Electrophysiological tolerance: a new concept for understanding the electrical stability of the heart

Abstract

The co-ordinated electrical activity of ∼2 billion cardiac cells ensures stability of the heartbeat. Indeed, the remarkably low incidence (<1%) of ventricular arrhythmias in the healthy heart is only possible when the electrical event across this syncytium is closely controlled. In contrast, the diseased myocardium is associated with increased electrophysiological heterogeneity, unstable rhythm, and increased incidence of lethal arrhythmias. But what is the link between cellular and tissue level heterogeneity? Recent research has shown the existence of considerable cellular heterogeneity even in the healthy heart, suggesting that cell-to-cell variability in electrical (e.g. action potential duration) and mechanical performance (e.g. twitch amplitude) is a normal property. This observation has been previously unappreciated because the aggregated function in the form of QT-interval and cardiac output varies <1% on a beat-to-beat basis. This article describes the underlying cellular variability that is tolerated—and perhaps needed—by different regions of the heart for normal function and indicates why this variability is not apparent in function at the chamber and organ level. Thus, in contrast to the current dominant view, this article postulates that heterogeneity is normal and potentially endows various functional benefits. This new view of how the component parts of the heart come together to function also suggests novel mechanisms for cardiac pathologies, namely that dysfunction may emerge from changes in the extent and/or nature of heterogeneity. Once understood, restoring normal forms of heterogeneity could be a novel approach to treatment.

Graphical Abstract

Graphical Abstract.

Introduction

Increased electrical heterogeneity in any one part of the heart is generally associated with increased likelihood of arrhythmic events.^1,2 Yet, clinical and experimental electrophysiologists using a range of techniques have established that these differences across the heart’s chambers are the basis for normal cardiac function.³ For example, at the level of the ventricle, regional differences in action potential duration (APD) minimize repolarization gradients that arise from activation delays.⁴ This outcome has clear benefits for electrical stability, as it minimizes local (∼0.5 cm) voltage gradients that arise from the electrical resistance between heart cells (∼5 MΩ).⁵ The claim in this viewpoint article is that tolerance of heterogeneity goes deeper, and that heterogeneity at several levels of cardiac excitation–contraction coupling is a key property in the heart`s flexibility and tolerance to perturbations. We provide examples from recent work that describe ‘hidden’ heterogeneous cellular properties that underlie the electrical activity of both atria and ventricles, and the fundamental subcellular events that control contraction. We use these examples to explore the concept of beneficial heterogeneity, and present new hypotheses linking abnormal heterogeneity to pathology, with potential implications for treatment.

While this viewpoint specifically focuses on the role of spatial heterogeneity across the heart’s chambers, it should be noted that beat-to-beat variability in APD and/or peak contraction also occurs. This temporal variability is normally constrained in the healthy heart (<1% in both single cells and multicellular preparations^6,7) and, like spatial heterogeneity, increased in arrhythmia-prone pathological conditions.⁷ Although undoubtedly a related phenomenon, this beat-to-beat form of heterogeneity is not featured in this article.

Tolerance to large between-cell differences in cellular electrophysiology generates stable syncytial electrophysiology

Until relatively recently, the range of action potential (AP) waveforms in atrial and ventricular myocytes was considered relatively limited.^8,9 Indeed, the AP and the subsequent excitation–contraction (E–C) process in a single cell were thought to represent a micro-version of the same process in the whole chamber. With a few exceptions, the same subgroup of ion channels was found responsible for the range of AP shapes. Therefore, the physiology and pathophysiology of E–C coupling in single cells were used to predict the behaviour of the whole chamber. In support of this, the heterogeneity of the APD in a few cubic millimetres of myocardium is minimal with a coefficient of variation (CoV = SD/Mean) < 5%. However, this pars pro toto view has recently been challenged by a study from one of the authors (G.L.S.) that showed that cells isolated from a limited sub-region of the left ventricle have in fact very large differences in AP shape and time course (CoV ∼40%) when ‘disconnected’ (Figure 1A).¹⁰ Indeed, the variation in APD in cells from one individual was larger than that seen across the population of hearts (CoV 20–30%, across species).¹⁰ The paper also showed that even within a subgroup of isolated cells with similar APD, cells can react very differently to ion channel blockers, indicating that two cells with comparable AP waveforms might have very different underlying ion channel conductance values. This between-cell variation is so large that regional differences in APD are minor in comparison, and only appear in mean values from a population. The implication of this finding is that ensemble cellular electrophysiology and ion channel pharmacology cannot be accurately estimated with data from <∼30–40 cells/heart-region, even when methodological factors that contribute to cellular variation are minimized.¹⁰

Figure 1.

(A) In the intact myocardium, strong cell-to-cell coupling limits heterogeneity in action potential configuration between neighbouring cells (left panel). The variation in action potential configuration is uncovered when cardiomyocytes are studied after isolation and decoupling. (B) Visual representation of the principles quantified in Table 1. The cumulative relative variability at each level of excitation–contraction coupling is represented in the grey figures. To the left, variability in the normal heart is represented with the range of ‘electrical tolerance’ and ‘mechanical tolerance’ above and below the figure. The three figures to the right represent the impact of pathology at one of three levels of excitation–contraction coupling, indicated by the red horizontal arrow. The down- or upsloping arrows between the three levels indicate dependency and impact of pathology at any one level on the other two. Green and red dotted lines indicate variability within or outside of the ‘tolerance’ at each level in the specific example.

The paper goes on to model this variation in silico and concludes that the electrophysiological cell-to-cell variation of normal hearts arises because the electrical conductance of individual classes of ion channels varies with CoVs of 50–80%. The reason that this very large variation is ‘tolerated’ is because the expression of ion channels that prolong the APD such as the L-type Ca channel is positively correlated with the expression of ion channels that shorten APD, such as the rapidly inactivating inward rectifier potassium channel (I_Kr). This correlation has been shown to operate across a population of human hearts.¹¹ The single cell data suggested that a range of relationships exist between the expression of subgroups of key types of ion channels regardless of their absolute values. The result is a very heterogeneous response to specific ion channel inhibitors in single cells, the aggregate of which represents the average behaviour of the syncytium.

Why is there such variability in the basic functional unit of the heart? Interestingly, similar variability and co-relationships have been seen in the normalized conductance of ion channels of neurons¹² and heart cells.^13–15 This raises the possibility that between-cell variability is a common physiological phenomenon, perhaps reflecting the ‘Good Enough’ engineering principle.¹⁶ Indeed, a cellular machinery that monitors and maintains membrane ion channel expression within narrow limits would require significantly more metabolic energy than one with a higher tolerance for variation. Thus, one could envision a survival benefit of a functional electrical syncytium that, by design, tolerates significant cell-to-cell electrical variability, particularly when such heterogeneity is minimized by strong cell-to-cell electrical coupling.⁶ An alternative explanation for the between-cell variation may be that optimal chamber haemodynamic function requires regional differences in cellular E–C coupling. For example, recent work employing diffusion tensor magnetic resonance imaging has shown that there is a complex 3D arrangement of cardiomyocytes in the heart, with overlaid laminar ‘sheetlets’ containing cardiomyocytes in distinct orientations.¹⁷ Thus, the mechanical cues that cells experience may be quite different, even for cells that are quite nearby each other, resulting in distinct AP configuration, Ca²⁺ homeostasis, and myofilament response. Indeed, cardiomyocytes are highly mechanosensitive, due largely to the presence of integrins in the cell surface and their associated signalling network.¹⁸ This signalling has been linked to an ability of the cells to fine-tune their subcellular structure and function to meet mechanical demand.¹⁹ Thus, between-cell/regional contractile heterogeneity may be aimed at optimizing contractile power in the complex structure of the syncytium.

Regardless of the reason for the cell-to-cell variability in the healthy heart, the obvious down-side is that any pathological process that reduces between-cell coupling and/or increases cellular electrophysiological heterogeneity may exaggerate the variability inherent in the myocardium and increase the pro-arrhythmic substrate.¹¹ For example, it has been shown that in the border zone (BZ) adjacent to a healed infarct, wall stress is markedly elevated and linked to distinct subcellular remodelling compared to the remote zone (RZ).²⁰ Amoni et al. further showed that within the BZ, the increase in wall stress was highly variable, and was linked to a more variable cellular APD. Resulting cell-to-cell APD90 variability was almost 2× values observed in the RZ, despite no observed difference in mean/median value, and was correlated with propensity for arrhythmias.²¹ Thus, larger than normal electrophysiological heterogeneity cannot be ‘tolerated’ within the myocardium without accompanying pro-arrhythmic changes. In the case of genetic diseases such as Brugada syndrome, where electrophysiological heterogeneity is thought to play a role in the pro-arrhythmic status,²² cellular heterogeneity as described in this article may be a predisposing factor.

One important implication of this paradigm is that novel anti-arrhythmic treatments could aim to restore the normal pattern of cellular electrical heterogeneity. This approach would be beneficial in comparison with those that simply alter the mean electrical signal via an action on a single ion channel. For example, while L-type calcium channel inhibition may reduce heterogeneity by shortening APD, it carries a risk of shorter QT with altered repolarization gradients. Instead, novel approaches may correct cellular heterogeneity by either identifying the underlying cause(s) (e.g. altered transcription patterns and/or mechanical conditions) and normalizing these, or by manipulating the electrophysiological consequences on multiple ion channel conductances directly. For example, actions to enhance repolarization through inhibition of late sodium current or enhanced delayed rectifier currents could be balanced by drug-induced calcium channel agonism to reduce the spread of APD in a population. Such manipulations would thus aim to maintain QT duration without affecting repolarization gradients across the myocardium.

Tolerance to cell-to-cell heterogeneity safeguards contraction

Heterogeneity in cellular electrophysiology has clear consequences not only for the electrical stability of the heart but also the heart’s contractility. During E–C coupling, the AP evokes calcium entry into the cytosol via the activation of the L-type calcium channels, and the release of calcium from the sarcoplasmic reticulum (SR). Intracellular calcium triggers the contraction event, and the extent and time course of the contraction–relaxation cycle are determined by the summed extrusion from the cytosol by the SR calcium ATPase (SERCA) and the sarcolemmal sodium/calcium exchanger (NCX). This is another point where the tolerance to heterogeneity should be assessed. Indeed, marked differences in the intracellular calcium transient, and therefore the strength of contraction, between cells or groups of sarcomeres coupled in series would mean that the stronger cell/sarcomere group would stretch the weaker cell/group. The result of this tug-of-war would be a reduction in the efficiency and extent of contraction.^23,24

What mechanisms protect against such contractile inefficiency? One mechanism appears to be the greater uniformity of between-cell electrophysiology observed in the coupled, intact heart compared to isolated cells. In support of this view, early studies using cell pairs with different APD and calcium transients showed minimization of these differences by electrical coupling.⁶  In silico modelling further indicated that with electrotonic coupling between 30 cardiomyocytes, the resultant unified AP minimizes cell-to-cell differences in the subsequent calcium transient, despite variation of ion channel expression.¹⁰ However, an experimental study looking at calcium transients in adjacent cells in the intact myocardium still noted considerable variability of calcium transient amplitude and time course.²⁵ Both parameters showed between-cell CoV values of ∼10%²⁵; values that are too high for the efficiency of contractility to be safeguarded. Thus, the final process that could absorb such natural cell-to-cell heterogeneity is the contraction mechanism itself, mediated via the sliding filaments of the sarcomere. Somewhat surprisingly, recent work by one of the authors (W.E.L.) showed that isolated cardiomyocytes exhibit heterogeneity in the shortening/stretch of individual sarcomeres, and that this can impair contractile force in cases of mutations in contractile proteins.²⁴  In vivo variability of this process has also been studied, using some remarkable measurements of single cell sarcomeres visualized in mouse heart during sinus beating.²⁶ Importantly, Li et al. observed that for sarcomeres with identical local calcium transients and diastolic sarcomere lengths, the variation in fractional shortening was very low, with a CoV of ∼2% (Li and Louch, unpublished). Thus, the final step in the E–C process helps to minimize contractility differences between adjacent cells, as the precise molecular geometry and regulation of the sarcomere provide a final control point to ensure a functional syncytium.

Table 1 summarizes experimentally observed variability at each level of E–C coupling: the AP, the calcium transient, and contraction. A corresponding graphical representation is shown in Figure 1B. We assumed that CoVs add numerically for each downstream process (downward arrows in table), eventually summing to predict the CoV of contraction. Notably, while the intrinsic CoV for the AP is directly manifested in isolated cardiomyocytes, the CoV is much smaller in the intact healthy heart (functional syncytium), yielding a markedly lower total variability for in vivo contraction than in isolated cardiomyocytes (13 vs. 52).

Table 1.

Intrinsic or natural variability at each step of E–C coupling and the influence on the behaviour of the syncytium in health and disease (in red)

Values represent estimates of the relative CoV of individual processes, and the cumulative effect on the electrophysiology and contractility of the cell, both in isolation or in a functional syncytium. The ‘Electrical tolerance’ represents an estimate of the threshold variability beyond which is considered pro-arrhythmic, the ‘Mechanical tolerance’ a threshold for mechanical dysfunction. This is based on the summed variability values of the three stages on E–C coupling. See text and Figure 1B for further explanations and examples.

‘Pathology’ has been estimated in the table as a doubling of the natural CoV for a given step in E–C coupling, based on reported increases in the spread of data for APD90,²¹ intracellular calcium signalling,²⁵ and sarcomere contraction/stretch due to a titin mutation.²⁴ However, the picture is complicated somewhat in the cases of defects in calcium signalling or myofilaments, since these functions feedback on the upstream processes (indicated by bidirectional arrows). Indeed, changes in calcium homeostasis directly influence AP configuration, and contractile function feeds back on calcium handling. To include these mechanisms, we have arbitrarily set the upstream effect at 50% of the magnitude of the initiating deficit. These calculations show that deficits in different steps of E–C coupling (AP, calcium, or myofilaments) have quite distinct effects on cumulative CoV (see Table 1 and Figure 1B).

What are the implications of these calculations for the electrical and mechanical ‘tolerance’ of cell-to-cell variability? Assuming that tolerance is set only slightly beyond the normal range of variability, we see that deficits in calcium homeostasis result in cumulative CoV that is markedly beyond this threshold, which may promote arrhythmia and/or mechanical impairment. In contrast, deficits in AP configuration or sarcomere function appear to be less consequential. Another important takeaway message is that it is difficult to predict the consequences of pathological increases in variation when only isolated cardiomyocytes are investigated, since electrophysiological variability is markedly lower in the intact heart.

Conclusion

Traditionally considered a pro-arrhythmic phenomenon, our claim is that a modest level of electrical heterogeneity is a normal feature of the healthy heart. Based on recent publications and emerging concepts, we have given examples of heterogeneity at different levels of E–C coupling. The interdependency of these processes appears to be a weakness in the heart’s design, as a fault in any part of the mechanism will introduce a higher than normal level of heterogeneity throughout the system, and increase propensity to arrhythmia and contractile inefficiency. The recognition of these tolerance limits in E–C coupling, and the extent to which pathology breaches these limits of variability, may serve as a novel biomarker for disease. Understanding tolerance limits may also support novel approaches that restore normal variability by identifying the source (electrophysiological, intracellular calcium, or myofilaments) in the myocardium, and direct treatments to the primary target.

Funding

MKS received funding from the Simon Fougner Hartmann Family Fund, Rakel og Otto Kr. Bruuns legat, Joh. H. Andresens medisinske fond and the South-Eastern Norway Regional Health Auhority.

Data Availability

No original raw data were produced for this article.