1 The Cardiac Mechano-Electric Regulatory Loop

The function of the heart is to provide the driving forces needed for the circulation of blood to meet the metabolic demand of the body. This is achieved through an elegant coordination of cardiac electrical excitation and mechanical pumping action. The body’s haemodynamic state is continuously adapting, being modulated, for instance, by breathing, postural changes, exercise or circadian oscillations in blood pressure. To facilitate the matching of blood output from the heart to changes in systemic demand, an auto-regulatory system has evolved that intrinsically tethers electrical and mechanical function of the heart through feed-forward (‘excitation-contraction coupling’, ECC) [1] and feed-back (‘mechano-electric coupling’, MEC) pathways [2,3,4,5,6,7]. Together, ECC and MEC form the cardiac mechano-electric regulatory loop, which allows the heart to acutely sense and respond to physiological changes in its intrinsic and external environment and to maintain adequate pump function [1, 5]. In physiological settings, MEC can have anti-arrhythmic properties, for instance, by reducing dispersion of repolarisation across the ventricles [8, 9]. However, MEC-mediated response to pathophysiological mechanical perturbations, disease-related alterations in the active or passive mechanical state of the heart or changes in factors driving MEC can destabilise cardiac rhythm and contribute to arrhythmogenesis [7, 10,11,12,13,14,15,16,17].

This chapter considers how pathophysiologically altered MEC, combined with changes in the cardiac mechanical environment, contributes to mechanically induced alterations in ventricular electrophysiology that can lead to arrhythmias (hereafter referred to as ‘mechano-arrhythmogenesis’). It then discusses cellular and subcellular components of MEC in ventricular cardiomyocytes (CM) that may contribute to ventricular arrhythmogenesis, including mechano-sensitive ion channels (MSC), biophysical signal transmitters (i.e. microtubules, MT) and mechano-sensitive biochemical signalling pathways (i.e. changes in cytosolic free calcium concentration, [Ca2+]i, or reactive oxygen species, ROS). Finally, it describes clinical manifestations of these cellular and subcellular MEC mechanisms in a local (acute regional myocardial ischaemia) and a global (hypertension) pathophysiological setting. Of note, this chapter will not address mechanical modulation of contractility (mechano-mechanical coupling) [18] or stretch effects on sinus rhythm [19, 20], atrial arrhythmias [21] or tissue conduction [22], as the focus is on MEC in ventricular CM.

2 Clinical Evidence and Experimental Studies of Ventricular Mechano-Arrhythmogenesis

Ventricular mechano-arrhythmogenesis requires a sufficiently large and critically timed mechanical stimulus that directly or indirectly activates MSC, and that may be modulated by mechano-sensitive biophysical transmission or biochemical signals. This can be influenced by pathological changes in ventricular MEC, which alter the likelihood that a given stimulus will overcome the threshold for AP induction in CM [5]. While this chapter is focused on CM, it warrants noting that mechano-arrhythmogenesis also involves hetero-cellular interactions with electrotonically coupled cardiac non-myocytes [23] that also express MSC [24]. The contribution of hetero-cellular coupling will be enhanced upon disease-related changes in non-myocyte levels and phenotypes (e.g. myofibroblast conversion), which increase hetero-cellular interactions (for instance, through enhanced intercellular connectivity by increased fibroblast connexin expression) [25] or affect their MSC expression or function [26, 27].

2.1 Clinical Evidence

One of the most dramatic clinical examples of ventricular mechano-arrhythmogenesis is Commotio cordis, during which a non-contusional external impact, usually to the precordium, results in rhythm disturbances of variable nature, including ventricular fibrillation [28]. In fact, mechanically induced ventricular fibrillation is one of the leading causes of sudden cardiac death among young athletes and has been one of the primary reasons for which, in many countries, automated external defibrillators are now increasingly located at sporting facilities where potentially deadly impacts to the chest are most likely to occur [29, 30]. Ventricular mechano-arrhythmogenesis is also common during clinical procedures in which contact of medical instruments (e.g. cardiac catheters) with the myocardium causes local tissue deformations that may lead to ectopic excitation and arrhythmogenesis [31,32,33,34,35,36,37,38,39,40,41,42]. As with precordial mechanical stimulation [43, 44], such device-tissue contact can also revert pre-existing arrhythmias to normal sinus rhythm [5]. Arrhythmogenic changes in ventricular electrophysiology are additionally seen during acute increases in ventricular mechanical load, such as that occurring with balloon valvuloplasty [45] or during surgical manipulations [46,47,48,49].

The risk of an abnormal mechanical stimulus triggering an arrhythmia can be enhanced by cardiovascular disease. In patients with established structural heart disease, acute fluctuations in ventricular volume or pressure are more likely to result in premature excitation and sustained tachyarrhythmias [7, 17, 50]. Indeed, arrhythmia incidence in these patients is affected by blood pressure alterations, including those caused by pharmacological interventions [50], circadian oscillations [51] or day-to-day variations [52], such that acute increases in ventricular load (both preload and afterload) are associated with higher rates of arrhythmogenesis. There is also evidence suggesting that local changes in myocardial mechanics play a role in mechano-arrhythmogenesis, as wall motion abnormalities in patients with ischaemic heart disease are associated with an increased prevalence of premature excitation during acute fluctuations in load [16]. Interestingly, this effect also works in reverse: in patients suffering from chronic ventricular volume overload and tachyarrhythmias, acute ventricular unloading can result in termination of ventricular tachyarrhythmias, but alas, often only for as long as the reduction in load can be maintained [53,54,55,56,57].

2.2 Experimental Studies

Existing clinical evidence of ventricular mechano-arrhythmogenesis has been corroborated in experimental studies that have helped to elucidate potential mechanisms. For instance, in the case of Commotio cordis, a pig model of baseball impacts to the chest has confirmed historic observations [58, 59] on critical factors for the induction of ventricular fibrillation (such as pre-cordial location and impact with small and hard projectiles) while adding new information on the link between impact timing and electrophysiological outcomes: while single extra beats can be triggered in diastole, ventricular fibrillation is seen only for impacts in a narrow time-window, ~15–30 ms before the peak T wave of the ECG [60,61,62]. Isolated heart experiments [63] and computational modelling [64, 65] have further shown that ventricular excitation, induced by a focal mechanical stimulus [66], depends on the degree of myocardial tissue deformation and that it is a local phenomenon (i.e. triggered at the contact site, rather than resulting from the impact-induced intraventricular pressure surge). If a mechanically induced focal excitation overlaps with the trailing wave of the previous regular excitation, it may initiate ventricular fibrillation, highlighting the critical spatio-temporally defined nature of the vulnerable window for mechano-arrhythmogenesis [63].

In addition to elucidating mechanisms of Commotio cordis, experimental studies have demonstrated the possibility for acute changes in ventricular load to lead to premature excitation and tachyarrhythmias. Transient increases in intraventricular volume during diastole cause cellular depolarisation, which triggers excitation if supra-threshold [67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84]. Ectopic excitation has also been observed in experiments involving rapid increases in ventricular pressure (due to an acute increase in afterload) [85,86,87], which has been suggested to be a consequence of enhanced late-systolic myocardial deformation [88]. When an acute ventricular load is applied during systole, it tends to heterogeneously alter repolarisation and refractoriness, furnishing an arrhythmogenic substrate [68,69,70, 89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111]. In fact, in the isolated heart, an acute increase in intraventricular pressure has been shown to be as arrhythmogenic as the substrate created by the catecholamine-rich milieu and electrical remodeling associated with heart failure [112]. It is important to note that, while changes in intraventricular volume or pressure affect the entire ventricle (whether applied in diastole or systole), the response is generally spatially heterogeneous [113]. This is presumed to be related to variations in myocardial stiffness across the ventricle, such that stretch and the resulting electrophysiological changes are non-uniform [83, 90], with excitation originating from areas of the largest stretch (for instance, in the right ventricular outflow tract) [67, 83, 90].

3 Cellular and Subcellular Mechanisms of Ventricular Mechano-Arrhythmogenesis

Ventricular mechano-arrhythmogenesis in the whole heart arises out of a complex interplay of intra- and intercellular MEC-mediated effects. This next section is focused on the intracellular effects—the cellular and subcellular mechanisms of MEC in ventricular CM, including MSC [24], biophysical signal transmitters [114,115,116] and biochemical signals [117,118,119,120,121]—and how they shape the effects of MEC at the tissue and organ level.

In ventricular CM, stretch occurring during the resting phase of the cardiac cycle (‘electrical diastole’) depolarises the sarcolemma. If depolarisation is supra-threshold, this will result in excitation (i.e. it will trigger an action potential, AP). On the other hand, stretch occurring during the AP (‘electrical systole’) can alter the plateau or affect repolarisation dynamics. As a result, depending on stretch timing, MEC will hasten early phase 3 repolarisation, delay later repolarisation or cause early or delayed afterdepolarisation-like events [68, 69, 89, 90]. Importantly, pathophysiological changes in CM mechanics, electrical activity or factors contributing to MEC may alter this timing dependence while also enabling smaller mechanical stimuli to initiate electrical disturbances [5].

3.1 Mechano-Sensitive Ion Channels

Most acute effects of stretch on ventricular tissue- and CM-level electrophysiology (e.g. diastolic depolarisation, excitation or altered repolarisation) can be explained by MSC [24, 122]. While stretch activated channels (SAC) have been defined exclusively as channels whose activity is directly activated by a mechanical stimulus, MSC also include ion channels whose current is primarily activated by another mechanism (e.g. ligand- or voltage-activated channels), but whose activity can also be mechanically modulated [24, 122]. In fact, it has been suggested that all channels whose opening and closing mechanisms involve changes in their in-plane dimension within the lipid bilayer are MSC and may therefore contribute to MEC [123]. While the exact molecular identity of many MSC in the heart remains a topic of debate, they are generally divided into two groups based on their principal ion permeability. These groups include cation non-specific MSC (MSCNS, which cause depolarisation or altered repolarisation) and potassium (K+)-selective MSC (MSCK, which will promote repolarisation [5, 24] and conceivably cause hyperpolarisation in pathological states that are associated with depolarised resting membrane potential, such as ischaemia). There are additional MSC, conducting chloride (Cl−) ions, which are activated by changes in cell volume (for instance, with osmotic swelling, MSCswell) rather than stretch. They are not believed to contribute to acute MEC-mediated responses, due to a pronounced lag-time in their activation (~ 1 min) and the presumed conservation of cell volume during externally applied deformations of CM [24]. However, in disease states associated with changes in cell volume, such as during cell swelling (e.g. in ischaemia and reperfusion) [124] or CM hypertrophy (where Cl− channels have been shown to be constitutively active) [125], the contribution of cell volume-sensitive MSC to mechano-arrhythmogenesis may become relevant, for example, by acting as modifiers of the electrophysiological background upon which MEC operates.

3.1.1 MSCNS

With a reversal potential (EREV) between −20 and 0 mV [24], the timing of stretch stimuli (causing MSCNS activation) relative to the AP [122] determines whether MSCNS pass an ‘inward’ depolarising current (at membrane potentials < EREV of the MSC) or an ‘outward’ re-/hyperpolarising current (at membrane potentials > EREV) [70, 126, 127]. Accordingly, MSCNS can depolarise resting CM. Even though experimental and computational evidence demonstrates stretch-induced depolarisation and AP-induction [67, 71,72,73,74, 122], which can be prevented by pharmacological block of MSCNS [69, 73, 75, 76, 92, 128], no single channel patch clamp recordings of MSCNS have been reported in adult ventricular CM. This is possibly due to localisation of MSC in membrane folds (including caveolae and transverse tubules) or at the Z-disc, locations where their direct measurement by the patch clamp method is difficult [24]. It may also relate to a dependence of MSC activation on cytoskeletal elements, the required deformation of which is not engaged in the available experimental systems (i.e. by negative patch pressure) [129]. While this has hindered identification of MSCNS in adult ventricular CM, two groups of ion channels are thought to make important contributions: Piezo and Transient Receptor Potential (TRP) channels [5, 24].

The discovery of Piezo channels generated a great deal of excitement as a potential key MSCNS in CM [130], because Piezo channel kinetics match macroscopic functional observations [24, 131]. However, while it appears that Piezo channels are expressed in macrophages [132] and cardiac fibroblasts [26, 133], only small amounts of Piezo mRNA have been found in cardiac tissue (from mice) [130], and (at the time of writing) no functional Piezo channels have been shown in ventricular CM from any species. Therefore, there is no direct evidence at present for a role of Piezo in ventricular mechano-arrhythmogenesis.

TRP channels are a ubiquitously expressed group of ion channels comprised of six subfamilies, many of which are found in the heart [27]. These channels have garnered interest regarding their role in the pathological progression of hypertrophy, heart failure and ischaemia, as well as in arrhythmogenesis [134, 135]. As TRP channels can pass inward or outward currents at physiological membrane potentials, and some have been shown to be inherently mechano-sensitive [136] (although this is not uncontested [137]), they may yet be found to play an important role in ventricular mechano-arrhythmogenesis.

3.1.2 MSCK

The identity of several MSCK in the heart is well established. TREK-1 and 2 are outwardly rectifying K+ channels [24]. Being K+-selective, they pass repolarising currents at all membrane potentials that are naturally experienced by CM. Accordingly, they do not trigger stretch-induced excitation, and their relative contribution to global cardiac MEC phenomena in physiological conditions is generally believed to be low compared to MSCNS [5]. However, in some diseases, additional mechano-sensitive K+ currents become available for mechanical activation, making contributions of MSCK more relevant as the ‘net reversal potential’ of mechanically induced whole cell current will depend on the relative contributions of MSCNS and MSCK. In ischaemia, for example, mechano- and adenosine triphosphate (ATP)-sensitive K+ channels (KATP) are pre-activated by the reduction in ATP (and increase in ADP) and thus alter MEC-mediated effects on cardiac electrophysiology (considered further in Sect. 4.1) [5, 138,139,140].

3.2 Biophysical Signal Transmitters

Biophysical signal transmitters, such as MT, relay mechanical cues across the cell, affecting mechano-sensitive cellular components, including MSC and elements responsible for the release (e.g. Ca2+ from the sarcoplasmic reticulum, SR, via ryanodine receptors, RyR [118], and from mitochondria [141,142,143,144,145]) or production (e.g. ROS by NADPH oxidase 2, NOX2 [120, 121]) of biochemical signals, which elicit and/or modulate electrophysiological responses [114, 116]. In ventricular CM, MT are particularly important for mechanical transmission [114,115,116, 118, 120, 146, 147] (although sarcomeric proteins and other cytoskeletal elements not discussed here, such as titin, focal adhesion proteins or integrins, will also be involved [148, 149]).

MT form physical links at the Z-disc through interactions with membrane-associated and intermediate proteins (e.g. desmin) [114]. These links create a rigid scaffold, conferring structural integrity to ventricular CM and facilitating their re-lengthening after contraction [116], as illustrated by transient buckling of MT during cell shortening [150]. MT mechano-transmission is enabled by their load-bearing ability, which is enhanced by their lateral reinforcement [151]. MT load-bearing involves interactions between the intermediate protein desmin and detyrosinated MT [150, 152]. Detyrosination is a post-translational modification that confers MT stability, simultaneously preventing MT degradation and promoting the formation of tight junctions with desmin, resulting in a shift from low-energy sliding to energy-costly buckling of MT during cell contraction [150]. Other, less explored post-translational modifications, such as acetylation, also increase the load-bearing capabilities of MT [146, 153]. Acetylation confers resistance to MT breakage during repetitive mechanical stimulation, resulting in more long-lasting (‘aged’) MT, which are then available for further post-translational modification [147, 154]. Thus, a denser, more detyrosinated (physically anchored) and/or acetylated (aged) MT network would be expected to enhance mechano-transmission in CM. Indeed, increased detyrosination and acetylation have been shown to enhance mechano-dependent biochemical signalling, such as an increase in the stretch-induced release of Ca2+ from the SR (i.e. Ca2+ sparks) and NOX2-dependent ROS production [115, 155]. Thus, MT network properties (including post-translational modifications) warrant exploration for their role in mechano-arrhythmogenesis (explored in Sect. 4.2).

3.3 Mechano-Sensitive Biochemical Signals

Mechano-sensitive biochemical signals are by-products of intracellular mechano-transmission and modulate the electrophysiological response of CM to a mechanical stimulus [117, 119, 121]. Principal mediators known to be important for ventricular MEC include mechano-sensitive changes in [Ca2+]i, which is determined by (1) trans-sarcolemmal Ca2+ in-/efflux [156, 157]; (2) Ca2+ release from/re-uptake into the SR [118, 158, 159]; (3) cytosolic Ca2+ buffering (such as by dissociation from/binding to troponin C, TnC [160, 161], or release from/re-uptake into mitochondria [142,143,144,145]); and (4) mechano-sensitive changes in ROS [120, 121] (although other important factors exist [119]).

These biochemical signals modulate the activity of ion channels, including TRP ankyrin-1 channels (TRPA1) [162, 163] and K+-selective Ca2+-dependent channels of big conductance (BKCa) [26, 164] or other trans-sarcolemmal ion flux pathways, such as the sodium/calcium (Na+/Ca2+) exchanger (NCX) [156, 157]. At the same time, biochemical signals also affect post-translational modification of MT (e.g. acetylation) [165], which will then affect mechano-transmission (see above) [155]. In this way, whether through interactions with MSC, other trans-sarcolemmal ion flux pathways or biophysical signal transmitters, biochemical signals can modulate ventricular MEC.

3.3.1 Mechano-Sensitive Ca2+ Handling

In ventricular CM, the total amount of intracellular Ca2+ is affected directly by mechano-sensitive trans-sarcolemmal Ca2+ flux or secondarily by the effects of trans-sarcolemmal Na+ flux on NCX activity [156, 157]. ‘Free’ cytosolic [Ca2+]i is additionally affected by the mechano-sensitivity of intracellular Ca2+ handling. For instance, increases in [Ca2+]i occur with mechanically induced Ca2+ release from mitochondrial stores (which is independent of sarcolemmal MSC or NCX-mediated Ca2+ influx) [142, 143, 145], through a process that is dependent on MT integrity. As a result, MT disruption leads to mitochondrial disorganisation, irregular Ca2+ propagation and arrhythmogenic Ca2+ waves (which may also involve Ca2+ release from the SR) [144]. During contraction, [Ca2+]i is affected by the mechanical modulation of the affinity of TnC for Ca2+, which is increased during stretch (such that more Ca2+ binds to myofilaments) [160]. Upon release of stretch, the rapid dissociation of Ca2+ from TnC can cause an arrhythmogenic surge of [Ca2+]i [161]. In diastole, stretch causes an acute increase in Ca2+ spark rate through a MT-dependent mechanism (reduced by MT disruption [118, 158]), suggesting that RyR themselves are in fact (non-sarcolemmal) MSC. This stretch-induced increase in Ca2+ spark rate would be diminished upon cell shortening during contraction and hence result in a negative-feedback mechanism that aids initiation and termination of ECC [166].

In physiological conditions, mechano-sensitive Ca2+ handling processes interact, presumably synergistically, and control [Ca2+]i. However, pathological alterations in these processes, such as in RyR open probability (for instance, due to oxidation or nitrosylation) [120, 167,168,169] or in myofilament Ca2+ binding affinity [170], may disturb the control of [Ca2+]i, resulting in Ca2+-mediated arrhythmogenesis. This is particularly relevant for diseases associated with altered mechanics [171] or Ca2+ overload, such as acute regional ischaemia [156, 172] and hypertension (considered in Sect. 4) [115, 155].

3.3.2 Mechano-Sensitive ROS Production

The stretch-induced increase in Ca2+ spark rate is modulated by another mechano-sensitive biochemical signal: ROS [120]. In the context of ventricular MEC, ROS (and any ROS-mediated Ca2+ release) is dependent on MT, as increasing MT stability enhances mechanically induced ROS production [115]. This makes it difficult to distinguish between direct and indirect roles of the MT in mechanical modulation of RyR-mediated Ca2+ release [118, 120].

Mechanically induced ROS production is graded by stretch magnitude and is more responsive to cyclic than static stretch, which is in keeping with a role during regular cardiac activity [173]. This contributes to the intracellular tuning of Ca2+ signalling in response to stretch [120] and to complex intercellular adjustments of contractility (mechano-mechanical coupling [18]) that has been observed in paired muscle (“duplex”) studies [174]. This duplex research showed that stretch effects on cellular Ca2+-balance enable CM to adjust their contractility to changes in external demand. It is conceivable that (at least some of the) MEC-mediated responses are a ‘side-effect’ of this autoregulation of CM mechanical performance. In any case, physiological (e.g. transmural) or pathophysiologically exacerbated heterogeneity in ventricular electro-mechanics (e.g. during ischaemia) may promote arrhythmogenesis via effects on Ca2+-dynamics [175]. For example, a pathological (including mechanically induced) increase in ROS production could contribute to mechano-arrhythmogenesis by sensitising MSC [162, 163] and promoting RyR Ca2+ leak [120] while also stabilising MT (through an increase in acetylation) [165]. The latter may be part of a compensatory response (as stiffer CM will be stretched less), but this would come at a cost (requiring higher forces for contraction).

4 Ventricular Mechano-Arrhythmogenesis in Cardiac Disease

One critical determinant of mechano-arrhythmogenesis, as highlighted in Sect. 2, is the spatial nature of a mechanical disturbance. To illustrate this point, we consider two cardiac disease states with spatially divergent alterations in electro-mechanics and different effects on cellular MEC: (1) acute regional myocardial ischaemia, which is focal in nature and involves acute metabolic, electrophysiological and ionic changes that directly alter MEC, and (2) chronic hypertension, which is global in nature and involves chronic cellular remodelling, with secondary effects on MEC.

4.1 Acute Regional Myocardial Ischaemia

Acute regional myocardial ischaemia is a mismatch between the supply and demand of blood in a localised area of the heart, for example, due to occlusion of a coronary artery. Ischaemia is characterised by three hallmark pathophysiological changes to the cellular milieu that drive pro-arrhythmic electrophysiological changes and MEC-mediated responses: (1) reduced interstitial oxygen content (hypoxia); (2) increased extracellular K+ concentration (hyperkalemia); and (3) decreased intracellular pH (acidosis). This ischaemic milieu (and reduced or lack of blood flow) causes a cardiometabolic shift, which leads to metabolite accumulation and increased osmotic pressure (Fig. 1) [176]. During progression of acute regional myocardial ischaemia, arrhythmias occur in a bi-modal fashion in periods termed ‘phase 1a’ (up to ~10 min following artery occlusion) and ‘1b’ (~15–60 min after occlusion).

Fig. 1
A flow diagram of the mechanisms of ventricular mechano-arrhythmogenesis has 3 parts. 1. metabolic, electrophysiological, and ionic changes. 2. changes in myocardial mechanics. 3. changes in intracellular C a 2 plus handling.

Cellular and subcellular mechanisms of ventricular mechano-arrhythmogenesis during acute regional myocardial ischaemia. Following coronary artery occlusion, key metabolic, electrophysiological and ionic changes include hypoxia, acidosis and hyperkalemia (black borders, top panel). These changes lead to altered myocardial mechanics through (1) decreased contractility of ischaemic tissue, resulting in stretch of the ischaemic region in phase 1a and of the ischaemic border in phase 1b, and (2) metabolite accumulation, resulting in osmotic swelling (blue borders, middle panel). Stretch activates cation non-specific and potassium (K+)-selective mechano-sensitive ion channels (MSCNS and MSCK, including ATP-sensitive K+ channels, KATP) and alters mechano-sensitive calcium (Ca2+) handling processes, leading to an increase in free cytosolic Ca2+ concentration ([Ca2+]i). At the same time, osmotic swelling activates cell volume-sensitive ion channels (MSCswell). Combined, these mechano-sensitive elements contribute to arrhythmogenic triggering (red borders or shading) and/or sustaining (green borders or shading) effects. Key changes in intracellular Ca2+ handling are summarised in the bottom panel. In phase 1a, ATP levels are relatively preserved, so despite increased Ca2+ flux into the cytosol, [Ca2+]i is maintained by normal ATP-dependent extracellular Ca2+ extrusion and intra-organelle Ca2+ uptake. In phase 1b, ATP levels decrease, so cytosolic Ca2+ efflux can no longer balance influx, and there is an increase in [Ca2+]i, which can also trigger and/or sustain arrhythmias. ADP, adenosine diphosphate; ICa,L, L-type Ca2+ current; INa, fast sodium current; [K+]o, extracellular potassium concentration; LTCC, L-type Ca2+ channel; mNCX, mitochondrial NCX; NCX, sodium-Ca2+ exchanger; NHE, sodium-hydrogen exchanger; O2, oxygen; RyR, ryanodine receptors; SERCA, sarcoendoplasmic reticulum ATPase; SR, sarcoplasmic reticulum; TnC, troponin C

Full size image

The arrhythmias occurring in phase 1a appear to be re-entrant in nature and relate to cellular hyper-excitability. This hyper-excitability is driven by diastolic membrane depolarisation secondary to hyperkalemia, combined with stretch of ischaemic tissue (clinically evident as paradoxical segment lengthening) [177], which activates MSCNS, driving further depolarisation and contributing to premature excitation [176, 178]. In phase 1b, excitability is reduced below normal levels in the central ischemic zone [176, 178], which at the same time begins to stiffen and, thus, resist stretch [179,180,181,182,183]. The particularly high incidence of arrhythmias in this phase has been suggested to instead involve systolic stretch of myocardium at the border between the stiffened ischaemic core and healthy contractile tissue, causing MSCNS and KATP channel activation [179,180,181,182,183]. In low-flow ischaemia, activation of volume-sensitive MSC through cell swelling caused by increased osmotic pressure may also contribute to electrophysiological changes that promote arrhythmogenesis in both phases (Fig. 1) [124, 125].

4.1.1 Metabolic, Electrophysiological and Ionic Changes

4.1.1.1 Hypoxia

Following coronary artery occlusion, reduced blood flow results in tissue hypoxia, causing an increase in anaerobic glycolysis that is concomitant with reduced mitochondrial pyruvate oxidation [26, 184, 185]. This metabolic change leads to an increase in lactate production, metabolite accumulation and a gradual decrease in ATP availability (following exhaustion of cytosolic phosphocreatine reserves that are engaged to initially buffer this change) [186,187,188,189] with a simultaneous increase of adenosine diphosphate (ADP). The resulting decrease in the ATP:ADP ratio removes inhibition of mechano-sensitive KATP channels (Fig. 1) [176]. This has two effects: depolarisation of CM resting membrane potential as a consequence of hyperkalemia and APD shortening due to an increase in outward repolarising K+ currents. Osmotic swelling, secondary to metabolite accumulation, may also activate volume-sensitive MSC, such as Cl− channels [124, 125], leading to further resting membrane potential depolarisation and APD shortening. The APD shortening is more pronounced than the reduction in Ca2+-transient duration that also occurs during ischaemia [178], resulting in a period in late repolarisation during which [Ca2+]i is still high when CM have repolarised. This can facilitate Ca2+-induced re-excitation [13, 190,191,192]. The degree of APD shortening will be spatially heterogeneous, due to differences in expression [193] and stretch-induced activation of KATP across the heart [138,139,140], resulting from regionally differing ventricular mechanics (discussed in more detail below).

4.1.1.2 Extracellular K+ Accumulation

The combination of KATP activation, reduced wash-out of the extracellular space and decreased activity of the Na+/K+-ATPase results in hyperkalemia (Fig. 1) [176]. This increase in extracellular K+ causes a shift in EREV for K+ toward less negative values, resulting in depolarisation of the resting membrane potential of CM. Hyperkalemia also occurs in a bimodal fashion, as in phase 1a, the initial preservation of Na+/K+-ATPase activity partially counteracts the KATP-mediated K+ efflux. During this period, CM become hyper-excitable, as their slight depolarisation brings the resting membrane potential closer to the threshold for activation of the fast Na+ channels that underlie AP initiation.

Progressive ATP reduction in phase 1b, however, interferes with Na+/K+-ATPase activity, resulting in a greater, secondary rise in extracellular K+. This leads to further membrane depolarisation, which partially inactivates fast Na+ channels, thereby reducing cell excitability and prolonging the effective refractory period (‘post-repolarisation refractoriness’) [178, 194]. In the regionally ischaemic whole heart, the level of extracellular K+ accumulation differs between the ischaemic core and healthy tissue, due to gradients in extracellular wash-out and in oxygen availability, but in part also due to heterogeneous stretch-induced activation of KATP in the ischaemic tissue [177]. This gradient results in injury currents that flow from the depolarised ischaemic tissue to the (still electrotonically coupled) healthy tissue, reducing the electrical sink in this region (by shortening the gap between the resting and threshold potential of the electrically coupled healthy myocardium) and contributing to hyper-excitability of tissue at the ischaemic border zone [195, 196].

4.1.1.3 Intracellular Acidosis

The hypoxia-induced metabolic shift to anaerobic glycolysis, combined with ongoing fatty acid beta-oxidation, enhances proton accumulation and leads to intracellular acidosis. As a result, an initial compensatory efflux of hydrogen through the Na+/H+-ATPase (while ATP levels are still sufficient) causes an increase in intracellular Na+, which reduces Ca2+-extrusion by NCX (or even causes it to operate in reverse mode), leading to intracellular Ca2+ accumulation. At the same time, acidosis causes inhibition of fast Na+ and L-type Ca2+ channels [178, 195], along with closure of gap junctions (Fig. 1) [197].

4.1.1.4 Electrophysiological Changes

At the cellular level, the effects of hypoxia, hyperkalemia and acidosis (as well as osmotic swelling-induced increases in Cl− currents [124, 125]) manifest as arrhythmogenic changes to the AP: (1) shortened plateau and more rapid repolarisation (resulting in decreased APD and AP triangulation); (2) less negative resting membrane potential; and (3) reduced AP amplitude and upstroke rate [176, 178]. In the regionally ischaemic heart, changes in APD are spatially heterogeneous, with an additional increase in beat-to-beat variability of repolarisation within the peri-infarct region compared to remote, healthy myocardium [198]. This results in dispersion of cell excitability, refractoriness and repolarisation timing, which increases the vulnerability to re-entrant arrhythmias [178, 195, 198].

This arrhythmic risk is further enhanced by increasing the ‘excitable gap’, i.e. the distance between activation wave-front and wave-end. This occurs though a reduction in APD, and it is enhanced by slowed conduction (due to reduced intercellular coupling and decreased fast Na+ channel availability). Ventricular conduction may be further slowed by stretch effects, which result in an increase in membrane capacitance [199, 200], caused by the unfolding of membrane invaginations, T-tubule deformation and the incorporation of caveolae into the surface and T-tubule membranes [201, 202].

4.1.1.5 Alterations in Ca2+ Handling

In ischaemia, there is a net increase in cytosolic [Ca2+]i driven by several mechanisms, including (1) reduced forward-mode NCX activity secondary to the acidosis-induced increase in intracellular Na+ via the Na+-H+ exchanger or stretch-induced (see below) Na+ entry (and in extreme cases, reverse-mode NCX activity causing Ca2+ influx); (2) decreased Ca2+ (re-)uptake by the sarco-endoplasmic reticulum ATPase due to reduced ATP levels; (3) an increase in Ca2+ leak from the SR, driven by an increase in the open probability of RyR [194]; and (4) stretch effects on Ca2+ handling (considered in Sect. 3.3.1). The increase in RyR opening is a result of several effects [176], including Ca2+-induced opening [203], stretch [118, 158] and increased ROS [120]. This overall net gain in intracellular Ca2+leads to an increase in SR Ca2+ load. As ischaemia progresses, any compensatory effect of this Ca2+ sequestration into the SR is exhausted, further promoting Ca2+ leak through RyR and increased [Ca2+]i (Fig. 1).

4.1.2 Stretch of Ischaemic Myocardium

Altered metabolic, ion channel and Ca2+ handling activity in CM during acute regional ischaemia affect myocardial contractility, resulting in regions of diastolic and systolic tissue stretch (Fig. 1).

In phase 1a, there is general stretch of weakened tissue in the central ischaemic region, with additional stretch during mechanical systole. In phase 1b, however, the central ischaemic tissue stiffens, and its stretch is reduced, resulting in stretch of weakened tissue at its border by contraction of adjacent healthy tissue [179,180,181,182,183]. The mechano-arrhythmogenic relevance of ischaemic myocardial stretch is supported by the correlation between wall motion abnormalities and the incidence of ventricular fibrillation seen in patients with coronary artery disease [16]. Experimentally, it has been shown that the onset of tissue stretch and ventricular fibrillation in phase 1a of acute regional ischemia is related [204] and that the degree of tissue stretch is a strong predictor of ventricular fibrillation occurrence [10,11,12, 177].

In phase 1b, arrhythmia incidence is ventricular load-dependent [13, 14, 80], and aberrant excitation tends to originate at the ischaemic border, where experimental evidence [13, 14] and computational modelling [205] suggest an involvement of stretch-induced depolarisation mediated by MSCNS. Mechanically induced arrhythmias arising from the ischaemic border will be facilitated by heterogeneous slowing of conduction, due to more pronounced gap junction uncoupling within the ischaemic core than at the ischaemic border [197].

Changes in mechano-sensitive biochemical signalling during ischaemia further facilitate mechano-arrhythmogenesis. For instance, the stretch-induced increase in Ca2+ spark rate [118] and ROS production [120] that occurs in healthy tissue is enhanced by ischaemia [172], a response potentially caused by an ischaemia-induced increase in SR Ca2+ load [194] or RyR sensitivity [120, 168] or a deficit in the antioxidant capacity of CM [206] (for instance, due to a reduction in the key antioxidant glutathione [207]). In fact, enhancement of mechano-sensitive biochemical signalling in regions of stretched myocardium has been shown to cause focal Ca2+ waves [208], which, if occurring at the ischaemic border where the injury current tends to reduce the excitation threshold [195], may contribute an additional trigger or serve as a substrate for sustained mechano-arrhythmogenesis.

4.1.3 Tissue-Level Considerations for Mechano-Arrhythmogenesis in Acute Regional Ischaemia

Mechano-arrhythmogenesis in acute regional ischaemia involves the localised interaction of ischaemia- and stretch-induced electrophysiological effects on CM, generating regionally heterogeneous changes in tissue excitability, refractoriness and electrical conduction. As the acute outcome of myocardial stretch is dependent on its magnitude and timing relative to the background electrical and mechanical activity of CM, ischaemic effects that alter CM electrophysiology—for example, shortening of APD—increase the likelihood that stretch will occur during a period of cellular excitability and, thus, provoke an arrhythmia.

In addition to the MEC effects driven by localised tissue stretch during ischaemia described above, there is also evidence for an increase in MEC itself. This includes an increase in mechano-sensitive biochemical signals (Ca2+ sparks, ROS) [172], increased RyR sensitivity (and ensuing Ca2+ release) [206] and pre-activation of mechano-sensitive KATP channels [138,139,140]. Heterogeneous stretch may therefore lead to regionally differing changes that furnish arrhythmogenic triggers and enhance the substrate for arrhythmia sustenance [162, 163, 209]. The heterogeneous expression of mechano-sensitive [139, 140] KATP channels [193] (or other MSCK [210]) across the heart will additionally drive dispersion of repolarisation or cause conduction block [211], favouring re-entrant electrical activity. Finally, mechano-arrhythmogenesis in ischaemia may involve heterogeneous changes in the relative dynamics of membrane voltage and Ca2+ transients, which facilitate Ca2+-mediated arrhythmias during late repolarisation [13, 190,191,192].

4.2 Chronic Hypertension

Chronic hypertension, defined as a sustained increase in systolic (≥130 mmHg) or diastolic (≥80 mmHg) blood pressure [212], results in increased ventricular afterload (i.e. the load against which ventricular CM must contract) [213]. This increase in afterload results in an increase in systolic intraventricular pressure (which maintains ejection), but may also lead to an increase in ventricular preload (i.e. intraventricular volume) if ejection is reduced. The elevated mechanical load experienced by ventricular myocardium in hypertension results in the stimulation of compensatory CM remodelling (although sustained overload ultimately results in decompensated heart failure) [214, 215], which may enhance MEC and contribute to increased mechano-arrhythmogenesis.

Enhanced MEC, secondary to structural and functional remodelling of CM in hypertension, is thought to contribute to the increased incidence of ectopic ventricular excitation in hypertensive patients [17] that occurs during acute fluctuations in ventricular load, including circadian [51] or pharmacologic [50] modulation of blood pressure. This increase in arrhythmogenic triggers may interact with remodelled myocardium, which acts as an arrhythmia-sustaining substrate, contributing to sustained tachyarrhythmias and/or sudden cardiac death [112, 216,217,218,219,220,221]. Further, mechano-arrhythmogenesis may be enhanced when hypertensive hypertrophic remodelling is complicated by diffusely distributed ischaemic regions (due to wall thickening and microvasculature remodelling [222]) involving mechanisms described in the previous section.

4.2.1 Structural Remodelling

Tissue-level remodelling in hypertension is characterised by concentric thickening of the ventricular wall (hypertrophy), which counteracts the increase in systolic wall stress caused by an increase in systolic intraventricular pressure with elevated afterload. The benefit of this change in chamber geometry is explained by Laplace’s law: (P ~ [T × m]/R). An increased intra-ventricular peak pressure (P) can result either from an increase in CM force production (thus raising tissue tension, T)—which, within normal physiological limits, will be afforded by length-dependent activation of force generation (Frank-Starling law of the heart [223])—or from an increase in ventricular wall thickness (m, hopefully in the absence of an increase in chamber radius, R, which would worsen the situation). Increases in sarcomeric force production are limited in scope, so in pathological settings the increase in ventricular wall thickness accounts for the necessary increase in intra-ventricular pressure (P). This is the result of the parallel addition of sarcomeres in CM, thereby increasing cell diameter, as opposed to their addition in series—which would increase cell length and increase chamber radius (although with prolonged hypertension this may also occur)—or the addition of new cells (CM division is exceedingly uncommon in the adult mammalian heart [224]). The degree of concentric hypertrophy thus scales with the increase in pressure, resulting in a normalisation of wall stress, along with an increase in the ratio of wall thickness to inner chamber diameter [214, 215].

4.2.1.1 Microtubule Network

In addition to ventricular wall thickening, chronic hypertension is associated with remodelling of the cytoskeleton, characterised by changes in the density (via polymerisation) and stability (via post-translational modifications, altered intermediate protein linkages and changes in expression of MT-associated proteins) of the MT network [225, 226]. These changes to the MT network may work to reduce CM stretch, which would occur if ventricular preload is increased.

While clinical observations [227, 228] and experimental models [229, 230] have demonstrated that an increase in MT density occurs in hypertension, it has remained unclear whether MT proliferation begins during compensatory hypertrophy or is a consequence of the transition to decompensation (discrepancies in published reports may relate to inconsistencies between hypertensive models used, both in terms of species and method to induce hypertension) [231]. Regardless of the moment of onset, the MT network has been consistently shown to become denser and more stable in hypertension [227,228,229,230, 232, 233]. MT network remodelling appears to contribute to altered CM shortening and relaxation: hyper-polymerisation of MT in healthy cells decreases contraction and relaxation, while MT de-polymerisation in failing CM results in an improvement [116, 234, 235]. The relative contribution of an increase in the rate of MT polymerisation (which would present as an overall increase in tubulin content) versus an enhanced stability of existing MT (which would manifest as post-translationally modified ‘aged’ MT) to these effects remains to be clarified [114, 116, 153, 228].

4.2.1.2 Microtubule Post-Translational Modifications

During contraction, MT buckle at wavelengths corresponding to the distance between adjacent sarcomere units [150]. This process, which also enhances viscoelastic resistance, is dependent on the level of detyrosination, a post-translational modification of α-tubulin that involves the removal of the C-terminal tyrosine, which exposes a glutamate at the newly formed C-terminus [146, 236]. This suggests that the load-bearing (and, by extension, contribution to mechano-transmission) and re-lengthening (which is resisted by viscous forces) [116] of MT are affected by detyrosination, rather than MT density alone. Therefore, pathological increases in the level of MT detyrosination in hypertension may have profound effects on chamber relaxation (which would affect passive ventricular filling during early diastole) [237, 238], ventricular MEC and, as a result, mechano-arrhythmogenesis.

MT detyrosination confers stability to the MT network by preventing the breakdown of existing MT and by facilitating cross-linking with intermediate proteins [114, 239, 240] (e.g. desmin [150, 152]). This results in a shift from low resistance MT sliding to high-energy buckling during contraction, which modulates the effect of MT on mechano-transmission and segment re-lengthening (by increasing viscoelastic resistance) [116, 150, 228]. In human [228] and murine CM [115], inhibition of detyrosination increases the wavelength and disrupts the organisation of MT buckling (suggesting an attenuation of their resistance to compression and load-bearing capability), which is associated with an increase in the velocity of cellular contraction and relaxation, presumably due to reduced viscoelastic resistance. Conversely, enhancing detyrosination (without a change in MT density) increases both cell stiffness and viscosity, with an associated decrease in contraction and relaxation velocity [115, 150]. This suggests that in ventricular CM subjected to pressure overload, enhanced detyrosination, rather than enhanced MT density, accounts for increased stiffness and impaired contraction and relaxation kinetics. Indeed, in addition to the overall increase in MT detyrosination in failing human CM, there is upregulation of a gene encoding a pre-detyrosinated tubulin [228], as well as of the detyrosination promoting MT-associated protein 4 (MAP4) [241]. As a consequence, suppression of detyrosination in CM from pressure overload patients improves contractile and relaxation function, and the degree of such improvement scales with initial CM stiffness [150, 228].

MT detyrosination is promoted by interactions with MT-associated and intermediate proteins, such as MAP4 [228, 241, 242] and desmin [150, 152]. In pressure overload, there is upregulation of MAP4 (which occurs early in hypertrophy and stabilises MT by preventing their degradation [241]) and desmin, both of which contribute to the pathologic increase in detyrosination levels. Desmin is known to bind specifically to detyrosinated tubulin and facilitate its buckling behaviour [150, 152]. The increase in desmin expression in pressure overload, however, is partially comprised of an isoform prone to misfolding and aggregation [228], which probably explains observations of MT and desmin misalignment [114, 228, 243]. Such MT disorganisation would contribute to decompensated heart failure, including detrimental effects on trafficking of sarcomere precursors, thereby limiting compensatory hypertrophy [114, 244,245,246].

4.2.2 Proposed Mechanisms of Mechano-Arrhythmogenesis in Hypertension

One function of the MT network is the transmission of mechanical signals across CM [116], which is enhanced by increases in its density or stability (e.g. through detyrosination or acetylation) [114]. Pathophysiological alterations in the MT network may be arrhythmogenic if they modulate MSC activity (which has been demonstrated experimentally by modulation of detyrosination [115] and acetylation [155]), by changing the degree of mechanical stimulation MSC experience (either by reducing the dampening effect of the MT or by enhancing mechano-transmission) [24, 129, 247] or by sensitising MSC [134, 162, 163] (through an increase in MT-dependent biochemical signal production [115, 155]).

In hypertension, remodelling of the CM cytoskeleton results in a laterally reinforced MT network (via interactions between detyrosinated tubulin, desmin and MAP4) [114]. Increased lateral reinforcement facilitates the load-bearing and, thus, biophysical signal transmission capability of the MT network [151]. The remodelled cytoskeleton will thus increase MEC, through an increase in mechano-sensitive biochemical signal production with increased levels of detyrosination [115] and contribute to the constitutive activation of volume-sensitive Cl− channels following transition to failure [248]. An increase in MEC would be expected to reduce the magnitude of CM stretch needed to cause excitation and trigger an arrhythmia. A role for MT in increased mechano-arrhythmogenesis has been corroborated in several experimental models, including one in which an acute pharmacologically induced rise in MT proliferation and detyrosination in the whole heart (with paclitaxel) was associated with an increased prevalence of acute volume pulse-induced mechano-arrhythmogenesis [81]. However, it should be noted that in this study, potential confounding effects of increased peak intraventricular pressure, associated with volume injections into stiffened ventricles, cannot be excluded (relevant data was not reported).

Enhanced mechanically induced ROS production in hypertension, promoted by elevated levels of detyrosination [115], would be further increased by upregulation of NOX2 [249]. This elevated level of ROS may promote an arrhythmogenic intracellular milieu (i.e. elevated [Ca2+]i levels) through, for example, increased Ca2+ leak from the SR [120, 169]. ROS and Ca2+ also modulate the activity of MSC, [162, 163] whose intracellular trafficking and stretch activation may be affected by alterations of the MT network in hypertension [24]. As such, pathological remodelling of biophysical signal transmitters and the resultant effect on ventricular MEC may contribute to formation of arrhythmic triggers and a substrate for mechano-arrhythmogenesis in hypertension.

5 Conclusion

The heart is an electrically controlled mechanical pump with intricate feedback mechanisms that dictate the heart’s response to acute changes in its mechanical environment. At the cellular level, these MEC effects involve mechano-sensitive components, including MSC [24], biophysical signal transmitters (e.g. the MT network [114]) and mechano-sensitive biochemical signals [117, 119, 121] (e.g. intracellular ROS and Ca2+ [118, 120]). In this chapter, we have discussed how these cellular and subcellular components of MEC in ventricular CM contribute to tissue-level ventricular mechano-arrhythmogenesis.

The potential for a mechanical change to elicit an electrical response depends on the timing of the associated mechanical stimulation relative to the AP and on its magnitude and rate of rise [5]. Therefore, disease states in which regional (e.g. ischaemia [172]) or global (e.g. hypertension [114, 236]) mechanical alterations occur and which are associated with changes in AP dynamics (altering the relative duration of systolic and diastolic periods [13, 134]) or changes in MEC (altering the threshold for stretch-induced excitation [7, 15, 17, 112]) may make the heart more prone to mechano-arrhythmogenesis. In these cases, mechanically induced excitation constitutes an arrhythmogenic trigger, which interacts with a disease-mediated pro-arrhythmic substrate and converts into sustained arrhythmic activity [17, 134].

The nature of metabolic, electrophysiological and ionic changes, as well as MT network remodelling affecting MEC, will vary across cardiac diseases. When considering the relative contribution of cellular and subcellular mechanisms to ventricular mechano-arrhythmogenesis identified in this chapter, it is essential to assess the precipitating factors of the disease being studied, as well their structural and functional manifestations. Comparison of diseases with differences in mechanical loading (e.g. ventricular volume versus pressure overload) or structure-function remodelling (heart failure with preserved versus reduced ejection fraction) will provide critical insight into how disease-induced alterations in cellular MEC drive ventricular mechano-arrhythmogenesis.

For instance, while pressure overload is characterised by densification and increased detyrosination of the MT network, resulting in an increase in MEC, in volume overload there is a reduction in detyrosination-desmin interactions and a disruption of cytoskeletal integrity [243], such that the influence of the MT network on MSC and biochemical signals may be attenuated. Effects of volume overload on MEC may in fact more closely resemble those seen in acute ischaemia (tissue stretch- [177] and osmotic CM swelling-induced [124, 125] MSC activation), but over the entire ventricle (globally), rather than regionally varying (local). In both cases it would be attractive to determine whether a stretch-induced increase in biochemical signalling (i.e. Ca2+ sparks and ROS production) involves alterations in post-translational modifications of MT and whether this increases MSC sensitivity.

In failing human CM, both contraction and relaxation are improved by suppression of MT detyrosination [228]. CM from patients with heart failure with preserved ejection fraction, however, show a greater improvement than those with reduced ejection fraction, suggesting there is a greater contribution of detyrosination to the reduced mechanical function in those cells. Thus, it would be informative to compare MEC and mechano-arrhythmogenesis between those two forms of heart failure.

In summary, this chapter has discussed cellular and subcellular mechanisms contributing to mechano-arrhythmogenesis in ventricular CM, with a focus on two disease states that highlighted local (e.g. acute regional ischaemia) and global (e.g. hypertension) pathological manifestation relevant for MEC. The intricate interaction between the various components of ventricular MEC contribute to formation of both triggers and substrate for arrhythmogenesis. Understanding the precise interactions of these components is necessary to facilitate the development of a mechanistic framework for understanding how therapies targeting MEC may contribute to the prevention of ventricular mechano-arrhythmogenesis.