations in other genes are also associated with BrS. In addition to Na⁺ channel α-subunits (SCN5A and SCN10A; Ref. 453) and β-subunits (SCN1B, SCN2B, and SCN3B) (see sect. VG, 2–4), these include glycerol-3-phosphate dehydrogenase 1-like (GPD1-L) (see sect. VIIIC). Other genes involved encode proteins associated with the Na⁺ channel. These include GTP-binding nuclear protein guanine nucleotide release factor (RANGRF), sarcolemma associated protein (SLMAP), and plakophilin 2 (PKP2). There are also K⁺ (ABCC9, KCNE3, KCNJ8, HCN4, KCND3, and KCNE5) and Ca²⁺ channel associations (CACNA1C, CACNB2B, CACNA2D1, and TRPM4) (1003). Recently, murine knockouts of the naturally occurring protein inhibitor of Kv4.3 which underlies I_to, SEMA3A, have been identified with an arrhythmic BrS-like phenotype (130). Table 4 summarizes murine models used in studies of BrS phenotypes.

Table 4.

Murine variants recapitulating conduction abnormality

Gene/Protein	Genotype	Phenotype	Reference Nos.
Na+ channel Nav1.5 (α-subunits)
Scn5a/Nav1.5; Na+ channel α-subunit	Scn5a+/−	BrS including fibrotic conduction changes	734, 867, 1095
	Scn5a-1798insD	Overlap BrS/LQT3 syndrome	208, 939, 1184
	Serine of Nav1.5 SIV motif replaced by premature stop codon causing truncated ΔSIV	Loss of binding to PDZ domain of lateral membrane α1-syntrophin; reduced lateral membrane NaV1.5 and INa	1063
Na+ channel Navβ (β-subunits)
Scn2b/Navβ2; Na+ channel, intercalated disk, β2-subunit	Scnb2−/−	Seizures; reduced neuronal Na+ channel density; cardiac phenotypes not yet examined.	185
Scn3b/Navβ3; Na+ channel, transverse tubular, β2-subunit	Scn3b−/−	Reduced Na+ channel function; monomorphic or polymorphic VT on programmed stimulation	304, 402, 403, 452, 713
Na+ channel-associated proteins
Ankyrin-G (affecting intercalated disk Nav1.5 and and spectrin)	Cardiac-selective ankyrin-G−/−	Reduced INa; altered AP waveform; bradycardia; reduced ventricular (QRS) conduction; ventricular arrhythmias following Na+ channel antagonism	716, 787, 790, 889, 1006
Synapse-associated protein-97 (SAP97) (affecting intercalated disk Nav1.5)	αMHC-Cre/Sap97-fl/fl: cardiomyocyte-specific SAP97 deletion	Reduced IK1, Ito, IKur; intact INa; slightly increased Nav1.5 expression	355
Desmosomal proteins
Plakophilin-2	Pkp2+/−	Desmosomal abnormalities; reduced and modified INa; ventricular arrhythmias	173, 174
Desmoglein-2	Dsg2-N271S	Reduced INa and conduction velocities	952.
Kv4.3 channel mediating Ito
Semaphorin 3a (Sema3a).	Sema3a−/− (KO of naturally occurring protein inhibitor of Kv4.3 (underlies Ito)	Arrhythmic BrS-like phenotype	130, 476

Note: For desmosomal proteins, see also Table 8.

2. The clinical BrS electrophysiological phenotype

BrS is characterized electrocardiographically by right precordial ST elevation, negative T waves, RV delay, and one of three types of repolarization pattern. The diagnostic type 1 pattern shows a coved ST segment elevation >2 mm followed by a negative T wave. Type 2 refers to a saddleback appearance producing a >2 mm ST segment elevation followed by a biphasic or positive T wave. Type 3 is characterized by a saddleback or coved ST elevation of <1 mm. There is also variability in clinical severity. In some cases the classical electrocardiographic type I characteristics can be elicited by administration of the I_Na blockers flecainide, ajmaline, or procainamide or appear under particular clinical or pharmacological circumstances including pyrexia (554).

BrS symptoms mainly appear during adulthood. The mean age for sudden death is around 40 yr, but there is a wide range of ∼2 days to 84 yr (33). Both the syncopal episodes and SCD are often preceded by fast polymorphic VT or VF. These often originate from the RV outflow tract (RVOT) (800). They are more common in males than females. This has been suggested to partly reflect sex-related differences in I_K and I_CaL (987, 1057). Males consequently show a more marked epicardial I_to-mediated notch often related to testosterone levels (318). In contrast, females have longer QT intervals and therefore APDs that may protect against BrS-associated arrhythmia (847) (see sect. VE).

Detailed mechanisms by which BrS leads to ventricular arrhythmia remain under discussion. Experimental studies in canine pharmacological ventricular preparations (1288) and some small clinical studies implicated a primary repolarization disorder (1051). The deep phase 1 notch, particularly in the RV epicardial AP, makes it susceptible to effects of reduced I_Na. This produces a steep RV transmural voltage gradient. This potentially causes epicardial reactivation by neighboring regions of myocardium with longer APs. The outcome would be a functional, phase 2, reentry. Alternatively, reduced I_Na might slow AP conduction and appears to do so particularly in the RVOT, in a depolarization disorder hypothesis (619, 811). This could implicate the RVOT in both the electrocardiographic abnormalities and the delayed epicardial conduction associated with potentially fatal ventricular arrhythmia in BrS (442, 766, 907). The RVOT might then act as a site for initiation of arrhythmias. The resulting regional differences in epicardial conduction velocity between the RVOT and RV might then trigger epicardial re-entrant excitation waves (766). This is discussed in detail in section VD1.

3. Anatomical abnormalities associated with BrS

Cardiac structure in BrS patients was initially thought to be normal. Several studies then reported progressive myocardial structural abnormalities (117, 940). Some patients showed a fatty replacement and fibrosis in the RVOT, despite normal LV anatomy (220). Nav1.5 changes, exemplified by SCN5A-D1275N, were also associated with ventricular dysfunction and structural changes including DCM (346, 758). The latter is consistent with known Nav1.5 interactions with cytoskeletal components (see sect. VA).

Nav1.5 abnormalities have also been associated with a heterogeneous group of inherited progressive conduction diseases (PCCD) often of unknown cause (367). The Lev-Lenegre variant manifests as progressive His-Purkinje conduction slowing with left (LBBB) or right bundle branch block (RBBB) and QRS complex widening. This can advance to complete atrioventricular block. PCCD is sometimes associated with syncope or SCD. Postmortem studies revealed diffuse fibrotic degeneration through the fibrous skeleton of the heart (644) or localized conduction tissue fibrosis (640). The earliest association of PCCD with Nav1.5 was observed in a family whose members showed progressive RBBB, LBBB, left anterior or posterior hemiblock, and long PR intervals (1021). Subsequent studies revealed a wide range of associated SCN5A mutations, as well as mutations in a range of other genes. These were implicated in loss of channel function that could result from abnormalities in channel trafficking into the membrane or gating, activation, an d/or inactivation behavior once the protein is inserted into the membrane, some of whose effects additionally were modified by Cx gene polymorphisms (737).

C. Experimental Models for Scn5a Insufficiency

1. Genetic models for Scn5a haploinsufficiency

BrS has been modeled in canine RV wedge preparations. However, these experiments entailed use of potentially nonspecific pharmacological manipulations to replicate the phenotypes under examination. Furthermore, they did not permit localization of arrhythmias between intact cardiac chambers or regions, including in particular the RVOT. Experimental systems directly replicating genetic modifications in BrS include the murine heterozygotic Nav1.5 haplo-insufficient Scn5a+/− mouse. The Scn5a+/− heterozygotes were generated by replacing exon 2 of the Scn5a gene with a splice acceptor (SA)-Gfp-PGK-neomycin cassette (867). Homozygous Scn5a−/− embryos died at midgestation. Scn5a+/− heterozygotes survived but demonstrated compromised atrial and atrioventricular conduction even at age 8–10 wk. Myocytes from 8–10 wk mice showed a ∼50% I_Na reduction compared with WT. Both the Scn5a+/− and the Scn5a-1798insD variants (see sect. VIB7) (1091) also proved useful as physiological models for both atrial and ventricular arrhythmic properties in human BrS (408, 632, 1095).

Loss of Nav1.5 function following Scn5a mutations resulted in a range of murine phenotypes that directly recapitulated clinical findings. These included SND, atrial arrhythmia, and progressive conduction disorders in addition to the ventricular phenotypes. Thus Scn5a+/− hearts replicated the ventricular, atrial, and sinoatrial arrhythmogenicity, and age-dependent fibrotic, properties observed in clinical BrS (408, 503, 632, 867, 968, 1095). Programmed electrical stimulation techniques demonstrated that Scn5a+/− hearts showed increased susceptibility to ventricular arrhythmias (867). These were accompanied by accentuations of the increases in EGD normally observed with progressively shortening S1S2 intervals (1095). The latter feature had been previously clinically associated with reentrant substrates (461, 1011, 1012). These investigations also recapitulated clinical observations of increased and decreased arrhythmic incidences following challenge by the normally antiarrhythmic agents flecainide and quinidine, respectively. These were correspondingly accompanied by accentuations or reductions in the EGD alterations (743, 1095). Studies on atrial arrhythmic susceptibility then suggested that 50% of Scn5a+/− but 0% of WT hearts had spontaneous atrial arrhythmia. This was eliminated by both flecainide and quinidine (240) (see sect. VF1). Their accompanying alterations in electrophysiological (740) and pharmacological (742) properties thus throw light on mechanisms for arrhythmia particularly in relationship to altered AP conduction.

2. Biophysical features of the murine Scn5a+/− system

Combined biophysical and molecular biological analyses separately examining the LV and RV of Scn5a+/− ventricles confirmed their selective loss of Nav1.5 function. They established a basis for clarifying the arrhythmic mechanisms that follow (Figure 5) (741). Thus Scn5a+/− ventricles showed reduced Nav1.5 mRNA and protein expression. These changes were more marked in the RV than the LV. These findings concurred with results from patch-clamp analyses of I_Na. These revealed reduced maximum I_Na (Figure 5, A and B). Scn5a+/− myocytes also showed correspondingly decreased AP upstroke velocities [(dV/dt)_max] and maximum I_Na densities than WT, changes again more noticeable in the RV compared with the LV (Figure 5C). There were concordant alterations in the density of the late Na⁺ currents (I_NaL). Despite normal steady-state dependences of activation upon test voltage (Figure 5D), I_Na showed a negative shift in the voltage dependence of its inactivation function (Figure 5E). These findings are consistent with depolarization abnormalities that would be expected to slow action potential conduction.

Figure 5.

Sodium (Na⁺) currents in Scn5a+/− hearts. A: Na⁺ currents (I_Na) normalized to cell capacitance from myocytes from left (LV) and right ventricles (RV) of wild-type and Scn5a+/− hearts. Corresponding current-voltage relationships (B), maximum I_Na (C), and activation (D) and inactivation (E) curves with Boltzmann fits are shown. *Effect of genotype. #Effect of cardiac ventricle. [From Martin et al. (741).]

In contrast, both Scn5a+/− and WT RVs showed similarly higher Kv4.2, Kv4.3, and KChIP2 expression levels than the LV. Both Scn5a+/− and WT ventricles accordingly showed greater RV than LV transient outward current (I_to) densities (741).

These findings correlate with studies at the organ level in Scn5a+/− hearts as described below. These further clarified the clinical evidence for BrS being primarily a RV condition. They additionally distinguished contributions to this property from alterations in ion channel expression and consequent electrophysiological abnormalities. These could result in abnormal depolarization (1091) or repolarization (39), fibrotic change, or reduced Cx expression (see sect. VE; Refs. 501–503, 1180).

3. Electrocardiographic features of Scn5a+/− hearts

Intact anesthetized Scn5a+/− mice showed ECG features also compatible with conduction abnormalities that could be directly attributed to reduced Na⁺ channel function. The initial studies observed a second-degree atrioventricular block and increased PR intervals (742). The changes were directly translatable to clinical observations in BrS patients with identifiable SCN5A mutations (717, 1052, 1148). Flecainide increased these PR intervals and the QRS durations particularly in the Scn5a+/− mice. This also paralleled the clinical effects of flecainide in normal and BrS patients (899, 1052). Scn5a+/− mice also showed accentuated QT dispersions, suggesting an increased repolarization heterogeneity. These were further increased by flecainide but not quinidine. This recapitulated their respective clinical proarrhythmic and antiarrhythmic effects, and their accompanying temporal QT dispersion, in clinical BrS (477, 899). The prolongations in PR and QRS intervals increased with age (968), particularly in male in contrast to female Scn5a+/− mice (501–503, 1180).

4. Action potential waveforms in Scn5a+/− ventricles

Endocardial and epicardial MAP recordings in isolated Langendorff-perfused Scn5a+/− hearts demonstrated similar RV and LV endocardial APDs that nevertheless exceeded their corresponding epicardial APDs. APDs were shorter in the Scn5a+/− compared with WT particularly in the RV. This resulted in greater RV than LV transmural repolarization gradients as quantified through ΔAPD₉₀ values, differences particularly marked in Scn5a+/− hearts. Flecainide and quinidine exerted similar patterns of pharmacological effects in Scn5a+/− and WT hearts. In both cases, flecainide shortened whereas quinidine lengthened the APDs. It did so for all four, endocardial and epicardial, and right and left, ventricular regions but particularly so for the RV epicardium. These changes accentuated the reduced RV transmural gradients that occurred even before pharmacological challenge. They thereby resulted in strongly positive RV but not LV transmural gradients in the flecainide-treated Scn5a+/− hearts which has been suggested as being potentially arrhythmogenic (see sect. VIA2). In contrast, quinidine lengthened the APDs in all ventricular regions in the Scn5a+/− hearts, particularly in RV epicardium. This decreased the RV transmural gradients which were now similar to the gradients observed in untreated WT ventricles (743).

5. Conduction and refractoriness in Scn5a+/− ventricles

Increased RV transmural repolarization gradients in Nav1.5-deficient ventricles combined with localized AP shortening could potentially reduce the AP dome. This would potentially increase the likelihood of the early epicardial phase II reexcitation previously implicated in arrhythmia in BrS (1288). However, such early reentry mechanisms would require corresponding reductions in epicardial refractory periods, if the reexcitation were to take place. Normal canine (243, 329) and human (624) hearts indeed showed a concordance between changes in APDs and changes in ventricular effective refractory periods (VERPs). However, VERPs can be selectively affected by factors that spare APD. Thus contrasting changes in VERPs in BrS were reported in clinical (50) and cell expression studies (118). Contrasting APD and VERP changes occur under other arrhythmogenic circumstances including hypokalemia (977, 978, 980).

In murine Scn5a+/− ventricles, increased arrhythmogenicity accompanied shortened APDs and increased RV repolarization gradients. The corresponding VERPs were increased rather than decreased, particularly in the presence of arrhythmogenic Scn5a+/− phenotypes (734). Both WT and Scn5a+/− hearts also showed regional VERP heterogeneities, consistent with both clinical reports (703) and known I_to differences (1124). Computational analysis confirmed that these findings were consistent with the reduced numbers of Na⁺ channels available for reexcitation in Scn5a+/− hearts (930, 1090). Refractory period would depend on recovery of a critical number of then reactivatable Na⁺ channels. This level would be achieved at a greater interval following repolarization when total channel numbers were reduced. Furthermore, notwithstanding their contrasting proarrhythmic and antiarrhythmic effects, flecainide and quinidine both increased VERP. This is in agreement with their known effects in canine, rabbit, and human hearts (148). Quinidine exerted the greater effects, likely reflecting its K⁺ channel blocking actions in contrast to the predominant Na⁺ channel blocking effect of flecainide (1231). However, increased VERPs along the patterns found in Scn5a+/− ventricles are inconsistent with an arrhythmia initiated by early local phase 2 reexcitation (1288).

Nevertheless, the findings remained compatible with roles for VERPs, particularly their spatial variations, and the consequences of these for conduction of excitation, in reentrant substrate. The Scn5a+/− hearts, particularly their RVs, in addition to the greater VERPs, showed a greater conduction slowing following S2 stimuli than did WT, as expected from their reduced I_Na. This effect was accentuated by flecainide, and, although less so, by quinidine challenge, concordant with clinical findings (907). Thus, whereas flecainide produced a more marked conduction slowing, quinidine exerted greater effects in increasing VERP. Either could potentially produce conditions of a slowed conduction or even conduction block through such partially refractory tissue with heterogeneously increased VERPs sufficient to result in arrhythmic substrate. These results paralleled other findings in LQTS (59), SQTS (775) and ischemic experimental models (487), and modeling studies (1141). LQTS (1175) and arrhythmic postinfarct patients (781) similarly show increased refractory period dispersions and greater heterogeneities in tissue excitability and refractoriness (1231).

6. Steady-state analyses of action potential stability in Scn5a+/− ventricles

These differences in AP properties between Scn5a+/− and WT translated into differing ventricular electrical instabilities during protocols applying progressively incremented steady-state pacing rates. This could be observed with decreasing BCLs ranging between 130 and 30 ms (see sect. IIA4) (754). Such protocols ended with either an onset of refractoriness and a 2:1 block, or VT or VF. Scn5a+/− ventricles showed more frequent arrhythmic end points leading to VT, particularly in the RV (752, 754). Flecainide increased the frequency of such arrhythmic events. These events then took place earlier in the pacing protocol and therefore at longer BCLs. In contrast, quinidine diminished the frequencies of such events.

Scn5a+/− and WT ventricles also demonstrated differing onsets of alternans in MAP magnitude and duration at RV and LV, epicardial and endocardial sites before and after flecainide and quinidine challenge in the course of the pacing procedures. Scn5a+/− ventricles accordingly demonstrated longer BCLs and DIs corresponding to the onset of either transient (tr) or sustained (ss) APD (DI*) and AP amplitude (DI') alternans. Following flecainide challenge, the RV epicardium showed the greatest evidence for potential instability. It then showed an onset of alternans at longer BCLs and larger amplitudes than the corresponding endocardium. This would generate transmural heterogeneities and potential discordant alternans. These findings thus provided an empirical basis for implicating the RV in the observed clinical arrhythmogenesis (752, 754).

Quantitative analysis of the corresponding restitution properties (see sect. IIA4) similarly selectively implicated the RV epicardium in the Scn5a+/− proarrhythmic phenotype. Refractory properties, reflected in the DI_ERP associated with onset of either 2:1 block or arrhythmia, were indistinguishable between corresponding recording sites in WT and Scn5a+/− before pharmacological manipulations. However, DI_ERP was increased in both RV epicardium and RV endocardium in flecainide or quinidine-treated Scn5a+/− hearts. Onsets of instability, reflected in DI values corresponding to unity slope, DI_crit, in APD versus DI plots, were similar in WT and Scn5a+/−, at each recording site. However, Scn5a^+/− RV ventricles showed steeper restitution functions than the corresponding WT. Flecainide then specifically destabilized Scn5a+/− RV, selectively increasing DI_crit (752, 754). This corroborated the alternans findings implicating RV epicardium in Scn5a+/− arrhythmia. This is consistent with suggestions that APD restitution is a predictor for arrhythmic tendency (274, 817).

In the above analyses, the relationship between alternans magnitude and restitution curve slope was a continuous second order one. This would contrast with the direct linear function containing an abrupt instability threshold predicted by voltage feedback mechanisms produced by AP repolarization and refractoriness properties (838). This prompted more complete analyses of electrophysiological instabilities which additionally explored for contributions from abnormalities in conduction velocities θ. Previous studies had implicated changes in the broadness of restitution functions representing effects on θ of similarly decreasing BCLs and consequently of DIs, as indicators of observed arrhythmia and alternans (929). These studies did not yield such findings. Nevertheless, they demonstrated that maximum achievable magnitudes of θ and DI corresponding to conduction failure correlated well with arrhythmic tendency, particularly in the Scn5a+/− RV epicardium. The latter was the case both before or following flecainide challenge. They also demonstrated θ alternans. Although less frequent than APD alternans, it abruptly appeared at pacing rates close to producing refractoriness. Finally, flecainide increased θ alternans specifically in Scn5a+/− RV epicardium (752).

7. Wavelength restitution analysis in Scn5a+/− hearts

The restitution analysis could then be further extended to consider indexes of excitation that reflected both θ, and recovery as represented by APD. This utilized the excitation wavelength, λ = APD × θ (see sect. IC3). Such wavelengths, λ, then fell nonlinearly with falling DI in agreement with previous reports (695). However, the variables in the plots of λ against DI were differently dimensioned in distance and time, respectively. They were consequently not amenable to feedback instability analysis. Nevertheless, a systems restitution analysis was made possible by expressing DI in terms of a resting, as opposed to excitation wavelength, λ₀ = DI × θ. Comparisons of λ against λ₀ gave rise to identically dimensioned plots of active and resting wavelengths. In these, the BCL variable was similarly transformed into a basic cycle distance, BCD = BCL × θ. It was then possible to analyze electrophysiological stability in terms of the wavelength parameters λ and λ₀, thus incorporating the effects of BCL upon APD, latency, and DI. The resulting, more general, instability analysis then demonstrated differing, epicardial and endocardial, maximum λ in untreated hearts at low heart rates. Quinidine decreased these in WT endocardium and abolished endocardial-epicardial differences in Scn5a+/− hearts. These findings were consistent with quinidine exerting proarrhythmic effects in WT but antiarrhythmic effects in Scn5a+/− hearts (753).

Increased heart rates then resulted in an onset of λ-alternans, failed wave propagation, and re-entrant foci corresponding to a positive feedback condition at unity gradients in the λ-λ₀ plots. This critical condition further corresponded to similar critical values of resting (λ_0crit) and total, basic cycle distance (BCD) wavelengths through both genotypes and all recording sites and all conditions of flecainide or quinidine challenge. This analysis thus yielded a common critical instability point which would thereby correspond to achievement of both necessary and sufficient conditions for initiating arrhythmia. Common instability points also implicate particular intrinsic cellular mechanisms underlying excitation and resting wavelengths. These could arise from interactions between channel opening and voltage change (1236), or other signaling that may modulate channel properties potentially through altered Ca²⁺ (204, 1137) and ATP homeostasis.

Pacing rates at which instability occurred were indistinguishable between untreated Scn5a+/− and WT hearts. However, they were decreased by both flecainide and quinidine specifically in Scn5a+/− RV epicardium, and by quinidine challenge in WT RV endocardium. Finally, over the studied pacing rates, alternans magnitude agreed more closely with the gradient of the λ than the APD restitution function. Thus analysis of λ rather than its component variables of an isolated θ or APD provides an analysis closer to mechanisms underlying alternans, thereby better predicting arrhythmia (753).

These findings are consistent with simple schemes directly relating changes in wavelength parameters to the occurrence or otherwise of wavebreak reentry leading to arrhythmia when the propagating AP encounters heterogeneities. With a long wavelength, as the AP passes over the heterogeneity, the back of the propagating wave blocks and extinguishes any retrograde excitation while orthograde excitation continues. With a short wavelength, the back of the AP wave may traverse the heterogeneity before retrograde excitation passes through the unidirectional block. This generates a new propagating retrograde, potentially reentrant excitation wave (see sect. IC3).

8. Direct visualization of reentrant circuit formation in Scn5a+/− ventricles

The two main BrS arrhythmogenic mechanisms involving slowed conduction and repolarization heterogeneity proposed in the experiments above could be both directly visualized and integrated by contact multielectrode (0.5 mm) array (MEA) mapping techniques applied to the RV and LV epicardial surfaces (736). Use of spontaneously beating hearts before and following flecainide or quinidine challenge provided more realistic indications of in vivo epicardial electrophysiological patterns following physiological activation than possible with imposed pacing. Each electrode array site recorded biphasic electrograms comprising negative followed by positive deflections. Spontaneously beating WT hearts showed neither ventricular ectopic (VE) nor VT phenomena before, and only low frequencies of these following, pharmacological challenge. The low incidences of ventricular ectopic events and VT in Scn5a+/− markedly increased with flecainide but not quinidine challenge.

Electrophysiological features derived from MEA recording could be integrated with those obtained from MAP studies. Activation (ATs) and recovery times (RTs) corresponded to points of maximal negative slope (dV/dt)_min of the initial and second negative deflections, respectively. Activation recovery intervals (ARIs), obtained from the AT-RT interval, correlated linearly with MAP measures of APD. Their dispersions were expressed as the AT, RT, and ARI differences (ATD, RTD, and ARID, respectively) between values of the first and last values in each set of electrograms of the arrays. Isochronal maps of these measures were then used to clarify the interactions between depolarization and repolarization abnormalities and the appearance and evolution of arrhythmic substrate through reentrant circuits.

The mapping studies directly demonstrated slowed conduction with dispersion of activation and lines of block in Scn5a+/− ventricles. LV and RV MEA recordings of excitation in both WT and Scn5a+/− hearts usually showed single planar wavefronts arriving at the epicardium with the intraepicardial vector directed from apex to base. However, WT ventricles showed similar LV and RV ATDs. Both increased in Scn5a+/− ventricles which additionally showed longer RV than LV ATDs. Flecainide, but not quinidine, increased the ATDs in the RVs of WT and LVs and particularly RVs of Scn5a+/− ventricles.

The mapping studies also directly demonstrated evidence for shorter, and increased dispersions of recovery in Scn5a+/− ventricles. Scn5a+/− ventricles thus showed shorter but more heterogeneous ARIs. ARIs and ARIDs indicating recovery time courses were similar in LVs and RVs of WT. However, ARIs were shorter and ARIDs increased in RVs of Scn5a+/− hearts. Flecainide decreased the ARIs whilst increasing the ARIDs. In contrast, quinidine increased ARIs and spared ARIDs in both RVs and LVs of both WT and Scn5a+/− hearts. Flecainide thus resulted in greater RV ARIDs in Scn5a+/− than WT hearts.

WT ventricles showed ARI maps closely similar in pattern to the corresponding activation maps. The longest and shortest ARIs, respectively, occurred in regions of earliest and latest regions of activation. In contrast, Scn5a+/− ventricles showed disorganized ARI maps without clear gradients. There was also greater RT heterogeneity with RTs spread over wider time courses in Scn5a+/− ventricles. Finally, WT hearts showed similar RTDs in the LV and RV that were spared by flecainide. In contrast, Scn5a+/− hearts showed increased RTDs in the LV and RV, with the RTDs additionally greater in the RV than LV. Flecainide increased these RTDs in the Scn5a+/− but not WT. Quinidine decreased them in both Scn5a+/− and WT.

These features promoted an initiation of VT, originating in the RV and arising from lines of conduction block leading to induction of reentrant circuits. Figure 6 illustrates the generation of a premature ventricular beat following arrival of activation at the RV epicardium. Results are shown for a flecainide-treated, spontaneously beating, Scn5a+/− heart leading to initiation of polymorphic VT. Such ventricular ectopics often occurred with increased RTDs in their preceding sinus beats. Figure 6, A–F, summarizes the sequence of events, mapped onto the ECG (Figure 6G, letters A–F). The last sinus beat shows close isochronal contours reflecting delayed arrival of epicardial activation (Figure 6A). Its corresponding repolarization map shows increased repolarization heterogeneities in all cases where premature beats led to the reentrant VT circuits that followed (Figure 6A″). A superimposed, premature ventricular activation produces a line of block. Impulse propagation flows around this (Figure 6B). A subsequent ventricular ectopic event initiates a circuit running anticlockwise (Figure 6C). This persists into the following beat, initiating VT (Figure 6D). VTs most frequently followed premature ventricular beats that were specifically synchronized with the T wave (Figure 6J). The continually changing line of block produces a nonstationary vortex accounting for the polymorphic nature of the VT (Figure 6, E and F). The VT then propagates as a wavefront across the LV from its initiation site in the RV (Figure 6I) (736).

Figure 6.

Re-entrant circuit initiation of ventricular arrhythmia in Scn5a+/− ventricle. A–F: right ventricular (RV) isochronal propagation maps in flecainide-treated Scn5a+/− heart illustrating initiation of ventricular tachycardia (VT). Thick black lines denote propagation block. Thin arrows denote lines of propagation. G: ECG trace with ventricular ectopic initiating polymorphic VT. H: part of the same ECG trace, with 8 electrogram traces, at the point of VT initiation. Electrogram numbers correspond to the channel numbers of the array marked in maps A–F. A: crowded isochronal lines in the last sinus beat and area of conduction slowing. A”: repolarization map of the last sinus beat with increased repolarization heterogeneity in the same area. B: premature ventricular beat superimposed on this leads to line of block with impulse propagation flowing around it. C: a second ventricular ectopic (VE) resulting in a reentrant circuit. D: the circuit continuing into the next beat to initiate VT. E and F: changes in the line of block that create a nonstationary vortex, causing polymorphic arrhythmia. I: propagation map, ECG, and electrogram traces of the VT propagating as a wave front across the LV from the RV. J: ECG trace of a VE occurring after the T wave. [From Martin et al. (736).]

D. Initiation Sites for Ventricular Arrhythmia in Scn5a+/− Hearts

1. The right ventricular outflow tract as a site for arrhythmia initiation in clinical BrS

In addition to outlining the basis for formation of arrhythmic substrate in BrS models, the previous section included evidence for preferential initiation of arrhythmia in the RV as opposed to the LV. Further evidence more specifically implicated the right ventricular outflow tract (RVOT) in this initiation of clinical arrhythmia in BrS. It also indicated that the murine Scn5a+/− system models both the human arrhythmia and its possible underlying mechanisms. The localization of the background depolarization and repolarization changes in RVOT as opposed to the remaining RV, and the LV, in BrS may reflect their differing embryological origins (1312). The RVOT originates from cells that also contribute to the slow conducting atrioventricular region (797). This may relate to suggestions for an existence of slow conducting tissue in the RVOT dependent on L-type Ca²⁺ channel rather than Nav1.5 activation (766), or of a further reduced Nav1.5 expression in the RVOT particularly with flecainide challenge (1183).

The RVOT has been implicated in both the clinical ECG abnormalities (see sect. VB) and a delayed epicardial conduction potentially associated with initiation of ventricular arrhythmia in clinical BrS (102, 218, 538, 766, 801, 907, 1148). These in turn were attributed to RVOT structural abnormalities such as fibrotic or fatty endocardial infiltration in BrS. Noninvasive clinical studies such as echocardiographic investigations for wall motion abnormalities (1148), body surface mapping (287, 491, 1148), signal-averaged ECGs (433), or tissue Doppler echocardiography indeed posed evidence for RVOT conduction delays (1148).

Some invasive RVOT endocardial studies demonstrated increased activation delays (531), electrogram prolongations, narrower ARI, and steeper restitution curves (619). A more restricted number of epicardial investigations demonstrated activation delays (811) and shortened RVOT ARIs on catheter recording in the great cardiac vein during ST elevation, as well as epicardial, but not endocardial spike-and-dome AP waveforms in BrS hearts during open chest surgery (605). Finally, explanted human BrS hearts showed VF associated with RVOT subendocardial rather than subepicardial reentry and RV activation slowing. In contrast, there was an absence of transmural repolarization gradients (218, 442). They also showed a reduced basal RV subepicardial local activation associated with the ST segment elevation (441). Both these studies also reported fibrosis and fatty infiltration in the RV. This could accentuate the current-to-load mismatch between the abnormal RVOT and more normal myocardial tissue already established by altered I_Na. This could explain the reduced conduction velocities and ECG patterns (441).

2. Murine Scn5a+/− hearts model the RVOT as an initiation site for arrhythmia in BrS

Comparisons of isolated Langendorff-perfused murine Scn5a+/− and WT ventricles provided a physiological basis for a possible RVOT involvement in the VT associated with BrS. These comparisons systematically compared arrhythmogenic tendency and the related electrophysiological parameters of VERPs, response latencies, electrogram durations (EGDs), APDs, the presence or absence of concordant or discordant alternans, and restitution curves. This involved making bipolar electrogram and MAP recordings at multiple RV and LV epicardial locations before and following flecainide and quinidine challenge (735). The bipolar electrogram measurements implicated the RVOT as an arrhythmogenic focus in Scn5a+/− hearts. Thus imposition of extrasystolic stimuli at progressively shorter intervals following regular pacing spikes elicited earlier occurrences of VT in the RVOT than the base of the LV in the Scn5a+/− but not the WT heart, which further showed only one VT episode. The same protocol demonstrated a greater degree of electrogram fractionation, previously used to assess the existence of arrhythmogenic reentrant pathways, in Scn5a+/− than WT ventricles. This was particularly so at the RVOT compared with the remaining recording sites.

The basis of these findings emerged from the MAP assessments of VERPs and conduction velocities. Both Scn5a+/− and WT showed marked apical-basal VERP gradients. However, these gradients were greater in Scn5a+/− than WT ventricles, specifically in the RV. Scn5a+/− ventricles also showed greater response latencies reflecting slowed conduction velocities than in WT. Both these differences were accentuated by flecainide challenge. Furthermore, the RVOTs showed shorter APDs and VERPs and steeper gradients of these values than elsewhere in both WT and Scn5a+/− hearts.

RVOTs of WT and Scn5a+/− hearts showed similar onsets of concordant alternans in incremental pacing protocols whether before or following pharmacological intervention. However, RVOTs of Scn5a+/− hearts showed more frequent transitions from concordant to discordant alternans. This in turn led to arrhythmogenesis, during pacing protocols involving progressively increased frequencies of stimulation. This tendency was exacerbated by flecainide and relieved by quinidine. In parallel with this, RVOTs of Scn5a+/− hearts showed the greatest heterogeneities in response latencies, and the largest APD and VERP gradients. The discordant alternans were always observed prior to any observed transitions into VT. Scn5a+/− RVOTs also showed steeper restitution curves. In these curves, the diastolic interval at which the gradient reached unity strongly correlated with the diastolic interval corresponding to the onset of discordant alternans (735).

3. Arrhythmic tendency, fibrotic change, and Nav1.5 expression in the RVOT of murine Scn5a+/− hearts

Further experiments correlated these arrhythmic features of the RVOT with key determinants for both arrhythmic tendency and conduction velocity. They demonstrated reduced Nav1.5 protein expression in both the RV and RVOT of Scn5a+/− compared with WT hearts. They also observed increased levels of fibrosis in the RVOT than the RV in both WT and Scn5a+/− (1336). Corresponding electrophysiological studies then first confirmed that stimulation at the RVOT as opposed to the remaining RV elicited increased arrhythmic incidences in Scn5a+/− but not WT hearts. Yet, Scn5a+/− showed greater VERPs than WT in both the RVOT and remaining RV. Furthermore, VERPs in the RVOT were greater in Scn5a+/− than WT, but were similar in the remaining RV in both Scn5a+/− and WT. These findings were subsequently paralleled in clinical reports (810). Both are at variance with a phase II reentry-induced arrhythmia, particularly if this were to take place in the RVOT of Scn5a+/− hearts (702, 1182) (see also sect. VC5).

In contrast, positive findings emerged from correlations of the arrhythmic properties with MEA assessments of AP conduction. First, overall AP conduction velocities were determined from simple regressions to planar fits to the local activation times of each MEA recording site. This correlated a reduced velocity with both the Scn5a+/− as opposed to the WT genotype, and in recordings from the RVOT as opposed to recordings from the RV. Second, the magnitudes of conduction velocities were averaged over the MEA recording sites by a local vector analysis. These magnitudes were reduced in the Scn5a+/− compared with WT, without variation between RV or RVOT recording sites. Third, dispersions in the conduction direction were assessed from the same vector analysis. The resulting dispersions were greater in the RVOT than in the RV, but there was then no influence of or interaction with, Scn5a+/− versus WT, genotype. Finally, these results were correlated with biochemical and histological results. Nav1.5 expression was correspondingly reduced in both RVOT and the remaining RV of Scn5a+/− as opposed to WT hearts. But interstitial fibrosis was greater in the RVOT than the remaining RV in both Scn5a+/− and WT hearts (1336).

E. The Scn5a+/− Genotype, Age-Dependent Fibrotic Change and Arrhythmogenesis

1. Age dependence of the BrS arrhythmic phenotype

BrS thus includes additional phenotypic features not directly explicable solely in terms of Nav1.5-related biophysical changes. As indicated above (sect. VB2) its clinical manisfestations are often postponed until adulthood giving an average age of SCD of 40 years (33). There is also a greater male prevalence (749). Nav1.5 haploinsufficiency is also associated with the fibrotic changes in the related condition of PCCD (Lev-Lenegre disease), which can also exist in an overlap syndrome with BrS (919) (see sect. VB3). Experimental studies in Scn5a+/− murine hearts recapitulated coexistent Nav1.5 deficiencies with fibrotic changes dependent on age and sex that would accentuate conduction and therefore arrhythmic disorders (500).

2. Sex- and age-dependent fibrotic phenotypes in Scn5a+/− hearts

Aged murine Scn5a+/− hearts demonstrated prolonged electrocardiographic P wave, PR and QRS durations, and evidence of spontaneous arrhythmic tendency (Figure 7A) (392). This accompanied a significant degree of ventricular fibrosis, decreased Cx43 expression, increases in the hypertrophic cardiac markers β-MHC and skeletal α-actin, and reductions in atrial levels of the atrial-specific gap junction protein Cx40 (642, 968). They showed elevated microarray and real-time PCR assays of Atf3 and Egr1 transcription factor (1180).

Figure 7.

Conduction and arrhythmic properties in ageing male Scn5a+/− hearts. A: lead II electrocardiographic traces obtained from anesthetized aged Scn5a+/− mouse showing spontaneous nonsustained ventricular tachycardia (VTs indicated by arrow). B: chest lead ECG complexes from young (a) and old (b) intact anesthetized male Scn5a+/− mice. The latter shows patterns of fragmented QRS complexes, indicating bundle branch block most frequently observed with Scn5a+/−. C and D: activation maps from five successive cardiac cycles in young male WT (C) and old male Scn5a+/− hearts (D). E: picrosirius red staining demonstrating ventricular fibrosis in 85-wk-old WT (a) and Scn5a+/− with mild (b) and severe fibrosis which appears red (c). F: corresponding I_Na records in ventricular myocytes from 12-wk-old WT (a) and mildly (b) and severely affected Scn5a+/− mice (c). G: frequency distributions of activation times in young male WT (a) and old male Scn5a+/− mice (b). [From Jeevaratnam and co-workers (501–503) and Leoni et al. (642).]

With epicardial MEA mapping (Figure 7, C and D), even young (3–4 mo) Scn5a+/− showed slowed RV conduction compared with the corresponding WT hearts, in an absence of demonstrable fibrosis or altered Cx43 expression. These findings were therefore primarily attributable to the direct biophysical consequences of a Nav1.5 haploinsufficiency (736). However, old WT mice showed slowed RV conduction associated with interstitial fibrosis. Old Scn5a^+/− ventricles showed slowed conduction in both the RV and LV, severe reactive fibrosis, and downregulated Cx43 expression (1180). The Scn5a+/− mice then could additionally be stratified into subgroups with severe (QRS >18 ms) and mild (QRS ≤ 18 ms) ventricular conduction defects. This stratification correlated with the extent of fibrotic change and arrhythmic severity. These findings paralleled heterogeneities also observed in human SCN5A-mutated patients (Figure 7, E and F). Ageing conserved and accentuated this stratification, together with its related reductions in Scn5a mRNA expression and I_Na. QRS durations with even the mild Scn5a+/− phenotype then exceeded that of WT (642).

Systematic, quantitative studies then analyzed the effects of, and interactions between, genotype, age (3 vs. >12 mo) and sex upon both conduction and fibrotic change in murine Scn5a+/− hearts. Electrocardiographic PR intervals were longer in young female compared with young male WT and Scn5a+/− anaesthetized mice. However, these intervals increased with age with a development of evidence for RV conduction block on chest lead recording in male but not female Scn5a+/− hearts (Figure 7B) (502). In contrast, old female Scn5a+/− and WT hearts showed similar PR intervals (503).

In isolated Langendorff-perfused WT hearts, MEA mapping of activation times demonstrated sequential propagations of RV epicardial activation from a localized region. This suggested orderly propagation of such excitation either through the plane of the epicardium or from endocardium to epicardium. In contrast, Scn5a^+/− hearts showed a more fragmented activation pattern (Figure 7, C and D; Ref. 501). This often included multiple random firing points consistent with greater likelihoods of reentrant proarrhythmic events between neighboring epicardial sites where these showed strongly different activation times (501).

Quantification of these activation times in terms of average and maximum activation delays indicated interacting effects on these of all three factors of genotype, age, and sex, with the greatest effects consequently observed in old male Scn5a^+/− hearts. It was possible first to quantify temporal variations in activation time at any given recording site. These depended on genotype and not on age and sex. They specifically gave larger variations in activation times in old male Scn5a+/− than the corresponding old WT hearts. Second, it was possible to assess spatial dispersions of activation times among recording sites within any given cardiac cycle. These were influenced by interacting effects of all the three variates of genotype, age, and sex. Old male Scn5a+/− hearts then gave greater dispersions than either old male WT or the remaining Scn5a+/− groups (501).

Analysis of frequency distributions in these activation times attributed these differences to alterations in an initially bimodal activation time distribution reflecting distinct early and late activation phases (Figure 7G). Old male Scn5a+/− showed a reduced early and a correspondingly increased late component relative to distributions shown by young male Scn5a+/−, old female Scn5a+/−, and old male WT. The corresponding morphometrically determined patterns of fibrosis then closely matched these electrophysiological findings among the experimental groups when these were sorted by genotype, age, and sex (502).

3. Regulatory consequences of Nav1.5 deficiency

Studies in murine Scn5a+/− hearts also suggested that the BrS condition itself accentuates anatomical, fibrotic, change. The latter could potentially contribute to its typically adult (∼40 yr) rather than juvenile presentation and primarily male prevalence (749). Na⁺ channels might thus exert regulatory functions distinct from their mediation of electrophysiological excitability (3, 408). As indicated previously, Na⁺ channels occur clustered with other proteins. Alterations in their properties or expression could therefore impact on expression or properties of other molecules. This could potentially include agents affecting fibroblast development. Nav1.5 haploinsufficiency correlated with upregulation of the stress inducible gene, Atf3, and the early growth response gene Egr1 (642, 968, 1180). In atrial preparations, Nav1.5 deficiency was associated with increased TGF-β₁ and vimentin transcript expression. These could potentially increase collagen and fibroblast abundance resulting in interstitial fibrosis (408). Such effects could accentuate the biophysical consequences of Nav1.5 downregulation particularly in Scn5a+/− RV (741).

4. Possible combined effects of Nav1.5 haploinsufficiency and fibrotic change in ventricular arrhythmogenesis

Nav1.5 haploinsufficiency in the murine Scn5a+/− system thus appears to result in both biophysical and structural change, particularly in the RV. This is consistent with its overlapping proarrhythmic clinical manifestations in the form of BrS and PCCD. This would account for the clinical evidence. Table 5 summarizes the resemblances between human BrS and the murine Scn5a+/− model. The experimental findings suggest that it is the overlap of these changes that together resulted in accentuated arrhythmic substrate sufficient to produce spontaneous arrhythmic events particularly in aged male Scn5a+/−hearts.

In such a scheme, the Nav1.5 deficiency would produce a background electrophysiological ion channel abnormality compromising conduction of excitation. This itself produces arrhythmic substrate if unmasked by flecainide or ajmaline challenge. This underlying phenotype could then interact with a coexisting tissue disruption arising from fibrotic change with age. This would be the effect of Na⁺ channels exerting regulatory or metabolic functions in addition to those of electrophysiological excitability particularly in male Scn5a+/− resulting in preferential expression of slowly at the expense of rapidly conducting pathways. Fibrotic change would compromise ephaptic, gap junction-mediated, myocyte-myocyte coupling. It could thereby compromise AP conduction between cells particularly with the accompanying Nav1.5 downregulation (741). It would also distort tissue geometry dispersing conduction through alternative conducting branches prolonging conduction pathlengths (see also sect. IXA3).

Arrhythmic tendency would then be the consequence of a combination of Nav1.5 haploinsufficiency and an evolving structural change in Nav1.5-associated arrhythmic disease including both progressive cardiac conduction defect and BrS. The background Nav1.5 haploinsufficiency would produce arrhythmic substrate typically unmasked by flecainide or ajmaline challenge that accentuates an underlying conduction defect (500). Cardiac fibrotic changes with age, accentuated in males, could then accentuate the arrhythmic substrate to an extent sufficient to lead to arrhythmic events (Figure 8). Such suggestions based on studies in murine hearts have proven translatable into clinical findings (810).

Figure 8.

Development of arrhythmia resulting from a combination of Nav1.5 haploinsufficiency and structural change in progressive cardiac conduction defect (PCCD) and Brugada syndrome (BrS). Nav1.5 haploinsufficiency produces a background electrophysiological defect in conduction. This results in arrhythmic substrate typically unmasked by flecainide or ajmaline challenge. Cardiac fibrotic changes occur with age, particularly in males. This further compromises action potential propagation. Superimposition of the two factors sufficiently compromises conduction, thereby accentuating arrhythmic substrate to lead to arrhythmic events. There is thus a combination of biophysical and structural change with age, particularly in males, that results in arrhythmia. [From Jeevaratnam et al. (500).]

A recently developed SCN5A-E558X porcine model provided a large animal system with features that were comparable with the arrhythmic phenomena in murine hearts (868). The porcine model showed heart rates, cardiac size, AP shape, and autonomic innervations as well as cardiac depolarization and repolarization kinetics similar to that of humans. Langendorff-perfused porcine SCN5A-E558X heart reproduced many clinical loss-of-Nav1.5 function phenotypes. They thus confirmed many features modeled by murine Scn5a+/− hearts. They shared increased incidences of pacing-induced or spontaneous VF as well as diminished Nav1.5 protein expression and I_Na. However, the porcine hearts did not show the macroscopic and microscopic structural changes shown by murine Scn5a+/− hearts (see sect. VB3). Such a system could nevertheless complement the mouse models. Nevertheless, their ease in genome manipulation, and shorter reproductive cycles (∼20 vs. 111 days gestation), will favor murine models particularly in disease occurring in later adulthood given longer porcine (between 15–17 yr) compared with murine (∼48 wk) life expectancies.

F. Nav1.5 Haploinsufficiency and Abnormal Sinoatrial and Atrial Electrophysiology

1. Physiological changes accompanying atrial and sinus node disorder

Loss of Na⁺ channel function is also associated with SND and atrial tachyarrhythmia. SAN pacemaker function in both humans and other mammals normally declines with age, decreasing intrinsic heart rates and increasing SAN conduction times (272, 731). Sinus node dysfunction (SND) causes sinus bradycardia, pause or arrest, atrial chronotropic incompetence, and SAN exit block. Its incidence increases exponentially with age, accounting for ∼50% of the million permanent pacemaker implants per year worldwide (272, 731). It may involve a range of ion channel changes (see sect. IIIA). Ageing guinea pig SANs showed reduced Cav1.2 expression reflected in a greater sensitivity of their electrical activity to L-type Ca²⁺ channel block. They also showed decreased Cx43 expression in the vicinity of the SAN (517). This could increase SAN conduction time and incidences of SAN exit block (518). Ageing rat SANs showed decreased Kv1.5 expression that might explain increased SAN APDs (1115). They also showed altered RyR2 expression. This could also compromise SAN pacemaking through actions on its Ca²⁺ clock mechanism of pacemaking (1115, 1298)

2. Sinus node disorder in murine Scn5a+/− hearts

Familial SND is also associated with ∼13 human Nav1.5 mutants (634, 637). These were variously associated with apparently normal (e.g., SCN5A-L212P, P1298L, DelF1617, and R1632H), reduced (e.g., E161K, T220I, and D1275N), or undetectable (e.g., T187I, R878C, G1408R, and the truncated variants W1421X, K1578fs/52, and R1623X; Refs. 379, 1339) peak I_Na in expressed heterologous systems.

Physiological evidence implicated both neuronal Nav1.1 and cardiac Nav1.5 channels in normal SAN function, suggesting that Nav1.1 was involved in pacing activity, and Nav1.5 was primarily involved in AP conduction through the SAN and from the SAN to surrounding atrial myocytes. It also suggested that these separate but related functions predominated in the smaller central pacing cells and larger conducting peripheral SAN cells, respectively. Nevertheless, Na_v1.5-mediated I_Na can modify heart rate by modifying impulse propagation within the SAN and from the SAN to the atrium (632).

Voltage-clamp studies of I_Na, and SAN tissue staining with anti-Nav antibodies distinguished TTX-sensitive I_Na associated with Na_v1.1 and TTX-resistant I_Na associated with Na_v1.5 expression (632, 635). In intact WT mouse hearts, selective Nav1.1 block by low (50 nM) TTX concentrations reduced pacemaker rates in isolated SAN and isolated SAN pacemaker cells (635, 714). The TTX-sensitive I_Na could be demonstrated by action potential clamp experiments to normally occur late in diastolic depolarization and during the AP upstroke (635). Block of both Nav1.1 and the relatively TTX-resistant Nav1.5 by mM TTX concentrations additionally slowed or even blocked AP conduction through the SAN periphery from its leading pacemaker site, and to surrounding atrium (635). Additionally, although Scn5a+/− mice retained normal circadian variations in telemetrically determined heart rates, they replicated the reduced overall mean rates and SA block observed in clinical SND (52).

Langendorff-perfused Scn5a+/− hearts showed sinus bradycardia, slowed SA conduction, and sinoatrial exit block. Isolated Scn5a+/− SAN and atrial preparations showed slowed SA conduction and frequent SA conduction block in mapping experiments (Figure 9A) (632). These findings correlated with results from patch-clamped Scn5a+/− SAN cells which showed normal steady-state Nav1.5 activation and inactivation characteristics but reduced maximum I_Na compared with WT. These whole animal, tissue, and cellular findings could then be unified by modeling of the resulting roles for Nav1.5 in AP propagation through the SAN and from SAN to atria. These replicated the changes in pacing rate and demonstrated that these changes were brought about by effects of loss of Nav1.5 function upon the electrical coupling between peripheral SAN and atrial cells (632).

Figure 9.

Variations in conduction properties through the isolated sinoatrial node of Scn5a+/− hearts with age. A: electrograms and activation mapping in young WT (a) and young Scn5a+/− (b) as well as old WT (c) and old Scn5a+/− atria (d). B and C: sinoatrial node (SAN) cycle lengths (B) and sinoatrial conduction times in the four experimental groups (C). D–F: fibrosis-regulating, TGF-β1, vimentin, and Nav1.5 gene expression and interacting effects of ageing and genotype. D: mRNA abundance assessed by real-time polymerase chain reaction. E: correlations between gene expression of TGF-β1 and Nav1.5 (a), vimentin and Nav1.5 (b), and vimentin and TGF-β1 (c). F: TGF-β1 gene expression initiated by Nav1.5 inhibition with Nav1.5 antibody. TGF-β1 gene expression was increased by Nav1.5 antibody treatment at the transcriptional (a) and protein levels (b) in both human neonatal myocytes and fibroblasts. [From Hao et al. (408).]

Ageing and associated fibrotic processes interacted with Nav1.5 haploinsufficiency in modifying SAN electrophysiological properties. This finding paralleled the associations between Scn5a+/− and ventricular fibrotic changes outlined above (see sect. VE4). Nav1.5 disruption increased P-wave duration, RR, PR, and QRS intervals, and SAN recovery times. It interacted with ageing in decreasing heart rate variability with the greatest effects in old Scn5a+/− mice. Aging and Nav1.5 insufficiency exerted both individual and interacting effects in increasing cycle lengths and SA conduction times in ex vivo, isolated, SAN preparations (Figure 9, B and C) (408), as well as in Langendorff-perfused hearts (391).

Both the Scn5a+/− phenotype and ageing correspondingly resulted in downregulated Nav1.5 expression to the following sequence of extents: old Scn5a+/− > old WT > young Scn5a+/− > young WT. They interacted in increasing collagen and fibroblast levels and the corresponding expression of the fibrotic modulator transforming growth factor-β₁ (TGF-β₁) and the fibroblast marker vimentin. The greatest effects were observed in aged Scn5a+/− (Figure 9D, a and b). Nav1.5 expression levels consequently negatively correlated with TGF-β₁ and vimentin levels (Figure 9E). Increased TGF-β production by both cardiac myocytes and cardiac fibroblasts could be replicated by acute application of either Nav1.5-E3 antibody or Na⁺ channel blocker (Figure 9F). This was consistent with previous reports that pharmacological block of electrical activity in Na⁺ channels upregulated TGF-β activity in rat primary myotubes (1159). Such activity could form a cell signaling pathway leading to fibrotic change and atrial electrophysiological pathology in other clinical circumstances (454). Nav1.5 haploinsufficiency in combination with aging was also accompanied by altered expression in a wide range of ion channels and regulatory genes in the SAN associated with ageing. Ageing itself accentuated the reduction in Nav1.5 expression found in Scn5a+/−. There were accompanying reductions in the tbox transcription factor Tbx3. This is known to regulate the SAN pacemaker gene expression program and decrease expression of a wide range of SAN ion channels.

3. Atrial arrhythmia in murine Scn5a+/− hearts

Finally, BrS (10–30%) is associated with increased tendency to AF. This often clinically precedes its ventricular manifestations in patients with a severe type I phenotype (26, 607). Langendorff-perfused Scn5a+/− hearts similarly showed an increased atrial arrhythmogenicity (240). This was greatest in young Scn5a+/− atria. Scn5a+/− atria particularly from aged hearts showed slower AP conduction whether measured by MEA mapping or PR intervals (391, 503). Young Scn5a+/− hearts showed the most prolonged monophasic APDs (240). However, the aged Scn5a+/− atria showed increased AERPs and therefore showed the smallest APD₉₀/AERP. Atria from young Scn5a^+/− hearts also showed the greatest prolongations in electrogram duration following extrasystolic stimulation. Slowed conduction and increased APD₉₀/AERP ratio thus correlate with the increased AF tendency particularly shown by young Scn5a+/− hearts (390).

G. Arrhythmic Changes With Abnormalities in Na⁺ Channel β-Subunits

1. Features of Navβ subunits

SCN5A forms macromolecular or signaling complexes with a number of further proteins (3, 852). These could directly or indirectly modify I_Na, Nav1.5 localization, or both (759). Of these, Navβ subunits are prominent integral, single-spanning, membrane glycoprotein components of Nav signaling complexes (844). Their extracellular NH₂-terminal region contains a V-type immunoglobulin domain with a short ”neck“ connected to a single α-helical transmembrane domain and intracellular COOH-terminal region (170, 484, 759, 1111). Navβ1 and Navβ3 have the more closely related primary sequences. They noncovalently bind to the α-subunit possibly at similar COOH-terminal sites close to domain IV. This contrasts with the disulfide bonding suggested for Navβ2 and Navβ4 (239). Atomic resolution X-ray crystallographic studies of human Navβ3 and Navβ4 suggest that Navβ3 and therefore possibly Navβ1 could trimerize. If so, Nav α-subunits could correspondingly coassemble into oligomeric complexes stabilised by Navβ3 (or Navβ1) subunits. This assembly could then potentially have effects upon electrophysiological properties through possible interactions between component Nav1.5s (765, 814).

Navβ subunits do not themselves mediate ion permeation. They rather perform multiple functions in cell adhesion, recruitment of scaffolding or anchoring proteins including ankyrins (485, 759). They may also be involved in trafficking Na⁺ channels to the cell membrane, thereby increasing peak I_Na, and negatively shifting steady-state activation and inactivation consistent with biophysical effects of providing screening charge (138, 231, 403). Navβ subunit mutations are associated with inherited conditions including epilepsy, neuropathies, cardiac conduction diseases, and some cancers (815).

Navβ1 and Navβ3 subunits occur in transverse tubular and surface membranes, whereas Navβ2 and Navβ4 subunits occur at cell membranes in the regions of the intercalated disks (713). Navβ2 and Navβ3 are discussed here in terms of the arrhythmic effects of Na⁺ channel loss of function. Loss of Navβ1 function has been associated with experimental LQTS, despite clinical BrS-like phenotypes, and is discussed in section VIB7. Navβ4 subunits produce negative shifts in activation (353, 1310). Their loss of function is associated with arrhythmogenesis accompanied by increased late I_Na (10). They are discussed in section VIB8 also in connection with LQTS.

2. Scn2b-encoded Navβ2 subunit variants

Scn2b-encoded Navβ2 subunits occur in intercalated disks particularly in the epicardium of human atria and ventricles (336, 713) in association with SCN5A (261). SCN2B-R28Q and SCN2B-R28W patients often show lone AF (1234). SCN2B-R28W patients additionally show prolonged PR intervals and right precordial ST elevation. Oocyte Navβ2 coexpression with neuronal Nav1.1 α-subunits increased I_Na, but left I_Na kinetics unchanged (484). It exerted negative shifts in voltage dependences of both activation and inactivation. These alterations were additive to the Navβ1 effects, to extents dependent on glycosylation (514). Navβ2 reduction or absence decreased I_Na. In CHO expression systems, SCN2B-R28G reduced I_Na density and shifted activation; SCN2B-R28W positively shifted inactivation. In both cases, I_NaL was not affected (1234). Scnb2−/− mice showed seizures accompanied by reduced neuronal Na⁺ channel density (185). However, the possible cardiac phenotypes have not been examined.

3. Scn3b-encoded Navβ3 subunit variants

SCN3B-encoded Navβ3 subunits occur in both ventricles (452) and atria of human hearts (852). Navβ3 subunits may function in both trafficking of SCN5A to the surface membrane and modifying its electrophysiological properties (482). Their presence increases, and absence reduces, I_Na. There are variable effects on I_Na voltage dependence and kinetics. Heterologous oocyte Navβ3 coexpression with SCN5A increased I_Na threefold, accelerated recovery from refractoriness, and positively shifted steady state voltage-dependent inactivation whilst sparing activation properties (304). Human SCN3B-L10P is implicated in the BrS7 variant (452), SCN3B-R6K, L10P and M161T in familial AF (852), and SCN3B-V54G in idiopathic VF (1165). Studies in TSA201, CHO-5, or HEK-293 expression systems related these conditions to reduced I_Na and reduced cell surface SCN5A expression. In the last case, these features were partly rescued by coexpression with Navβ1 (452, 1165). SCN3B-V36M and SCN3B-A130V are associated with SIDS and AF, respectively. Studies using HEK-293 expression systems related these to decreased I_Na with either increased late relative to peak I_Na (SCN3B-V36M) or without decreased SCN5A surface expression (SCN3B-A130V) (1110, 1221).

Scn3b-encoded Navβ3 subunits also normally occur in mouse ventricles (403, 452) and atria (402, 852) and sheep ventricular and Purkinje but not atrial tissue, where they are similarly associated with SCN5A (304). Homozygous Scn3b−/− hearts showed the expected absence of Scn3b mRNA but showed elevated Scn1b mRNA, and increased RV Scn5a mRNA. Isolated, Langendorff-perfused Scn3b−/− hearts showed increased incidences of monomorphic or polymorphic VT on programmed electrical stimulation. Their accompanying electrophysiological abnormalities showed both parallels with and contrasts from those of Scn5a+/− hearts (403). They showed a consistently reduced Na⁺ channel function compatible with BrS-like phenotypes (Figure 10). Whole cell patch-clamped Scn3b−/− myocytes showed reduced maximum peak I_Na (approximately −60 vs. −90 pA/pF in WT) (Figure 10A). This suggested reduced membrane expression or proportions of functional Na⁺ channels. Inactivation curves were shifted in the negative direction relative to those in WT (Figure 10B). However, the time courses of recovery from inactivation were not changed (Figure 10C).

Figure 10.

The Scn3b−/− exemplar for cardiac arrhythmogenesis. A: I_Na traces from WT (a) and Scn3b−/− myocytes (b) showing peak I_Na that is significantly smaller in Scn3b−/− myocytes. B: Boltzmann fits to steady-state voltage dependence of activation (squares) and inactivation (circles) in WT (filled symbols) and Scn3b−/− (open symbols). Differing voltage dependence of inactivation in Scn3b−/− compared with WT particularly between holding voltages of −70 and −40 mV. C: recovery from inactivation in WT (filled symbols) and Scn3b−/− (open symbols). D: bipolar electrogram waveforms from programmed electrical stimulation at the longest (a, c) and shortest (b, d) S1–S2 intervals in WT (a, b) and Scn3b−/− hearts (c, d). The latter hearts showed consistently longer waveforms. E: representative monophasic action potential (MAP) recordings from the WT (a) and their shortening in Scn3b−/− hearts (b). F: ECG recordings obtained from WT (a) and Scn3b−/− mice (b). Scn3b−/− mice showed slower heart rates and prolonged PR intervals. c: Some ECGs from Scn3b−/− mice showed ventricular QT complexes occurring independently of regularly occurring atrial P waves, i.e., third degree heart block. G: atrial tachycardia (AT) resulting in regular deflections at a higher frequency (a) and atrial fibrillation (AF) resulting in irregular deflections at a higher frequency following atrial burst pacing (ABP) in Langendorff-perfused Scn3b−/− hearts (b). The ventricular spikes result in the larger, and the atrial spikes in the smaller deflections. c: PES-induced VT beginning as a monomorphic then deteriorating into polymorphic VT in a Scn3b−/− heart preparation. [From Hakim and co-workers (402–404).]

Bipolar electrogram recordings obtained during programmed electrical stimulation yielded deflections whose latencies mapped onto demonstrated conduction curves, suggesting lengthened electrogram durations compared with WT, particularly at the shortest S1S2 intervals. This suggested reduced AP conduction velocities (Figure 10D). Scn3b−/− ventricles also showed shorter VERPs than WT. MAP recordings demonstrated reduced endocardial and epicardial APDs. However, the resulting endocardial-epicardial differences, ΔAPD₉₀s, were indistinguishable from those of WT (Figure 10E). Finally, both flecainide and quinidine selectively reduced arrhythmic incidence in Scn3b−/− relative to WT hearts without affecting ΔAPD₉₀, likely through increasing VERP (404).

Sinoatrial and conduction properties in Scn3b−/− complemented these findings. They also suggested a compromised Na⁺ channel-mediated excitability and conduction. Scn3b mRNA and protein were expressed in the atria of WT but not Scn3b−/− hearts. ECGs from Scn3b−/− mice showed slower heart rates, and prolonged P waves and PR intervals (Figure 10F). Spontaneously active Langendorff-perfused Scn3b−/− hearts showed abnormal atrial electrophysiological properties and partial or complete atrioventricular block. Scn3b−/− hearts also showed atrial tachycardia and fibrillation on atrial burst pacing and longer sinus node recovery times than WT hearts (Figure 10G, a–c) (402).

H. Nav1.5 and Scaffolding Proteins

Nav1.5 is also associated with scaffolding proteins. In addition, two PDZ domain-scaffolding proteins, SAP97 and α1-syntrophin, mediate reciprocal interactions upon cardiomyocyte membrane expression levels of Nav1.5 and Kir2.1; the latter underlies the inward rectifying I_K1 (1262). Recent evidence from effects of inheritable mutations suggest similar additional interactions involving plakophilin-2, ankyrin-G, dystrophin, syntrophin, and caveolin-3 (1262).

Of these, 1) caveolin-3 (Cav-3) promotes formation of cavelike membrane structures (65) and organization and concentration of particular molecules interacting with caveolin. Its precise association with Nav1.5 is uncertain (861). It may produce an adrenergic I_Na upregulation mediated by G proteins (1301). Cav-3 mutations are associated with HCM (415) (see also sect. VIC3). 2) Mutations in the cytosolic scaffolding protein α1-syntrophin (SNTA1) (12) increase I_NaL. They are associated with LQTS12 (1158) and SIDS (194). These are considered in detail in section VIC1.

3) The scaffolding protein ankyrin-G may connect ion channels, including Nav1.5, to the spectrin-actin cytoskeleton (99). Ankyrin polypeptides are known to be critical to ion channel and transporter targeting in both excitable and nonexcitable cells. Ankyrin-G interaction with Nav1.5 is disrupted by the SCN5A-E1053K mutation resulting in BrS (787). Studies on cardiac-selective ankyrin-G−/− mice suggested that ankyrin-G is a functionally important intercalated disc receptor for both Nav1.5 and βIV spectrin. The COOH-terminal domain of βIV spectrin in turn associates with the Nav1.5 regulator CaMKIIδ. Ankyrin-G-cKO mice were bradycardic and showed cardiac conduction abnormalities, QRS prolongation, and ventricular arrhythmias following challenge by Na⁺ channel antagonists. Their myocytes showed reduced I_Na and decreased AP amplitudes and (dV/dt)_max. Nav1.5 expression was decreased particularly at intercalated discs. This accompanied a reduced βIV spectrin recruitment to intercalated disc membranes. In common with mouse qv4J hearts carrying βIV spectrin lacking its COOH-terminal domain, they showed defective CaMKIIδ targeting and defective CaMKIIδ regulation of I_NaL. There was additionally a reorganization of the resident desmosome protein plakophilin-2. The latter is critical to intercalated disc integration with the intermediate filament-based cytoskeleton and lethal arrhythmias in response to β-adrenergic stimulation (716). However, such reduced ankyrin-G expression did not affect peripheral sarcolemmal Nav1.5.

4) The desmosomal proteins plakophilin 2 (PKP2) and desmoglein-2 (DSG2) are localized to intercalated discs associated with Nav1.5 (1007). Mutations in both are associated with human ARVC (see sect. IXD, 2 and 5) (251). However, VF and SCD often occur prior to the onset of morphological change. Experimental studies attributed this to cross-talk between desmosomes and the Na⁺ channel complex. Murine Pkp2+/− hearts showed ultrastructural but not histological or anatomical abnormalities, with sporadic or absent desmosomes, and unevennesses in intercellular spaces. I_Na was reduced in maximum amplitude in both Pkp2+/− and with a number of heterozygous missense mutations (173). Voltage dependences of activation were unchanged, but there were negative shifts in steady-state inactivation and slowed recoveries from inactivation. Flecainide affected electrocardiographic parameters in anesthetized animals and ventricular conduction in Langendorff-perfused hearts. It reduced myocyte I_Na to greater extents in Pkp+/− than WT. It triggered ventricular arrhythmias and death in anesthetized Pkp2+/− but not WT animals (174). Hearts from Dsg2-N271S mice similarly showed reduced I_Na and conduction velocities (952).

5) In common with α1-syntrophin, the membrane-associated guanylate kinase (MAGUK) protein synapse associated protein-97 (SAP97) includes PDZ domains. SAP97-M861T has recently been associated with BrS (1262). However, myocytes from a murine model with a cardiomyocyte-specific Sap97 deletion showed reduced I_K1, I_to, and I_Kur but unaltered I_Na yet slightly increased Na_V1.5 protein expression, indicating a requirement for further experiment (355).

I. Idiopathic Ventricular Fibrillation

A small proportion (∼5%) of sudden adult death cases had classically occurred in an absence of identifiable hereditary causes. However, recent reports indicate that a significant proportion of these share the J-point or ST-segment elevations with BrS (784). Some of these cases may be associated with abnormal Ca²⁺ (CACNA1C, CACNB2, and CACNA2D1) or K_ATP channel (KCNJ8) genes (156). Some of the remaining cases of idiopathic VF may involve the DPP6 gene which regulates I_to (932). No murine genetic models yet exist for these conditions.

VI. RECOVERY FROM EXCITATION AND ARRHYTHMIC DISORDER

A. Action Potential Repolarization and QT Prolongation

1. Arrhythmogenic clinical long QT syndromes (LQTS)

In addition to altered AP activation and conduction (see sects. IV and V), arrhythmic substrate can also arise from alterations in AP repolarization and refractoriness. The latter can arise from alterations in Na⁺ channel recovery from fast or slow inactivation and development of late Na⁺ currents, I_NaL. It could also result from changes in the magnitude or time course of Ca²⁺ and K⁺ currents that normally are activated following the initial Na⁺ channel activation that triggers the AP. AP repolarization and recovery from refractoriness can thus be prolonged by persistent or increased inward depolarizing, plateau I_Na or I_CaL, or compromised repolarizing, I_K, currents. The resulting delayed myocardial repolarization clinically results in LQTS (563, 692). The characteristic prolonged ECG QT intervals accompany an increased predisposition to atypical ventricular, torsades de pointes, arrhythmias, and SCD (549). The incidences of these phenomena appear to correlate with the extent of this QT prolongation (918). Some LQTS variants are also associated with early onset AF (515).

Acquired forms of LQTS can follow a number of clinical conditions as well as a wide range of medications (37, 313, 334). The latter often produce LQTS through reducing I_Kr (953). Hereditary LQTS were first described as the autosomal recessive Jervell and Lange Nielsen (JLN) syndrome accompanied by congenital sensorineural deafness (506) and the more common autosomal dominant Romano-Ward (RW) syndrome (549, 961, 1232). However, only ∼50% of LQTS patients have known mutations (1123a, 1197). The remainder may be associated with genetic conditions causing abnormal ion channel function often with incomplete penetrance and phenotypic overlaps (144, 224, 358, 1031, 1253). LQTS has been associated with genetic modifications in Na⁺ channel α-subunits (Nav1.5 encoded by SCN5A, giving LQTS3), their associated Navβ4 subunits (Navβ4 encoded by SCN4B: LQTS10), various K⁺ channel α-subunits mediating I_Ks (Kv7.1 encoded by KCNQ1: LQTS1), I_Kr (Kv11.1, encoded by KCNH2: LQTS2), I_K1 (among others, Kir2.1, encoded by KCNJ2: LQTS7 or Anderson-Tawil syndrome) and I_KAch (Kir3.4, encoded by KCNJ5: LQTS13), the K⁺ channel β-subunits MinK and MiRP (encoded by KCNE1 and KCNE2: LQTS5 and LQTS6), and Ca²⁺ channel α-subunits [Cav1.2 encoded by CACNA1C: LQTS8 or the Timothy Syndrome (TS)]. LQTS is also associated with abnormalities in the cyoskeletal proteins. The latter include caveolin-3 (encoded by CAV3: LQTS9), A-kinase anchor protein 9 (encoded by AKAP9: LQTS11), and α1-syntrophin (encoded by STNA1: LQTS12). All these variants can present as a Romano-Ward syndrome. Some LQTS1 and LQTS5 can also present with Jervell and Lange-Nielsen syndrome. The clinical prevalence of inherited LQTS is ∼0.01% of the population. The most common, LQTS1, LQTS2, and LQTS3, account for 45, 45, and 5%, respectively, of inherited and genotype-confirmed LQTS cases (1030).

2. Electrophysiological phenomena associated with LQTS arrhythmia

Clinical and experimental evidence associates LQTS-mediated arrhythmogenesis, exemplified in murine models of the kind listed in Table 6, with two potentially proarrhythmic electrophysiological abnormalities (827). First, the normally smooth time course of AP repolarization can be interrupted. Where this follows phase 2 plateaus, EADs result. EADs were experimentally observed following exposure to drugs implicated in acquired LQTS, or catecholaminergic or hypoxic challenge, and in genetically modified murine hearts modeling human LQTS, particularly under bradycardic conditions (1128). EADs have been clinically observed in both bradycardic hypokalemic conditions (1058, 1059) and in congenital LQTS patients (1058). EADs have been attributed to a prior conditioning by plateau currents under conditions of AP prolongation combined with L-type Ca²⁺ channel recovery from inactivation within the voltage-window range of such recovery. The resulting L-type Ca²⁺ channel reactivation then produces depolarizing current. If it involves sufficient numbers of cardiomyocytes, this produces ectopic beats. The latter can trigger polymorphic VT when they occur within a proarrhythmic myocardial substrate (498, 499). Other suggestions for mechanisms producing EADs invoked increased I_NaL (416) or I_NCX (1204).

Table 6.

Transgenic murine models associated with long QT syndromes

Protein/Gene	Genotype	Ion Current	Clinical Condition/Phenotype	Reference Nos.
Channel α-subunits: Na+ channel variants
Nav1.5/Scn5a	Scn5a-ΔKPQ/+	INaL	LQTS3; prolonged APD	416, 843, 1095, 1128, 1129
	Scn5a-F1759A	INaL	Atrial and ventricular arrhythmia	1214
	Scn5a−1798insD/+	INaL	Overlap syndrome: SND, conduction disease, BrS and LQTS3	115
Channel α-subunits: K+ channel variants
Kv4.2/Kcnd2	Kcnd2-DN	Ito,f	Prolonged APD	558, 687
Kv1.4/Kcna4	Kcna4−/−	Ito,s	Normal APD	558, 687
Kv4.2/Kcnd2 and Kv1.4/Kcna4	Kv4.2-DN × Kv1.4−/−	Ito,f and Ito,s	Prolonged APD	558, 687
Kv11.1/Kcnh2	Merg1−/−		Embryologically lethal	53, 688
	Merg1+/−	?IKr	LQTS2; prolonged apical and basal APDs and VERPs	994
	Merg1b−/−	?IKr	Normal APD; episodic sinus bradycardia	627, 1307
	hERG-G628S (dominant negative) expression	?IKr	Normal QT durations	53
Kv7.1/Kcnq1	Kcnq1−/−	IKs	LQTS1; Jervell and Lange-Nielsen syndrome; prolonged QT intervals in intact animal	167, 1138, 1222
	Human KCNQ1 isoform 2 overexpression	IKs	Prolonged QT interval (dominant negative effects on murine Kcnq1 isoform 1)	254
Kv1.5/Kcna5 (ventricular)	Cardiac Kv1.1 NH2-terminal fragment overexpression	Islow	Prolonged QT intervals; spontaneous nonsustained VT	321, 688
Kv1.5/Kcna5 (atrial)	Kcna5-E375X/+	IKur	Atrial AP prolongation, EADs, and AF	853
Kir2.1/Kcnj2	Kcjn2−/−	IK1,	Anderson-Tawil syndrome (LQTS7); broader APs; spontaneous APs	1144, 1323
Channel α-subunits: Ca2+ channel variants
Cav1.2/Cacna1c.	Cacna1c-G406R/+	ICaL	LQTS8;Timothy Syndrome (TS)	56, 192, 277, 1086
	Cacna1c-G406R/Akap−/−		Normal phenotype	192, 1150
Channel β-subunits
Navβ1/Scn1b	Scn1b−/−	INa and INaL	Human: BrS; mouse: prolonged QT and RR intervals (increased Nav1.5, INa and INaL; altered Ca2+ homeostasis)	694
Navβ4/Scn4b	Scn4b−/−	INa or INaL	LQTS10; resurgent INa/increased INaL	373, 713
MinK/Kcne1	Kcne1−/−	IKs	LQTS5; Jervell and Lange-Nielsen syndrome	64, 276, 1130, 1195
MirP/Kcne2	Kcne2−/−	IKslow1, Ito,f	LQT6: prolonged APD	956
Cytoskeletal proteins
Ankyrin-B	Ankyrin-B+/−; ankyrin-B +/E1426G		LQTS4; abnormal Ca2+ homeostasis; triggered events, prolonged QT intervals; VT; atrial arrhythmias	225, 361, 788, 790
	Ankyrin B−/−		Deficient Kir6.2 membrane expression and decreased IKATP	182, 649
Caveolin-3/Cav3	Cav3−/−		LQTS9 (transverse tubular abnormalities)	337, 401, 434, 779

In contrast, where the membrane potential is unstable following full AP repolarization, DADs result. DADs accompany conditions of disrupted Ca²⁺ homeostasis. The latter in turn can follow challenge with pharmacological agents including cardiac glycosides or catecholamines, or genetic conditions such as CPVT (see sect. VIIC). DADs have been attributed to transient inward currents (I_ti) resulting from increased inward I_NCX or Ca²⁺-activated Cl⁻ current [I_(Ca)Cl] following excessive SR Ca²⁺ release. In common with EADs, DADs potentially result in triggered beats initiating arrhythmia where the displacements in membrane potential they produce reach the threshold for reexcitation (498, 499).

Second, enhanced dispersion of repolarization can potentially generate arrhythmogenic reentrant circuits (19, 38). Normal cardiac repolarization takes spatial directions running from epi- to endocardium and apex to base. These reflect longer endo- than epicardial and basal than apical APDs (687, 1228, 1287). The latter are the consequence of differing myocardial repolarization rates reflecting varying local densities of ion channels mediating repolarizing K⁺ currents. The epi-/endocardial transmural gradients themselves contribute to a coordinated process of initiation of, followed by recovery from, electrophysiological excitation (558, 563, 981).

As indicated in section IIB3, mouse ventricles show shorter, triangulated, APs (∼30–80 ms), as opposed to the prolonged waveforms AP that contain distinct plateau phases in larger mammals including humans (∼150–400 ms, respectively) (Figure 1D) (236). AP repolarization is thus dominated by the fast, Kv4.3- and Kv4.2-mediated I_to,f and the more slowly inactivating Kv1.4-mediated I_to,s, components of the transient outward current I_to and includes smaller I_CaL contributions (Figure 1, B and D).

The dispersions of repolarization in murine hearts correspondingly arise from spatial variations in I_to giving gradients of excitation and recovery that similarly enhance the overall cohesiveness of cardiac activation (826, 989, 1228). Of I_to components, I_to(fast), mediated by the pore-forming Kv4.3/Kv4.2 (267) and its regulatory β-subunit KChIP2 (604, 876), are expressed along the apical-basal, epicardial-endocardial, and interventricular axes of adult murine hearts (1108). In contrast, I_to(slow), mediated by Kv1.4, appears confined to the interventricular septum (151, 1284). Region-specific myocardial Kv4.2 and KChiP2 distributions were assessed using both standard manual and laser capture microdissection giving 80- to 100-μm width myocardial slices particularly useful in analyses of small murine hearts. Quantitative RT-PCR measurements were assessed against control, uniform, transmural distributions of Kir2.1. There were no apex-base expression gradients but greater RV than LV, and greater epicardial than endocardial Kv4.2 expression levels in both LV and RV (1124). The laser capture microdissection went on to demonstrate similar, though less marked, LV and RV, KChIP2 transmural gradients. These observations contrasted with an absence of spatial gradients in the other selected K⁺ channel genes. These Kv4.2 and KChIP2 expression differences could thus underly I_to gradients between RV and LV, and epicardium and endocardium, but not apex and base in adult mouse ventricles (151). These findings parallel the overall transmural current density gradients found in ventricles of larger mammals (875), although human and canine I_to differences may primarily reflect KChIP2 expression gradients (964, 965). In both events, alterations in such gradients have been associated with myocardial disease states (863).

Nevertheless, manipulations involving I_to accomplished alterations in membrane potential recovery, refractory periods, and their spatial variations as well as the occurrence of triggered activity that paralleled those shown by the more prolonged APs shown in larger animals. Murine hearts thus remain valuable models for understanding mechanisms of arrhythmogenesis more broadly, including arrhythmic phenomena associated with AP repolarization.

Perfused, canine ventricular wedge preparations of larger mammalian hearts (36) have been reported to contain a further, midmyocardial cell (or M-cell) layer. This is located at varying locations within the thickness of the ventricular wall (35). Such M-cells showed the greatest repolarization times. This was attributed to a decreased I_Ks but increased I_NaL and the greatest APD sensitivity to bradycardia and I_Kr antagonists. This property results in an overall increase in transmural dispersion of repolarization (TDR). Thus the difference between longest and shortest APD would then be the difference between M cell and epicardial APDs (1056). This property had been implicated in arrhythmogenecity in BrS and both congenital (1053) and acquired LQTS (37, 772). However, the contribution of M cells to myocardial repolarization and arrhythmogenesis in intact hearts remains under discussion (40, 216, 859). Furthermore, although observed in isolated cells and ventricular tissue preparations, the existence of M cells in intact porcine (955), canine (859), or human hearts is uncertain (1106).

Langendorff-perfused mouse hearts did show regional APD heterogeneities, as reflected in differences between epi- and endocardial APD₉₀s in common with larger mammalian hearts. The availability of murine models with specifiable genetic modifications made it possible to directly relate mutations seen in LQTS patients to physiological mechanisms of cardiac arrhythmias (560, 563). These issues are explored in specific examples in section VI, B4, E4, F2, and F3.

B. Arrhythmic Properties and Na⁺ Channel Recovery From Excitation

1. The late Na⁺ current

Arrhythmogenic gain of Na⁺ channel function can result from SCN5A mutations that result in a slowing of I_Na inactivation and rates of its recovery from inactivation. This is thought to be the basis of clinical LQTS3 (98, 416, 843). It could also be acquired in association with epileptic conditions (125). In either case, the resulting increased APDs and QT intervals predispose the myocardium to EADs and arrhythmogenesis (235, 1223). Normally, a gradually increasing late Na⁺ current, I_NaL, accompanies the AP potential plateau. LQTS3 is associated with increases in this normally small I_NaL (806) whether in acute or chronic, or inherited or acquired, pathological settings. This type of I_NaL could result from alterations in the gating, trafficking, or cytoskeletal anchoring of the Na⁺ channel. It results in both triggering events and arrhythmic substrate (88).

Increased I_NaL thus can prolong the APD and increases its variability. This enhances the genesis of, and arrhythmic triggering by, EADs (444, 499). In addition, it produces increases in temporal and the spatial dispersions of repolarization. These in turn increase APD restitution curve gradients and the development of APD alternans, produce local conduction block, to result in reentrant arrhythmic activity (381, 891). The increased inward I_NaL also decreases repolarization reserve or causes diastolic depolarization. Both enhance automaticity in partially depolarized cardiomyocytes. Finally, increased I_NaL causes increases in [Na⁺]_i. This potentially induces reverse-mode NCX (1062). This could elevate diastolic [Ca²⁺]_i, and thereby promote spontaneous SR Ca²⁺ release, Ca²⁺ waves and DADs at more negative voltages, and EADs at more positive voltages (162, 444, 988). The consequent elevations in [Ca²⁺]_i potentially activate CaMKII and its protein phosphorylation effects. These could enhance spontaneous RyR2-mediated SR Ca²⁺ release (1203) and I_NaL (51, 467).

I_NaL itself can be activated by CaMKII. This action is localized to a Ser-571 phosphorylation site located in the Nav1.5 DI-DII linker. Studies in Scn5a knock-in murine models whose Nav1.5 carries a phosphomimetic Ser571 (S571E) mutation, or in which the phosphorylation site is ablated (S571A), suggested that Ser-571 specifically regulates I_NaL but not other channel properties linked to CaMKII. The resulting increased I_NaL altered repolarization and intracellular Ca²⁺ homeostasis. These had proarrhythmic consequences (363). I_NaL is also modified by more direct interactions with CaMKII through Nav1.5-associated proteins. CaMKII associates with a COOH-terminal motif in actin-associated βIV-spectrin protein involved in CaMKII-mediated phosphorylation of the subpopulation of voltage-gated Na⁺ channels residing in the cardiomyocyte intercalated discs (467, 716). This may modulate not only steady-state Na⁺ channel inactivation and recovery from inactivation, but also the magnitudes of I_NaL through separate Nav1.5 modulation pathways (467, 1209). Modeling studies suggested positive-feedback interactions in which CaMKII causes increases in often arrhythmogenic I_NaL in turn increasing [Na⁺]_i and [Ca²⁺]_i and resulting in further activation of CaMKII (105). CaMKII may also complex with SAP97. This may define its interaction with Kv4.3 α mediating I_to (295).

2. Long QT3 syndrome, LQTS3

The arrhythmic features of LQTS3 differ from those of BrS. This is exemplified in their responses to antiarrhythmic agents (914). The class IA cardiotropic agent quinidine is proarrhythmic and the class IC agent flecainide is antiarrhythmic in LQTS3 (805, 954). This directly contrasts with the respective antiarrhythmic effects of quinidine and proarrhythmic effects of flecainide in BrS (150). LQTS3 also differs from other LQTS variants in that 39% of fatal arrhythmias occur during sleep or rest (954). In contrast, arrhythmia in LQTS1 and LQTS2 is associated with stress or arousal (1029). Furthermore, in LQTS3, β-adrenoceptor agonists paradoxically exert antiarrhythmic effects. LQTS3 patients derive less benefit from β-adrenoceptor antagonist therapy than do LQTS1 and LQTS2 patients (915).

3. Murine LQTS3 exemplars: Scn5a+/ΔKPQ hearts

In WT murine hearts, the sea anemone, Anemonia sulcata, toxin and I_NaL agonist ATX-II produced effects resembling an acquired LQTS3. It slowed intrinsic heart rate, prolonged PR and QT intervals, increased MAP durations, and increased sinus node recovery times. Its application resulted in an appearance of EADs, DADs, and rapid, repetitive ventricular tachyarrhythmias and sinoatrial bradyarrhythmias. ATX-II slowed sinus node pacemaking. This resulted in in bradycardic arrhythmias in isolated sinoatrial preparations. The I_NaL antagonist ranolazine (10 μM) inhibited such effects (1275). This finding suggests potential translational, antiarrhythmic, therapeutic application (385).

Two murine genetic, LQTS3, models were independently produced by targeted deletion of ΔKPQ1505-1507 residues in the Nav1.5 inactivation domain (416, 843). Scn5a+/ΔKPQ ventricular myocytes showed prolonged AP waveforms (Figure 11A), increased peak I_Na despite normal Nav1.5 expression, increased I_NaL (Figure 11, B and C), and EAD events (Figure 11, D and E) (416). These features were replicated in patient-specific induced pluripotent stem cells harboring a SCN5A mutation for the condition (707). They were alleviated by the I_NaL antagonist ranolazine (333).

Figure 11.

Separation of contributions to arrhythmogenesis from EADs and transmural APD gradients in Scn5a+/ΔKPQ hearts. A: microelectrode recordings showing action potential prolongation in Scn5a+/ΔKPQ myocytes. B and C: prolonged recovery tail currents after 100 ms test pulses to +40 mV from a −120 mV holding potential in WT (B) and Scn5a+/ΔKPQ myocytes (C). D: action potential (AP) waveforms with early afterdepolarizations (EAD) generating inward late tetrodotoxin (TTX)-sensitive I_NaL in myocytes in an action potential clamp (E). F and G: epicardial monophasic AP recordings from spontaneously active Scn5a+/ΔKPQ hearts showing multiple EADs and nonsustained VT (F), abolished by 1 μM nifedipine (G). H and I: patch-clamp studies in isolated LV ventricular myocytes from Scn5a+/ΔKPQ hearts demonstrating that nifedipine (300 nM) completely suppressed the inward Ca²⁺ current following depolarizing steps from a holding voltage of −40 mV to 10 mV (H). In contrast, nifedipine had no effect on inward Na⁺ currents in response to depolarizing steps from −100 mV to −40 mV (I). J–L: number of hearts showing VT arrhythmia (J), percentage of monophasic action potentials showing EADs in Scn5a+/ΔKPQ hearts (K), and mean ± SE endocardial and epicardial APD₉₀ values and ΔAPD₉₀ in Scn5a+/ΔKPQ (L) at different nifedipine concentrations (0 nM, 1 nM, 10 nM, 100 nM, 300 nM, and 1 μM). M: effects of nifedipine (1 μM) on VERPs of Scn5a+/ΔKPQ and WT hearts. [From Head et al. (416) and Thomas et al. (1128).]

Isolated, Langendorff-perfused, Scn5a+/ΔKPQ hearts showed increased arrhythmogenecity on programmed electrical stimulation imposing extrasystolic S2 stimuli at progressively decreased S1S2 intervals following trains of pacing S1 stimuli (Figure 11, F and G). These studies yielded conduction curves for which paced electrogram fractionation analysis demonstrated increased EGDs compared with WT (416, 1095, 1096). Similar findings were previously associated with clinical reentrant substrate and arrhythmogenesis in both HCM (1009) and LQTS (1011) (Table 5). Abrupt rate accelerations transiently increased ventricular APDs with potentially arrhythmic effects (843). Pharmacological challenge altered arrhythmic tendencies in parallel with its clinical effects. Thus, in the murine Scn5a+/ΔKPQ hearts, flecainide exerted antiarrhythmic and quinidine exerted proarrhythmic effects. This directly contrasted with the effects of these agents in Scn5a+/− hearts (1095, 1096). Arrhythmogenesis persisted with propranolol perfusion which further increased EGD, reflecting accentuated reentrant substrate (416).

4. Spatial repolarization heterogeneities and triggering events in Scn5a+/ΔKPQ hearts

Monophasic AP waveform recordings in Langendorff-perfused Scn5a+/ΔKPQ hearts related the arrhythmias at the whole-heart level to abnormal ventricular myocardial EADs that could potentially act as arrhythmic triggers. Thus Scn5a+/ΔKPQ but not WT hearts showed EADs that were often followed by VT. They also showed spatial differences in electrophysiological recovery properties that could additionally potentially result in arrhythmic substrate. The latter took the form of abnormal transmural repolarization gradients.

Thus Scn5a+/ΔKPQ hearts showed increased epicardial APD₉₀ (∼60 vs. ∼47 ms in WT) despite normal endocardial APD₉₀ (∼52–54 ms in both Scn5a+/ΔKPQ and WT hearts). The increased epicardial APD possibly reflected regional differences in I_NaL in Scn5a+/ΔKPQ hearts. This c ould provide the necessary critical voltage ”window“ that has been described for L-type Ca²⁺ channel reactivation (1128). In addition, the preferential prolongation of epicardial APD₉₀ in Scn5a+/ΔKPQ resulted in abnormal transmural repolarization gradients. This was quantified by the difference, ΔAPD₉₀. Such ΔAPD₉₀ changes showed a pattern that was the inverse of that in WT. Epicardial APD now exceeded endocardial APD. Endocardial repolarization consequently preceded epicardial repolarization, the reverse of the transmural epicardial to endocardial repolarization sequence normally found in WT (1128). Hearts from larger mammalian models made to pharmacologically replicate LQTS3 similarly showed altered EADs and transmural repolarization gradients (1056).

It was possible separately to implicate EAD phenomena in the triggering of arrhythmia, and alterations in ΔAPD₉₀ in the generation of reentrant substrate by examining their respective sensitivities to pharmacological agents. First, nifedipine was useful as a specific dihydropyridine L-type Ca²⁺ channel antagonist (1190). It exerted antiarrhythmic effects in several experimental animal models (28, 821). This contrasts with the actions of phenylalkylamine Ca²⁺ channel blocker verapamil used in previous pharmacological investigations (16, 773) which also affects I_Na and I_Kr (1330). Nifedipine at progressively increasing concentrations (10 nM to 1 μM) correspondingly decreased the incidences of both EADs and arrhythmias evoked by programmed electrical stimulation in Scn5a+/ΔKPQ hearts in parallel (Figure 11, J and K). However, it did so without altering either epicardial or endocardial APD₉₀, the resulting ΔAPD₉₀ (Figure 11L), or VERP (Figure 11M) in either Scn5a+/ΔKPQ or WT. Thus suppression of EADs alone sufficed to suppress arrhythmogenesis without alterations of underlying re-entrant substrate. Patch-clamp studies in isolated ventricular myocytes confirmed that these effects of nifedipine were the result of alterations in I_CaL but not I_Na (Figure 11, H and I) (1128). This makes it likely that alterations in SR-Ca²⁺ loading might also contribute to the arrhythmic phenotype (664).

Second, treatment with varying concentrations of the β-adrenoceptor antagonist propranolol made it possible empirically to separate distinct contributions of EADs, and arrhythmic substrate arising from repolarization gradients, to arrhythmogenesis in Scn5a+/ΔKPQ hearts (1129). Thus low propranolol concentrations (100 nM) suppressed both spontaneous and provoked arrhythmias. It also both suppressed EADs and reduced epicardial APD. The latter change accordingly corrected the repolarization gradient abnormalities in Scn5a+/ΔKPQ hearts. In contrast, high (1 mM) concentrations eliminated both EADs and spontaneous arrhythmias. However, it increased susceptibility to provoked arrhythmogenesis, while also prolonging epicardial yet reducing endocardial APDs (1129). This would exacerbate the abnormalities in repolarization gradient shown by Scn5a+/ΔKPQ hearts. The latter findings concur with similar differential actions of propranolol upon ventricular epicardial and endocardial APs previously reported in canine ventricular wedge preparations (593). Together these findings provide a physiological basis for clinical reports that β-adrenoceptor block is relatively ineffective in suppressing arrhythmia in LQTS3. This is in contrast to their use to suppress arrhythmia in LQTS1 and LQTS2 (915).

Third, agents specific for K⁺ channel as opposed to Ca²⁺ homeostatic function furnished a final separation of the contributions of EADs acting to trigger, and repolarization gradients providing substrate for, persistent cardiac arrhthymia (445). Thus the K_ATP channel opener nicorandil exerts antiarrhythmic effects in clinical LQTS. It reduces QT intervals and spatial repolarization gradients (17, 324, 1056). It also suppreses APD alternans in idiopathic LQTS (335). Nicorandil at concentrations of ∼20 mM has been reported to increase time-independent outward currents while sparing L-type Ca²⁺, delayed rectifier, or transient outward currents (430). Such concentrations reduced the incidence of arrhythmia provoked by programmed electrical stimulation in Langendorff-perfused Scn5a+/ΔKPQ hearts during MAP recording. This was accompanied by a reduction of LV epicardial but not LV endocardial APD₉₀s in both Scn5a+/ΔKPQ and WT. This resulted in a restoration of normal transmural repolarization gradients in Scn5a+/ΔKPQ ventricles. Scn5a+/ΔKPQ hearts also showed greater epicardial critical intervals for reexcitation, reflected in APD₉₀-VERP differences, than WT hearts, but these were reduced by nicorandil (445).

Finally, alterations in electrophysiological properties of the Purkinje cell conducting system (1163) in the Scn5a+/ΔKPQ variant were consistent with increased incidences of triggering effects. Isolated ventricular and Purkinje cell myocytes could be distinguished using offspring from crosses of Scn5a+/ΔKPQ or WT, with contactin2-Egfp BAC. The latter murine variants express a fluorescent reporter gene within the Purkinje fiber network. Whole cell patch-clamp studies have proven possible in murine cardiac Purkinje cells (780). These demonstrated increased I_NaL in both Scn5a+/ΔKPQ (Egfp+) Purkinje cells and Scn5a+/ΔKPQ (Egfp−) ventricular myocytes. Scn5a+/ΔKPQ Purkinje cells showed the greater increases in I_NaL and APD, as well as frequent EADs not observed in the ventricular myocytes, features reversible by mexiletine. The latter events correlated with repetitive oscillations of microfluometrically measurable [Ca²⁺]_i (490).

5. Temporal heterogeneities and arrhythmic phenomena in Scn5a+/ΔKPQ hearts

Electrophysiological activity in Scn5a+/ΔKPQ hearts also showed greater temporal heterogeneities in the form of epicardial and endocardial APD₉₀ alternans, particularly at increased pacing rates, than WT. The derivation of epicardial and endocardial restitution curves from such experiments made it possible to derive the values of DI_crit at which the slopes of such curves exceeded unity. The latter point is known to correlate with an onset of arrhythmic phenomena (see sect. IIA4). In control WT hearts, both quinidine and flecainide increased DI_crit. This would agree with the observed correspondingly increased incidences of arrhythmic activity. Both Scn5a+/ΔKPQ and Scn5a+/− hearts similarly showed greater DI_crit in parallel with their increased arrhythmic activity. Furthermore, quinidine (1 μM) and flecainide (1 μM) challenge then exerted contrasting effects on DI_crit. Thus, in Scn5a+/ΔKPQ hearts, quinidine increased and flecainide decreased the DI_crit values in parallel with their observed proarrhythymic and antiarrhythmic effects. In contrast, in Scn5a+/− hearts, quinidine correspondingly decreased and flecainide increased DI_crit in parallel with decreased and increased arrhythmic activity respectively (983). Nicorandil reduced the epicardial and endocardial DI_crit values to levels indistinguishable from untreated WTs through an overall flattening of their restitution curves, it correspondingly reduced arrhythmogenicity in Scn5a+/ΔKPQ hearts (445).

6. Atrial phenotypes in Scn5a+/ΔKPQ hearts

Murine Scn5a+/ΔKPQ hearts also showed age-dependent atrial arrhythmic phenotypes. Both young and aged Scn5a+/ΔKPQ hearts showed slowed sinus rates compared with corresponding WT. They thus showed an overlap in phenotype with Scn5a+/− hearts. Atrial arrhythmic tendencies in young Scn5a+/ΔKPQ were indistinguishable from or even reduced relative to young WT. However, they were increased in aged Scn5a+/ΔKPQ compared with WT (241, 392). These findings were explicable in terms of alterations in the properties of atrial AP initiation, propagation, and recovery. MEA recordings thus demonstrated slowed intra-atrial conduction in Scn5a+/ΔKPQ (1276). Furthermore, atrial Scn5a+/ΔKPQ cardiomyocytes showed prolonged APD₉₀ values and frequent EADs particularly with slow pacing (127, 241). Both these effects were rescued by ranolazine (<2 Hz) (639, 812). Atrial APDs and P wave durations were similarly prolonged in regularly paced Scn5a+/ΔKPQ relative to WT particularly with age (241, 392). AERPs were similar in young WT and Scn5a+/ΔKPQ but increased with ageing in WT but not Scn5a+/ΔKPQ. This left aged Scn5a+/ΔKPQ with the greatest APD₉₀/AERP ratios and consequent arrhythmic substrate. These findings correlated with observations of a greater Nav1.5 expression in young Scn5a+/ΔKPQ relative to young WT, but increases in Nav1.5 expression with age in WT but not Scn5a+/ΔKPQ (241, 392).

Murine Scn5a-F1759A atria also showed incomplete Nav1.5 inactivation, resulting in increased I_NaL, prolonged APD, and spontaneous and prolonged AF episodes. It was possible to demonstrate atrial rotors and wave and wavelets in parallel with human AF. There was an accompanying fibrosis, atrial and ventricular enlargement, myofibril disarray, and mitochondrial injury. Both the atrial and ventricular arrhythmias were inhibited by acute inhibition of the NCX, likely through its effects on the increased Na⁺ entry (1214).

7. Multiple phenotypes associated with mutations involving Nav1.5: overlap syndromes

A final and important group of Nav1.5 variants appear to be associated with overlapping combinations of phenotypes. These were then associated with both loss of SCN5A function producing BrS and conduction disease (1111), and gain of function mutations producing increased I_NaL resulting in LQTS3 (115, 941, 1215). First, a number of albeit uncommon monogenic SCN5A mutations (SCN5A-D1114N and SCN5A-delF1617) are clinically associated with a overlapping combination of LQTS3 and BrS (191, 914, 1085) or impaired cardiac conduction (1322). A SCN5A-delK1500 in the intracellular DIII-DIV linker close to the LQTS3-associated mutation SCN5A-delKPQ1505-1507 similarly results in a heterogeneous clinical LQTS, BrS, and conduction disease (370). Finally, members of a family carrying SCN5A-1795insD variously showed ECG evidence for isolated or combined features of SND, conduction disease, BrS and LQTS3 (115). Expression studies demonstrated reduced I_Na, and negatively shifted inactivation-voltage curves in an absence of I_NaL, in Xenopus oocytes (115). Alternatively, there was a contrasting, increased I_NaL resulting from disrupted fast inactivation, and increased slow inactivation that reduced maximum I_Na at high pacing frequencies in HEK-293 cells (1184). However, computational analysis suggested that both sets of characteristics were required to reproduce a combined LQTS and BrS, depending on heart rate (208).

Nevertheless, heterozygous Scn5a-1798insD/+ mice recapitulated the corresponding human SCN5A-1795insD/+ phenotype (939). They showed both signs of SND, of prolonged PQ, QRS, and QTc intervals on ECG recording, and of slowed conduction, especially in the RV, on epicardial mapping in isolated hearts. Patch electrode recordings demonstrated prolonged AP repolarization phases particularly at low, and reduced I_Na particularly at high pacing frequencies. Accompanying marked reductions in peak I_Na there was a prolonged time course of fast inactivation, and an increased I_NaL (208, 1184).

Second, other expression system studies similarly correlated multiple biophysical defects in Nav1.5 with overlapping clinical consequences. Scn5a+/delK1500 enhanced I_Na inactivation but induced an increased I_NaL (370). Scn5a+/delF1617 was associated with reduced peak I_Na, slowed recovery from inactivation. It showed an increased I_NaL only at positive membrane potentials, and reduced open times at negative potentials. Both variants were associated with an overlapping LQTS3/BrS phenotype with conduction disease.

A third group of SCN5A-V1777M and SCN5A-V1763M mutations are associated with LQTS3 in the presence of conduction disease. These displayed an increased I_NaL, yet showed no decrease in peak I_Na or other properties suggesting loss of Nav1.5 function (180). This may reflect further contributions besides biophysical I_Na properties, possibly involving structural changes (see sect. VE) in producing the overlap syndromes. Nav1.5 changes, exemplified by SCN5A-D1275N, are also associated with SND and atrial arrhythmias (238, 614, 941).

C. Arrhythmic Consequences of Genetic Modifications in Na⁺ Channel-Associated Proteins

1. Mutations in Scn1b subunits of the Na⁺ channel

The general features of Navβ subunits as well as the Navβ2 and Navβ3 variants associated with BrS-like phenotypes have been discussed above (see sect. VG). Of the remaining variants, consequences of Navβ1 changes differed when studied in coexpression in vitro systems, murine hearts, and clinical studies.

Thus 1) Navβ1 coexpression with Nav1.5 in HEK293 cells and Xenopus oocytes resulted in overlapping increases in I_Na and decreases in I_NaL (842, 1166). It increased rates with which I_Na recovered from inactivation (304). Conversely, in CHO cells, coexpression with SCN1B-D153N and SCN1B-R85H decreased I_Na relative to findings with WT Navβ1 (1234).

2) Scn1b−/− mice showed prolonged QT and RR intervals expected for LQTS as opposed to BrS phenotypes. Their acutely dissociated ventricular myocytes showed increases in both peak I_Na and I_NaL (∼1.6-fold) and in Nav1.5 expression (∼1.3-fold). Gating and kinetic properties were unchanged, and AP repolarization phases were prolonged (694). Nav1.5, Navβ, ankyrin B, ankyrin G, N-cadherin, and Cx43 showed normal membrane localization on immunostaining. In juvenile mice cardiac specific for Scn1b−/−, macropatch and scanning ion conductance microscopy methods demonstrated increases in TTX-sensitive I_Na specific to the midsection of isolated ventricular myocytes that could even precede full transverse tubule formation. Finally, ventricular myocytes from adult Scn1b−/− mice showed indications of increased transverse tubular expression of TTX-sensitive I_Na as well as the increased Scn5a mRNA.

3) Clinical loss of Navβ1 function was associated with BrS rather than LQTS phenotypes, despite increased I_NaL observed in the experimental systems. Thus a truncated SCN1B-p.Trp179X and a SCN1B-1E87Q mutation was associated with clinical BrS and cardiac conduction disease (1235). Furthermore, SCN1B-R85H and D153N was associated with lone AF (1234) and SCN1B-R85H with the right precordial electrocardiographic ST elevation diagnostic of BrS.

The different, experimental and clinical, situations thus yielded important differences in results. Levels of I_Na expression were decreased in 1 yet increased in 2, while both 1 and 2 reported increased I_NaL. Such findings would be predictive of clinical overlap and LQTS phenotypes, respectively. The available clinical evidence 3 agreed with predictions from 1 but differed from those arising from 2 in reporting a BrS rather than a LQTS phenotype. Findings in 3 also differed from those in both 1 and 2 whose observed increases in I_NaL would have predicted a LQTS clinical phenotype. Further investigations into these differences might then yield fundamental insights clarifying the relationship between expression systems, murine models, and clinical effects. They would be of particular interest in view of a number of additional features associated with Scn1b−/−.

Thus 1) Scn1b-encoded Navβ1 subunits were associated with not only Nav1.5 α-subunits but also with ankyrin-B (787) in cardiac and neuronal tissue (926, 927, 1234). They also occurred as alternative Navβ1 or Navβ1b splice variants (926).

2) Scn1b−/− myoctes showed increased mRNA staining for neuronal Scn3a. This was consistent with increased TTX-sensitive Nav1.3 protein nevertheless excluded from expression around the intercalated disc.

3) Scn1b−/− hearts showed evidence for altered cellular Ca²⁺ homeostasis. They showed triggered beats, delayed Ca²⁺ transients, and frequent spontaneous Ca²⁺ release events accompanying the increased susceptibility to polymorphic ventricular arrhythmias. These abnormalities in Ca²⁺ homeostasis were prevented by TTX at concentrations (100 nM) expected to block the TTX-sensitive, Nav1.3-mediated I_Na (663).

4) Murine Scn1b−/−, but not Scn1b+/−, mice also showed neurophysiological changes. This took the form of an epileptic phenotype. They mimicked the severe myoclonic epilepsy of infancy normally associated with heterozygous SCN1A mutations in Dravet syndrome. Hippocampal slice recordings of CA3 neurons exhibited higher peak voltages in APs of correspondingly increased amplitude compared with WT Scn1b+/+ while not showing the altered I_Na exhibited by the Nav1.1, Scn1a+/−, variant of Dravet syndrome (877).

2. Mutations in Scn4b subunits of the Na⁺ channel: LQTS10

LQTS10 was first associated with a genetic, SCN4B-L179F, variant (138, 761). Subsequently a SCN4B-S206L mutation was associated with SIDS (1110). Both conditions were associated with increased I_NaL and prolonged APDs when overexpressed in rat ventricular myocytes (1110). SCN4B-V162G and SCN4B-I166L were then associated with AF (656). Navβ4 subunits have largely been studied experimentally in connection with neuronal function, permtting resurgent I_Na generation leading to high-frequency firing by nerve and muscle Nav isoforms (67, 159, 645). Navβ4 subunit coexpression with Nav1.2 and Nav1.4 in tsA201 cells negatively shifts their activation (1310). Navβ4 subunits also occur in mouse ventricle (713). The arrhythmic effects of Navβ4 mutations may thus result from resurgent I_Na (373) or increased I_NaL. It coimmunoprecipates with SCN5A in the HEK293 expression system. However, in the latter situation it did not affect I_Na kinetics or density, or I_NaL (761).

3. Caveolin-3 (Cav3) mutations: LQTS9

Caveolae are rounded, 50- to 100-nm-diameter, surface membrane invaginations abundant in ventricular, atrial, and nodal cells (65). Their structure depends on interactions between their contained cholesterol and specific scaffolding proteins, exemplified by caveolins 1–3. The latter are encoded by CAV1, CAV2, and CAV3, respectively. Caveolin-3 is also associated with the transverse tubules during skeletal muscle development (870). Human caveolin-3 mutations are classically associated with skeletal muscle phenotypes. These take the form of limb girdle muscular dystrophy, rippling muscle disease, distal myopathy, and hyperCKemia (elevated serum creatine kinase) (1273). Gene mutations involving caveolin-3 are also associated with LQTS9.

Cav-3, via its scaffolding domain, likely interacts with numerous signaling molecules. These include G protein-coupled receptors (GPCRs), ion channels, and receptor tyrosine kinases. CAV3 may be important for G protein-mediated adrenergic cardiac I_Na upregulation (1301). Cav-3 inhibits neuronal nitric oxide synthase (nNOS) (1186) and SCN5A nitrosylation by such nNOS increases I_NaL (195, 1158). Expression of WT Cav-3 with SCN5A in HEK293 cells did not affect either peak I_Na or I_NaL (223, 1177). However, the clinical CAV3 mutations CAV3-F97C and CAV3-S141R, that first demonstrated a LQTS, resulted in an increased I_NaL in HEK293 cells (1177). So did the CAV3-V14L, CAV3-T78M, and CAV3-L79R mutations found in a SIDS cohort (223). The CAV3-F97C mutation was also associated with increased SCN5A nitrosylation and increased I_NaL and APD in HEK293 cells (223). Cav3 mutations have also been linked to HCM through an as yet unclear mechanism (415, 1272).

Murine knockout or knock-down studies implicated Cav3 in cardiomyocyte caveolus formation (337). These may integrate ion channel and exchangers, including pacemaker Hcn4, Cav1.2, Kv1.5, Kir6.2/Sur2a and Nav1.5 channels, and NCX, into specific signaling complexes (65, 1082, 1303). Two independent groups reported generation of caveolin-3-deficient mice by targeted exon 2 disruption involving the Cav3 caveolin-scaffolding, transmembrane domains, and the COOH-terminal region (337, 401). The mice also showed mild myopathic changes resembling those of limb girdle muscular dystrophy-1C in patients with CAV3, exon 2 mutations (434, 779).

Cav3−/− mice showed an abnormal, diffuse, tubular dihydropyridine receptor DHPR1-α and RyR1 localization. Their transverse tubular systems were disorganized. They exhibited dilated and longitudinally oriented transverse tubules (337). The mice developed a progressive cardiomyopathic phenotype. Thus magnetic resonance imaging and transthoracic echocardiographic studies demonstrated hypertrophy, dilation, and reduced fractional shortening. Histological examination revealed cardiomyocyte hypertrophy, cellular infiltratation, and interstitial fibrosis. Such changes were evident even at age 4 mo. There was an accompanying increase in activity in the p42/44 mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK1/2) cascade. The latter is normally inhibited by Cav3 (1272). Conversely, cardiac-specific Cav-3 overexpression resulted in lower resting heart rates at rest, prolonged PR interval but shortened QTc intervals, and alterations in ion channel expression in an absence of alterations in metabolic background activity (727, 728, 1017).

4. Arrhythmic consequences of genetic abnormalities in α1-syntrophin (SNTA): LQTS12

The rod-shaped scaffolding protein α1-syntrophin (SNTA) forms part of the dystrophin glycoprotein complex that is involved in linking proteins to the cytoskeleton. It is associated with multiple further proteins including other syntrophins and dystrophin (12). It may be associated through its PDZ domain to the three specific residues, Ser-Ile-Val, at the distal end of the COOH terminus (1063) of those SCN5A channels that are located in the lateral membranes of cardiac myocytes (345, 889, 1158). SNTA mutations are associated with LQTS12 (1158) and SIDS (194). The LQTS12, SNTA-A390V, mutation selectively dissociates the plasma membrane calcium transporter and nNOS inhibitor PMCA4b from this complex (846). This increases SCN5A nitrosylation and thereby increases I_NaL (1158). In contrast, mice with genetically modified α1-syntrophin lacking the COOH-terminal motif (ΔSIV) showed altered targeting of Nav1.5 to the lateral membrane, normal Nav1.5 targeting to intercalated discs, and reduced I_NaL (1063).

D. Arrhythmic Properties and Altered I_CaL: Timothy Syndrome (LQTS8)

1. Clinical features of Cav1.2 abnormality

Slowed Ca_v1.2 inactivation constitutes a further source of AP plateau prolongation that may promote lethal arrhythmias. Timothy Syndrome (LQTS8, TS) results from rare, autosomal dominant gain-of-function missense CACNA1C-G406R mutations involving the junction between DI/S6 and the I–II loop of the L-type Ca²⁺ channel (Ca_v1.2). It may occur with or without an accompanying CACNA1C-G402S mutation (1086). AP prolongation may be the result of abnormal stabilization of the open state of Ca_v1.2 involving the scaffolding protein AKAP79 (AKAP150 in rodents). The enhanced Ca²⁺ influx may be further accentuated by increased channel activity resulting from physical interactions between neighboring channels through their C-tails (268). The condition is often accompanied by congenital heart disease, syndactyly, immunodeficiency, cognitive abnormalities, and autism.

2. Murine models for Timothy syndrome

Two distinct TS mouse lines have been used respectively to illustrate its resulting neurological (56) and cardiac phenotypes (192). In the cardiospecific transgenic mouse LQTS8 model of TS, ∼30% of the Cav1.2 channels were Cav1.2-TS. The LQTS8 hearts were hypertrophied relative to WT. Their ventricular myocyte I_CaL showed a slowed inactivation. This likely accounts for the prolonged AP waveforms and longer QT intervals. It likely also accounts for the increased incidences of exercise-induced arrythmogenic events including premature ventricular depolarizations and torsades de pointes, in intact hearts. Cell-attached patch recordings attributed this to greater opening probabilities, open time durations, and coupled gating frequencies in the Ca_v1.2 (192). Ca_v1.2-TS channel expression also appeared to increase background sarcolemmal Ca²⁺ leak in resting ventricular myocytes, increasing diastolic intracellular [Ca²⁺] relative to WT. This was attributed to an increased SR Ca²⁺ load and Ca²⁺ spark activity. The latter also resulted in larger amplitudes of evoked Ca²⁺ transients and Ca²⁺ wave frequencies (277). Crossing LQTS8 with AKAP150−/− giving LQTS8/AKAP150−/− rescued all these phenotypes (192, 1150).

E. Hypokalemic Murine Models for Acquired LQTS

1. Arrhythmic consequences of acquired LQTS

Inheritable LQTS-mediated arrhythmias account for only 1–2% of lethal clinical ventricular arrhythmias. The remaining LQTS are associated with acquired rather than congenital risk factors. These include metabolic conditions, electrolyte abnormalities, particularly hypokalemia, and bradycardic situations (37, 288, 953, 1037, 1171). A significant number of drugs not initially associated with cardiac effects also exert significant cardiotoxic profiles that include QT prolongation. In some cases these predispose to polymorphic VT and torsades de pointes (953). Even therapeutic levels of some cardiac drugs such as dofetilide and ibutilide can noticeably (∼50 ms) increase QT interval and risk of torsades de pointes (513, 953). However, even agents slightly increasing QT interval (5–10 ms) can increase risk of torsades de pointes (288, 910). These effects often have been attributed to inhibition of K⁺ current, particularly I_Kr (1171).

2. Murine hearts as models for clinical hypokalemia

Experimental murine cardiac models permitted a modeling of acquired LQTS following reversible imposition of their causative conditions in physiological studies of often common, clinically important, situations (562). For example, both increased and decreased [K⁺]_o clinically predispose to major arrhythmias (230). Decreased [K⁺]_o increases risks of torsades de pointes (37). This may involve a delayed repolarization prolonging the QT interval that predisposes to EADs in turn triggering premature beats (953). These appeared within a substrate resulting from heterogeneous transmural ion channel distributions altering transmural dispersions of repolarization (TDR) and therefore refractoriness, particularly when accompanied by increased APD (494). Reduced [K⁺]_o has been used to assess arrhythmic effects of pharmacological agents implicated in acquired LQTS or generation of arrhythmias (288, 772, 774).

Mouse hearts replicate the [K⁺]_o sensitivity shown by the major voltage-dependent repolarizing K⁺ currents shown in hearts of larger mammals (561, 829, 1294). In particular, I_K1 similarly contributes membrane current through the voltage sensitivity arising from its inward rectification properties. These reduce K⁺ conductance at membrane potentials positive to −20 mV during phases 0–2 of the action potential, but permit outward currents following repolarization to negative to −40 mV late in phase 3, and stabilize the diastolic, phase 4, resting potential (357, 1020). In ventricles, these effects are augmented by currents arising from its tubular localization (211, 322). Ventricular transverse tubules form a restricted extracellular space permitting K⁺ accumulation following depolarization where this results in outward K⁺ current activation. The tubular K⁺ accumulation shifts the local tubular K⁺ Nernst potential, activating I_K1, thereby modulating AP recovery. This activation was demonstrated in single adult mouse ventricular myocytes in the form of transient inward tail currents observed following termination of depolarizing voltage steps sufficient to elicit outward current by repolarizations to holding potentials close to the normal cell resting potential (211, 317). The latter effect was modified by both increased and decreased [K⁺]_o (see sect. VIE). Murine hearts thus therefore can usefully physiologically model the effects of clinical hypokalemia (301, 559, 563, 979, 982, 983).

3. Triggering and arrhythmic substrate associated with hypokalemia

Studies that examined the electrophysiological effects of decreasing [K⁺]_o in murine cardiomyocytes (559–561, 978) confirmed previous reports in other systems of decreased I_to, I_Kr, and I_K1 despite increased driving forces on outward K⁺ currents (164, 317, 1001). Under normokalemic (5.4 mM [K⁺]_o) conditions, patch-clamped epicardial myocytes showed greater I_to than endocardial myocytes (561) (Figure 12A). This was in keeping with the higher epicardial than endocardial levels of their corresponding proteins (151, 967). The resulting transmural I_to differences were in turn predictive of the normally observed transmural APD differences. Hypokalemic conditions decreased epicardial (Figure 12A, a and b) but not endocardial I_to (Figure 12A, c and d). This would predict preferential epicardial over endocardial increases in APD, and diminished repolarization gradients. In contrast, epicardial and endocardial myocytes under normokalemic conditions showed similar I_K1 obtained in response to hyperpolarizing steps. These were reduced to similar extents by hypokalemia. This would be consistent with effects in increasing APD but not to the resulting transmural repolarization gradients.

Figure 12.

Electrophysiological features contributing to arrhythmogenesis in hypokalemic murine ventricles. A: outward and inward K⁺ currents (a, c) and representations of their respective maximum currents (b, d) from whole cell patch-clamped epicardial (a, b) and endocardial myocytes (c, d) under normokalemic (dark lines) and hypokalemic (3 mM) conditions (pale lines). Under normokalemic conditions, epicardial myocytes (a) showed greater early outward I_K than endocardial myocytes (c). However, whereas hypokalemia reduced early outward I_K in epicardial but not endocardial cells (b), it reduced inward I_K1 in epi- and endocardial cells by similar extents (b, d). B: monophasic AP recordings from LV endocardial and epicardial Langendorff-perfused WT murine hearts paced at 125 ms BCL, under control (5.2 mM [K⁺]; a) and hypokalemic conditions (3 mM [K⁺]; b). C: steady-state epicardial (white columns) and endocardial APD₉₀s (gray columns) and the resulting ΔAPD₉₀ (black columns) at [K]_o = 5.2 mM (a), 4 mM (b), and 3 mM (c), respectively. D: programmed electrical stimulation (PES) of isolated, WT Langendorff-perfused mouse hearts under normokalemic (a) and hypokalemic 4 mM (b) and 3 mM [K⁺]_o (c) conditions did not induce VT in (a) but induced VT in 2 of 7 hearts in 4 mM [K⁺]_o (b) and 9 of 11 hearts in 3 mM [K⁺]_o (c). E: LV epicardial monophasic action potential (MAP) recordings during intrinsic pacing (a), and programmed electrical stimulation (PES) (b) in the presence of 3 mm [K⁺]_o and 2 μm KN-93. KN-93 reduced the occurrence of early afterdepolarizations (EADs), triggered beats and ventricular tachycardia (VT) in spontaneously beating hearts (a). It failed to protect against arrhythmia provoked by PES in 6 of 6 hearts (b). [From Killeen and co-workers (559, 561).]

AP recordings in intact, Langendorff-perfused mouse hearts fulfilled these predictions from this cellular analysis of changes in I_to and I_K1 (Figure 12, B and C) (561). Under normokalemic conditions, they thus showed longer ventricular endocardial than epicardial MAPs. This led to transmural APD gradients of ∼14 ms. This difference closely correlates with previous studies of in vivo murine MAP recordings (Figure 12Ba). Reduced [K⁺]_o prolonged both the epicardial and the endocardial MAPs. However, it did so to differing extents. This led to a reduction in the transmural APD gradients as quantified by ΔAPD₉₀ values (Figure 12, Bb and C, a–c). In addition, the more bradycardic of intrinsically active hearts showed EADs even with reductions in [K⁺]_o, from 5 to 4 mM. They also showed a 29% incidence of VT (Figure 12D). These findings directly parallel clinical findings in male hypertensive patients on diuretic therapy in which each 1 mM reduction in [K⁺]_o increases ventricular arrhythmic risk by 28% (214). Further reduced, 3 mM, [K⁺]_o additionally resulted in triggered beats followed by episodes of nonsustained VT (Figure 12Dc). This would implicate L-type Ca²⁺ channels as a possible depolarizing charge carrier at low slow stimulation rates (1325).

Pharmacological maneuvers using the L-type Ca²⁺ channel-blocker nifedipine and CaMKII inhibitor KN-93 made it possible to separate the contributions made by EADs and transmural repolarization gradients to the arrhythmic process. They suggested that the mechanisms responsible for triggering and those producing reentrant substrate leading to persistent arrhythmia were separate (Figure 12E) (559). Thus both nifedipine (100 nM) and KN-93 reduced EADs and spontaneous arrhythmias (Figure 12Ea). However, neither prevented arrhythmias provoked by programmed electrical stimulation (Figure 12Eb). Nor did they alter either epicardial or endocardial APD. They thus preserved the abnormal repolarization gradients associated with the hypokalemic conditions. In contrast, higher (1 μM) nifedipine concentrations abolished all arrhythmic phenomena, whether spontaneous or provoked. It now correspondingly shortened the epicardial APDs and restored the transmural repolarization gradients ΔAPD₉₀ to control values. This implicated the alterations in ΔAPD₉₀ in arrhythmic substrate (559). Together these findings were also compatible with a prolongation of the AP repolarization thereby inducing EADs through Ca²⁺ channel reactivation. This in turn would trigger premature APs and triggered beats (Figure 13, A and B) (301, 497). These findings concurred with findings in murine (KChIP2−/−) cardiomyocytes lacking KChIP2, which encodes three auxiliary subunits of Kv4.2 and Kv4.3. These cells similarly lacked I_to, showed increased APD, and were susceptible to induced, but not spontaneous ventricular arrhythmias (604).

Figure 13.

Development of arrhythmia resulting from a combination of triggering activity and recovery abnormality exemplified by the gain of Nav1.5 function Scn5a+/ΔKPQ, or loss of K⁺ channel function through genetic modification or hypokalemic challenge in murine exemplars. Emerging from studies of murine exemplars (A), triggered activity results from the prolonged APD predisposing to L-type Ca²⁺ channel reexcitation, producing triggered action potentials (B). Recovery abnormality, particularly in the epicardium, produces a background electrophysiological defect measurable as a LQTS, resulting in arrhythmic substrate involving alterations in recovery gradients arising from APD and VERP changes in epicardial and endocardial myocardium (C). Together, B and C predispose to arrhythmic events and SCD (D).

Other pharmacological manipulations similarly culminating in altered APD and the appearance of EAD phenomena yielded concordant results. For example, NS1643 (30 μM) is known to reduce HERG inactivation (169). It reduces APD in isolated guinea pig ventricular myocytes (407). The K_ATP channel activator nicorandil (20 μM) was reported as antiarrhythmic in LQTS (1056). Both of these agents reduced the occurrences of both spontaneous, unprovoked VT, and VT following premature stimulation in murine hearts studied under hypokalemic conditions. This was accompanied by reduced incidences of EADs and reductions in epicardial APD that thereby restored the normal transmural repolarization gradients (562).

Evidence from murine systems thus attributed arrhythmic tendencies in both Scn5a+/ΔKPQ (Section VIB) and hypokalemic hearts to alterations in AP repolarization and refractoriness. This would be in direct contrast to the altered conduction velocity implicated in Scn5a+/− hearts (see sect. V, B and C). Such a distinction was confirmed in investigations of arrhythmic or refractory end points in adaptive programmed electrical stimulation protocols that consisted of regular pacing (S1) stimulus trains that were followed by S2 extrastimuli. S1S2 intervals were progressively decreased, but S2 amplitudes were progressively increased to maintain stimulus capture. Control normokalemic WT, test hypokalemic WT, and Scn5a+/ΔKPQ hearts all showed similar relationships between conduction velocity and S1S2 coupling interval and between reexcitation thresholds and S1S2 coupling interval. Yet the onset of arrhythmia did not correspond to a given change in conduction velocity. When arrhythmias were induced by extrasystolic APs in the LQTS, hypokalemic WT, and Scn5a+/ΔKPQ groups, these took place at higher AP conduction velocities and lower S2 stimulus amplitudes than in the normokalemic WT controls. Furthermore, where there were arrhythmic as opposed to refractory outcomes, the APs initiating such arrhythmias showed lower conduction velocities than when there were refractory outcomes. These conduction velocities were higher and occurred at longer S1S2 coupling intervals and smaller stimulus amplitudes in the LQTS groups compared with controls (281).

4. Transmural heterogeneities and arrhythmic substrate in hypokalemic hearts

Section IIA2 discussed situations in which substrate for reentrant excitation arose from relative changes in parameters describing AP repolarization and refractoriness. This suggested that windows of reexcitation could result from an appearance of critical intervals produced by positive time differences between measures of action potential repolarization such as APD₉₀, and recovery from refractoriness as exemplified by VERP. Such differences could either take place within a given cardiac region or involve adjoining, electrotonically coupled regions of myocardium (978). These predictions were demonstrated in isolated hypokalemic (∼3 mM [K⁺]_o) Langendorff-perfused hearts (Figure 13, C and D). These made it possible to test the effects of conditions under which alterations in transmural repolarization gradients would give rise to reentrant excitation in LQTS models (Figure 12C).

In this analysis, hypokalemia increased the epicardial but not the endocardial APD₉₀. It also decreased both the epicardial and endocardial VERPs. However, it left differences in endocardial and epicardial AP latencies unchanged. Lignocaine abolished the arrhythmogenecity associated with the hypokalemia. It also abolished these alterations in recovery parameters. The differences in the endocardial and the epicardial intervals between stimulation and 90% repolarization remained unchanged. These findings gave rise to quantititave predictors of arrhythmic behavior involving AP latency, APD₉₀ and VERP. Risks of local reexcitation in either the epicardium or endocardium could thus be obtained from critical intervals that were based on their respective (APD₉₀-VERP) differences. Risks of transmural reentrant epicardial excitation by endocardium, or endocardial excitation by the epicardium could be predicted by critical intervals given by (endocardial APD₉₀ + Δlatency − epicardial VERP) and (epicardial APD₉₀ + Δlatency − endocardial VERP), respectively. The Δlatency term corrected for delays between endocardial and epicardial excitation. These findings were corroborated by computer modeling of AP waveforms using established data on the effects of hypokalemia on ionic conductivities in ventricular myocytes (978).

The relevance of incorporating a VERP term reflecting recovery from refractoriness was further justified by extending the analysis of arrhythmic tendency in hypokalemic hearts through a range (80–180 ms) of, as opposed to single, BCLs. These extended to bradycardic conditions. This gave a graphical analysis of local and transmural relationships plotting epicardial and endocardial APD₉₀, appropriately corrected for conduction times, against VERP. Hearts studied under normokalemic (5.2 mM [K⁺]) conditions then yielded straight line loci close to lines of equality. These lines thus subtended an angle ∼45° with the abscissa. Hypokalemia shifted the APD₉₀ values above these reference lines. These differences increased with BCL. The latter was correctly predictive of correspondingly increasing arrhythmic tendency. It also direct paralleled the corresponding increases in critical (APD₉₀-VERP) and in the subtended angle. The antiarrhythmic effects of lignocaine directly correlated with its effect in modifying this relationship. It decreased the subtended angle and (APD₉₀-VERP) in hearts studied under both normokalemic and hypokalemic conditions (977).

5. Temporal heterogeneities and arrhythmic substrate in hypokalemic hearts

Hypokalemia also contributed a dynamic component to the arrhythmic substrate. This was apparent first in the features of extrasysolic APs that were obtained in response to premature (S2) stimuli. The APs that followed subsequent (S3) stimuli then showed alterations in their transmural repolarization gradients ΔAPD₉₀ under hypokalemic conditions. This effect was not observed under normokalemic conditions (980). Furthermore, in parallel with its antiarrhythmic effects both in murine preparations and in clinical situations (938), lignocaine abolished these ΔAPD₉₀ changes. Second, hypokalemia also contributed changes to alterations in restitution properties with progressively decreasing BCL. Isolated, Langendorff-perfused hearts studied under hypokalemic (3 mM) but not normokalemic (5.2 mM [K⁺]) conditions showed APD alternans and arrhythmia with progressive increases in heart rates. This was associated with increased restitution curve slopes. Both such effects were prevented by lidocaine (10 μM). Hypokalemia increased the values of both epicardial and endocardial maximum gradients and DI_crit corresponding to unity gradient identified with the onset of arrhythmia. In contrast, lidocaine decreased such gradients and DI_crit under hypokalemic but not normokalemic conditions (979, 982).

F. Genetic Modifications in Voltage-Dependent K⁺ Channels and Their Associated Subunits

1. Loss of function in transient outward currents, I_to

The pore-forming Kv4.3 α-subunit coassembles with a range of modulatory β-subunits, particularly KChIP2, but also KCNE2 (MiRP1) and DPP6 (2, 294). Coassembly of Kv4.3/KChIP2 channels with MiRP1 slowed activation and inactivation in whole cell patch-clamped CHO cells (933). Murine Kcne2−/− ventricles showed prolonged APDs, accompanied by 25% reductions in I_to,f and 50% reductions in I_K,slow1, implicating associations between Kcne2 and Kv1.5 for the first time. There was an accompanying coimmunoprecipitation of ventricular MiRP1 protein with native Kv1.5 and Kv4.2 but not Kv1.4 or Kv4.3 (956). Kcne2 knockdown by RNA interference similarly prolonged APD and decreased I_to in both neonatal and adult mouse myocytes. In contrast, Kcne2 overexpression produced by adenoviral gene delivery shortened APD and increased I_to in neonatal but not adult myocytes (679).

The magnitude of I_to normally falls with the LV epicardial-endocardial gradient. This transmural gradient is likely important in normal ventricular repolarization (151, 383, 604). Studies in Nfat−/− mice, deficient in nuclear factor of activated T cell, suggest that this may reflect control of Kv4 expression by calcineurin and NFATc3 (967). APD is consequently shorter in their LV epicardial than their LV endocardial myocytes (209, 604). Transgenic mice with altered I_to components can show APD and QT prolongation. This may correlate with arrhythmic tendency where these involve regional I_to,f and I_to,s variations. The latter in turn may alter regional apex-base, endocardium-epicardium, or septum-ventricular free wall patterns of ventricular repolarization (70, 151, 383, 691, 1284).

The consequences of reduced I_to for APD, apical-basal APD gradients, and arrhythmogenecity varied with the extent to which these reductions affected I_to,f, I_to,s, or both. Thus Kv4.2 (KCND2) dominant negative transgenic (Kv4.2-DN) mice lacking I_to,f, but not Kv1.4 (KCNA4)−/− mice lacking I_to,s, showed increased APD and prolonged QT. Neither showed spontaneous arrhythmias. However, double transgenic (Kv4.2-DN × Kv1.4−/−) mice lacking both I_to,f and I_to,s showed both AP and QT prolongation, and spontaneous ventricular tachyarrhythmias. With optical mapping using di-4-ANEPPS, Langendorff-perfused Kv4.2-DN, Kv1.4−/− and Kv4.2-DN × Kv1.4−/− hearts showed similar activation patterns and conduction velocities, but longer APD₇₅ (∼28, ∼28, and ∼34.3 ms) than WT (∼20.3 ms). However, Kv4.2-DN hearts showed reduced apicobasal dispersions of refractoriness compared with WT. They did not show VT on premature stimulation. Voltage-clamped Kv4.2-DN myocytes showed 30% greater apical than base I_to,f. In contrast, both Kv1.4−/− and Kv4.2-DN × Kv1.4−/− hearts showed increased dispersions, and VT, on premature stimulation (558, 687).

2. Loss of function in the rapid delayed rectifier K⁺ current, I_Kr: LQTS2 and LQTS6

The Kv11.1 (KCNH2)-mediated I_Kr is important both for completing phase 3 repolarization and preventing arrhythmias induced by EADs in hearts of larger mammals (1070). Phase 0 depolarization rapidly activates initially resting I_Kr channels into their open state. But this is rapidly succeeded by channel inactivation. However, repolarization produces a rapid recovery from inactivation involving transitions through the open state preceding return to the resting state. The latter produces a tail current (696, 697, 1000, 1070, 1170). Loss-of-function KCNH2 (formerly termed HERG, the human form of ERG) mutations affect I_Kr (229), particularly in the S5/pore region (472, 1342). The consequent clinical LQTS2 results in low-amplitude bifid T waves in the ECG (806), reflecting increased transmural dispersions of repolarization, cardiac arrhythmic events following sudden arousal, and episodic sinus bradycardia (1002, 1307).

I_Kr is also the dominant repolarizing current in fetal (∼day 18) mouse hearts. I_Kr block by dofetilide then caused EADs and failure of repolarization. I_Kr block in fetal mice produced bradycardia and arrhythmia resulting in lethal circulatory failure (1219). Simlarly, complete knockout of all the ERG1 isoforms is intrauterine lethal.

However, in adult mice, I_Kr may only play minor roles in ventricular repolarization. Nevertheless, it likely retains a role in SAN pacemaker function (see sect. IIIA). Thus results in mice in which the B transcript was eliminated while preserving the A transcript suggested that in adult myocytes, Erg1B is necessary for I_Kr surface membrane expression and that its knockout predisposes to episodic sinus bradycardia (627). However, QT intervals were normal notwithstanding loss of the rapidly decaying component of I_Kr. Murine LQTS2 models carrying the G628S dominant negative mutation showed normal QT durations but minor abnormalities in QRS and T waveforms. There was a loss of I_Kr and increased APDs in isolated myocytes, but normal AP waveforms in ventricular strips (53). Merg1b−/− mice showed episodic abrupt sinus bradycardia. This finding parallels clinical observations in some LQTS2 families where this can trigger arrhythmias (1307). However, they also showed normal QT intervals despite an abolished I_Kr in isolated myocytes (626), while not excluding a persistence of repolarizing K⁺ currents produced by other mERG isoforms (690). Nevertheless, epicardial optical mapping studies related an increased VT propensity in a Merg+/− mouse to prolonged apical and basal APDs and VERPs, despite conduction velocities similar to those found in WT (994).

Studies in murine models have also yielded important results bearing on roles of alternatively spliced isoforms. Human and mouse hearts express two, A and B, isoforms in the underlying ERG1. These have differing NH₂ termini and show slow and rapid deactivation properties, respectively (627). The A transcript is relatively abundant in both human and mouse hearts. The B transcript is abundant in mouse but not human hearts (627). Most reports bearing on LQTS2 appear to reflect properties of the A isoform. Studies in Xenopus oocyte expression systems suggest that functional I_Kr in mouse hearts requires heteromeric assembly of different, mERG1a, mERG1a', and mERG1b isoforms. Currents from expressed mERG1a alone resembled those obtained following HERG expression. Currents from mERG1a', lacking the first 59 NH₂-terminal amino acids, and mERG1b currents showed a 10-fold more rapid deactivation. Coexpression of mERG1a and mERG1b gave currents closely resembling murine I_Kr (690). hERG1b is also likely to be critical to normal repolarization with its deficiency resulting in proarrhythmic effects in human cardiomyocytes. sh-RNA-induced knockdown of the 1b subunit halved its corresponding mRNA and protein levels as well as I_Kr, increasing AP duration and variability, with an appearance of EAD phenomena in cardiomyocytes from human induced pluripotent stem cells. These features were replicated by converting hERG heteromers to hERG1a homomers by expressing a fragment representing the HERG1a-specific NH₂-terminal Per-Arnt-Sim domain absent in hERG1b (516).

I_Kr channels may require further coassembly of auxiliary modulatory subunits with different channel isoforms to reproduce their in vivo function. Thus native I_Kr differs from currents recorded from hERG channels in heterologous expression systems in their gating, external K⁺ regulation, and drug sensitivity properties (207). Normal in vivo I_Kr function thus also depends on association of KCNH2 with the accessory KCNE2 (minK-related, MiRP1) protein (1, 2, 294). This results in a current with +5 to +10 mV shifts in steady-state activation, enhanced rates of deactivation, and reduced unit channel conductance from 13 to 8 pS compared with electrophysiological findings associated with expression of KCNH2 alone (696, 697). Loss-of-function KCNE2 mutations are associated with LQTS6 (1085). This has phenotypic consequences similar to those of LQTS2. Conversely, the gain-of-function KCNE2-R27C mutation has been associated with familial AF (1295).

3. Loss of function in the slowly activating delayed rectifier K⁺ current, I_Ks: LQTS1 and LQTS5

The slow, I_Ks, current observed in hearts of larger mammals is mediated by the K⁺ channel protein KCNQ1. This protein is modulated by its association with the single transmembrane domain β-subunit KCNE1 (Mink) (69, 999). Adult mouse hearts do express Kcnq1 (276), but their Kcne1 expression declines with age (276, 308) from age ∼1 wk (253). Adult mouse myocytes consequently lack typical I_Ks (53, 253). In addition, alternative splicing of KCNQ1 results in an isoform 1 with a long NH₂-terminal end (69) and an isoform 2 with a short NH₂-terminal end of only two amino acids (253). Isoform 2 does not form functional channels but exerts strong dominant-negative effects on isoform 1.

KCNQ1 mutations are associated with clinical LQTS1. Cardiac incidents typically occur during exercise (1222). Kcnq1−/− mice recapitulated the morphological inner ear abnormalities, auditory defects, and the shaker/waltzer and cardiac phenotypes of Jervell and Lange-Nielsen syndrome, with which clinical LQTS1 often presents. Intact Kcnq1−/− mice showed abnormal in vivo electrocardiographic P and T waves, prolonged QT intervals, and episodic nonsustained arrhythmias. In contrast, isolated perfused murine Kcnq1−/− hearts showed normal QT intervals and APDs. In Kcnq1−/− but not WT hearts, these were nevertheless prolonged by nicotinic or sympathomimetic stimulation. This implicates extrinsic cardiac innervation in reproducing the clinical phenotype (1138). Nevertheless, mice overexpressing human KCNQ1 isoform 2, which exerts dominant negative effects on murine Kcnq1 isoform 1, showed marked QT prolongation. They also showed SND and atrioventricular block. The latter was associated with prolonged atrial-His but normal His-ventricular intervals in His bundle recordings. Their patch-clamped mutant cardiomyocytes showed increased APD. However, there was also evidence for additional electrophysiological remodeling. This took the form of reductions in I_to and I_ss through downregulation of Kv4.2 and Kv1.5 and upregulation of Kv4.3 (254). Thus complex physiological changes add to those directly arising from the genetic changes in vivo.

Loss-of-function mutations involving the β-subunit KCNE1 are associated with LQTS5 (1085, 1087). Its phenotype resembles that of LQTS1, with prolonged APs and increased arrhythmic risk (121, 999). Homozygous Kcne1−/− mice with a deletion in the entire KCNE1 coding sequence recapitulated both the sensorineural and the arrhythmic effects clinically associated with LQTS5 (64, 1195). They showed circular movements, repetitive falling, head nodding, and bilateral deafness attributed to deficient transepithelial endolymphatic K⁺ transport (1195).

Kcne1−/− mice showed increased QT intervals at slow heart rates that paradoxically shortened at increased heart rates (276). Spontaneously beating Langendorff-perfused Kcne1−/− hearts showed frequent EADs, closely coupled triggered beats and spontaneous VT. Regular pacing demonstrated increased bipolar electrogram durations. Premature stimuli during programmed electrical stimulation frequently provoked monomorphic VT episodes (64, 447). Prolonged epicardial and endocardial APDs resulted in reduced critical intervals as given by APD₉₀-VERP differences. They also resulted in reduced transmural repolarization gradients, consistent with spatial reentrant substrate (1130). These changes took place in an absence of epicardial apical-basal APD₉₀s measured by optical mapping (994). Temporal re-entrant substrate appeared as increased APD₉₀ alternans and steeper epicardial and endocardial APD₉₀ restitution curves during dynamic pacing. Nifedipine suppressed the EADs, triggered beats, and repolarization alternans. It selectively shortened epicardial APD, restoring the transmural repolarization gradient (1130).

Pharmacological manipulations directed at K⁺ channel rather than Ca²⁺ homeostatic function using nicorandil yielded concordant results (562). Nicorandil increases K⁺ conductance (478, 528), enhancing repolarization reserve (478), and shortening APD₉₀ in atrial (1289) and Purkinje fibers (478), and isolated guinea pig and rabbit ventricular myocytes (528). These effects were independent of external [Na⁺] (478). They occurred despite normal I_Ca (528), (dV/dt)_max (478, 528), or maximum diastolic membrane potentials. Nicorandil (20 μM) rescued both the spontaneous and provoked arrhythmogenic phenomena. It also restored the AP parameters reflecting reentrant substrate towards the normal values in untreated WT, further shifting them in WT hearts in similar directions (447).

4. Loss of function in I_slow

In addition to rapidly decaying outward I_to, mouse ventricular myocytes show gradually decaying contributions from a rapidly activating but slowly inactivating, 4-aminopyridine-sensitive I_slow. This is likely at least partially carried by Kv1.5. There is also a constant current component (1348). Murine studies demonstrated potentially arrhythmic contributions arising from abolition of I_slow. These employed a dominant negative transgenic mouse overexpressing a cardiac Kv1.1 NH₂-terminal fragment controlled by the α-myosin heavy chain promoter (688). The truncated channel fragment coassembles with WT Kv1.x. This traps heteromeric channels in the endoplasmic reticulum. The latter accounts for the dominant negative reduction in Kv1.x currents in heterologous expression systems (321).

The transgenic mice showed prolonged QT intervals and spontaneous nonsustained VT on ambulatory telemetry. Programmed stimulation applied to the RV induced polymorphic VT in anesthetized, open-chest preparations. The patch-clamped ventricular myocytes showed prolonged APDs attributable to reduced I_slow. Studies using voltage-sensitive dyes and optical mapping techniques demonstrated the normal apical-basal activation in apically paced hearts. Mean AP conduction velocities (∼0.5 m/s) were also similar to those showed by WT. Transgenic hearts showed increased APD₇₅ and APD₉₀ (59). They showed altered restitution properties. Thus APD did not show the normal decrease with decreased BCL. There was instead a flattened dependence on decreased BCL. These findings implicate I_slow in the adaptation of APD to altered heart rate, particularly at short BCLs. These properties together culminated in marked reentrant substrate. Premature apical stimuli triggered prolonged (≥30 min) reentrant VT episodes in transgenic but not WT hearts.

5. Other K⁺ channel variants

Deficiencies in inward rectifying K⁺ current, I_K1, resulting from loss-of-function KCNJ2 mutations encoding Kir2.1 are associated with Anderson-Tawil syndrome (LQTS7) (1143, 1144). In common with Jervell and Lange Nielsen syndrome, there are accompanying extracardiac phenotypes. In LQTS7, these include periodic paralysis, dysmorphic facial features, and syndactyly (1144). Kcjn2−/− mouse neonates lacking Kir2.1 did not demonstrate ectopic beats or reentry arrhythmias but were bradycardic on ECG recording. Their ventricular myocytes lacked detectable I_K1 while showing persistent sustained outward K⁺ currents and L- and T-type channel-mediated Ba²⁺ currents. They showed broader APs and more frequent spontaneous APs. Kcjn12−/− mice lacking Kir2.2 showed 50% reductions in I_K1 (1323). This contrasts with the association between the gain of I_K1 function KCNJ2-V93I mutation and familial AF (1278).

Murine models with decreased ultrarapid delayed rectifier potassium current (I_Kur), mediated by Kv1.5, resulting from loss of Kcna5 function showed atrial AP prolongation, increased incidences of EADs, and AF (853).

G. K⁺ Channels and Scaffolding Proteins

1. Genetic abnormalities in ankyrin-B (Ank2): LQTS4

In common with Na⁺ channels, K⁺ channels are associated with scaffolding proteins (1071). Of these, ankyrin-B, ANK2, is critical in regulating cardiac membrane protein expression. ANK2 dysfunction can follow myocardial infarction (377, 469). Genetic ANK2 variants are clinically associated with LQTS4 but may correlate with arrhythmic susceptibility in general (1034). ANK2 loss-of-function results in complex phenotypic abnormalities including sinus node and conduction disorder, ventricular arrhythmia, and SCD (788–790, 1033). LQTS4 patients showed increased QT intervals and arrhythmias with additional features including sinus bradycardia and paroxysmal AF (1023).

Ankyrin-B+/− mice recapitulated the latter findings. Mice with heterozygotic, loss-of-function, ankyrin-B +/E1426G mutations showed reduced transverse tubular targeting and protein levels in Na⁺-K⁺-ATPase, NCX, and inositol 1,4,5-trisphosphate receptors. All the latter proteins are known to bind ankyrin-B (789). Abnormalities in Na⁺-K⁺-ATPase and NCX expression would potentially increase intracellular Na⁺ overload and reduce Ca²⁺ extrusion. The consequent SR Ca²⁺ overloading might then account for the abnormal SR Ca²⁺ release events following adrenergic stimulation. The adult cardiomyocytes correspondingly showed potentially arrhythmic abnormal intracellular SR Ca²⁺ transients and EADs and DADs following adrenergic challenge (788). In vivo ECGs showed prolonged QT intervals and polymorphic VT episodes.

Subsequent studies of further mutational variants reported a wide range of clinical phenotypes. These included bradycardia, and AF, although not always QT interval prolongation (790). Ankyrin-B also associates with Kir6.2, in the complex that includes Na⁺-K⁺-ATPase (579). This is consistent with its possible regulation of I_KATP gating which in turn may cardioprotect from ischemia. Ankyrin B−/− hearts and cardiomyocytes showed deficient Kir6.2 membrane expression and decreased I_KATP. (649). Ankyrin-B in turn is recruited to the cardiac dyad by the actin associated βII spectrin. The ankyrin-B-p.R990Q mutation disrupts this interaction causing a severe human arrhythmia phenotype. Correspondingly, βII spectrin−/− mice showed SND and ventricular arrhythmia associated with afterdepolarization phenomena and abnormal Ca²⁺ waves (1072).

Disruption of ankyrin B may also clinically predispose to atrial arrhythmias. Ambulant ankyrin-B+/− mice developed spontaneous atrial arrhythmias. This was attributed to altered interactions with Ca_v1.2 reducing I_CaL and APD (225). They also show evidence of altered pacemaker function (361).

2. Genetic abnormalities in A-kinase anchoring protein-9 (AKAP-9): LQTS11

Scaffolding A-kinase anchoring protein subtypes (AKAPs) structurally and functionally link particular enzyme molecules and their end targets. The latter include ion channels, thereby affecting cardiac myocyte excitation and contraction (755). Genetic abnormalities in AKAP9, also known as Yotiao, are associated with LQTS11. AKAP9 complexes with KCNQ1, the type II regulatory subunit (RII) of PKA, and protein phosphatase 1 (PP1). It thereby conveys PKA to the vicinity of the channel where PKA phosphorylates serine residue S27 on the KCNQ1 NH₂ terminus. AKAP9 possess two KCNQ1 binding sites located respectively close to its NH₂ and COOH termini. Thus the AKAP9-G589D mutation near its LZ motif disrupted both AKAP9-KCNQ1 interaction and functional regulation of I_Ks by PKA-mediated phosphorylation. A clinically observed AKAP9-S1570L mutation in the latter region similarly reduced KCNQ1 binding and phosphorylation. It also diminished cAMP-mediated enhancement of I_Ks channel activity and prolonged APD on computational modeling particularly following isoproterenol stimulation (188). AKAP9 thus mediates sympathetic regulation of I_Ks channel regulation. Disruption of this produces pathological effects that may include LQTS (187, 746).

H. Short QT Syndromes

Short QT syndromes (SQTSs) (388) are characterized by often familial occurrences of shortened electrocardiographic QT intervals (less than ∼320 ms) and peaked T-waves despite normal cardiac anatomy (874). They are clinically associated with syncope, shortened AERPs and VERPs, and increased atrial and ventricular arrhythmogenicity potentially causing SCD (359). The known genetic SQTS variants complement some of the genetic changes in LQTS. SQTS1-3 result from gain-of-function K⁺ channel mutations in KCNH2, KCNQ1, and KCNJ2 encoding α-subunits mediating I_Kr (149), I_Ks (93), and I_K1, respectively (257). SQTS4-SQTS6 result from loss-of-function CACNA1C, CACNB2b (34), and CACNA2D1 mutations (1116). These involve L-type Ca²⁺ channel α1C, β2b, and α2δ-1-subunits, respectively. Future studies on these will likely yield important insights on arrhythmogenic mechanisms complementing those derived from LQTS models. Available physiological analysis applying K⁺ channel openers in LV wedge preparations implicated increased transmural repolarization dispersions and shortened VERP as arrhythmogenic substrate (873).

A murine SQTS model with increased Kir2.1 expression and therefore increased I_K1 showed shortened QT intervals and APDs (841). This was accompanied by bradycardia, extrasystoles, atrioventricular block, and atrial flutter (651) as well as ventricular arrhythmias (893). These included prolonged episodes of high-frequency VT (841). Recent computational analysis replicated QT interval shortening in hERG-N588K SQTS1 and Kir2.1-D172N SQTS3 mutations, modeling experimental data from recombinant WT and hERG-N588K channels (7, 8, 257, 464, 911, 1326, 1328). These predicted increased tissue vulnerability to premature stimuli and an increased tendency to formation of reentrant excitation waves (7).

VII. ARRHYTHMIC CONSEQUENCES OF ALTERED Ca²⁺ HOMEOSTASIS

A. Effects of Cellular Ca²⁺ Homeostasis on Myocyte Electrophysiology

1. Modulation of the excitation-contraction coupling process

The process of contractile activation following membrane excitation brought about by the ion channel function discussed in previous sections begins with a voltage-triggered, L-type Ca²⁺ channel-mediated, influx of extracellular Ca²⁺ (Figure 1, A and C). This locally increases [Ca²⁺]_i in the vicinity of each individual L-type Ca²⁺ channel and its nearby RyR2s. A Ca²⁺-induced Ca²⁺ release then initiates Ca²⁺ release sparks. The amplitudes of these are graded with that local Ca²⁺ current flow (193, 1092). A nonregenerative, spatial and temporal, graded, summation of such unit events makes up the resulting [Ca²⁺]_i transient (480, 1226). The consequent increase in [Ca²⁺]_i turn drives the Ca²⁺ binding to troponin that initiates myofilament mechanical activity (see sect. IB2).

The released Ca²⁺ is subsequently returned from the cytosol to the SR for resequestration by calsequestrin. This transport process is mediated by a phospholamban (PLN)-regulated SERCA2a. Ca²⁺ is also returned to the extracellular space by two major routes. The electrogenic NCX provides a low-affinity and a high-capacity transporter (129, 265). In contrast, the PMCA, for which translocation of one Ca²⁺ is coupled to hydrolysis of one ATP, acts as a high-affinity (K_m 100–200 nM) low-capacity transporter effective even at the normally very low [Ca²⁺]_i (143) (see sect. IB3).

The molecules involved in the above sequence can undergo modifications in response to extrinsic physiological demands. Kinase-mediated protein phosphorylation exerts strategic regulatory effects on myocyte excitation-contraction coupling, metabolism, intracellular Ca²⁺ homeostasis, mitochondrial function, and protein transcription. This can be driven by upstream β-adrenergic signaling. This stimulates G proteins (G_s) that activate adenylate cyclase. The latter in turn increases cellular cAMP levels. cAMP dissociates the inhibitory regulatory (R) subunit (R subunit) from PKA. The resulting PKA activation phosphorylates multiple proteins involved in cardiac excitation, contraction, and relaxation dynamics (see sect. VIIB1) (544). This may account for the multiple, inotropic, chronotropic, and lusitropic effects of cAMP signaling (103) (Figure 14A).

Figure 14.

Signalling pathways underlying excitation-contraction coupling. β-Adrenergic receptor (βAR) activation through stimulatory guanine nucleotide-binding (G_s) proteins increases cellular cAMP levels (A). This in turn drives a phosphokinase-A (PKA)-mediated phosphorylation and activation of L-type Ca²⁺ channels and cardiac ryanodine receptor (RyR2) SR-Ca²⁺ release channels (A) and the exchange protein directly activated by cAMP (Epac) pathway producing a calmodulin kinase II (CaMKII)-mediated RyR2 activation (B). Either action on the Ca²⁺-induced Ca²⁺ mechanism impinges on the level of SR Ca²⁺ release (C), and consequent alterations in cytosolic Ca²⁺ (D). Experimentally used agonists (+) and antagonist (−) agents used on these pathways include isoproterenol, H-89, 8-pCPT-2-O-Me-cAMP (8-CPT), and KN-93.

Thus PKA-mediated phosphorylation of the COOH-terminal tail region of the Cav1.2, L-type Ca²⁺ channel modifies both the ventricular AP plateau phase as well as SAN pacemaker potentials. That of RyR2 reduces the binding of a highly expressed regulatory cis-trans peptidyl prolyl isomerase, FK506 binding protein type 12.6 (FKBP12.6), whose binding to RyR2 normally stabilizes its closed state. This FKBP12.6-RyR2 dissociation increases the Ca²⁺ sensitivity of RyR2-mediated release of SR Ca²⁺. Hyperphosphorylation inducing leaky channels may occur in pathological situations of sympathetic overactivity, as may occur in cardiac failure (see sect. VIIC7) (729, 730, 747, 1239, 1241). Phosphorylation of PLN removes its inhibition of SERCA2 in its reuptake of previously released cytosolic Ca²⁺ (544).

The protein phosphatases PP1 and PP2A conversely mediate protein dephosphorylation. This often takes place at the same protein substrates and serine/threonine sites as catalyzed by PKA (704). PP2A acts preferentially on L-type Ca²⁺ channels reversing the β-adrenergic effects of PKA (405). Its actions are upregulated with stimulation of inhibitory G proteins, G_i, through other signaling processes (547). In saponin-permeabilized cardiomyocytes, both PP1 and PP2A increased frequencies of spontaneous Ca²⁺ sparks. This was followed by a disappearance of such events, reflecting the consequent SR Ca²⁺ store depletion. These effects were inhibited by the PP1 and PP2A phosphatase inhibitors okadaic acid and calyculin A (89, 1119). PP1 caused dephosphorylation of the SERCA2a inhibitor PLN with a consequent reduction in SERCA2a activity. PP2A also reduced Cx43 conductivity (see sect. VIIG) (14, 636).

An alternative or coexistent, PKA-independent, mechanism of cAMP-dependent catecholaminergic signaling may offer a further level of adrenergic control. This occurs downstream of β-adrenergic receptor-dependent cAMP generation but upstream of Ca²⁺-induced Ca²⁺ release (see sect. VIIB2) (Figure 14B). It involves signaling by cAMP-dependent, exchange proteins directly activated by cAMP (Epac) (541, 962). Of Epac isoforms, Epac1 contains a single cAMP-binding domain and appears to be ubiquitously expressed. Epac2 contains two cAMP-binding sites and occurs preferentially in brain, pituitary, and adrenal gland (134, 962). Cardiac Epac1 forms part of the macromolecular regulatory complex for RyR2 possibly through RyR2-Epac functional coupling along with the muscle-specific A-kinase anchoring protein (mAKAP) and PKA (437). Use of a novel fluorescent Epac analog has differentially localized Epac1 and Epac2 to the nucleus and the Z lines, respectively, in cardiomyocytes, consistent with their having separate intranuclear and calcium homeostatic roles (885).

2. Direct and indirect physiological effects of modified Ca²⁺ homeostasis

Modifications in [Ca²⁺]_i and Ca²⁺ homeostasis in turn affect strategic molecules that may further modify the Ca²⁺ homeostatic processes themselves, alter potentially electrogenic transmembrane ion fluxes, or modify particular enzymic signaling sequences. These in turn potentially directly or indirectly modify membrane excitation and stability with potential arrhythmogenic consequences (556, 1210). These modifications can result from direct actions of Ca²⁺ itself. Alternatively, they could involve mechanisms directly or indirectly involving a wide possible range of intracellular targets. Besides PKA and protein kinase C (PKC), the latter include calmodulin (CaM) and calcium/calmodulin kinase II (CaMKII) (166). Both CaM and CaMKII may sense local [Ca²⁺] through Ca²⁺ binding to EF-hand motifs at the NH₂ and COOH terminals of CaM. The binding results in formation of a Ca²⁺/CaM complex. This then binds to the CaMKII regulatory domain. The binding relieves the baseline action of the CaMKII autoinhibitory domain. Thr-287 autophosphorylation then preserves the resulting activated state even following return of [Ca²⁺] to its resting level, until a phosphatase-mediated dephosphorylation (29).

CaMKII is a multifunctional serine/threonine kinase that exists in cardiomyocytes as various splice variants of its δ isoform, CaMKIIδ (29). These variants have a wide range of downstream targets. Cytoplasmic CaMKIIδ2 (CaMKIIδC) is localized close to, and associates with, L-type Ca²⁺ channels and RyR2. Further CaMKII subpopulations occur in the intercalated disc and mitochondria (1103). These act upon PLN, Nav1.5, and multiple voltage-gated and ATP-sensitive K⁺ and Cl⁻ channels (468, 898). Kinase-mediated protein phosphorylation thus exerts strategic regulatory effects in myocyte excitation-contraction coupling, metabolism, intracellular Ca²⁺ homeostasis, mitochondrial function, and protein transcription (see sect. VII, B5 and F). As detailed below, these actions perturb the balance between inward I_CaL and outward I_K currents determining cardiac AP plateau durations (24) and modify intracellular Ca²⁺ in turn modifying CaM and CaMKII activity (898).

3. Effects on Ca²⁺ release and its triggering

These direct and indirect regulatory mechanisms potentially alter both the triggering and the Ca²⁺ release process itself through effects on I_Ca (see sect. VIIB1) and on the RyR2-Ca²⁺ release channel (Figure 14C) (see sect. VIIB3). Increased subsarcolemmal [Ca²⁺] exerts rapid negative feedback inhibitory effects on I_CaL. These take place in addition to, and over much more rapid time courses compared with, the slow voltage-mediated inactivation of I_CaL following its initial activation by the AP upstroke (1067). The feedback effects likely entail Ca²⁺ binding to CaM itself prebound to the L-type Ca²⁺ channel COOH terminal (641). In contrast, Ca²⁺-free CaM reduces L-type Ca²⁺ channel inactivation (897). The resulting Ca²⁺/CaM then interacts with a closely located IQ domain. This in turn blocks the inner pore of the L-type Ca²⁺ channel (712). This mechanism potentially provides a mechanism by which Ca²⁺ either entering from the extracellular fluid or released from the SR exerts a normal negative feedback inactivating L-type Ca²⁺ channel activity. Such an action would normally prevent Ca²⁺ overload.

Increased [Ca²⁺]_i may also potentiate I_CaL. This may also involve a Ca²⁺ binding to CaM. This in turn activates CaMKII which then promotes phosphorylation of the COOH terminal of the L-type Ca²⁺ channel α1-subunit (27, 1311). Moderate increases in [Ca²⁺]_i following repetitive stimulation do produce such a calmodulin-mediated CaMKII activation (1311). Recovery from Ca²⁺-mediated I_CaL inactivation then requires a fall in subsarcolemmal [Ca²⁺] typically brought about by the NCX, whose activity is also voltage-dependent. These mechanisms together result in complex interactions between L-type Ca²⁺ channel and NCX activity, membrane potential, and intracellular Ca²⁺ (341).

RyR2s also show Ca²⁺-dependent adaptation and inactivation properties dependent on separate mechanisms sensing both local cytosolic and SR intraluminal [Ca²⁺]. These may involve luminal (L-) (K_D = ∼60 μM), cytoplasmic (A-) (K_D = ∼0.9 μM) activation sites, and a cytoplasmic Ca²⁺ inactivation (I2-) site (K_D = ∼1.2 μM). Regulation by luminal and cytoplasmic [Ca²⁺] are closely related. L-site activation produces brief (1 ms) <10/s openings. These nevertheless give luminal Ca²⁺ access to the A-site whose occupancy by Ca²⁺ produces considerable longer openings. They also give access to the I2 site, which inactivates the RyR2 at high Ca²⁺ turnovers. In addition, luminal Mg²⁺ (at ∼1 mM) is essential for the control of SR excitability in inhibiting cardiac muscle RyR2 but not skeletal muscle RyR1. It competes with Ca²⁺ for the L-site (review in Ref. 622). In consequence, alterations in cytosolic Ca²⁺ influence the triggering of RyR2-mediated Ca²⁺ release. SR luminal Ca²⁺ levels also influence the Ca²⁺ available for, and the extent of SR Ca²⁺ release, and the likelihood of spontaneous and propagated release phenomena. RyR2-Ca²⁺ release channels also show inactivation/adaptation phenomena dependent on the level and the rate of Ca²⁺ increase (394). Finally, Ca²⁺-free CaM decreases RyR2 open channel probabilities by ∼60%, effects reversible by Ca²⁺ itself (898), and CaMKII action results in RyR2 phosphorylation promoting release of SR-Ca²⁺ (1243).

Finally, significant increases in [Ca²⁺]_i (to >320–560 nM) could close gap junctions mediating intercellular coupling (389, 756, 839, 1074).

4. Effects on cellular Ca²⁺ efflux processes

As indicated above, NCX extrudes much of the L-type Ca²⁺ channel-mediated Ca²⁺ entry. A smaller part is transported by plasma membrane Ca²⁺-ATPase. NCX activity is electrogenic: it transports three Na⁺ in exchange for one Ca²⁺ (129, 265). The direction of its current flow depends on the Na⁺ and Ca²⁺ gradients and the membrane potential E_m. Thus I_NCX is inwards and depolarizing when Ca²⁺ is extruded, and outwards and hyperpolarizing with Ca²⁺ influx. Consequently, increased diastolic [Ca²⁺]_i potentially activates transient inward currents (I_ti) mediated by a forward-mode NCX-mediated Na⁺ influx (101, 354). This could cause premature sarcolemmal depolarization and a consequent triggered activity which causes systolic dysfunction and increases arrhythmic tendency (see sect. VII, B4 and C4) (903). It is also possible that increased SR Ca²⁺ leak into the restricted space of the dyadic cleft could increase local NCX activity removing the resulting increased local Ca²⁺. This in turn would correspondingly increase local [Na⁺]_i potentially reducing the driving force for Na⁺ entry through nearby Na⁺ channels, and the resulting I_Na (124, 666).

Both heart failure and hypertrophic cardiomyopathies are associated with increased spontaneous SR Ca²⁺ leak (13, 226, 864). An increased SR Ca²⁺ would then activate inward depolarizing NCX-mediated current. If sufficiently large, such transient inward currents (I_ti) could initiate proarrhythmic spontaneous DAD phenomena (106). Human cardiac heart failure is also associated with a decreased I_Na (1164). The consequent reductions in CV could similarly increase susceptibility to AF (362).

Finally, PMCA activity is enhanced by Ca²⁺/calmodulin and calmodulin binding (143).

5. Effects on Nav1.5 function and expression

Several studies suggest that altered cytosolic Ca²⁺ also more directly modifies either Nav1.5 function (15, 51, 1112) or expression (196, 282, 1114) (see sect. VIIC6). The functional effects may involve direct and/or indirect Ca²⁺ binding to Nav1.5 itself. Direct Ca²⁺ binding can take place at an EF hand motif close to the Nav1.5 COOH terminal. This shifted the voltage dependence of Na⁺ channel inactivation in a positive direction, potentially increasing Na⁺ channel activity (1265). Indirect Ca²⁺ binding involves an additional binding site, the ”IQ“ domain, for Ca²⁺/CaM in the Nav1.5 COOH-terminal region, or multiple phosphorylatable sites, including serines 516 and 571, and threonine 594, in the DI-II linker region which is targeted by CaMKII (368, 799, 1211).

These two mechanisms both require prior Ca²⁺ binding to the EF hand motifs of Ca²⁺/CaM or CaMKII. They shift I_Na activation in a positive direction along the voltage axis (51). They also enhance slow I_Na inactivation (1112). Thus increased pipette [Ca²⁺] reduced I_Na density and (dV/dt)_max in patch-clamped myocytes (168). Increased and decreased [Ca²⁺]_i following increasing extracellular [Ca²⁺] and the acetomethoxy ester, BAPTA-AM, respectively increased and decreased I_Na densities in cultured neonatal rat myocytes while sparing unit Na⁺ channel conductance or gating (196). The Ca²⁺ channel antagonist verapamil and the Ca²⁺ ionophore calcimycin respectively increased and decreased Nav1.5 mRNA and Nav1.5 protein expression in rat cardiomyocytes (282, 850, 1114). Such properties appeared specific to cardiac as opposed to skeletal muscle Na⁺ channels (260, 1308). However, recent rapid Ca²⁺ photorelease methods did not induce Ca²⁺ modulation in Nav1.5, but did so in Nav1.4 (96).

Furthermore, more indirect interactions, which may in particular influence I_NaL (see sect. VIB1), may involve protein subunits associated with the Na⁺ channel. CaMKII associates with a COOH-terminal motif in actin-associated βIV-spectrin protein. The latter is strongly homologous to a binding domain within the L-type Ca²⁺ channel β subunit that has been invoked in the increased open probability and mean open time following channel phosphorylation (105, 1103). The CaMKII/βIV-spectrin interaction also operated in CaMKII-mediated phosphorylation of those voltage-gated Na⁺ channels residing at the cardiomyocyte intercalated disc (467, 716). This may mediate the modulation of steady-state Na⁺ channel inactivation, recovery from inactivation, and the magnitudes of I_NaL components observed in heterologous cells and primary myocytes. These actions likely involve separate Nav1.5 modulation pathways (467, 1209).

6. Effects on repolarization processes

Alterations in both [Ca²⁺]_i and its consequent phosphorylation levels thus modulate both I_CaL and I_Na and therefore action potential propagation and plateau duration, respectively. These alterations also affect membrane processes related to recovery from excitation. These may similarly have potential arrhythmic effects. Guinea pig ventricular APs do not involve K⁺-mediated I_to: its function may be replaced by a Ca²⁺-activated Cl⁻ current in turn activated by SR Ca²⁺ release (1067). In addition, I_Ks is enhanced by increased [Ca²⁺]_i (1136), α- or β-adrenergic stimulation, or PKC activation (1213). I_Kr is increased by PKA likely through increased [Ca²⁺]_i and PKC activation, primarily through altering its inward rectification properties (418). Modeling studies suggested that changes in I_Ks modify APD restitution to a greater extent than changes in I_Kr, although the ratio between the two currents is likely important. Spatial variations in the magnitude of these currents across the myocardium may also influence their restitution properties. This would be particularly the case in areas with steeper AP repolarization slopes including subendocardial M-cell layer, with possible implications for altered Ca²⁺ homeostasis upon restitution properties (752, 753, 1201).

Of membrane transporters, the Na⁺-K⁺-ATPase is maximally activated at physiologically occurring K⁺ concentrations. Instead, it is intracellular Na⁺ concentration, in turn dependent on I_Na and NCX activity, that may limit its activation. However, Na⁺-K⁺-ATPase activity is also increased by its phosphorylation by PKC following adrenergic stimulation or increased Ca²⁺ (340). Na⁺-K⁺-ATPase activity is thus stimulated by β- and α-adrenergic receptor activation and increased [Ca²⁺]_i (340). NCX in turn is energetically dependent on transmembrane Na⁺ gradients and therefore Na⁺-K⁺-ATPase activity.

B. Acquired Arrhythmic Disorders Associated With Altered Cellular Ca²⁺ Homeostasis

1. Ventricular arrhythmic effects of β-adrenergic activation

Abnormal cellular Ca²⁺ homeostasis is accordingly implicated in a wide range of arrhythmic pathologies. Cardiac failure and hypertrophy exemplify acquired conditions associated with arrhythmic phenomena (1244, 1299). Of these, cardiac failure is associated with chronically elevated adrenergic receptor-mediated cellular signaling (1238, 1242). In common with the inherited condition of CPVT, it has been analyzed in terms of schemes involving a PKA-mediated phosphorylation of RyR2 serine 2809 (or serine 2808, depending on species). This increases its open probability (P_o) by dissociating FKBP12.6 from RyR2 (1238). However, there remains uncertainty in the exact phosphorylatable RyR2 sites and the individual protein kinases involved, whether involving CaMKII or protein kinase G (PKG), and the functional roles of such phosphorylation (109, 1279).

At all events, murine systems provide useful models for the acute arrhythmic effects of altered Ca²⁺ homeostasis. This extends their utility in studies of the arrhythmic effects of altered ion channel expression (see sects. V and VI). For example, isolated murine hearts recapitulated the effects of β-adrenergic stimulation and L-type Ca²⁺ channel blockade both on ventricular arrhythmogenesis in intact Langendorff-perfused hearts and Ca²⁺ homeostasis in isolated myocytes. Programmed stimulation thus demonstrated arrhythmogenic tendencies predisposing to VT following both isoproterenol challenge and elevated extracellular [Ca²⁺]. Programmed electrophysiological fractionation analysis suggested that this took place with normal AP conduction, in contrast to the increased EGDs at shortened S1–S2 intervals in arrhythmic Kcne1−/− and Scn5a+/− hearts. The latter had then suggested reentrant as opposed to triggered arrhythmia (64, 416, 867, 1095).

These comparisons were consistent with the existence of an arrhythmogenic tendency that could result from triggered activity. Accordingly, in isolated ventricular myocytes, both isoproterenol and elevated extracellular [Ca²⁺] increased the amplitudes of electrically evoked Ca²⁺ transients as reported in other cardiac systems (104). Pretreatment with the dihydropyridine or benzothiazepine L-type Ca²⁺ channel blockers nifedipine or diltiazem both suppressed the VT and decreased, and prevented the isoproterenol-induced increases in Ca²⁺ transient amplitudes (63). Murine hearts thus recapitulate the increased inward I_CaL and consequently SR Ca²⁺ uptake, in turn producing spontaneous SR Ca²⁺ release and increased I_ti with β-adrenergic stimulation (905).

2. Ventricular arrhythmic effects of Epac activation

Murine systems also permitted an examination of the effects of Epac pathway activation on cardiac Ca²⁺ homeostasis, electrophysiological properties, and arrhythmic tendency (see sect. VIIA1). Of Epac isoforms, Epac1 is localized and functionally involved in nuclear signaling. Epac2 is located at the transverse tubules and may regulate arrhythmogenic SR Ca²⁺ leak (885). Baseline basal cAMP concentration and Epac activation levels appeared low in adult ventricular myocytes in the absence of β-adrenergic activation. Murine Epac1−/−, Epac2−/−, and CaMKIIδ−/− hearts showed normal in vivo baseline cardiac structure, ratios of heart to body weight, or early pressure overload-induced hypertrophy. They also showed normal physiological indexes of ventricular ejection, heart rate, and LV pressure indexes. All these indexes furthermore showed normal responses to dobutamine challenge. Western blot studies revealed normal SERCA and NCX expression levels. Isolated Epac1−/− and Epac2-/1 myocytes showed normal Ca²⁺ transient amplitudes and decline time courses, SR Ca²⁺ content, NCX function, and diastolic Ca²⁺ spark frequency when normalized to SR Ca²⁺ load (883).

RyR2 activation produced by Epac activation that is independent of PKA action was attempted using 8-(4-chlorophenylthio)-2′-O-methyladenosine 3′,5′-cyclic monophosphate (8-CPT). Low (1 μM) as opposed to high (100 μM to 1 mM) 8-CPT concentrations provide a specific >300-fold efficacy in activating Epac rather than PKA (437). In Langendorff-perfused mouse WT hearts, this produced arrhythmic effects often resulting in triggered activity and VT (446). 8-CPT also increased the frequencies of spontaneous cytosolic Ca²⁺ transients in neonatal rat cardiac myocytes (798), amplitudes of electrically evoked Ca²⁺ transients in mouse ventricular myocytes (849), and frequencies of Ca²⁺ sparks in isolated rat ventricular myocytes (884). There were also associated Ca²⁺ homeostatic abnormalities observable in isolated cardiomyocytes (446). There was thus a propensity to generate the spontaneous Ca²⁺ waves previously associated with cardiac arrhythmogenesis triggered by I_ti resulting in DADs (Figure 15, A and B) (101, 903). There were also increased incidences of ectopic Ca²⁺ release in both paced and resting cells (Figure 15C, a and b). These were abolished by treatment with the CaMKII inhibitor KN-93 (Figure 15Cc).

Figure 15.

Epac-induced RyR2 activation as an exemplar for Ca²⁺-mediated arrhythmia. A: propagation of a Ca²⁺ wave from one end of the cell to the other following 8-CPT challenge shown in successive confocal microscope frame scan images. B: Ca²⁺ fluorescence signal from six successive 2 μm × 2 μm regions of interest (a–f), placed along the long axis of a myocyte demonstrating progressively increasing delays in onset of the Ca²⁺ transient. C: Ca²⁺ signals in regularly stimulated ventricular myocytes (triangles mark timing of pacing stimuli) showing irregularly occurring ectopic Ca²⁺ transients during 8-CPT treatment (a, b) abolished by KN-93 (c). D: persistent ventricular tachycardia (VT) following programmed electrogram stimulation observed during perfusion with 8-CPT prevented by CaMKII inhibition with KN-93. E: triggered activity (*) during intrinsic activity observed during perfusion with 8-CPT, prevented by KN-93 pretreatment. F: epicardial (a) and endocardial APD₉₀ (b), ΔAPD₉₀ (c), and VERP (d) under control conditions (clear bars), during 1 μM 8-CPT treatment in the absence (black bars) and following 1 μM KN-93 pretreatment (striped bars). [From Hothi et al. (446).]

At the level of intact Langendorff-perfused murine hearts, challenge by either 8-CPT or a combination of isoproterenol and the PKA inhibitor H-89 (808) increased the incidences of VT provoked by programmed electrical stimulation (Figure 15D). It also increased incidences of triggering events following AP activation (Figure 15E). In contrast, spatial electrophysiological ventricular heterogeneities whether in the form of epicardial or endocardial APD₉₀, transmural or apico-basal gradients of repolarization, or of VERP, remained unchanged (Figure 15F). Temporal AP properties in the form of maximum APD₉₀ restitution gradients also remained normal (446). These findings provide clearcut contrasts with the altered, negative, ΔAPD₉₀s, and altered restitution properties thought to mediate arrhythmogenesis through the resulting reentrant substrate in the murine congenital LQTS and hypokalemic models described above (see sect. VI, B–E) (561, 1128, 1130).

Both the cellular effects of Epac on Ca²⁺ homeostasis and its arrhythmic effects in intact hearts were reduced by CaMKII inhibition using 1 μM KN-93. This was consistent with a dependence of Epac action on CaMKII activity (446). Finally, genetic ablation of Epac2, β1-AR, CaMKIIδ, and RyR2-S2814 phosphorylation all abolished Epac-dependent arrhythmogenic effects. These findings implicate Epac2 in β1-adrenergic arrhythmias through mechanisms involving CaMKIIδ and RyR2-S2814 phosphorylation (883).

3. Ventricular arrhythmic effects of direct RyR2-SR Ca²⁺ release channel activation

Interventions directly opening RyR2-Ca²⁺ release channels (1139) employed either caffeine or the immunosuppressant FK506 known to alter RyR2-FKBP12.6 binding (142, 527, 618, 1282, 1300). These also altered myocyte Ca²⁺ homeostasis and exerted arrhythmogenic effects in murine hearts (62). This agrees with previous reports in canine and rabbit hearts in vivo (481, 762). Programmed electrical stimulation confirmed that these proarrhythmic effects took place under conditions of normal conduction (62).

Caffeine either conserved or reduced the amplitudes of regularly evoked cytosolic [Ca²⁺] transients in isolated mouse myocytes consistent with a partial SR Ca²⁺ depletion (62). The latter would parallel the reduced myocyte SR Ca²⁺ in human heart failure (665). Both caffeine and FK506 increased the frequencies of spontaneous periodic peaks resulting from propagating Ca²⁺ waves (62) associated with SR Ca²⁺ loading by increased extracellular [Ca²⁺] (263, 1139, 1258). The latter effect is thought to result from increased Ca²⁺ spark frequency and to trigger arrhythmia (325) through activated NCX (760). These effects were abolished by the RyR inhibitor tetracaine.

However, diltiazem but not nifedipine prevented VT following programmed electrical stimulation in caffeine-treated hearts (1300). It also abolished the effects of both caffeine and FK506 upon the frequency of spontaneous Ca²⁺ release events. These findings correlate well with findings that diltiazem but not nifedipine inhibited RyR2-mediated SR Ca²⁺ leak in canine SR vesicles after addition of FK506 (1300) in addition to their L-type Ca²⁺ channel block (64). Such Ca²⁺ wave phenomena have also been observed in skeletal muscle under conditions of dihydropyridine receptor-RyR1 decoupling. These showed similar pharmacological sensitivities to caffeine, L-type Ca²⁺ channel block, and tetracaine challenge (183).

Cyclopiazonic acid (CPA) inhibits the activity of SR Ca²⁺-ATPase (SERCA) (1035) by blocking its Ca²⁺ channel (794). It inhibits SERCA activity in myocardial vesicle preparations (1032) while sparing Ca²⁺ sensitivity in contractile myofilaments, Ca²⁺ currents, and NCX activity (1107, 1302). It would therefore be expected to produce delayed reductions in SR Ca²⁺. CPA (100 and 150 nM) accordingly produced delayed rather than immediate reductions in Ca²⁺ transients in regularly stimulated fluo 3-loaded isolated myocytes. It correspondingly antagonized arrhythmic effects of isoproterenol, elevated extracellular [Ca²⁺], or caffeine. However, it did so only when the latter agents were added following rather than simultaneously with introduction of CPA. Both this introduction of arrhythmia and its rescue occurred without appearance of reentrant substrate, MAP waveforms, or VERPs (351).

4. Ventricular arrhythmic effects of purinergic receptor activation

Murine models also modeled the arrhythmic effects of extracellular fluid alterations resulting from metabolic change in surrounding tissue. A number of these may produce arrhythmia through Ca²⁺-dependent mechanisms. For example, both in normal and ischemic cardiac cells, sympathetic nerve terminals release ATP (934). Conversely, ventricular myocytes express both ionotropic P2X receptors and G protein-coupled P2Y receptors (1176). Cardiac ischemia is often accompanied by elevated interstitial [ATP] (165) and is often also associated with ectopic beats and VT. Extracellular ATP induced oscillatory contractions and ectopic APs (203) in isolated ventricular myocytes which additionally then showed afterdepolarizations with isoprenaline challenge (1079). Similarly, Langendorff-perfused murine hearts showed ectopic beats and VT (386) with challenge by extracellular ATP levels (∼100 μM) close to those previously reported under pathological conditions. Such findings did not occur with adenosine and UTP (213).

These findings implicated P2 purinergic rather than adenosine-activated P1 or UTP-activated P2 receptors. Thus ATP correspondingly induced sustained increases in diastolic Ca²⁺ and triggered Ca²⁺ waves that possessed relatively early (10–20 s) rather than prolonged (minutes) onsets. Single-cell voltage-clamp experiments demonstrated ATP-activated multiple cationic inward currents. These showed positive reversal potentials as expected of the Ca²⁺ (1 mM) rather the monovalent (Na⁺ and Cs⁺) ion concentrations in which the isolated ventricular myocytes were studied. All these arrhythmic, Ca²⁺ homeostatic, and electrophysiological effects were antagonized by the P2 receptor antagonists suramin and pyridoxalphosphate-6-azophenyl-2′,4′-disulfonic acid (PPADS). These observations are consistent with a Ca²⁺ influx following P2X receptor opening. These currents could induce arrhythmogenic changes in resting membrane potentials. APs induced by repetitive (2 Hz) current-clamp steps in the presence of ATP (100 μM) were followed by oscillatory resting membrane potential changes, DADs, and triggered ectopic APs and EADs, and themselves were frequently prolonged (386).

In contrast, metabotropic P2Y receptors are mainly coupled to G_q proteins. They then activate PKC through phospholipase C-dependent production of diacylglycerol (DAG) (157). PKC in turn can target L-type Ca²⁺ channels, RyR2s, and voltage-gated Na⁺ and K⁺ channels. ADP (100 μM), associated with P2Y receptor activation, did not produce even delayed (>3 min) diastolic Ca²⁺ elevations. However, its application was followed by gradual declines in diastolic Ca²⁺ and [Ca²⁺]_i transients. Its sustained application (>3–5 min) resulted in prolonged waveforms of Ca²⁺ transients induced by field stimulation. There was then an ultimate failure of subsequent stimulation to induce [Ca²⁺]_i transients. These changes persisted even following ADP washout (386).

5. Ventricular arrhythmic effects of acidotic stress

Murine hearts also recapitulated the acute arrhythmic effects that followed the imposition and withdrawal of acidotic challenge. These effects showed features consistent with altered Ca²⁺ homeostasis involving CaMKII action (860). This may be important in ischemia/reperfusion injury. Intracellular acidosis is then a component of the ischemia. The reperfusion that follows produces rapid return of pH to its background value. Cell acidification is also an early accompaniment to acute myocardial ischemia (1168). Such acidication produced both arrhythmogenic effects (860) and increased CaMKII activity (466, 586).

MAPs recorded from Langendorff-perfused rat and mouse hearts subject to imposition followed by withdrawal of respiratory acidosis showed ectopic beats. Their frequency was reduced by the CaMKII kinase inhibitor KN-93, the SR uptake inhibitor thapsigargin, and the RyR2 SR Ca²⁺ release blockers ryanodine and dantrolene (992). They were not observed in transgenic mice expressing a CaMKII inhibitory peptide targetted to cardiac longitudinal SR (507). The acidosis increased phosphorylation of the Thr¹⁷ site of phospholamban (PT-PLN) increasing SR Ca²⁺ load. KN-93 inhibited the latter effect. The ectopic activity was triggered by DADs, particularly in epicardial myocytes. The latter was also prevented by KN-93.

Acidification also produced spontaneous SR Ca²⁺ release, spontaneous AP firing, and triggered arrhythmias (264). It increased the amplitudes of caffeine-induced Ca²⁺ transients that were abolished by KN-93 (586). These were attributed to increased NCX activity in response to Ca²⁺ leak from an overloaded SR likely due to CaMKII-dependent increase in SR Ca²⁺ uptake (860). Thus arrhythmias following acidosis result from a CaMKII activation that increases SR Ca²⁺ load occurring mainly through from increased PT-PLN activity.

MAP recordings from Langendorff-perfused whole murine heart preparations could be performed in parallel with single whole-ventricular myocyte AP and Ca²⁺ fluorescence measurements under acidotic stress conditions. These made it possible to separate CAMKII-dependent and CAMKII-independent contributions to the arrhythmic changes following metabolic acidification (879). Lactate challenge induced spontaneous arrhythmogenesis in both intrinsically active and regularly paced Langendorff-perfused murine hearts. It also induced spontaneous Ca²⁺ waves in isolated intact fluo 4-loaded myocytes. Both these effects were reversibly abolished by KN-93. Acidification also produced spontaneous AP firing and membrane potential oscillations in isolated ventricular myocytes whose patch pipette solutions permitted [Ca²⁺]_i to increase. Both KN-93 and use of EGTA-buffered pipette solutions holding cytosolic [Ca²⁺]_i constant during the acidosis abolished the latter effects. Together these findings implicate CaMKII-dependent waves of SR Ca²⁺ release in spontaneous arrhythmic events during metabolic acidification.

However, programmed electrical stimulation similarly induced VT during metabolic acidification. This correlated with appearance of an arrhythmic substrate resulting from reduced LV transmural repolarization gradients that resulted from prolongations of epicardial but not endocardial APD₉₀. However, these latter parameters were KN-93-resistant both before and after acidification. Thus agents inhibiting CaMKII action reduced AP triggering while sparing the associated arrhythmic substrate produced by cell acidification (879).

6. Roles of Ca²⁺ in acute atrial arrhythmogenesis

Mechanisms initiating AF are incompletely understood. Nevertheless, significant contributions to these may arise from alterations in cellular Ca²⁺, particularly diastolic, signaling. These could involve release of Ca²⁺ from nevertheless finite SR Ca²⁺ stores, a process induced by an initial extracellular Ca²⁺ entry (1202). The resulting increase in NCX activity might then result in EAD or DAD phenomena initiating focal firing. Acute episodes of acquired atrial arrhythmias can follow cardiothoracic, particularly coronary artery bypass grafting, surgery, typically at ∼2 but rarely beyond 7–15 postoperative days. They are reduced by either β-adrenoceptor or Ca²⁺ channel antagonism (60). Sustained arrhythmia may require reentrant substrate that might arise from altered balances between conduction velocity (θ) and refractory period (269, 819, 822). Such substrate may develop following physiological and anatomical remodeling (184, 269, 270, 451, 658, 923, 1194).

Murine WT Langendorff-perfused hearts recapitulated these roles of altered Ca²⁺ homeostasis in producing acute acquired atrial arrhythmias. They related these arrhythmias to an occurrence of diastolic Ca²⁺ events in isolated atrial myocytes. It was then possible to pharmacologically separate the contributions from altered SR Ca²⁺ storage or release, and extracellular Ca²⁺ entry, in this initiation (1338). These studies thus implicated an increased Ca²⁺-induced Ca²⁺ release from a nevertheless finite SR Ca²⁺ store ultimately dependent upon entry of extracellular Ca²⁺ in such arrhythmogenesis.

Thus, in isolated Langendorff-perfused hearts, caffeine exerted atrial proarrhythmic effects immediately following but not >5 min after application. Regularly stimulated atrial myocytes then correspondingly showed diastolic Ca²⁺ waves and progressive reductions in their evoked Ca²⁺ transients and these diastolic events with prolonged exposure (1338). Both findings are directly explicable in terms of increased Ca²⁺ release through sensitized SR RyR2-Ca²⁺ release channels from a finite and therefore depletable intracellular Ca²⁺ store (1187). The arrhythmic phenomena associated with caffeine challenge could thus be abolished by pretreatment with either the SERCA inhibitor CPA or the I_CaL blocker nifedipine (1338). At the cardiomyocyte level, CPA applied by itself produced successive reductions in the amplitudes of evoked Ca²⁺ transients. Use of nifedipine as a I_CaL blocker implicated a dependence of extracellular Ca²⁺ entry on these effects (1048, 1142). Thus nifedipine by itself promptly reduced the evoked Ca²⁺ transients. These again increased then declined in amplitude with caffeine challenge but in an absence of diastolic Ca²⁺ events (1338).

The dependence of acute atrial arrhythmogenesis upon extracellular Ca²⁺ entry was contrastingly accentuated by the Ca²⁺ channel opening agent FPL-64176. This is known to prolong depolarization-induced L-type Ca²⁺ channel opening and slow repolarization-induced L-type Ca²⁺ channel closure (306, 1347). FPL-64176 challenge elicited diastolic Ca²⁺ events in isolated myocytes and atrial arrhythmogenesis intact perfused hearts. It did so without changing AERP. Both effects were inhibited by pretreatment with nifedipine, caffeine, or CPA. Enhanced extracellular Ca²⁺ entry thus exerts acute atrial arrhythmogenic effects nevertheless dependent on diastolic Ca²⁺ release (1335).

C. Genetic Abnormalities in the RyR2-SR Ca²⁺-Release Channel: CPVT

1. Clinical features of CPVT

Genetic disorders have also provided useful systems to study how abnormal Ca²⁺ homeostasis may cause cardiac arrhythmogenesis (676, 917). One such exemplar of these is offered by the CPVT typically associated with RyR2 or calsequestrin (CASQ) mutations (786, 1241). An autosomal dominant CPVT is associated with cardiac RyR2 (916). A recessive variant is associated with homozygous cardiac calsequestrin (CASQ2) gene mutations (610). CPVT-like conditions are also associated with ankyrin-B mutations (790). RyR2 mutations are also implicated in ARVC type 2 (82, 234, 621, 1135).

CPVT presents as an occurrence of acute VT on adrenergic stimulation that may degenerate into potentially fatal VF (175, 913). It typically presents earlier, at age ∼7–9 yr, than BrS (625). CPVT patients show relatively normal resting ECGs. However, exercise or catecholaminergic challenge elicits arrhythmic features that include bidirectional VT with 180 degree beat-to-beat alternations in the QRS axis. Clinical management involves β-blocker therapy but ∼30% of patients require implantable cardioverter defibrillators (ICDs) (913). Left cardiac sympathetic denervation is considered for intractable arrhythmias (1261). CPVT has a poor prognosis, particularly in males with a RyR2 mutation. It likely contributes ∼15% of SCD in structurally normal hearts (678, 1123). A third of CPVT patients have family histories of premature sudden death or stress-related syncope (625). Of these, 50–70% show a genetic abnormality (58, 913).

2. Human and murine RyR2 mutations in CPVT

Two groups of investigators made the initial identifications of RyR2 mutations in CPVT (613, 916). Five clinically affected carriers proved to have RyR2-S2246L, RyR2-R2427S, RyR2-N4104K, and RyR2-R4497C. Three unrelated Finnish families carried RyR2-P2328S, RyR2-Q4201R, and RyR2-V4653F. The latter abnormalities were accompanied by a 30–33% mortality by age 35 yr. RyR2-P2328S was associated with normal resting ECGs including normal QTc intervals. Ten of 12 such patients showed exercise-induced VT. Two were asymptomatic but exhibited abnormalities on clinical testing.

A large proportion of the >70 RyR2 mutations reported subsequently in CPVT, a number of which have been replicated in murine models (Table 7), have similarly been base-pair substitutions involving highly conserved residues. The mutations mainly cluster in three, NH₂-terminal (176–433), central (2246–2504), and COOH-terminal regions (4104–4653) of the RyR2 gene (1240). Approximately 20 have been studied in vitro (676). All three regions are represented in murine studies of ventricular properties associated with the RyR2-R420W (851), RyR2-R176Q/+ (534), RyR2-R4496C/+ (172), RyR2-R2474S/+ (1157), and heterozygotic and homozygotic RyR2-P2328S exemplars, respectively (364, 1340). All the murine variants showed normal structure and histology, despite known associations of RyR2-R176Q with human ARVC2. Finally, studies in HEK293 expression systems indicate that large deletions encompassing exon 3 in the NH₂-terminal RyR2 region remain compatible with continued RyR2 stability and function through insertion of a flexible loop into the β-trefoil domain (119). Ca²⁺ release then terminates at a lower luminal Ca²⁺ (1113).

Table 7.

Transgenic murine models for catecholaminergic polymorphic ventricular tachycardia (CPVT)

Gene	Genetic Modification	Phenotype	Reference Nos.
RyR2 (NH2-terminal region)	RyR2-R176Q/+	CPVT (ARVC2)	534
	RyR2-R420W	CPVT	851
RyR2 (central region)	RyR2-P2328S/+ and RyR2-P2328S/P2328S	CPVT	364, 1340
	RyR2-P2328S/P2328S	AF	569, 832, 996, 1334, 1337
	RyR2-R2474S/+	CPVT	585, 629, 1156
RyR2 (COOH-terminal region)	RyR2-R4496C/+	CPVT	172, 673
Casq2	Casq2-D307H overexpression	CPVT	266
	Casq2−/−	CPVT and cardiac hypertrophy	581
	Casq2+/−	No phenotype	199
	Casq2-D307H/D307H	Spontaneous arrhythmias	1078
	Casq2-ΔE9/ΔE9	Spontaneous arrhythmias; cardiac hypertrophy with age	1077
	Casq2-R33Q/R33Q	Spontaneous arrhythmias	951

3. Electrocardiographic features of murine models with genetically modified RyR2

Genetically modified RyR2 mouse models recapitulated electrophysiological phenomena associated with initiation of arrhythmogenesis in CPVT. In vitro studies associated gain-of-function RyR2, and loss-of-function Casq2, mutations with increased SR Ca²⁺ leakage thereby increasing [Ca²⁺]_i (509, 631). This would give a situation resembling clinical digitalis toxicity (625). In either case there would be an expected increase in NCX activity. This would increase I_ti that in turn drives DADs. The latter have been suggested to potentially cause triggered activity where they produce membrane depolarizations attaining the I_Na reactivation threshold (581, 673). This would particularly affect Purkinje fibers owing to their greater susceptibility than ventricular myocytes to Ca²⁺ overload due to their greater Na⁺ loads and longer APDs (1174). Clinical arrhythmic activity in both CPVT and digitalis toxicity is thus accompanied by bidirectional VT reflected in alternating 180° rotations of the electrocardiographic QRS axis between beats (625, 674). This can then be followed by degeneration into rapid polymorphic VT and VF.

Electrocardiographic properties of a significant number of these mice expressing variant RyR2 recapitulated these clinical arrhythmic and ECG features. Polymorphic VT and/or bidirectional VT was observed in ambulant RyR2-R4996C/+ (172, 673) and RyR2-P176Q/+ mice (534) upon combined or individual epinephrine and caffeine administration, and in RyR2-R2474S/+ mice during treadmill testing (585, 629, 1157). Endocardial optical potentiometric mapping localized the arrhythmic foci in the RyR2-R4496C/+ to triggered firing alternating between the left and left Purkinje bundles. Thus chemical ablation of the RV Purkinje network converted the bidirectional VT to a monomorphic, wide-QRS, VT (175).

Episodes of bigeminy and BVT also occurred in anesthetized heterozygotic RyR2-P2328S/+ (RyR2^S/+) hearts following isoproterenol challenge. They further occurred in anesthetized homozygotic, RyR2-P2328S/P2328S (RyR2^S/S) hearts in association with multiple ventricular ectopic beats, both before and following isoproterenol challenge (1334). These findings potentially implicated an involvement of Purkinje system conduction in CPVT-associated arrhythmia (1016). Thus, before pharmacological challenge, there was only a single BVT episode among 12 RyR2^S/S mice. Pharmacological, epinephrine and caffeine, challenge selectively increased the number of RyR2^S/S, but not RyR2^S/+ or WT mice showing arrhythmias in the form of bigeminy, bidirectional, monomorphic, or polymorphic VT. In common with clinical findings, the bidirectional VT appeared as salvos of four or more beat-to-beat 180° QRS axis alternations and bigeminy. They formed repeated occurrences of a ventricular premature complex succeeded by a normal beat. Polymorphic VT or VF appeared as series of three or more ventricular ectopics with irregular QRS axes. Monomorphic VT was identified in runs of abnormally shaped ECG complexes nevertheless occurring with regular rates and rhythms (1340).

4. Altered Ca²⁺ homeostasis in murine models of altered RyR2

The murine models for altered RyR2 consistently showed evidence of abnormal Ca²⁺ homeostasis. Murine RyR2-R176Q/+ cardiomyocytes showed higher incidences of spontaneous, nontriggered Ca²⁺ oscillations compared with WT both before and following isoproterenol stimulation. Isoproterenol (100 nM) induced a range of Ca²⁺ transients. These included slowed decays, spontaneous diastolic transients, and prolonged evoked transients (534). A comparison of Ca²⁺ transients from regularly stimulated WT, heterozygotic (RyR2^S/+), and homozygotic (RyR2^S/S), RyR2-P2328S myocytes similarly suggested altered Ca²⁺ release patterns and a resulting SR Ca²⁺ depletion (364). Thus WT, RyR2^S/+, and RyR2^S/S myocytes showed indistinguishable peak amplitudes in their evoked Ca²⁺ transients. RyR2^S/S myocytes showed subsidiary events, as well as irregular Ca²⁺ signaling patterns. Isoproterenol (100 nM) elicited ectopic peaks and subsidiary events in WT and RyR2^S/S but not RyR2^S/+. However, it reduced the amplitudes of evoked peaks in the RyR2^S/+ myocytes. Isoproterenol-treated RyR2^S/S myocytes additionally showed spontaneous Ca²⁺ propagating waves (Figure 16A). Such waves have been previously associated with arrhythmogenic properties under conditions where they were produced by caffeine (62). RyR2-R2474S/+ myocytes showed markedly greater Ca²⁺ spark frequencies with isoproterenol addition (584). They also showed SR Ca²⁺ thresholds for spontaneous Ca²⁺ transients, cAMP-induced Ca²⁺ sparks and waves following PKA-mediated RyR2 phosphorylation that were greatly reduced compared with WT (1157).

Figure 16.

Proarrhythmic features in RyR2-P2328S hearts. A: sequential gallery of confocal microscope images showing Ca²⁺ waves in a single fluo3-loaded ventricular homozygotic RyR2-P2328S (RyR2^S/S) myocyte following isoproterenol (100 nM) challenge. Arrowed: path taken by typical Ca²⁺ wave. B–D: epicardial monophasic action potentials in intrinsically active RyR2^s/s hearts. B: spontaneous early afterdepolarizations (EADs) (*) followed by episodes of sustained monomorphic VT (sVT). C: coupled beats. D: persistent ventricular fibrillation (VF) following the cessation of regular S1 pacing after two intrinsic MAPs. E: S2 extra-stimuli during PES typically producing limited episodes of nonsustained VT in the absence (left trace) but sustained (>30 s) VT in the presence of 100 nm isoproterenol (right trace). F–I: electrophysiological assessment of Na⁺ channel function in RyR2^S/S, Scn5a+/−, and WT atria. F: left atrial intracellular APs showing conduction latencies from WT, Scn5a+/−, and RyR2^S/S myocytes and their corresponding (dV/dt)_max (G). H and I: loose patch-clamp recordings of currents during a 100 mV 50-ms activation step following 50-ms prepulses between 20 and 100 mV for WT (H) and RyR2^S/S atria (I). [From Goddard et al. (364) and King et al. (568).]

5. Spontaneous and provoked arrhythmia related to electrophysiological properties

Electrophysiological studies in Langendorff-perfused RyR2^S/+ and RyR2^S/S hearts went on to distinguish between phenotypes reflecting tendencies to provoked, and both provoked and spontaneous arrhythmia. These employed conditions of either regular or programmed electrical stimulation. They were performed both before and following isoproterenol challenge (364) (Figure 16, B–E). Unchallenged RyR2^S/S though not RyR2^S/+ hearts showed extrasystolic events. These were often followed by spontaneous sVT during intrinsic activity (Figure 16, B and C). RyR2^S/+ hearts correspondingly showed nonsustained VT with programmed electrical stimulation but not regular pacing. However, following isoproterenol treatment, RyR2^S/+ hearts showed both nonsustained and sustained VT with both regular pacing and programmed electrical stimulation. RyR2^S/S hearts then showed higher incidences of nonsustained VT before and mainly sustained VT during both regular pacing and programmed electrical stimulation, particularly at higher pacing frequencies (Figure 16, D and E). Correspondingly, isolated RyR2-R4496C/+ but not WT myocytes showed DAD and triggered activity. These increased in frequency with isoproterenol administration and were abolished by ryanodine (673). This increased Ca²⁺ release could extend to compromising SR Ca²⁺ stores sufficiently to compromise both atrial and ventricular contractility (314).

6. Evidence for reentrant substrate in RyR2-P2328S hearts

In addition to demonstrating such potential triggering events, Langendorff-perfused RyR2^S/S hearts additionally demonstrated evidence for arrhythmic substrate (816, 1340). Regularly paced RyR2^S/+ and RyR2^S/S hearts showed ventricular AP recovery characteristics of ventricular MAP duration, refractory period (364), and AP restitution properties indistinguishable from WT. This was the case whether before or following catecholaminergic challenge (984). Ventricular AP conduction velocities, assessed by vectorial analysis of local epicardial activation times using multiarray electrode array recordings, were indistinguishable in untreated WT and RyR2^S/S. However, isoproterenol and caffeine challenge that would elicit arrhythmia specifically in RyR2^S/S ventricles correspondingly slowed these conduction velocities. Intracellular microelectrode recordings demonstrated correspondingly reduced values of (dV/dt)_max. The latter is known directly to correlate with conduction velocity and peak I_Na (567, 1047)_. Such findings directly parallelled alterations in membrane excitability following manipulations of Ca²⁺ homeostasis in skeletal muscle (1161).

Further studies corroborated this existence of potential arrhythmic substrate. They revealed evidence for reduced membrane Nav1.5 expression and a correspondingly reduced I_Na. These could account for the slowed ventricular conduction velocities in RyR2^S/S ventricles (832). Thus Western blot analyses demonstrated that both whole tissue and membrane fractions of RyR2^S/S ventricles showed similar Cx43, but reduced Nav1.5 protein expression compared with WT. Similarly, loose patch-clamp studies demonstrated a reduced I_Na. Extrasystolic S2 stimulation following 8-Hz S1 pacing at progressively decremented S1S2 intervals correspondingly demonstrated increases in ventricular arrhythmic tendency despite unchanged VERPs in Langendorff-perfused RyR2^S/S hearts. Similarly, dynamic pacing employing progressively reduced BCLs produced greater reductions in conduction velocity at equivalent BCLs and diastolic intervals in RyR2^S/S compared with WT ventricles. Yet both showed comparable changes in APD₉₀ and its alternans.

These findings were compatible with both short- and long-term consequences of abnormal cytosolic Ca²⁺ homeostasis for Nav1.5 function. Thus ectopic activity and possible effects on Nav1.5 activation may follow transient [Ca²⁺]_i elevations consequent upon catecholaminergic stimulation. Chronic elevations in [Ca²⁺]_i might downregulate either Nav1.5 expression or activity. Either effect would reduce AP conduction velocity potentially generating arrhythmic substrate. Furthermore, the short-term events could potentially initiate triggering arrhythmic events then perpetuated by a longer term preexisting arrhythmic substrate reflected in decreased conduction. Such hypotheses merit further exploration. They have broader implications for potential therapeutic interventions involving Ca²⁺ homeostasis in management of potentially Nav1.5-dependent arrhythmic conditions (832).

Other situations related to disturbed Ca²⁺ homeostasis also caused conduction disturbance and major cardiac arrhythmia related to compromised I_Na. FKBP12 shares an 85% homology with FKBP12.6. However, it has mainly been associated with a modulation of skeletal rather than cardiac muscle RyR1-mediated excitation-contraction coupling (1133). Nevertheless, FKBP12 has reported effects on RyR2. Its haploinsufficiency is lethal in utero, and this is accompanied by ventricular hypertrabeculation, noncompaction, and septal defect (1061). Transgenic αMyHC-FKBP12 mice overexpressing FKBP12.6 showed high incidences (38%) of SCD. This accompanied ECG, electrophysiological, and optical mapping evidence of slowed conduction, slowed AP upstrokes, and prolonged APDs. Whole cell patch-clamp analyses attributed these findings to an 80% reduction in peak I_Na, its slowed recovery from inactivation, positive shifts of its steady-state activation and inactivation curves, and enhanced I_NaL. Conversely, cardiomyocyte-restricted conditional FKBP12 knockout (FKBP12-f/f/αMyHC-Cre) ventricular cardiomyocytes showed faster AP upstrokes and greater than twofold increased peak I_Na. These effects were reversed by dialysis of exogenous recombinant FKBP12 protein. This resulted in I_Na changes seen in the overexpression αMyHC-FKBP12 myocytes (744).

7. RyR2 phosphorylation and arrhythmic tendency

Murine hearts have also been used to model the mechanisms relating RyR2 protein abnormalities to alterations in RyR2-Ca²⁺ release function that would potentially trigger arrhythmia, although these remain controversial. These include the arrhythmic changes in acquired conditions including cardiac failure, which have been attributed to increased RyR2 open probabilities. However, the RyR2 is a large and complex molecule. Its monomer is associated with numerous potentially regulatory proteins including CaM, FKBP12.6, PKA, PP1 and PP2A, and CaMKII (107). Hence, altered regulation of RyR2 gating by processes such as phosphorylation is likely critical to altered diastolic SR Ca²⁺ release. One initial line of inquiry explored suggestions that PKA could alter RyR2-Ca²⁺ release function through a RyR2 hyperphosphorylation that dissociates a normally bound FKBP12.6 (745, 747, 856). This specifically implicated RyR2-S2808, as opposed to other possible PKA and CaMKII targets in cardiac failure. Mice with the nonphosphorylatable substitution RyR2-S2808A ablating the RyR2-S2808 phosphorylation site were correspondingly not susceptible to effects of adrenergic agonists or other pathological stressors known to induce congestive cardiac failure (747, 1242). Furthermore, selectively preventing PKA-mediated RyR2-S2808 hyperphosphorylation improved postinfarction or exercise cardiac performance in the genetically modified mouse (630, 1039, 1241, 1242).

These findings led to suggestions for a physiological role for PKA-mediated RyR2-S2809 phosphorylation in the increased cardiac contractility following sympathetic stimulation during exercise. Persistent pathological increases in such adrenergic activity would cause RyR2-S2809 hyperphosphorylation. This would dissociate FKBP12.6 normally bound to and stabilizing the RyR2 and thereby cause an arrhythmogenic diastolic leakage of SR Ca²⁺ (747). However, subsequent studies along similar or independent lines did not confirm critical predictions of this RyR-S2809 hyperphosphorylation hypothesis (449).

A first group of experiments concerned the physiological consequences of RyR2-S2809 phoshorylation. Channel function in recombinant phosphorylated and dephosphorylated RyR1 and RyR2 (RyR1-S2843D and RyR2-S2809D, and RyR1-S2843A and RyR2-S2809A, respectively) expressed in HEK293 cells did not show alterations in their responses to applied Ca²⁺, Mg²⁺, and ATP in single-channel and [³H]ryanodine binding measurements, or to caffeine in Ca²⁺ imaging measurements, that differed from findings in WT. Coimmunoprecipitation and Western blot analysis demonstrated similar binding of FK506-binding proteins. Metabolic labeling generated evidence for RyR2 phosphorylation sites other than S2809 (1089). Excitation-contraction coupling in whole hearts and isolated cardiomyocytes from a RyR2-S2808A mouse model showed normal β-adrenergic responses with a persistent maladaptive cardiac remodeling following chronic stress. Single-channel RyR2-S2808A activity showed only minor alterations and then only in their responses to applied Ca²⁺ (97, 356, 708).

Second, FKPB12.6 binding or dissociation did not modify RyR2 sensitivities to applied Ca²⁺ or caffeine, their tendencies to spontaneous or store overload-induced Ca²⁺ release, or the effect upon these of FK506, in HEK293 cells either coexpressing RyR2 and FKBP12.6 or expressing RyR2 alone. In these particular studies, FKBP12.6-null mice did not show altered tendencies for stress-induced ventricular arrhythmias. The properties of their RyR2 channels were indistinguishable from WT (1281).

Third, a series of experiments in cardiomyocytes indicated that PKA phosphorylation did not dissociate FKBP12.6 from RyR2. FKBP12.6 occurred in lower abundance than RyR2 therefore binding only to ∼15% of available RyR2s, in contrast to the more abundant FKBP12. However, PKA-mediated RyR-S2809 phosphorylation dissociated neither FKBP12.6 nor FKBP12 from RyR2 (382).

Fourth, Ser-2808 phosphorylation was often relatively insensitive to either PKA activation or inhibition (13, 109, 657). PKA-dependent increases in Ca²⁺ spark frequency in ventricular myocytes were instead attributable to enhanced SR Ca²⁺ load following phospholamban B phosphorylation rather than altered resting RyR2 function even under conditions of controlled [Ca²⁺]_i (657). CaMKII- rather than PKA-dependent phosphorylation of RyR2 was implicated in the enhanced SR diastolic Ca²⁺ leak and reduced SR Ca²⁺ load consequent upon adrenergic stimulation or cardiac failure (13, 227). Finally, where CaMKII, or S-nitrosylation, mediated phosphorylation did enhance RyR2-mediated Ca²⁺ sparks and SR Ca²⁺ leak (232, 382, 884), it did so possibly at an alternative, S2814, site (943).

Nevertheless, phosphorylation events in general could still be expected to increase RyR2 Ca²⁺ sensitivity (747, 1241). This would then increase SR Ca²⁺ leak, with potential arrhythmogenic triggering effects (325). This would also ultimately deplete SR Ca²⁺ and thereby compromise Ca²⁺ transients and the resulting contraction in ventricular myocytes (411, 665, 892, 895). Thus both RyR2-P2328S and RyR2-Q4201R mutations were associated with decreased RyR2-FKBP12.6 binding, increased Ca²⁺ opening and positive shifts in half-maximal inhibitory Mg²⁺ concentrations after PKA phosphorylation. In contrast, the benzothiazepine derivative JTV519 (K201) appeared to both enhance RyR2-FKBP12.6 binding and normalize channel function (631). The RyR2-G230C mutant showed similar biophysical defects and demonstrated decreased FKBP12.6 binding and in common with RyR2-P2328S, normal luminal Ca²⁺ activation, but a negative shift in the cytosolic Ca²⁺ dependence of activation (763). Similar hyperphosphorylation hypotheses have been postulated for the triggering of arrhythmia in other systems, including a rabbit LQTS2 model (1118).

Furthermore, murine hearts with a heterozygous FKBP12.6 knockout correspondingly showed polymorphic VT with adrenergic challenge. This was rescued by JTV518 (631, 1241). RyR2-R2474S hearts also demonstrated reduced FKBP12.6-RyR2 binding on adrenergic stimulation. Rycal (S107) stabilized the binding, and therefore also stabilized the RyR2 closed state. It thus rescued the VT (629). The RyR2-S2246L, RyR2-R2474S, and RyR2-R4497C mutations similarly impaired RyR2-FKBP12.6 binding in parallel with their CPVT phenotype (1241). Ca²⁺ release by RyR2-P2328S, RyR2-Q4201R, and RyR2-V4653F expressed in HEK293 cells similarly showed rightward shifts in half-maximal inhibitory [Mg²⁺] and decreased FKBP12.6 binding (631, 1239).

8. Store overload-induced Ca²⁺ release

A store overload-induced Ca²⁺ release mechanism has been suggested as an alternative or complementary mechanism to altered FKBP12.6 binding in producing CPVT phenotypes. Murine RyR2-R4496C, equivalent to human RyR2-R4497C (511), showed enhanced Ca²⁺ and caffeine sensitivitivity when expressed in HEK293 cells. Lipid bilayer and HL1-cardiomyocyte studies similarly suggested increased diastolic SR Ca²⁺ release with sympathetic activation (348, 631, 1241). Despite this, interaction between RyR2-R4496C and FKBP12.6 was normal both before and following such adrenergic challenge (348, 509). Furthermore, JTV519 did not inhibit either in vitro DADs or in vivo ventricular arrhythmias induced by adrenergic challenge (673). In lipid bilayers, JTV519 did suppress both spontaneous Ca²⁺ release and [³H]ryanodine-RyR2 binding. However, it did so independently of FKBP12.6 association (470), likely through a JTV519 binding site in the RyR2 2114–2149 domain (1286). The latter may further bind the central RyR2 protein domain (2234–2750) and thereby stabilize the RyR2.

These findings suggested that the increased SR-Ca²⁺ release was alternatively attributable to an increased RyR2 channel sensitivity specifically to SR luminal Ca²⁺ that would thereby decrease cytosolic Ca²⁺ release thresholds (509, 510). Such a store overload-induced Ca²⁺ release mechanism would be consistent with observed diastolic Ca²⁺ leaks associated with the RyR2-R176Q/T2504M and RyR2-L433P mutations that involve the channel domains of RyR2 (1131). It would also be compatible with the loss of such luminal activation with the RyR2-A4860G mutation (508).

It is nevertheless possible that these detailed differences in arrhythmogenic mechanisms could reflect genuine differences between phenotypic consequences of the various genetic RyR2 variants. Thus in vitro assays associated mutations in the RyR2 central region not only with decreased (RyR2-R2427S, RyR2-S2246L, and RyR2-P2328S) but also with increased (RyR2-N2386I and RyR2-Y2392C) FKBP12.6 binding affinities (1135, 1241). COOH-terminal RyR2 mutations (RyR2-Q4201R, RyR2-R4496C, and RyR2-V4653F) were associated with decreased FKBP12.6 affinities (1241).

D. Genetic Abnormalities in the RyR2-SR Ca²⁺-Release Channel: Atrial and SAN Arrhythmic Phenotypes

1. Genetic RyR2 abnormalities associated with acute AF phenotypes

A number of RyR2 mutations are associated with SND and atrial arrhythmic in addition to CPVT phenotypes. Atrial arrhythmia can occur as a first clinical manifestation of CPVT (623). Both murine and clinical reports suggest that this could increase the risk of VT in this condition (302). The RyR2 exon 3 deletion is associated with a broad range of atrial, ventricular, and SND phenotypes. Some single RyR2 mutations associated with CPVT are also associated with exercise-associated atrial arrhythmias. These include RyR2-P2328S (613, 1099, 1104), RyR2-G3946A (900), RyR2-S4153R (542), and RyR2-A7420G. Rapid AF culminating in CPVT has been associated with RyR2-M4109R and RyR2-I406T (835). VT and AT following isoproterenol infusion has been associated with RyR2-W4645R (86).

Murine hearts provided realistic models for the acute initiation of atrial arrhythmias in the presence of genetic RyR2 abnormalities classically associated with CPVT. They similarly related these to specific SR Ca²⁺ release abnormalities (1337). Studies in a selected group of mouse models with modified RyR2 genotypes demonstrated that, in common with the ventricular arrhythmia, intracellular diastolic SR Ca²⁺ leak via RyR2 can trigger, and inhibiting this leak can prevent, AF (184, 824, 1040, 1098, 1334).

Increased RyR2-dependent Ca²⁺ leakage due to enhanced CaMKII activity may be an important downstream effect of CaMKII that generates susceptibility to AF. Thus rapid atrial pacing in RyR2-R176Q/+ mice increased atrial arrhythmic tendency compared with WT mice (184). This atrial arrhythmia in RyR2-R176Q/+ hearts accompanied increased CaMKII-mediated RyR2 phosphorylation. Both the arrhythmia and increased SR leak of Ca²⁺ were prevented by pharmacological inhibition of CaMKII. Similarly, preventing CaMKII-mediated RyR2 phosphorylation by a RyR2-S2814A mutation prevented AF in a vagotonic experimental model (534). Atrial biopsies from mice with atrial enlargement and spontaneous AF showed increased CaMKII-mediated RyR2 phosphorylation (184).

AF can also follow alterations in RyR2-associated proteins. These include FKBP12.6 deletion (655, 1081), overexpression of the Ca²⁺ binding junctate 1 (438, 439), or the sarcoplasmic transmembrane protein junctin (438), as well as deletion of the Ca²⁺- and GTP-dependent membrane-associated annexin Aa7 (Anxa7) (423). Intraesophageal burst pacing induced AF and diastolic SR Ca²⁺ leaks at the myocyte level in RyR2-R2474S+/−, RyR2-N2386I+/−, and RyR2-L433P+/− mice (1040). The RyR2-R2474S+/− RyR2s showed an increased oxidation and calstabin2 depletion compared with WT. S107 is known to stabilize RyR2-calstabin2 interactions by inhibiting oxidation/phosphorylation-induced calstabin2 dissociation from RyR2. It rescued both the increased SR-Ca²⁺ release and AF phenotype. It contrastingly failed to do so in calstabin2-deficient mice (1040).

The RyR2-P2328S mutation is clinically associated with atrial in addition to ventricular arrhythmia (613, 1104). RyR2^S/+ and RyR2^S/S mice showed normal ECG parameters whether before or following isoproterenol challenge apart from decreased cycle lengths. However, both regular and programmed stimulation resulted in a greater atrial arrhythmogenic incidence in isolated perfused RyR2^S/S, though not RyR2^S/+, compared with WT. Isoproterenol challenge increased arrhythmic incidences in all three groups. This was despite unchanged APD₉₀ and AERPs on MAP recording. Regularly stimulated, fluo 3-loaded RyR2^S/S, but not RyR2^S/+ or WT atrial myocytes, showed diastolic Ca²⁺ release phenomena. These were made more frequent by isoproterenol (1334). Floating microelectrode recordings in isolated perfused RyR2^S/S but not WT atria correspondingly demonstrated DADs and ectopic APs potentially providing arrhythmic trigger (569).

2. Altered atrial AP conduction in RyR2-P2328S hearts

In parallel with findings reporting on their ventricular arrhythmic properties (1340), murine RyR2^S/S atria also demonstrated potential arrhythmic substrate potentially important for perpetuation of atrial arrhythmias. MEA recordings showed reduced epicardial AP conduction velocities despite normal MAP recovery measurements of both APD₉₀ and AERP. Floating intracellular electrode recordings demonstrated increased interatrial conduction delays and correspondingly reduced (dV/dt)_max (Figure 16, F and G). In contrast, AP recovery parameters of atrial effective refractory periods (AERPs), APDs, and AP amplitudes were indistinguishable from WT. Furthermore, despite persistent CV differences with extrasystolic S2 stimulation at progressively decreasing S1S2 intervals, the onset of arrhythmia coincided with similar CVs in WT and RyR2^S/S (569).

These findings accompanied loose-patch clamp findings of compromised Nav1.5 activity in both RyR2^S/S atria and positive Scn5a+/− controls. These took the form of reduced peak I_Na despite similar normalised current-voltage relationships (Figure 16, H and I) and reduced Nav1.5 but normal Cx40 or Cx43 protein expression. WT hearts showed similar reductions in I_Na under conditions of acutely increased [Ca²⁺]_i resulting from increased extracellular [Ca²⁺], caffeine, or CPA. These findings together suggest a slowed conduction resulting from downregulated Nav1.5 expression that could be further exacerbated by the effects of elevated [Ca²⁺]_i upon Na⁺ channel function in RyR2-P2328S atria (568).

The RyR2-P2328S system thus provided an experimental platform for an examination of recent clinical reports suggesting that flecainide may provide an approach to antiarrhythmic therapies through actions mediated through the RyR2 (1233). These reported that flecainide exerted marked antiarrhythmic effects in CPVT. This was attributed to tetracaine-like actions resulting in its acting as a potent RyR2 inhibitor (429, 455, 458, 473, 1233, 1252). As indicated above, flecainide acts both as a class 1C I_Na (66, 675, 1064) and RyR2 inhibitor (IC₅₀ values of respectively 5–11 and 2–7 μM) (338, 417).

MEA recording confirmed decreased CVs in isolated intact RyR2^S/S atria. This directly paralleled findings of a decreased I_Na relative to WT and the higher incidences of atrial arrhythmia observed in these mice. Low flecainide (1 μM) concentrations both paradoxically restored CV and reduced the frequency of arrhythmia. This sharply contrasted with the effect of flecainide in decreasing CV and promoting arrhythmia in WT. Thus low flecainide concentrations correspondingly restored I_Na to normal levels in intact RyR2^S/S atria from a likely inhibited baseline level. These effects were replicated by the RyR2 blocker dantrolene (228, 996). These findings suggested that by inhibiting the increased SR Ca²⁺ leak, flecainide reduces Ca²⁺-dependent inhibition of I_Na. This could restore conduction velocity in RyR2^S/S hearts. Such findings target diastolic SR Ca²⁺ leak in potential therapeutic antiarrhythmic strategies for heart failure (1318).

3. Ca²⁺ signaling and remodeling associated with chronic AF

Established as opposed to acute episodic clinical AF is associated with additional, anatomical or functional, atrial remodeling and congestive cardiac failure. There is often an accompanying intracellular Ca²⁺ overload and increased spontaneous diastolic SR Ca²⁺ release (309, 451, 1088, 1269). However, precise cause-and-effect relationships between these changes remain uncertain. It is also difficult to isolate contributions of disease entities associated with AF to the changes in Ca²⁺ signaling in AF. Thus short-term, sustained high atrial activation rates in a rabbit model appeared to produce an antiarrhythmic “Ca²⁺ signaling silencing.” This was accompanied by normal CaMKII expression levels but reduced CaMKII-mediated RyR2 phosphorylation and [Na⁺]_i (372).

Atrial myocytes from chronic AF patients show potentially proarrhythmic, increased frequencies in spontaneous Ca²⁺ release events (61, 451, 1194). These took place despite reduced I_CaL (202, 1274). These features might reflect increased single RyR2-Ca²⁺ channel open probabilities as observed in canine hearts with persistent AF. There was then an increased protein kinase A-dependent RyR2-S2808 phosphorylation and decreased FKBP12.6-RyR2 binding. Such changes would be expected to compromise RyR2 channel closure (1194, 1238, 1244). However, the direct relevance of these mechanisms to AF remains uncertain (290).

Murine hearts provide experimental paradigms appropriate to the analysis of remodeling changes accompanying and potentially exacerbating atrial arrhythmia. The transgenic mouse model overexpressing the cardiac transcriptional repressor cAMP-response element M, CREM-IbΔC-X (CREM), recapitulates this clinical progression (572). Telemetric ECG recordings first showed spontaneous atrial ectopy at ∼3 mo. This then progressed to paroxysmal and chronic AF. These changes followed CaMKII-dependent increases in diastolic Ca²⁺ release (184). The latter could be attributed to increased RyR2 open probabilities and Ca²⁺ spark frequencies. There were accompanying atrial dilatation and atrial conduction abnormalities (654). The latter conduction changes were related to Cx40 downregulation (572). RyR2-S2814 phosphorylation accompanied, and the RyR2-S2814A mutation prevented, this remodeling. There was no accompanying fibrotic change. Increased RyR2-S2814 phosphorylation correspondingly increased AF inducility by programmed electrical stimulation in mutant mouse models (655). These findings together paralleled clinical observations of increased CaMKII-activity and RyR2-S2814 phosphorylation in long-term AF patients.

Thus abnormal RyR2-mediated SR Ca²⁺ release, related to RyR2-S2814 phosphorylation in CREM-mice, contributes to remodeling in AF. It results in arrhythmic substrate, in addition to triggering events. These changes directly correlated with hyperactive RyR2 channels directly stimulating the Ca²⁺-dependent hypertrophic nuclear factor of activated T cell (NFAT)/Rcan1-4 pathway, detected through Rcan1-4 mRNA expression, implicating it in developing substrate for longlasting AF (654).

Chronic AF is also clinically associated with reduced sarcolipin (SLN) mRNA and protein expression (1041). The transmembrane protein SLN regulates SERCA2a activity in common with PLN (54, 777). It likely does so through CaMKII rather than PKA-mediated mechanisms (120). SLN inhibits SERCA at low [Ca²⁺] (1151). Aged homozygous Sln−/− mice (54) developed atrial fibrosis and spontaneous AF under anesthesia (1283), paradoxically suggesting that increased SERCA function promotes AF.

4. RyR2 function and SAN pacemaker activity

Continuing discussions bear on possible important roles for RyR2-mediated Ca²⁺ release in SAN pacemaker activity (see sect. IIIA) (615, 723, 1121). Enhanced SR Ca²⁺ release should activate inward I_NCX (522, 616). This would potentially facilitate AP firing in a “Ca²⁺ clock” mechanism (132). Ryanodine-mediated RyR2 block thus reduced SAN activation frequencies (132, 522, 972). Roles of SR-Ca²⁺ release in regulating normal SAN pacemaker function thus merit closer study (615, 723).

However, human CPVT is associated with reduced SAN automaticity resulting in basal bradycardia, sinus pauses, and impaired chronotropic β-adrenergic responses (625, 909). In vivo telemetry revealed that RyR2-R4496C mice did not show basal bradycardia. Nevertheless, they similarly showed sinus pauses followed by atrial and junctional escapes and impaired SAN automaticity with isoproterenol challenge. This did reproduce findings in CPVT patients following exercise (824). RyR2-R4496C SAN cells showed markedly reduced I_CaL on whole cell patch-clamping. Spontaneous [Ca²⁺]_i transients studied by confocal microscopy in isolated SAN cells correspondingly showed slowed pacemaker activity, impaired chronotropic β-adrenergic responses, and sinus pauses. Isoproterenol produced 5- to 10-fold increases in the frequencies of Ca²⁺ sparks and Ca²⁺ waves (824).

E. Genetic Abnormalities in Cardiac Calsequestrin 2: CPVT

1. Action of CASQ2 on Ca²⁺ homeostasis

The calsequestrin 2 (CASQ2) protein buffers SR Ca²⁺. This action would result in a reduction of free SR Ca²⁺ concentrations potentially below those expected to inhibit SERCA2a. This would then enhance SR Ca²⁺ reuptake. CASQ2 thus acts both as the major SR Ca²⁺ reservoir and a local intra-SR Ca²⁺ buffer. CASQ2 also functions as a Ca²⁺-dependent RyR2 channel regulator in cardiac myocytes (108). Thus CASQ2 is anchored to RyR2 via junctin and triadin. It may modulate RyR2 sensitivity to luminal Ca²⁺ through functional interactions with these molecules. It has been suggested that CASQ2 binding to triadin and junctin inhibits RyR2 channel activity at low luminal [Ca²⁺]. Increasing luminal [Ca²⁺] would then relieve this inhibition. This binding would be accompanied by a CASQ2 transition from a monomeric to a polymeric form with the greater Ca²⁺ binding capacity. The resulting occupation of CASQ2 Ca²⁺ binding sites would then weaken the interactions between CASQ2 and triadin and/or junctin. This would increase RyR2 sensitivity to luminal Ca²⁺. The result would be an increased open probability in the RyR2-Ca²⁺ channel (581).

2. CASQ2 mutations and the CPVT phenotype

Around 3% of CPVT patients show one of seven homozygous CASQ2 mutations (836). Four such mutations constitute premature stop codons. This generates truncated forms of the protein, lowering CASQ2 levels. This would be expected to shorten the time required for functional SR Ca²⁺ store recharging thereby increasing the likelihood of diastolic RyR2-mediated Ca²⁺ leaks and premature RyR2 activation. This would produce phenotypic similarities to CPVT (1120).

The remaining three mutations are single nucleotide replacements involving single amino acid substitutions. CASQ2-D307H was first described in seven consanguinous Bedouin families (611). This results in a substitution of the negatively charged aspartic acid for histidine in a highly conserved putative Ca²⁺ binding region between the second and third thioredoxin-like domains. This likely affects the CASQ Ca²⁺ affinity in its monomeric form and impairs its Ca²⁺ binding and its interactions with junctin and triadin (448, 565). The CASQ2-R33Q mutation was first demonstrated in a CPVT patient with history of syncope and death of a male sibling at a young age (908). When overexpressed in rat cardiac myocytes it resulted in abnormal spontaneous diastolic Ca²⁺ release. However, it did not appear to alter SR Ca²⁺ store capacity. CASQ2-R33Q is associated with impaired CASQ2 dimerization (565) or polymerization which consequently only takes place at a higher [Ca²⁺] (1167). This reduces the CASQ2-mediated RyR2 inhibition at low luminal Ca²⁺ (924, 925). Finally, CASQ2-L167H was identified in a young patient with stress-induced ventricular arrhythmia and cardiac arrest. In rat cardiomyocyte expression systems, CASQ2-L167H appeared to behave like a functionally inert protein (565, 924, 925).

These features are consistent with similar mechanisms for CPVT arrhythmias whether these are associated with RYR2 or CASQ2 mutations. They suggest similar parallels for the malignant hyperthermia associated with RYR1 or CASQ1 mutations in skeletal as opposed to cardiac muscle. One could thus postulate that arrhythmia would be initiated when levels of free SR luminal [Ca²⁺] exceeded a threshold for a store-overflow-induced Ca²⁺ release (SOICR). In such a hypothesis, RyR2 mutations would reduce the threshold for such SOICR. SOICR would then take place with the increases in the SERCA-mediated Ca²⁺ uptake that results from β-adrenergic stimulation. In contrast, the reduced CASQ2 Ca²⁺ binding and buffering resulting from CASQ2 mutations would increase the free luminal Ca²⁺ levels beyond those of a normal RyR2 SOICR threshold (711). However, studies in murine systems suggest greater complexities in relating the genetic changes to their phenotypic consequences.

3. Murine hearts with modified Casq2

Features of mice with genetic Casq2 modifications, exemplified in Table 7, were consistent with reductions in Casq2 being involved in their observed catecholaminergic stress-induced polymorphic VT and bidirectional VT phenotypes. Casq2−/− mice also showed increased tendencies to AF. This was associated with increased diastolic spontaneous Ca²⁺ elevations and DADs. All these features were inhibited by the RyR2 inhibitor R-propafenone (303). However, whereas even heterozygous (Casq2+/−) mice showed positive phenotypes (199), heterozygous carriers of premature CASQ2 truncations were asymptomatic. Furthermore, the precise mechanisms by which the Casq2 deficiency causes the causative diastolic SR Ca²⁺ leak leading to premature spontaneous Ca²⁺ release and arrhythmogenic triggered beats remains uncertain. Casq2−/− mice appeared to show normal SR Ca²⁺ content (581). However, they showed compensatory increases in SR volume and posttranscriptional reductions of junctin and triadin. Casq2 thus appears not to be essential for maintaining SR Ca²⁺ stores (581). Yet the Casq2 deficiency appeared central to the stress-induced polymorphic and bidirectional VT phenotype. In bilayer experiments, removal of CASQ2 at fixed intraluminal [Ca²⁺] increased RyR2 open probabilities in bilayer experiments (393). In heterozygous Casq+/− mice, even 25% decreases in Casq2 protein increased diastolic SR Ca²⁺ leak despite unchanged triadin and junctin levels and SR volume (199).

Two transgenic models with positive CPVT phenotypes showed evidence for increased expression of both RyR2 and the fetal SR Ca²⁺ binding protein calreticulin. The latter might then produce a compensatory increase in SR Ca²⁺ binding capacity. In common with Casq2−/− and homozygous Casq2-D307H/D307H mice, the homozygous deletion carrier Casq2-ΔE9/ΔE9 showed a 95% reduction in Casq2 protein. However, this accompanied other changes including increased RyR2 and calreticulin though normal calstabin levels (1077). Casq2-D307H/D307H mice were used to explore verapamil as a possible therapeutic antiarrhythmic agent (540). Homozygous, Casq2-R33Q/R33Q, mice also showed stress-induced polymorphic VT and bidirectional VT. Their myocytes showed diastolic depolarization phenomena and decreased SR Ca²⁺ (951). However, despite normal Casq2 mRNA, they showed a 50% reduction in Casq2 protein possibly reflecting a greater degree of Casq2-R33Q protein digestion by trypsin (951). However, calreticulin and RyR2 protein levels then were normal.

F. CaMKII Dysfunction, Altered Ca²⁺ Homeostasis, and Cardiac Arrhythmogenesis

Murine systems have thrown light on CaMKII function and consequences of its modification under conditions of both genetic and pharmacological manipulation. In the latter connection, the CaMKII inhibitor KN-93 has been useful as a blocker of CaMKII activation. However, KN-93 may also act upon voltage-gated K⁺ and L-type Ca²⁺ channels. Inhibition through endogenous CaMKII, autocamtide-3 inhibitor (AC3-I), and autocamtide-2 inhibitor proteins (AIP) may offer greater specificity and therefore future promise (882).

Transgenic mice overexpressing either CaMKIIδ2 or CaMKIIδ3 developed DCM and cardiac failure (1331, 1332). CaMKII overexpression also altered Ca²⁺ homeostasis and cellular excitability. This resulted in proarrhythmic AP afterdepolarizations (291). In contrast, CaMKII deletion prevented the heart failure that follows transaortic constriction (55, 667). Similarly, overexpression of the CaMKII inhibitory peptide AC3-I enhanced postinfarction cardiac function and Ca²⁺ handling. It reduced LV hypertrophy and dysfunction following isoproterenol challenge. It also decreased PLN phosphorylation at Thr17 and SR Ca²⁺ load (1329). Murine AC3-I overexpression systems also demonstrated feedback relationships between CaMKII, PLN, and sarcolemmal K⁺ channels. They showed upregulated inward rectifier I_K1 and transient outward currents I_to,f and consequent APD shortening. This was in contrast to the decreased I_to and I_K1 and APD prolongation in cardiac failure (295). These changes accompanied normal channel RNA and protein expression consistent with either altered membrane expression or posttranslational modification. This remodeling was prevented with crossing of Ac3-I and Pln−/− (650). Finally, both KN-93- or AC3-I-mediated CaMKII suppression in a murine calcineurin overexpression cardiac failure model reduced spontaneous ventricular arrhythmias and improved LV function (557).

CaMKII dysfunction has also been associated with AF and SND (654, 1277, 1297), CPVT (677), LQTS3 (588), and the ankyrin-B related (LQTS4) (249) and Timothy (LQTS8) syndromes (1125). A recent study in transgenic mice with cardiomyocyte-specific expression of the SR-targeted CaMKII inhibitor aryl-hydrocarbon receptor-interacting protein (AIP) implicated RyR2 phosphorylation produced by ROS-activated CaMKII, in arrhythmogenic changes following impairment of glucose tolerance (857, 1076). Studies in permeabilized cardiomyocytes suggested that whereas the small proportion of CaM variants associated with a CPVT could evoke arrhythmogenic perturbations in Ca²⁺ homeostasis, this did not appear to be the case with variants associated with LQTS (474, 1306). Finally, mice haploinsufficient in Purkinje cell protein-4 (Pcp4) which may regulate CaM and CaMKII signaling specifically within His-Purkinje cells showed proarrhythmic ventricular phenotypes, increased I_Ca, positive shifts in I_Ca inactivation, and altered evoked Ca²⁺ transients accompanying CaMKII activation (566).

G. Genetic Modifications in Dephosphorylation Pathways

1. Cardiac expression of p21-activated kinase-1

Increasing evidence implicates phosphatases in dampening kinase-mediated cardiac excitation and contraction mechanisms. This potentially exerts cardioprotective effects (548). The serine/threonine protein kinase p21-activated kinase-1 (Pak1) is highly expressed in different cardiac regions. It is localized to the Z-disc, cell and nuclear membrane, and intercalated discs in rat ventricular myocytes (547). It is directly activated by the small GTPases, cell division control protein 42 homolog (Cdc42), and Ras-related C3 botulinum toxin substrate 1 (Rac1). Coimmunoprecipitation studies suggest that Pak1 is associated with the similarly localized cellular cardiac PP2A for which it might form an upstream activator in both SAN and ventricular cells (546, 1254). PP2A removes phosphate groups at PKA phosphorylation sites on the cardiac isoform of the cTnI troponin subunit that inhibits actin-myosin ATPase activity (547, 709), MYBPC3 (547), PLN (709), inwardly rectifying K⁺ channels (543), as well as L-type Ca²⁺ channels (242) and RyR2s (730), increasing Ca²⁺ spark generation (1119). It also regulates transcription factors including cAMP response element-binding protein (CREB) (1208) and nuclear factor kappa-light-chain-enhancer of activated B cells (NFκB) (1291). The latter effects influence cardiac remodeling (312, 327).

Both Pak1 and its upstream activators Cdc42 and Rac1 also stimulate cytoskeletal changes. They reduce stress fibers and focal adhesion complexes, filapodia and lamelliopodia formation, and membrane ruffling (544, 545, 1102, 1229). These features are compatible with a dynamic balance between phosphatase and kinase activity in regulating aspects of Ca²⁺ homeostasis ion channel activity influencing pacemaker and excitable activity in sinoatrial and ventricular cells, respectively, and their modulation by autonomic stimulation.

2. Loss of Pak-1 function and atrial arrhythmic phenotype

Mice with either cardiac-specific conditional (Pak1-cko) or total (Pak1−/−) Pak1 deletions showed increased sinus heart rates compared with control homozygous flox (Pak1-f/f) littermates. This was despite similar electrocardiographic, P, PR, QRS, and QT waveforms (1230). Conversely, infection of cultured guinea pig SAN cells with recombinant adenovirus expressing constitutively active Pak1 (CA-Pak1) depressed L-type Ca²⁺ and delayed rectifier K⁺ currents. It reduced repetitive AP frequencies in response to isoproterenol-induced β-adrenergic stimulation compared with findings in SAN cells infected with Ad-LacZ controls (545). The PP2A inhibitor okadaic acid partly reversed this suppression of L-type Ca²⁺ channel responses to isoproterenol in Ad-Pak1-infected cells (543). These findings led to suggestions for a dynamic balance between kinase and phosphatase activities in modulating both L-type Ca²⁺ channel and delayed rectifier I_K activity. This could furnish a potentially important mechanism for alterations in pacemaker activity following autonomic stimulation (545).

3. Loss of Pak-1 function, SERCA function, and ventricular arrhythmic phenotype

Pak1-cko and Pak1−/− hearts also showed increased incidences of polymorphic VT, despite normal refractory periods, following either acute or chronic β-adrenergic isoproterenol challenge, relative to controls (Figure 17A) (259, 1044, 1230). Pak1-cko and control, Pak1-f/f, ventricular myocytes showed comparable baseline I_CaL similarly reduced by chronic β-adrenergic stress. However, Pak1-cko myocytes showed higher incidences of irregular, alternating Ca²⁺ transients with repetitive stimulation, and of spontaneous Ca²⁺ waves, both before and particularly following imposition of chronic β-adrenergic stress (Figure 17B). This appeared at least partly to reflect decreased SERCA function. The latter would reduce SR Ca²⁺ content and increase diastolic [Ca²⁺], effects induced or accentuated by either chronic or acute β-adrenergic stress (Figure 17, C and D). The latter target offers potential future therapeutic targets for novel agents directed at SERCA activity through PLN such as istaroxime and CDN1163 (463, 533). Such changes were suggested by assessments of SR Ca²⁺ levels through observed time constants and integrals of I_NCX, as well as of SERCA2 activity from decay constants of systolic Ca²⁺ transients, following caffeine challenge. It also emerged from comparisons of the decays of systolic Ca²⁺ transients, and increases in diastolic and decreases in systolic [Ca²⁺], and reductions in systolic contraction amplitude, despite unchanged I_CaL, following regularly applied stimulation (1230).

Figure 17.

Physiological features mediating arrhythmogenesis in Pak1-deficient hearts. A: monophasic action potential (AP) recordings showing AP alternans (a), torsades de pointes (TdP) (b), and polymorphic VT (c) following burst pacing (horizontal line below trace) in ex vivo Pak1-cko hearts, features not observed in Pak1-f/f hearts. B: Ca²⁺ transients in field-stimulated Pak1-f/f (a) and Pak1-cko myocytes (b) at a 1-Hz stimulation frequency under baseline (left traces; i) and chronic β-adrenergic stress conditions (right traces; ii). Increased pacing frequencies increase the occurrence of Ca²⁺ waves to greater extents in Pak1-cko than Pak1-f/f myocytes particularly with chronic β-adrenergic stress. C and D: recovery of SR Ca²⁺ stores from after previous depletion by caffeine challenge, after which regular stimulation resumed. i: Ca²⁺ transients indicating recovery of SR Ca²⁺ in Pak1-f/f (C) and Pak1-cko myocytes (D) under baseline conditions. a–c: Comparison of increasing Ca²⁺ transients (ii), constant I_CaL (iii), and increasing I_NCX (iii) at different stages (a–c) of SR Ca²⁺ recovery. [From Wang et al. (1230).]

More direct assessments of effects on SERCA2 activity first interposed a caffeine challenge to deplete SR Ca²⁺ in regularly stimulated ventricular myocytes. They then followed the subsequent recovery of evoked Ca²⁺ signals to assess recovery of SR Ca²⁺ stores. Pak1-cko myocytes showed more prolonged recovery time courses compared with Pak1-f/f whether before or following chronic β-adrenergic stress (Figure 17, C and D, i). Such chronic β-adrenergic stress increased both these differences and the recovery times. In contrast, I_CaL and I_NCX magnitudes and time courses were indistinguishable through these conditions and the duration of recovery from caffeine challenge (Figure 17, C and D, ii–iv) (1230).

Pak1-cko hearts correspondingly showed reduced SERCA2a mRNA and protein expression. However, they showed normal PLN mRNA, protein and phosphorylation levels, Nav1.5 protein levels, and RyR2 and Cacna1c mRNA levels whether before or following imposition of chronic β-adrenergic stress. Pak1-cko contrasted with Pak1-f/f hearts in failing to upregulate NCX protein or NCX1 mRNA following hypertrophic stress. Control Ad-shC2 virus-infected neonatal rat cardiomyocytes correspondingly showed increased SERCA2a protein and mRNA expression with phenylephrine challenge. This was abolished by Pak1-knockdown in Ad-shPak1-infected neonatal rat cardiomyocytes (NRCMs). It was accentuated by Pak1 overexpression produced by Ad-CAPak1 (1230).

In parallel with these results, myocytes expressing constitutively active Pak1 (CA-Pak1) showed faster time to peak shortenings but slower [Ca²⁺]_i decays and relengthening times compared with LacZ-expressing controls. CA-Pak1 expression also inhibited the effect of isoproterenol in increasing [Ca²⁺]_i transient amplitude. It increased SR Ca²⁺ content, slowed [Ca²⁺]_i decay rates, and increased cell shortening. It also reduced Ca²⁺ spark amplitude and slowed propagation of spontaneous Ca²⁺ waves. Although it did not affect PLN phosphorylation, it reduced cardiac troponin I phosphorylation (1044).

Pak1 additionally downregulates NADPH-oxidase 2 activity in electron transfer from NADPH to O₂ to generate superoxide. Simulated ischemia thus increased [Ca²⁺]_i and induced spontaneous Ca²⁺ release and Ca²⁺ waves in Pak1^−/− ventricular myocytes. These effects were rescued not only by the reverse-mode NCX blocker KB-R7943 and the Ras-related C3 botulinum toxin substrate 1 (RAC1) inhibitor (NSC23766) but also the ROS scavenger TEMPOL. Furthermore, both Pak1−/− and WT myocytes treated with the Pak inhibitor IPA3 showed increased cellular ROS production not exhibited by either p47-phox−/−, NADPH oxidase 2 (NOX2) deficient, myocytes or WT myocytes exposed to the NADPH oxidase peptide inhibitor gp91ds-tat (258).

4. Loss of Pak-1 function and hypertrophic phenotype

Pak1-cko hearts showed an accentuated cardiac hypertrophy and heart failure following either chronic pressure overload or angiotensin II infusion. There was a reduced activation of c-Jun N-terminal kinase (JNK). The latter is known to phosphorylate and inactivate the transcription factor NFAT, important in activating hypertrophic genes (793). The Pak1 activator FTY720 prevented these effects in WT but not Pak1-cko hearts (681). FTY720 also exerted cardioprotective effects against ischemia/reperfusion injury in both rat and mouse models (545).

Pak1 activating peptide (PAP) is known to increase phosphorylated Pak1 levels in cultured neonatal rat ventricular myocytes. It both reduced such angiotensin II-induced hypertrophy in neonatal rat ventricular myocytes (NRVMs) and C57BL/6 ventricles and inhibited the arrhythmias that followed such ventricular hypertrophy (1225). Intriguing studies in Pak1−/− mice attributed such effects of Pak1 deficiency to decreased amplitudes and slowed kinetics in Ca²⁺ transients coinciding with decreased transverse tubular densities and capacitances. All these changes were rescued by overexpressing constitutively active Pak1. The resulting alterations in cytosolic Ca²⁺ homeostasis might then modify regulation of transcription factors involved in hypertrophic remodelling (258).

H. Causal Schemes Relating Altered Ca²⁺ Homeostasis to Cardiac Arrhythmia

The above exemplars can be organized into a hypothetical scheme (Figure 18) that could form a basis for further explorations of causal relationships between modifications in cellular Ca²⁺ homeostasis and cardiac arrhythmias. Such modifications might arise from acquired or genetic alterations that might increase RyR2-mediated release of SR Ca²⁺ or decrease SERCA2a-mediated Ca²⁺ reuptake from cytosol to SR Ca²⁺ store. These could take place both through direct modifications of the molecules concerned, or indirectly through other signaling mechanisms (Figure 18, A and B). This in turn could potentially alter NCX activity whose electrogenic effects might lead to diastolic triggering phenomena (Figure 18C). Results from some of the exemplars above also suggest that it could reduce synthesis or membrane trafficking and therefore membrane expression of Nav1.5 or acutely alter its biophysical properties (Figure 18D). These effects could either chronically or acutely slow AP conduction, resulting in arrhythmic substrate (Figure 18E) even under conditions of normal action potential recovery as reflected in normal APD/ERP ratios.

Figure 18.

Scheme exploring relationship between perturbations in Ca²⁺ homeostasis, triggering, arrhythmic substrate, and generation of arrhythmia. A: acquired or genetic perturbations resulting in increased release of sarcoplasmic reticular (SR) Ca²⁺ or decreased Ca²⁺ reuptake from cytosol to store both perturb cytosolic Ca²⁺ (B). This in turn alters (C) Na⁺-Ca²⁺ exchange (NCX) electrogenic activity leading to diastolic triggering phenomena. It can also (D) reduce Nav1.5 synthesis or membrane trafficking and therefore its membrane expression, or directly alter Nav1.5 biophysical properties. Both effects potentially slow conduction, resulting in arrhythmic substrate even under conditions of normal action potential recovery as reflected in action potential duration/effective refractory period (APD/ERP) ratios. Combination of C and D culminates in E, potentially fatal ventricular arrhythmia. Possible direct actions of intermediates arising from metabolic change on I_Na are not shown for simplicity.

VIII. CARDIAC ENERGETIC DYSFUNCTION AND ARRHYTHMIA

A. Energetic Consequences of Cardiac Disease

1. Metabolic stress and mitochondrial dysfunction

Murine exemplars containing specific membrane ion channel modifications directly bearing upon cell excitability and Ca²⁺ homeostasis thus offer relatively straightforward physiological insights into the relationship between biophysical channel modification and whole-heart arrhythmogenesis. In so doing, they usefully model a range of relatively rare arrhythmic conditions. However, the clinically more common causes of arrhythmia are often less clearly genetically characterized. Significant insights into their arrhythmia can nevertheless be obtained through a strategic physiological targeting of particular aspects of cell function. For example, both the oxidative stress associated with hypertrophic change, cardiac failure, ischemia-reperfusion (90, 201, 374, 1117), and metabolic conditions including obesity, insulin resistance and type 2 diabetes, and increasing age are accompanied by energetic, in particular, mitochondrial, dysfunction (600, 1005, 1196). Mitochondria constitute the main cardiac cellular source of ATP (845). Dysfunctional mitochondria have been implicated in ventricular arrhythmia with altered ion channel function, AP heterogeneity, and cell excitability (49, 146, 483). They are also associated with the oxidative stress, possibly associated with L-type Ca²⁺ channel activity in atrial tachyarrhythmias (155).

2. Mitochondrial dysfunction and biophysical properties

A number of mechanisms link such alterations in metabolism to altered membrane excitability. These often take place under conditions also associated with increased frequencies of Ca²⁺-mediated triggers (41, 42). Mitochondrial inner membranes sustain a ∼200 mV potential that drives electrons down the electron transport chain. The latter ultimately drives ATP synthesis. This process is tightly regulated to match energy supply to demand. Mitochondrial oxidation destabilizes this potential through a mitochondrial permeability transition. This is likely produced by activation of inner membrane anion channels (IMAC) among possible others (845, 881). This destabilization initiates ROS-induced ROS release producing alternans in both mitochondrial and surface electrical activity (41). Damaged or dysfunctional mitochondria can thus increase ROS production by up to 10-fold (374).

ROS exert numerous, potentially arrhythmogenic, actions. They decrease I_Na (671) and I_K (1217), activate sarcolemmal K_ATP channels (1249), modify Na⁺ and L-type Ca²⁺ channel inactivation kinetics, increase I_NaL, and oxidize RyR2. These result in an increased leak of SR Ca²⁺ leak and a consequent modulation of intracellular Ca²⁺ cycling (135, 146, 1117). They also reduce Cx43 trafficking and function (18, 1292, 1293). These actions together also increase [K⁺]_i, [Na⁺]_i, and [Ca²⁺]_i. This produces a wide range of potentially proarrhythmic physiological effects bearing upon cell-cell coupling (1074), AP conduction (671), repolarization (1217) and alternans (91), and Ca²⁺-mediated triggers (1117).

ATP depletion or increased ADP increases open probabilities in sarcolemmal ATP-sensitive K⁺ channels (sarcK_ATP) (18). These occur at relatively high densities in myocyte surface membranes. This opening may physiologically protect myocyte viability during ischemia (380, 532). However, even opening 1% of such channels shortens APD, and therefore ERP and AP wavelength. This can potentially produce proarrhythmic reentrant substrate (305, 323). In addition, opening of sufficient channels generates heterogeneous current sinks driving the cell membrane potential towards E_K. This can compromise cell excitability and AP propagation (18).

B. The Nuclear Receptor Family of Peroxisome Proliferator-Activated Receptors

1. Role of PGC-1 species

The nuclear receptor family of peroxisome proliferator-activated receptors (PPARs) include the PPAR-α, PPAR-γ, and PPAR-δ isoforms. Modifications in these offer one possible genetic access to studying actions of mitochondrial-related factors on cardiac function (944). All are subject to transcriptional coactivation by peroxisome proliferator-activated receptor gamma coactivator 1 (PGC-1) species. The latter include PGC-1α, PGC-1β, and PGC-1-related coactivator (922). All have been increasingly implicated in regulation of energy metabolism in both health and disease (316, 661).

Both PGC-1α and PGC-1β regulate expression of genes involved in oxidative phosphorylation (555, 825, 1014, 1015). High PGC-1α and PGC-1β expression levels thus occur in actively oxidative tissues often characterized by abundant mitochondria. These include brown adipose tissue and cardiac and skeletal muscle. Nevertheless, mouse models with deletions of either PGC-1α or PGC-1β showed normal baseline cardiac contractile function. However, mutations involving both are postnatally lethal through cardiac failure. This is consistent with redundancies or overlaps between the functions of PGC-1α and PGC-1β (612). Both PGC-1α and PGC-1β deletions accentuated cardiac failure following transverse aortic constriction (44).

2. Effects of genetic modification of PGC-1α

PGC-1α appears important to cardiac adjustments that require increased ATP and work output in response to physiological stimuli. Embryonic mouse hearts showed large increases in mitochondrial biogenesis and oxidative metabolism close to birth. This followed dramatically increased PGC-1α expression (628). Ectopic PGC-1α expression drives mitochondrial biogenesis in cultured cardiac myocytes. Oxidative phosphorylation in these mitochondria was tightly coupled to ATP production (628). One of the two available PGC-1α knockout models showed normal baseline cardiac contractility, and the other showed age-dependent contractile dysfunction (43, 628). Both showed normal mitochondrial volume density. However, both also showed impaired expression of genes involved in oxidative phosphorylation and fatty acid oxidation. These were associated with an impaired maximal mitochondrial oxygen consumption using palmitoyl-l-carnitine substrate and mitochondrial ATP synthesis accompanied by mitochondrial uncoupling (43). In vivo and Langendorff-perfused PGC-1α-KO hearts showed impaired inotropic and chronotropic exercise and dobutamine responses, respectively.

3. Effects of genetic modification of PGC-1β

The four independently generated PGC-1β-KO models (638, 1080, 1196) showed normal inotropic responses in contrast to, but reduced chronotropic responses in common with, PGC-1α hearts following dobutamine challenge (638). PGC-1β−/− models showed impaired expression of genes involved in oxidative phosphorylation but not fatty acid oxidation (945). PGC-1β−/− mice showed irregular heartbeats and increased incidences of polymorphic VTs following isoproterenol challenge. They showed an increased Scn5a, Kcna5, RyR2, and CaMKII RNA expression. Langendorff-perfused PGC-1β−/− hearts showed AP alternans, EADs, and VT on monophasic AP recording (Figure 19, A and B). Myocyte patch electrode studies (Figure 19, C and E–H) demonstrated unstable oscillatory resting potentials and early and delayed afterdepolarizations accompanying their APs (Figure 19C). Sustained current injection elicited burst firing (Figure 19H). These changes were accompanied by diastolic Ca²⁺ transients exaerbated by isoprenaline (Figure 19D), increased inwardly and outward rectifying K⁺ currents (Figure 19, F and G), and a negatively shifted Ca²⁺ current inactivation (Figure 19E) (387).

Figure 19.

Arrhythmogenic features of PGC1β−/− hearts. A: increased heart rate following isoproterenol challenge accompanied by polymorphic ventricular tachycardia (VT) in PGC1β−/− mice during ECG recording. B: monophasic action potential (AP) recordings of VT following programmed electrical stimulation in Langendorff-perfused PGC1β−/− hearts. C: APs from PGC1β−/− ventricular myocytes showing early (EADs) and delayed afterdepolarizations (DADs) and ectopic APs. Inset magnifies voltage trace 40-fold. D: abnormal Ca²⁺ homeostasis with intermittent elevations in diastolic Ca²⁺ and Ca²⁺ waves increased in amplitude and frequency in isoproterenol challenged PGC1β−/− ventricular myocytes. E: voltage-gated I_Ca and transient and sustained outward I_K (F) in response to depolarizing pulses from −40 to +50 mV altered in successive 10-mV increments from a −40 mV holding potential. G: inwardly rectifying currents obtained in response to hyperpolarizing steps from −40 to −100 mV incremented in 10-mV intervals. E–G shown for PGC1β−/− (a) and WT ventricular myocytes (b), with voltage step protocols shown in c. H: step current injections produced single APs with prolonged plateaus with burst AP firing in PGC1β−/− (a) but not WT myocytes (b). [From Gurung et al. (387).]

C. Effects of Cellular Redox Potential

1. Importance of NAD⁺/NADH

Studies in aging skeletal muscle suggested that reduced nuclear NAD⁺ levels variously attributable to altered NAD⁺/NADH ratios (921) or altered NAD⁺ consumption or synthesis (205) may also compromise mitochondrial function. This may involve mechanisms independent of PGC-1α/β. Thus restoring the NAD⁺ levels restored the mitochondrial function (365). In addition, these pyridine nucleotides themselves have independent cellular actions. Some of these could potentially interact with those of altered mitochondrial function. Thus NADH enhanced smooth muscle superoxide production which could result in and be further reinforced by PKC activation (221). NAD⁺ caused PKA activation in osteoblastic cells (960). In turn, PKA upregulated (1349) and PKC downregulated Nav1.5 function (809). Pyridines also modified voltage-dependent K⁺ channel gating (1134), Ca²⁺-activated K⁺ channel (869), and TrypC currents (420). NAD⁺ enhanced and NADH inhibited RyR2-mediated Ca²⁺ release (1352). In addition, NAD⁺ is convertible, by RyR2-associated CD38 (4, 1100), to cyclic ADP-ribose, known to trigger RyR2-mediated Ca²⁺ release (5).

Finally, mutations in GPD1-L protein, highly expressed in cardiac tissue, are associated with reduced I_Na and clinically with both BrS and SIDS. GPD1-L is >80% homologous in amino acid sequence with glycerol-3-phosphate dehydrogenase. The latter enzyme family is directly involved in NAD-dependent energy metabolism potentially altering cellular redox state as reflected in NAD⁺/NADH levels. This would bridge alterations in cellular energetic state with modifications in I_Na (689). Coexpression of Scn5a with the implicated Gpd1-l-A280V, E83K, I124V, and R273C reduced I_Na in expression systems and neonatal myocytes (689, 840).

2. Effects of NAD⁺/NADH on I_Na

Altered NADH might modify Nav1.5 function and expression giving changes that might explain the effects of GPD1-L mutations on I_Na. Such studies first demonstrated that additional expression of GPD1-L-A280V but not of WT-GPD-1L markedly increased [NADH]_i in HEK cell lines stably expressing human Nav1.5 alone. Second, [NADH]_i produced marked dose-dependent (20–100 μM) rapid and stable, ≤50%, reductions in maximum I_Na in HEK cells expressing human Nav1.5. It also did so to a lesser extent in rat neonatal cardiomyoctyes. In contrast, the overall voltage dependences of both activation and inactivation of Nav1.5 and Nav1.5 mRNA expression were preserved. These effects were reversed by NAD⁺, the PKC antagonist chelerythrine, and superoxide dismutase. They were mimicked by the PKC activator phorbol 12-myristate 13-acetate (PMA). The rescue effects of NAD⁺ were antagonized by PKA inhibition by PKAI_6–22 (100 nM) (360). Conversely, the PKA activator (1 μM) forskolin blocked the NADH effect on I_Na while not by itself affecting I_Na. Thus the mutually antagonistic effects of internally or externally applied NAD⁺ and NADH likely involved separate signaling pathways. In contrast, surface Nav1.5 expression was conserved through these manipulations (672). Finally, WT murine hearts exposed to increased [NADH]_i through perfusion using external 10:1 lactate:pyruvate showed a reversible arrhythmogenesis on programmed electrical stimulation. Conversely, application of [NAD⁺]_o perfusion reversibly abolished the arrhythmogenesis shown by Scn5a+/− hearts (672).

Figure 20 simplifies some of the factors outlined in section VIII, A–C, and their possible relationships to biophysical events leading to arrhythmogenic effects.

Figure 20.

Energetic dysfunction and arrhythmic phenotype. Possible simplified relationships between (A) energetic dysfunction associated with ischemic conditions, cardiac failure, ageing and diabetes, (B) mitochondrial dysfunction associated with ROS production, altered NAD⁺/NADH, and ATP/ADP and their possible consequences for (C) RyR2-mediated release of SR Ca²⁺ leading to increased [Ca²⁺]_i, and NCX and DAD triggering activity, and (D) Na⁺ and K⁺ channel activity affecting AP excitation, propagation, and recovery potentially resulting in (E) substrate that can be potentially triggered to give arrhythmic phenotypes.

IX. ARRHYTHMIC SUBSTRATE RESULTING FROM STRUCTURAL CHANGE

A. Fibrotic Change and Arrhythmic Phenotype

1. Cytokine signaling and fibrotic change

In addition to ion channel and cellular homeostatic changes, structural changes, whether at the subcellular, cellular, tissue, or even organ levels, could alter the conditions of excitation, its conduction, and recovery from such excitation. They thus would also contribute to arrhythmic phenotype. Table 8 lists examples to be considered below (465). Previous sections had alluded to the association between Nav1.5 haploinsufficiency in the murine Scn5a+/− system and structural in addition to biophysical change (see sect. VE). This particularly involved a cardiac fibrotic phenotype induced by increased levels of TGF-β₁ (see sect. VF). This may be a manifestation of a more general regulation of fibrotic change with possible implications for arrhythmia (432, 959).

Table 8.

Murine genetic models of structural disorders associated with ventricular or atrial arrhythmia

Gene Product	Model	Features	Reference Nos.
Fibrotic change
TGF-β1	Cardiac-specific TGF-β1 overexpression (MHC-TGFcys33 ser)	Atrial fibrosis and AF inducibility	1193
TNF-α	Cardiac-specific TNF-α overexpression	Atrial dilation, fibrosis, thrombosis, spontaneous AF in ambulant mice	598, 974
Angiotensin I converting enzyme (ACE)	ACE 8/8	Atrial dilatation, fibrosis, and spontaneous AF in ambulant mice	1280
	Cardiac-specific overexpression of ACE
Adenosine (A1 or A3) receptors	Cardiac-specific A3AR overexpression of at high (A3high) or low (A3low) adenosine levels	Bradycardiomyopathy	300
Ras homolog gene family member A (RHOA)	Cardiac-specific (αMhc promoter) overexpression of WT or activated form of RHOA	Atrial enlargement, cellular hypertrophy, interstitial fibrosis, AF susceptibility, atrioventricular block, premature death	991
Mitogen-activated protein kinase kinase 4 (Mkk4),	Atrial cardiomyocyte-specific conditional knock-out Mkk4-acko-Nppa-Cre4	AF, atrial fibrosis, upregulated TGF-β1 signaling and dysregulation of matrix metalloproteinases	245
Hypertrophic cardiomyopathy (HCM)
α-Myosin heavy chain (Myh6)	Mhc-R403Q/+	QT dispersion, conduction heterogeneity, inducible VT, sudden death, variable degree of hypertrophy	112, 347, 1266
Myosin binding protein C (Mybp-C3)	MyBP-C-t/+ and MyBP-C-t/t; (t = truncated Mybp-C3)	Hypertrophy, fibrosis, ventricular ectopic activity	113
Troponin T (Tnnt2)	Severity: Tnnt2-I79N >TnnT-F110I>TnnT-R278C	Altered Ca2+ sensitivity, spontaneous VT, reduced IK1	83, 580, 583
Myosin light chain (coded by Mlc17)	Mlc-17-M149V	Hypertrophy	1185
Ras (HRAS)	Targeted, chronic Ras-Raf–MAPK pathway activation and αMHC-loxp-GFP-loxp-H-RasV12	Hypertrophy, ventricular arrhythmia, atrial fibrillation, conduction block	969, 1346
Dilated cardiomyopathy (DCM): sarcomeric protein mutations
α-Myosin heavy chain (Myh6)	Mhc-F764L/+, Mhc-S532P/+, Mhc-F764L/F764L and Mhc-S532P/S532P)	Dilatation	1018
Mammalian-enabled protein and vasodilator-stimulated phosphoprotein (Vasp)	Cardiac myocyte-restricted, α-myosin heavy chain promoter-directed expression of the dominant-negative VASP-EVH1 domain	Dilatation, hypertrophy, bradycardia, sudden death	293
Neuron-restrictive silencer factor (Nsrf)	α-MHC-dnNRSF (dominant-negative knock-in)	Dilatation, heart block, sudden death	608
Dilated cardiomyopathy (DCM): cytoskeletal protein mutations
Lamin (coded by LMNA)	LMNA+/−	AV nodal disease, atrial arrhythmia, VT	1267
	Lmna-N195K	Dilatation, fibrosis bradycardia, SA node exit block	807
	Lmna-H222P/H222P	Models Emery-Dreifuss syndrome; dilatation, increased PR interval, variable conduction block, QRS prolongation, ventricular extrasystoles	45
Dilated cardiomyopathy (DCM): SR Ca2+-cycling protein mutation
Phospholamban (Pln)	Pln-R14Del	Dilation, fibrosis, VT sudden death	397
	Pln-R9C/R9C	Reduced protein kinase A activity, delayed SERCA-Ca2+ reuptake, dilated cardiomyopathy, congestive heart failure	1019
Dilated cardiomyopathy (DCM): developmental gene mutations
Cardiac homeobox gene product (Nkx-2.5)	Ventricle-restricted deficiency: Nkx2-5-f/f/cre+	Hypertrabeculation, noncompaction, conduction disturbance	871
Vinculin (VCL)	Cardiac-restricted Vclfl/fl MLC-2vCre/+ knockout	Dilatation, VT, sudden death	1324
Troponin I (TNNI3)	cTnI-R192H	Atrial dilatation, decreased LV volume, decreased EF, delayed relaxation	278, 279
Arrhythmogenic right ventricular dysplasia (ARVD): desmosomal protein gene mutations
Plakophilin-2 (Pkp2)	Pkp2−/−	Lethal	376
	Pkp2+/−	Normal cardiac phenotype	376
Plakoglobin (Pg)	Pg−/−	Lethal	647
	Cardiac specific Pgf/f-αMHCcre	RV enlargement, fibrosis, slowed RV conduction, spontaneous VT of right ventricular origin. Phenotype exacerbated by training	571, 647, 973
Desmoplakin (Dsp)	Dsp−/−	Embryonic lethal	339, 475
	Dsp+/−	Slowed conduction and inducible VT prior to adult structural changes	366
	Cardiac overexpression of flag-tagged Dsp cDNA carrying Dsp-R2834H-Tg COOH-terminal mutation	Apoptosis, fibrosis, RV and LV dilatation, reduced ventricular function, widened intercalated disks	1296
	Cardiac specific Dsp-f/f- (MLC2v-Cre)	Cardiomyopathic changes, ventricular arrhythmias accentuated by exercise and catecholaminergic challenge, slowed RV conduction	706
Desmoglein (Dsg2)	Cardiac overexpression of flag-tagged dominant Dsg2-N271S mutant (homologous with human DSG2–N266S ARVC mutation)	Myocyte necrosis, calcification and fibrous tissue replacement, spontaneous ventricular arrhythmias, conduction slowing, ventricular dilatation and aneurysms, replacement fibrosis and SCD	896
Arrhythmogenic right ventricular dysplasia (ARVD): laminin receptor mutations
Laminin receptor 1 (Lamr1)	KK/Rvd (knock-down)	Right ventricular fibrosis, QRS prolongation	48
Left ventricular noncompaction cardiomyopathy (LVNC)
Cypher/Zasp	Mutations in Cypher/Zasp	LVNC with DCM and HCM phenotype, early postnatal death	1345, 1350

TGF-β₁ is a key regulator of fibrosis in adult hearts. Increased fibrosis, to which atria are particularly susceptible (949, 1305), reduces cardiomyocyte-cardiomyocyte coupling potentially resulting in conduction defects (21). Increased TGF-β₁ expression is clinically associated with chronic AF (222). In murine hearts, cardiac-specific TGF-β₁ overexpression was followed by atrial fibrosis, atrial conduction disturbances, and AF (1193). The latter was attributed to triggered activity following reductions in APD and an appearance of spontaneous Ca²⁺ release events (198). Angiotensin II and reactive oxygen species may also increase TGF-β₁ promoter activity, connective tissue growth factor (Ctgf), and collagen in fibroblasts causing a profibrotic environment (408, 1146). Similarly cardiac-specific TNF-α overexpression produced at rial dilation, fibrosis, thrombus development, and spontaneous AF in ambulant mice (598, 974). In some models, it resulted in Cx40 downregulation (1013).

2. G protein signaling and fibrotic change

Cardiac G protein-coupled receptors serve various signaling pathways involving epinephrine, angiotensin, endothelin, and adenosine as messengers. Thus Gα_q associates with α1-adrenergic, endothelin (ET_A), and angiotensin II type I (AT₁) receptors. They may thereby exert metabolic effects that impact on tissue fibrosis. Thus AF patients show increased atrial angiotensin II expression (519). Rapidly paced in vitro atrial myocytes show increased angiotensin II secretion (1145, 1280). Cardiac-specific overexpression of angiotensin I converting enzyme caused atrial dilatation, fibrosis, and spontaneous AF in ambulant mice (1280). Similarly, transgenic models overexpressing adenosine A1 or A3 receptors showed atrial bradycardia and atrioventricular block and increased susceptibility to AF dependent on the degree of bradycardia (300, 570). This could potentially lead to bradycardiomyopathy (300). Both available transgenic mouse lines overexpressing active Gα_q subunit under αMhc promoter control developed cardiac hypertrophy followed by diffuse atrial and ventricular fibrosis and heart failure (233, 764) and spontaneous AF in ambulant animals (431).

The membrane-associated GTPase, Ras homolog gene family member A (RHOA), is involved in actin cytoskeleton organization. Mice overexpressing either WT RHOA or an activated form of RHOA under cardiac-specific αMhc promoter control died prematurely. They then showed cardiac, particularly atrial, enlargement, cellular hypertrophy, and interstitial fibrosis. This accompanied AF, atrioventricular block, and heart failure (991). ECG performed under anesthesia suggested a susceptibility to AF and atrioventricular block (991). A murine overexpression system for the rho protein GDP dissociation inhibitor RhoGDIα showed atrial arrhythmias and mild ventricular hypertrophy and bradycardia, atrioventricular block, and atrial arrhythmias with a downregulation of Cx40. These were followed by cardiac hypertrophy and cardiac failure (1245).

3. Mitogen-activated protein kinases, fibrotic change, and atrial arrhythmic phenotype

Mitogen-activated protein kinases (MAPKs) are serine/threonine kinases central to diverse cellular stimuli including environmental stress, proinflammatory cytokine action, and developmental cues. Of these, mitogen-activated protein kinase kinase 4 (MKK4) and MKK7 are critical components in the stress-activated MAPK signaling pathway (1227). Both MKK4 and MKK7 are kinases for c-Jun NH₂-terminal kinase (JNK) (255), reported to suppress TGF-β1 gene expression (1188). MKK4 additionally activates p38 (139). Studies in genetically modified murine platforms suggested that attenuation of cardiac p38 activity resulted in a progressive growth response and myopathy particularly following aortic banding (141). MKK4 may also regulate Cx43 remodeling by either phosphorylation- or transcription-dependent mechanisms (890). MKK4 may thus inhibit maladaptive pathological hypertrophy through its activation of the JNK pathway (680).

Genetic modifications in Mkk4 expression offered an experimental platform for investigating the consequences of alterations in Cx43 expression. The kinetics of Cx43 synthesis, trafficking, and degradation results in a rapid turnover (half-life ∼1–5 h) (948). This in turn potentially influences its expression in directions that could contribute arrhythmic substrate. Both knockdown of Mkk4 expression by silencer siMkk4 and inhibition of its kinase activity by infection with a dominant negative, Ad-dnMkk4, adenovirus reduced the Cx43 expression induced by phenylephrine challenge in NRCMs. This likely involved two specific activator protein-1 binding elements in the Cx43 promoter region.

Cardiomyocyte-specific knockout mice (Mkk4-cko) showed reduced and heterogeneous Cx43 expression and zonula occludens-1 (ZO-1) protein content following transverse aortic constriction. Their ECG recordings demonstrated widened QRS durations (∼16 vs. 10 ms), and increased yet normal QT intervals. Programmed electrical stimulation induced VT in 6 of 13 Mkk4-cko but none of the studied WT mice. Finally, epicardial activation mapping indicated ventricular activation delays in Mkk4-cko hypertrophied hearts. Mathematical modeling simulated a slowed, fragmented, and potentially proarrhythmic ventricular conduction (1351).

MKK4-based systems also permitted analysis of the effects of fibrotic change on connectivity between atrial myocytes. The clinical prevalence of AF and that of atrial fibrosis increase in parallel with age. Clinical evidence implicated MKK4 in this AF pathogenesis. Thus qPCR comparisons of MAPK components and profibrotic signaling molecules in atrial tissues demonstrated MKK4, but not MKK7, P38 or JNK, downregulation, in AF relative to age-matched control patients in sinus rhythm. These findings were further confirmed as decreased MKK4 protein levels. qPCR analysis in the AF patients further suggested increased profibrotic gene expression. There were thus increased levels of Col1a1 and Col3a1 collagen, and the profibrotic connective tissue growth factor (CTGF), in contrast to downregulated matrix metalloprotein 2 (MMP2) expression. MKK4 mRNA downregulation was the more marked in older (>75 yr) compared with younger age AF groups.

These clinical observations could be correlated with findings in a Mkk4-acko conditional knockout mouse model with an atrial cardiomyocyte-selective deletion of Mkk4 using the natriuretic peptide precursor A (Nppa) promoter-driven Cre transgene. Young (3–4 mo), adult (6 mo), and old (1 yr) Mkk4-acko mice were compared with age-matched Mkk4-f/f controls. Aging Mkk4-acko mice were more susceptible to atrial tachyarrhythmias. These were associated with slowed and dispersed atrial conduction (Figure 21, Aa, Ab, and B). There were increased incidences of arrhythmia particularly in old Mkk4-acko hearts (Figure 21C). Yet, Langendorff-perfused Mkk4-acko and Mkk4-f/f hearts showed similar APD₉₀ and ERP values, and Cx levels and distribution through all age groups. Studies of atrial proteins that might contribute to slowed conduction and atrial arrhythmogenesis demonstrated no differences in mRNA or protein expression level expression of Nav1.5, RyR2, NCX, inositol 1,4,5-trisphosphate receptor (IP₃R), the transient receptor potential channel proteins, TRPC1, TRPC3, and TRPC6, and the gap junction proteins, Cx40 and Cx43 between young or old, Mkk4-f/f and Mkk4-acko mice.

Figure 21.

The arrhythmic mitogen-activated protein kinase kinase 4 knockout heart as an arrhythmic exemplar for fibrotic change. In vivo and ex vivo cardiac electrophysiological characterizations of mice carrying an atrial cardiomyocyte specific mitogen-activated protein kinase kinase 4 knockout, Mkk4-acko, compared with Mkk4-flox/flox, Mkk4-f/f, controls. A: representative in vivo ECG recordings showing (a) normal rhythm in Mkk4-f/f (a) in contrast to polymorphic atrial ectopic beats and spontaneous atrial tachycardic episodes in Mkk4-acko mice (b). B: ex vivo atrial epicardial monophasic action potential (AP) recordings in Langendorff-perfused hearts during programmed electrical stimulation interposing extrasystolic S2 stimuli following trains of pacing S1 stimuli. These show contrasting persistent sinus rhythm in Mkk4-f/f (a) with observations of frequent AF in Mkk4-acko (b). C: occurrence of atrial arrhythmic events (AT and AF) in young (3 mo) and old (12 mo), Mkk4-f/f and Mkk4-acko, hearts. D: picrosirius red-stained atrial tissue, fibrotic areas dark red, from 3- and 12-mo-old, Mkk4-f/f and Mkk4-acko, mice. E: percentage fibrotic area in 3- and 12-mo-old Mkk4-f/f (white) and Mkk4-acko atria (black bar). F: epicardial multielectrode array (MEA) activation maps resulting from differing AP conduction velocities following pacing in the center of the array in right (RA) and left atria (LA) of Mkk4-f/f and Mkk4-acko mice (a). G: computer modeling of the effects of fibroblast-cardiomyocyte coupling resulting in reentry following a premature beat. A standard S1S2 stimulation protocol is applied at the left edge of the 2D model containing randomly distributed fibroblast populations in which between one and five fibroblasts are coupled to any given cardiomyocyte. Subsequent snapshots demonstrate a breaking down of the wavefront of atrial excitation wave leading to formation of reentrant excitation waves. [From Davies et al. (245).]

The increased arrhythmogenecity thus more likely resulted from a development of an increased interstitial fibrosis (Figure 21D). This was accompanied by upregulated TGF-β₁ signaling and dysregulation of matrix metalloproteinases (Figure 21E). Mkk4 inactivation also increased the sensitivity of TGF-β₁ signaling to angiotensin II-induced activation in cultured cardiomyocytes. It thus increased the expression of profibrotic molecules in cultured cardiac fibroblasts. Thus Mkk4 likely downregulates the TGF-β₁ signaling associated with atrial remodeling and arrhythmogenesis with age (245). Mkk4 deletion thus appeared to increase atrial arrhythmic tendency by altering conduction properties especially in old Mkk4-acko mice. MEA recordings of atrial AP arrival times, following stimulus application at a point in the center of the MEA array, demonstrated similar conduction velocities in both LAs and RAs of young (3 mo) Mkk4-f/f and Mkk4-acko hearts. However, conduction velocity was slower in old (12 mo) Mkk4-acko atria than old Mkk4-f/f hearts (Figure 21F).

Modeling studies confirmed that such fibrotic processes could potentially lead to atrial arrhythmogenic phenotypes in Mkk4-acko mice through formation of electrical fibroblast-cardiomyocyte couplings (574). This decreased AP overshoots and (dV/dt)_max values, prolonged APDs, and depolarized resting potentials to extents dependent on the number of fibroblasts coupled to each myocyte. This would fulfil expectations from the coupled fibroblasts variously acting as passive electrical loads and electrical drivers through the cardiomyocyte AP time course. Two-dimensional modeling of the consequences of these loading features demonstrated that such fibroblast-cardiomyocyte couplings slowed the conduction velocities of atrial excitation waves. Furthermore, it did so to extents increasing with fibroblast density. It thereby replicated differences in conduction velocity in young and old Mkk4-acko or Mkk4-f/f atria, particularly the greater negative dependence of conduction velocity on fibroblast density in the old than the young group. It further demonstrated wavebreaks in propagated atrial excitation resulting from such coupling. Figure 21G illustrates such effects following S1S2 stimulation protocol applied at the left edge of the two-dimensional map representing the situation in which one to five fibroblasts had been coupled to any particular cardiomyocyte. Illustrations of the subsequent S2-induced excitation waves at a S1S2 coupling interval of 60 ms demonstrated breakdown of the excitation wavefront and re-entrant excitation.

B. Hypertrophic Cardiomyopathy

1. Mutations involving sarcomeric proteins

Cardiomyopathies are characterized by structural myocardial abnormalities not accountable for by acquired, ischemic, valvular, hypertensive, or congenital conditions (732). These conditions together account for most cases of SCD. Cardiomyopathies are broadly categorized into hypertrophic (HCM), dilated (DCM), arrhythmogenic RV, restrictive, and LV noncompaction cardiomyopathies (715, 732) with multiple genotypes occasionally existing through families (1220). The most common, HCM (prevalence 1:500), presents as asymmetric or symmetric myocyte hypertrophy, fibrosis, and myofiber disarray (1260). It is associated with a wide variety of often sarcomeric mutations (732). The condition is clinically heterogeneous. The different mutant alleles involved are associated with different hypertrophic severities and clinical outcomes (1237). The anatomical abnormalities provide potential reentrant excitation circuits leading to high arrhythmia risk (1010). Some sarcomeric mutations appear to result in high SCD risk even without cardiac hypertrophy (1173).

Seventy-five percent of HCM cases involve mutations in genes encoding myosin α- and β-chains, MYBPC3, and troponin T. The murine models available provide broad phenotypic parallels relating to their accompanying structural changes or alterations in Ca²⁺ homeostasis that, however, merit further more detailed study. Heterozygous, particularly male, Mhc-R403Q/+ mice that contain a clinically relevant, myosin α-heavy chain mutation, developed the classical histology of disarray, hypertrophic myocytes, and interstitial fibrosis between 5 and 15 wk (347). These changes accompanied prolonged ECG repolarization intervals and right axis deviation. There were also heterogeneous ventricular electrophysiological conduction properties, prolonged sinus node recovery times, and ventricular ectopic events (112). However, their arrhythmic susceptibility varied more with the extent of ventricular hypertrophy than it did with the extent or location of myocyte disarray or cardiac fibrosis (1266).

Heterozygous, MyBP-C-t/t, defects truncating myosin-binding protein C in its cardiac myosin heavy chain-binding and titin-binding domains also cause human HCM. Mice with analogous heterozygous, MyBP-C-t/+ and homozygous MyBP-C-t/t showed mild dilated and severe hypertrophic cardiomyopathic phenotypes, respectively. However, they showed normal ECG intervals, and sinus node, atrial, and ventricular, conduction, and refractoriness. MyBP-C-t/t but not MyBP-C-t/+ mice were more susceptible to induction of VT than WT and showed ventricular ectopic activity, to extents that, however, did not correlate with their degrees of interstitial fibrosis (113).

Mutations in TNNT2 which encodes cTnT involve the thin filament sarcomeric protein component troponin T. These are associated with ventricular arrhythmia and SCD. In the case of TNNT2-R92T, SCD occurs typically at ∼17 ± 9 yr. There is then minimal hypertrophy or fibrosis but significant myofilament disarray (795, 1173). Similarly, Tnnt2-I79N, clinically linked to human HCM, is associated with increased incidence of SCD despite a mild cardiac hypertrophic phenotype. However, TNNT2 mutations might cause their associated stress-induced VT through electrophysiological remodeling of Ca²⁺ transients and reducing I_K1 rather than hypertrophy and/or fibrosis. Ambulant Tnnt2-I79N mice were more susceptible to nonsustained VT produced by behavioral stress. Intact mouse hearts and ventricular myocytes carrying Tnnt2-I79N showed Ca²⁺ transients with slowed decay kinetics, suggesting increased Ca²⁺ sensitivities in their TnT. They also showed shortened APD₇₀, and normal I_Ca and I_to but decreased I_K1. The APD reductions were reversed by increasing intracellular Ca²⁺ buffering or blocking NCX function. Higher pacing rates or isoproterenol challenge resulted in diastolic Ca²⁺ levels higher than those found in control myocytes and increased susceptibility to ventricular ectopic events (583).

Phenotypic comparisons of HCM-linked murine TnT mutations resulting in strong (TnT-I79N), intermediate (TnT-F110I), and an absence of Ca²⁺ sensitization (TnT-R278C) respectively in human cardiac fibers directly implicated myofilament Ca²⁺ sensitivity in this arrhythmic susceptibility which takes place in the absence of hypertrophy, fibrosis, or myofibrillar disarray (83, 422, 580). All three groups showed similar low baseline incidences of ventricular ectopic activity that were increased by isoproterenol challenge. However, fast pacing triggered sustained VT with incidences in the sequence: TnT-I79N > TnT-F110I > TnT-R278C = WT. Epifluorescence imaging demonstrated greater spatial variability in activation times in TnT-I179N than TnT-R278C hearts. Finally, the Ca²⁺-sensitizing agent EMD 57033 and the myofilament Ca²⁺ desensitizing agent blebistatin respectively reproduced and diminished these arrhythmic susceptibilities. Ca²⁺ sensitization altered ventricular APD, increased its beat-to beat variation, and shortened VERPs. It also increased the dispersion of ventricular conduction velocities at high heart rates. All these changes potentially create arrhythmic substrate (83).

Mutations in other sarcomeric proteins, such as the regulatory (RLC) or essential (ELC) myosin light chain, constitute rarer causes of HCM (902a). However, even genetically modified mice carrying ELC or RLC mutations known to result in human HCM with midventricular cavity obstruction frequently did not show a hypertrophic phenotype, even in the presence of myofilament and cellular level increases in Ca²⁺ sensitivity and relaxation rates (997).

2. Abnormalities in nonsarcomeric cellular components

HCM can also follow abnormalities in nonsarcomeric cellular components. Syrian hamsters carrying spontaneous δ-sarcoglycan mutations show HCM, DCM (993), and stress-induced SCD (750) in common with human sarcoglycan mutation-related HCM. Gain-of-function mutations involving the Ras-Raf protein kinase pathway classically associated with development disorders including the Noonan and LEOPARD syndromes are also associated with HCM. This likely takes place through an upregulation of RAS signaling that leads to a pathological cardiomyocyte hypertrophy (865). Genetically modified mice with a targeted and chronic Ras-Raf-MAPK pathway activation developed diastolic dysfunction, SR Ca²⁺ defects, and increased early sudden cardiac mortality (1346). Inducible gene-switch activation of the hypertrophic Ras-Raf-mitogen-activated protein kinase pathway in adult murine hearts elicited ventricular hypertrophy and ventricular arrhythmias. The isolated ventricular myocytes showed an electrophysiological remodeling that paralleled changes frequently reported in hypertrophic and failing hearts. The latter took the form of AP prolongation, increased NCX activity, reduced I_K, altered SR Ca²⁺, and deficient PKA-dependent PLN phosphorylation. The changes accompanied induction of Gα-inhibiting subunit 1 (Giα1) expression and were partly rescued by G_i/o inhibitor pertussis toxin (969).

C. Dilated Cardiomyopathy

1. General features

DCM results in LV dilatation and consequent systolic dysfunction despite normal loading conditions and coronary function. Twenty-five to 40% of cases are familial, often as autosomal dominant characteristics with variable penetrances. DCM can also be acquired following infection, toxin exposure, or metabolic abnormalities. Its associated electrophysiological abnormalities include sinus node dysfunction, atrioventricular block, and atrial or ventricular arrhythmias often following tissue scarring (603, 904). DCM is associated with over 30 different cytoskeletal, sarcomeric protein/Z-band, nuclear membrane, and intercalated disc proteins (603). The murine systems replicating DCM may also demonstrate its accompanying alterations in contractile function and/or its arrhythmic phenotypes.

2. Ion channel genes involved in DCM

Of ion channel genes associated with DCM (504, 617), SCN5A has been implicated as a candidate gene in screening studies of DCM families with cardiac conduction disorders. These include the SCN5A-D1275N (758), and SCN5A-T220I, R814W, and D1595H missense and 2550-2551insTG truncation mutations (854). SCN5A-D1275N cosegregated with age-dependent DCM phenotypes with variable penetrances associated with AF, impaired automaticity, and conduction delay. A Chinese family carrying SCN5A-A1180V showed a mild in vitro I_Na phenotype accompanied by DCM preceded by atrioventricular block (346). Hypertrophic change has also been associated with Na⁺ channel-related, LQTS phenotypes (553, 1049). Murine Scn5a+/− hearts while similarly showing conduction defects, AF and ventricular arrhythmia, and progressive fibrotic change (see sect. VE) did not exhibit features of DCM. This finding is in direct contrast to the remaining close phenotypic similarities between murine Scn5a+/− and human BrS phenotypes. This may prompt further investigation for differences that might be related to the effects of the intraventricular systolic pressures developed within the small murine as opposed to the relatively large human ventricular chambers.

3. Sarcomeric protein mutations

A substantial proportion of clinical DCM cases are associated with sarcomeric protein mutations. However, the two murine myosin heavy chain (MYH6) models were not arrhythmic (1018). In contrast, murine mutations modulating myosin function showed arrhythmia, thus recapitulating the corresponding clinical phenotypes. Transgenic mice containing dominant-negative mutations altering cardiac-specific expression of the intercalated disk mammalian-enabled protein (Mena) and vasodilator-stimulated phosphoprotein (VASP) showed bradycardic and sudden death phenotypes. These were associated with reduced cell-to-cell interactions (293). Similarly, mice overexpressing a dominant-negative transgenic copy of the transcriptional myosin regulator, neuron-restrictive silencer factor, showed VT and heart block associated with upregulation of I_f and I_CaT, normally specific to pacemaker and Purkinje cells (608).

4. Cytoskeletal protein mutations

Of mutations disrupting cytoskeletal function, those involving the lamin A/C (LMNA) gene cause a wide range of clinical conditions. Lamin A/C, LMNA+/−, mutations producing 50% reductions in normal cardiac lamin A/C resulted in a development of atrioventricular node myocyte apoptosis and conduction loss, atrial arrhythmias, and VT by age 4 wk, and DCM with age (1267). Mice carrying the missense Lmna-N195K murine variant known to cause human DCM, similarly showed DCM, a profound and progressive sinus bradycardia reflecting sinoatrial exit block, and a range of conduction abnormalities. These included progressive PR interval prolongations with a premature, arrhythmic, death on continuous ECG monitoring. They also showed misexpression of transcription factor Hf1b/Sp4 and of Cx40 and Cx43, and disorganized sarcomeres and intercalated disks.

The cardiomyopathy thus likely arises from this disruption of internal cardiomyocyte organization and possibly an abnormal expression of transcription factors ensuring normal cardiac development (807). The LMNA-H222P missense mutation occurred in a family with the autosomal dominant Emery-Dreifuss muscular dystrophy. Homozygotic Lmna-H222P mice showed apparently normal embryonic and sexual development. However, adult male, and at a later stage, female, mice showed abnormal voluntary locomotor activity and an echocardiographically detected DCM. They also showed ECG evidence of increased PR intervals and QRS durations, episodes of sinoatrial block and ventricular extrasystoles, and premature mortality at ∼9 mo. There was an accompanying cardiac and skeletal muscle degeneration and fibrosis and an activated Smad signaling (45).

5. SR Ca²⁺-cycling protein mutations

Genetic abnormalities in SR Ca²⁺-cycling proteins, particularly those affecting the SERCA regulator PLN, also cause familial cardiomyopathies. PLN dephosphorylation reduces apparent SERCA2a Ca²⁺ affinity. The resulting SERCA2a inhibition can be abolished by PKA-mediated phosphorylation. Of naturally occuring PLN mutations, one carrying a homozygotic lysine-39 stop codon (PLN-L39stop) abolished PLN expression and caused a lethal DCM and heart failure at teenage. However, this clinical phenotype was not reproduced in Pln-deficient mice (398). A R9C mutation in the Pln coding region seemed to preclude PKA-mediated phosphorylation of even WT PLN producing a chronic inhibition of SERCA2a (1019). This was similarly associated with a clinical DCM recapitulated in transgenic mice overexpressing Pln-R9C.

Mutations affecting Pln transcription levels also result in inhibited SERCA2a activity. One of 87 of a series of hypertrophic patients showed an A to G mutation in this region (776). Infections of in vitro rat myocytes suggested that such a mutation markedly increased PLN promoter activity consistent with a role in producing DCM. Finally, a heterozygotic Pln-R14Del variant involving PLN coding demonstrated by genetic screening of DCM patients was clinically associated with LV dilatation, contractile dysfunction, episodic ventricular arrhythmias, and heart failure by middle age. These features were reproduced in transgenic mice overexpressing Pln-R14Del. WT and Pln-R14Del coexpression in HEK-293 cells demonstrated a SERCA2a superinhibition that might provide a mechanism for the DCM (397).

6. Developmental gene mutations

Of developmental genes, the cardiac homeobox gene Nkx2-5 is central to early cardiogenesis (409). Mutations produce a wide range of malformations including septal and outflow tract defects, cardiomyopathy and LV hypoplasia, and associated arrhythmias (1022). Neonatal mice with ventricle-restricted Nkx2-5 deficiency appeared structurally normal. They displayed no arrhythmias on ECG recording. They subsequently developed a trabecular muscle hypertrophy and progressive, first degree (by 12 wk), and complete heart block (at 6–12 mo) recapitulating clinical findings in some Nkx2-5 patients. This accompanied a cardiomyocyte dropout and fibrosis following their sarcomeric and cellular disorganization in the central conduction system. This produced a distinct form of cardiomyopathy characterized by extensive trabeculae and myocardial noncompaction. There was a progression to dilatation of both chambers, their marked trabeculation, and heart failure. There was a dysregulation of genes associated with myocardial cell proliferation and trabeculation. This included an increased expression of bone morphogenic protein-10 (BMP-10) and a downregulated cardiac homeodomain-only protein, HOP. These findings likely represented a distinct pathway for progressive cardiomyopathy associated with conduction defects in congenital heart disease (871).

7. Cytoskeletal protein mutations

Of cytoskeletal proteins, vinculin anchors the actin cytoskeleton to the cell membrane. It is thus likely required to preserve cell-cell and cell-matrix adhesion. Murine cVcl-KO models with Cre-loxP generated cardiac-myocyte-specific vinculin (Vcl) gene inactivation showed telemetric ECG evidence of VT leading to sudden death, and optical mapping evidence of reduced myocardial conduction even at <3 mo. This was followed by development of DCM often fatal at <6 mo. This accompanied abnormal adherens junction and intercalated disc structure, reduced cadherin and beta1D integrin expression, and mislocalization of Cx43 to the lateral myocyte border (1324).

D. Arrhythmogenic Right Ventricular Cardiomyopathy

1. Clinical features

ARVC is associated with ventricular arrhythmias and SCD, particularly in young athletes (79, 219), giving a phenotype unmasked by exercise testing (886). It is characterized by a fibro-fatty infiltration leading to RV dilatation. Dilatation may extend to areas of the LV but characteristically spares the septum (732). Desmosomal protein gene mutations occur in 50% of such patients (685). Desmosomes participate in end-to-end connections transmitting contractile force between cardiac myocytes (250). Abnormal junctional organization leads to myocardial damage and replacement fibrosis culminating in focal scars isolating cardiomyocytes within nonconducting fibrous tissue (80). This slows conduction, generating reentrant substrate.

2. Plakophilin-2 (Pkp2) deletions

Murine models exist for the transmembrane desmosomal cadherins, plakoglobin (γ-catenin) and plakophilin 2, and the plakin protein desmoplakin, which links the junctional complex to the intermediate filament, desmin, thereby maintaining mechanical integrity (685, 1046). Mutation of the gene encoding plakophilin-2 may occur in up to one third of ARVC families (349). However, Pkp2+/− mice with a heterozygotic plakophilin-2 targeted deletion showed no cardiac morphological phenotype. Nevertheless, human subjects with plakophilin-2 (PKP2) mutations showed more frequent VT episodes than those without detected PKP2 mutations (68). Furthermore, the homozygous deletion was lethal at postfertilization day 11 with evidence for endothelial discontinuities in heart and vessels (376).

3. Consequences of plakoglobin (Pg) deletion and overexpression

One of the two initially available heterozygous murine plakoglobin deletion models showed RV enlargement, increased spontaneous ventricular arrhythmias, and RV conduction slowing (123, 973). This took place despite an absence of replacement fibrosis, junctional remodeling, or alterations in Cs43 localization and distribution, at 10 mo. These abnormalities were exacerbated by exercise (571) and relieved by load reduction regimes (299). In contrast, homozygous targeted plakoglobin deletion was lethal between ED 9.5 and ED 16 with thinning of the cardiac walls, reduced trabeculation, and epicardial discontinuities. Nevertheless, a cardiac-specific targeted Pg deletion controlled by αMHCcre partially recapitulated human ARVC. Pg^f/f-αMHCcre mice showed ∼30% of WT protein but no Pg in cardiac sections. They showed a tendency to SCD, progressive RV and LV dilatation, and fibrosis though no cardiac fat deposition from ∼2 mo. Desmosomal ultrastructure was disrupted (647). There was evidence of myocyte apoptosis in addition to myocyte necrosis, in contrast to the necrosis observed with desmoglein (Dsg2)-N271S overexpression. The intercalated disks also lacked other desmosomal proteins, including Dsg2 (647). However, β-catenin expression appeared increased and may have partially rescued the sudden death phenotype. The latter was suggested in crosses of double-targeted mice carrying both floxed Pg and floxed β-catenin loci (Pg-f/f; β-catenin-f/f) loci with αMHC/MerCreMer mice. Tamoxifen challenge resulting in specific targeted deletion produced a strong arrhythmogenic SCD phenotype in all animals between 3 and 5 mo after tamoxifen injections (1105). Finally, mice with both WT or mutant Pg overexpression showed increased incidences of SCD (686).

Similarly, a two-base pair plakoglobin gene deletion is associated with the recessive Naxos syndrome, associated with further, keratotic, manifestations (757, 973). Yet overexpression of the resulting, truncated, Naxos-associated plakoglobin also produces an arrhythmogenic cardiomyopathic phenotype. However, it is likely that it is an insufficiency of either the normal or truncated plakoglobin that produces disease phenotypes. One of two knockin murine models expressing a Naxos-associated plakoglobin that bypassed the nonsense-mediated mRNA decay mechanism thereby producing normal levels of truncated plakoglobin showed normal cardiac function (1343).

4. Desmoplakin (Dsp) targeted deletion and overexpression

Homozygous loss-of-function desmoplakin (Dsp) mutations are clinically associated with a biventricular form of ARVC. Heterozygous Dsp+/− mice developed normally. However, adults then showed attenuated ventricular walls, increased LV diameters, reduced LV ejection fractions, spontaneous arrhythmias on ECG recordings, and extrastimulus-induced ventricular arrhythmias. The structural changes were preceded by increased activation times and arrhythmic tendency measurable in Langendorff-perfused hearts (366). Homozygous, Dsp−/−, desmoplakin targeted deletions were lethal at ED 6.5 (475). Postponing this lethality to ED 11 by tetraploid aggregation permitted demonstration of severe cardiac malformation though persistent desmosomal-like ultrastructures (339). Embryos with a cardiac-specific, αMHCcre induced, targeted Dsp−/− deletion revealed poorly formed hearts (342).

Mice with cardiac overexpression of flag-tagged Dsp cDNA carrying the Dsp-R2834H-Tg COOH-terminal mutation demonstrated increased heart-to-body weight ratios compared with mice whether with human WT (WT-Tg) Dsp cDNA or WT littermates (1296). Dsp-R2834H-Tg hearts also showed increased apoptosis and fibrosis, RV and LV dilatation, reduced ventricular function, widened intercalated disks, and disrupted plakoglobin-desmoplakin interaction. Overexpression of Dsp with NH₂-terminal, Dsp-V30M, and Dsp-Q90R mutations was lethal after ED 13.5, resulting in reduced ventricular wall attenuation and dilatation.

Finally, studies with crosses of desmoplakin-floxed mice (Dsp-f/f) with cardiomyocyte-specific ventricular myosin light chain-2-Cre recombinase (MLC2v-Cre) knock-in mice yielded viable animals. But these showed early ultrastructural desmosomal defects and cardiomyopathic changes including the cell death, ventricular fibro-fatty replacement, biventricular dysfunction, and failure and premature death. Hearts showed ventricular arrhythmias accentuated by exercise and catecholaminergic challenge. They also showed slowed RV conduction. These were attributable to reduced Cx40 and Cx43 expression, suggesting roles for desmoplakin in stabilizing Cx integrity (706).

5. Desmoglein (Dsg) overexpression models

Mice with cardiac overexpression of flag-tagged Dsg2-N271S mutant, homologous with the human DSG2-N266S ARVC mutation, showed spontaneous ventricular arrhythmias, conduction slowing, ventricular dilatation and aneurysms, replacement fibrosis, and SCD at age <2 wk. This was associated with myocyte necrosis, calcification, and fibrous tissue replacement. In contrast, mice with flag-tagged Dsg2-WT showed normal phenotypes indistinguishable from WT littermates even at age 2 mo (896). Mice with a targeted deletion in the extracellular adhesion domain of Dsg2 similarly showed RV and LV dilatation, fibrosis, calcification, and SCD (597). Hearts showed intercellular widenings at the intercalated disk and loss of desmosomal ultrastructure (536).

6. Laminin receptor mutations

A line of KK/Rvd mice with right ventricular dysplasia (RVD) contain a gene laminin receptor 1 (Lamr1) mutation containing an intron-processed retroposon whose gene product binds to heterochromatin protein 1 potentially modifying transcriptional regulation. The hearts showed a marked fibrosis in the RV free wall and cardiomyocyte cell death. Electrocardiography revealed prolonged QRS durations, suggesting intraventricular conduction disturbance potentially increasing susceptibility to arrhythmia (48).

E. Left Ventricular Noncompaction Cardiomyopathy

LV noncompaction cardiomyopathy (LVNC) is a rare, frequently autosomal dominant condition characterized by a histologically distinct noncompacted myocardial layer containing deep and prominent trabeculae (1248, 1320). It is associated with cardiac failure, angina pectoris and arrhythmia in the form of AF, ventricular tachyarrhythmias, and SCD in some series (848). The condition has been related to mutations in the neurogenic locus notch homolog protein (NOTCH) pathway regulator E3 ubiquitin-protein ligase (MIB1) (705) as well as various genes overlapping those associated with other cardiomyopathies. These include the cypher/Z-band alternatively spliced PDZ-motif protein in the sarcomeric Z line (1178). This binds to α-actinin likely contributing cytoskeletal structural integrity during contraction, and may be involved in PKC signaling through its LIM domains (1045). Mice with disrupted Z-line proteins as a result of mutations in the CYPHER/ZASP gene show phenotypes in common not only with LVNC but also DCM and HCM and suffer early postnatal death (1345, 1350).

X. SUMMARY AND CONCLUSIONS

Cardiac arrhythmias result from a breakdown of the ordered successive patterns of excitation through the SAN, atria, or ventricles essential for normal contractile function. They constitute a major public health burden (see sect. IA). Their analysis requires understanding of the process of normal electrical excitation, conduction, recovery, and membrane stability. At the molecular level, this process depends on normal function in a large range of individual ion channels with specific ion permeabilities and voltage dependences in their steady state and kinetic properties. Successful function at the cellular level requires a sequential and interdependent activation and inactivation of their resulting ion channel currents I_i (see sect. IB). At the tissue level, the resulting AP propagates through electric current flow along the intracellular resistance r_a to depolarize the membrane capacitance c_m. It thereby initiates excitation in hitherto quiescent connected membrane effectively along a multidimensional cable. Regularly paced APs thus acquire their characteristic duration, APD, and refractory period, ERP, separated by diastolic interval DI, through an integration of molecular mechanisms into cellular level events. The APs then propagate through myocardial tissue with velocity θ, accordingly forming active regions with wavelength λ = θ × APD, or θ × ERP, followed by a quiescent wavelength determined by θ and the diastolic interval DI, defined by an integration of cellular events and tissue properties. These fundamental characteristics define the conditions necessary for orderly excitation patterns initiated from the pacemaker and propagating through the remaining cardiac structures (see sect. IC).

Arrhythmias following perturbations of these parameters may thus take place through their temporal heterogeneities resulting from transient membrane potential change, or alterations in steady-state pacing rates. They may also be the consequence of spatial heterogeneities potentially resulting from varying tissue properties in different organ regions, or pathological anatomical change, particularly those impairing the processes of conduction or excitation (see sect. IIA). Analysis of cardiac excitation and its alterations producing arrhythmias thus requires experimental systems amenable to study at different levels of biological organization. Studies translatable to human arrhythmic conditions would require such systems to parallel human hearts in structure and function, and share a significant number of molecular species underlying their electrophysiological activation.

Mouse hearts fulfil a significant number of such criteria. In addition, they are amenable to genetic modification modifying the expression or function of particular, physiologically important, biological molecules. Murine systems further provide a range of translationally relevant exemplars for the clarification of arrhythmic processes arising from abnormalities in the initiation, propagation, and recovery of electrical activity or upon the potential for spontaneous triggering of arrhythmia (see sect. IIB). Each examplar was amenable to experimental and theoretical analysis at varied levels of biological organization and function through application and collation of results from a wide range of techniques. These extended from electrocardiography in entire organisms, through extracellular and intracellular AP recording from isolated hearts subject to varied transient perturbations or steady-state pacing conditions, biophysical ion channel and spectrophotometric Ca²⁺ homeostasis studies at the cellular level, to characterizations of protein and mRNA expression and localization at the molecular level (see sect. IIC).

Arrhythmic phenomena ultimately arise from biophysical events, beginning with ion channel-mediated processes in the cell surface membrane. A first group of studies examined a series of defined monogenic surface ion channel variants (see sects. III–VI). These recapitulated the arrhythmic phenotype in their corresponding clinical genetic arrhythmic conditions. This gave credence to their use as disease exemplars in physiological studies including explorations for possible therapeutic maneuvers. They fell into canonical groups that could be classified into effects relating to pacing (sect. III), cellular connectivity (sect. IV), AP excitation (sect. V), and recovery from depolarization as well as refractoriness (sect. VI). The last two recapitulated features of the BrS and the LQTS, respectively. The analysis accordingly began with the effects of compromised or prolonged Na⁺ channel function whether through alterations in the conducting Nav1.5 subunit itself, or its associated β-subunits or intracellular regulatory or structural regulatory proteins (sects. V and VI). These were followed by considering the physiological consequences of alterations in other ion channels carrying depolarizing Ca²⁺ or repolarizing K⁺ currents, and their different variants occurring variously in ventricular or atrial myocytes. The analysis of the murine systems recapitulating BrS further demonstrated longer term structural alterations recapitulating those found in human BrS and reproduced the time development of conduction change and arrhythmia in BrS and PCCD.

In each group, a systematic survey identified the observed arrhythmic properties with alterations in pacing, initiation, and conduction of, as well as, recovery from excitation. They quantified critical conditions in heart rate (HR), conduction velocity θ, APDs, and refractory periods ERPs. Detailed alterations in APD waveforms or recovery could modify the likelihood of reexcitation phenomena, potentially through effects upon Ca²⁺ current mediating EADs, and through effects on refractoriness. Arrhythmic substrate could arise from both temporal variations of these physiological parameters with extrasystolic provocation or with varying steady-state pacing rates, and spatial variations over the epicardial surface, across myocardial walls, and different cardiac regions and chambers. Genetic modifications of any particular molecular species impacted upon more than one of these physiological parameters. Conversely, any one physiological parameter was affected by alterations in more than one species of channel. These complex physiological dependences have translational implications for the application of therapeutic agents directed at specific molecular species in the management of arrhythmia.

A second group of canonical variants concerned intracellular, as opposed to surface ion channel, changes. Intracellular Ca²⁺ is the predominant ion species involved in excitation-contraction coupling. The consequent alterations in Ca²⁺ homeostasis in turn impact upon surface membrane channel action (sect. VII). Such alterations could take place through abnormal L-type Ca²⁺ channel-mediated Ca²⁺ entry, RyR2-mediated SR Ca²⁺ release, SERCA-mediated SR Ca²⁺ reuptake, and SR store levels. A range of acquired acute or chronic conditions could potentially alter the Ca²⁺ release process in both atria and ventricles. Similarly, murine systems with modified RyR2 or calsequestrin 2 provided genetic exemplars for an abnormal release of SR Ca²⁺. The resulting variations in [Ca²⁺]_i could in turn influence the stability of surface membrane potentials through actions on electrogenic NCX mediating DAD phenomena potentially triggering extrasystolic activity. It could also potentially modify Na⁺ channel activity either acutely or through downregulation of its membrane expression reducing AP conduction velocity and thereby causing arrhythmic substrate. Modified intracellular Ca²⁺ could in turn drive metabolic alterations through its action on intracellular Ca²⁺ activated enzymes.

A third group of variations concerned alterations in cell metabolic state, with alterations in levels of particular intermediates affecting surface K_ATP and Na⁺ channels as well as RyR2-Ca²⁺ release channels (sect. VIII). The latter intermediates include cellular levels of ATP, redox [NAD⁺]/[NADH] state, and altered mitochondrial function generating reactive oxygen species. These could be exemplified in genetic models modifying the mitochondrial upregulator PGC1, as well as models manipulating [NAD⁺]/[NADH].

The final group of examplars involve pathological anatomical, cardiomyopathic, changes that would be expected to modify passive membrane capacitative, c_m, or intracellular resisistive, r_a, cable properties determining electrophysiological conduction (sect. IX). In addition to the fibrotic changes in the Na⁺ channels outlined previously, these could be exemplified by hearts containing variant signaling mechanisms.

Study of murine systems themselves thus provide insights concerning the function of particular ion channels at the tissue and organ levels, and the effects of their deficiencies or overexpression. They thereby contribute substantially to our understanding of the physiological basis of normal electrical activity and its propagation. This leads to developments in our fundamental understanding of the genesis and perpetuation of arrhythmia in terms of perturbations of these processes. A substantial number of exemplars additionally show phenotypic features that parallel the corresponding human clinical conditions. In such situations, they may then be useful in fundamental explorations of both diagnostic and risk assessment criteria and possible therapeutic interventions. However, as with all physiological models, species-related differences would be expected between mouse and humans. This concerns not only normal physiology, but also the pathological changes either leading to or resulting from phenotypic changes at the whole animal level. This is reflected in the small but important group of murine exemplars whose phenotypes showed marked contrasts from their corresponding human conditions. These continue to offer opportunities for specific investigations of such differences, and assessing the applicability or otherwise of specific biomarkers useful in risk stratification in mice to clinical management in humans. Nevertheless, particular readouts from gene targeted mouse models provide insights into broad physiological mechanisms underlying electrophysiological stability or arrhythmogenesis. It is the latter that may be the more useful for the identification of approaches to direct translation.

GRANTS

I thank the Medical Research Council, Wellcome Trust, McVeigh Benefaction, the British Heart Foundation, Helen Kirkland Trust, and Sudden Arrhythmic Death Syndrome-SADS UK for research support.

DISCLOSURES

No conflict of interest, financial or otherwise are declared by the author.