Abnormalities of calcium metabolism and myocardial contractility depression in the failing heart

HF is characterized by chronic activation of neuro-hormonal pathways representing a compensatory response to overcome depressed cardiac function. However, over weeks to months, a chronic hyperadrenergic state ensues with elevated plasma catecholamine levels, which contribute to maladaptive cardiac chamber remodeling, progressive deterioration of pump function, and deadly arrhythmias. Due to the chronic hyperadrenergic state, desensitization of β-adrenoreceptors (β-ARs) and reduced global intracellular cAMP synthesis occur in the failing heart [1‐3]. However, stimulation of β-ARs and other signaling pathways is maintained and continues to affect the remodeled cells and proteins.

A broad range of molecular pathways are involved in the development of HF, and there is likely to be substantial overlap between these pathways. Typically, cell-surface receptors are activated by ligands leading to the activation of stress-response protein kinases and phosphatases such as cAMP-dependent protein kinase A (PKA), protein kinase C (PKC), protein kinase D (PKD), mitogen activated kinases (MAPKs), Ca²⁺ calmodulin-dependent kinase II (CaMKII), and calcineurin. Chronic activation of these stress pathways and their cellular effectors including transcription factors, which target multiple genes, result in changes in cellular structure, function, and overall regulation of the heart, collectively referred to as cardiac remodeling. Additionally, an imbalance between cell survival and cell death pathways results in a low rate of cardiomyocyte apoptosis which may contribute to cell loss in HF [4, 5]. A different form of cell death, necrosis, may contribute to HF, for example through mitochondrial damage mediated by increased cytosolic Ca²⁺ concentrations in cardiomyocytes.

HF is a syndrome which results from different insults, typically following myocardial infarction (MI), viral myocarditis, toxic cardiomyopathy, or other less frequent causes such as genetic defects. However, following chronic remodeling of the heart, cardiac phenotypes occurring from different causes show important similarities, including depressed intracellular Ca²⁺ metabolism.

Ca²⁺ is the central regulator of excitation-contraction (EC) coupling, which controls muscle contraction during each heartbeat. EC coupling is activated by an incoming action potential wavefront and the subsequent opening of voltage-dependent L-type Ca²⁺ channels (Ca_V1.2). A relatively small Ca²⁺ influx current (I _Ca) triggers a quantitatively larger intracellular Ca²⁺ release from sarcoplasmic reticulum (SR) Ca²⁺ stores through ryanodine receptor (RyR2) Ca²⁺ release channels by the process of Ca²⁺ induced Ca²⁺ release (CICR) (Fig. 1a). CICR is followed by re-uptake of Ca²⁺ into the SR by Ca²⁺ pumps (SERCA2) and removal of Ca²⁺ from the cytosol by the Na⁺/Ca²⁺ exchanger (NCX) (Fig. 1a). The activity of SERCA2 is regulated by phospholamban (PLN). Unphosphorylated PLN inhibits SERCA2 activity, and phosphorylation by PKA and/or CaMKII increases SR Ca²⁺ uptake. Similarly, RyR2 phosphorylation by PKA and/or CaMKII increases SR Ca²⁺ release [6, 7]. Within the approximately 12 nm wide compartment between the plasma membrane and the terminal SR membrane (junctional SR) where SR Ca²⁺ release occurs, CICR is organized within Ca²⁺ release units representing functional microdomains between T-tubuli and the terminal SR (junctional subspace). Strikingly, all components essential for CICR, including the SR Ca²⁺ release microdomain, the Ca²⁺ storage organelles, and the Ca²⁺ transport proteins themselves (Table 1), become significantly altered during remodeling of the failing heart [8].

The study of intracellular Ca²⁺ concentration, [Ca²⁺]_i, in heart function requires to monitor the dynamics of [Ca²⁺]_i in living cells. However, the free [Ca²⁺]_i is heterogeneous even in resting cells and ranges from nanomolar to micromolar concentrations in the cytoplasm and SR Ca²⁺ stores, respectively [9]. During electrical cell activation or stimulation by hormone receptors, the resultant changes in the cytoplasmic and organellar Ca²⁺ concentrations occur as spatially and temporally defined patterns within cellular microdomains [10]. Thus, understanding of changes in Ca²⁺ signaling during HF has been largely dependent on methods to monitor Ca²⁺ concentrations in living cells.

Ca²⁺ indicators form selective and reversible complexes with free Ca²⁺ ions. The physicochemical differences of the free and bound indicator allow for the fluorescence detection of changes in ∆[Ca²⁺]_i usually by the absorbance and/or emission of light [11]. Thus, the concentration of free Ca²⁺ is measured indirectly by monitoring the amount of the free versus the Ca²⁺ complexed indicator. The chemiluminescent photoprotein aequorin purified from jellyfish can be microinjected into tissues or cells and applied as a Ca²⁺ indicator [12]. Moreover, recombinant aequorin can be targeted to specific subcellular compartments through signal sequences [13]. However, recombinant aequorin emits a relatively small signal upon Ca²⁺ binding which compromises the measurement of Ca²⁺ concentrations in smaller cells at the cost of low time resolution [14]. Tsien and colleagues synthesized the first, rationally designed, fluorescent Ca²⁺ indicator for intracellular use based on the Ca²⁺-chelator EGTA [15]. The widespread and successful use of fluorescent polycarboxylate dyes started with the introduction of lipophilic acetoxymethyl (AM) ester derivatives allowing efficient and stable indicator loading of living cells without potentially damaging pipette injection techniques [16].

Pioneering studies used injection of the aequorin Ca²⁺ sensor into physiologically contracting muscle preparations from explanted human hearts. Aequorin loaded muscles from failing hearts showed a reduced capacity to restore nanomolar resting Ca²⁺ levels during diastole in conjunction with depressed muscle contraction and relaxation [17]. Different from healthy heart muscle and pronounced at faster rates, the aequorin injected failing heart muscle showed a reduced amplitude of the intracellular Ca²⁺ transient (∆[Ca²⁺]_i) together with diminished force production [18]. Additionally, myothermal measurements in isolated muscle strip preparations from failing and non-failing human hearts showed that the thermal equivalent of total Ca²⁺ cycling is reduced significantly in HF [19]. The rate of heat production was significantly reduced indicating reduced SR Ca²⁺ uptake [19].

Consistent with multicellular muscle preparations, dissociated single cardiomyocytes from failing human hearts displayed a prolonged relaxation, depressed systolic contraction, and elevated diastolic tension [20]. Contractile dysfunction of failing cardiomyocytes occurred in conjunction with abnormal [Ca²⁺]_i metabolism including reduced SR Ca²⁺ release, elevated resting [Ca²⁺]_i, and a reduced rate of Ca²⁺ removal [21, 22]. For late stage HF it is now accepted that cardiomyocytes and/or muscle preparations from explanted patient or animal hearts exhibit a reduced Δ[Ca²⁺]_i amplitude and a slowed decay of the global intracellular Ca²⁺ transient [17, 23‐25]. Thus, depressed contractility in HF appears to be associated with cellular signaling abnormalities at the level of the global intracellular Ca²⁺ transient.

Notably, the combination of an abnormal intracellular Ca²⁺ transient together with depressed contractile function all occur from quite different forms of cardiac insults. Such phenotypic changes have been documented in dilated human cardiomyopathy [23, 26], hypertension induced hypertrophy HF in salt-sensitive rats [24], rats with myocardial infarction [27], mice with muscle LIM protein knockout [28], mice with replication-restricted full-length Coxsackievirus B3 overexpression and myocarditis [29], and mice overexpressing either the catalytic protein kinase A (PKA)-Cα subunit [30] or the cytosolic CaMKIIδ_c splice variant [31]. Thus, different HF models appear to agree that qualitatively similar changes of the intracellular Ca²⁺ transient and subsequent alterations of EC coupling and contractile dysfunction occur.

While cardiac contraction and relaxation cycles are controlled by a cell-wide (global) signaling event, the Ca²⁺ transient, intracellular Ca²⁺ release occurs within local (subcellular) Ca²⁺ release units as evidenced by elementary Ca²⁺ release events measured with bright fluorescent polycarboxylate dyes [8, 32]. How do local, elementary Ca²⁺ release events, known as Ca²⁺ sparks, modulate the intracellular Ca²⁺ transient? At a diastolic [Ca²⁺]_i of approximately 100 nM, Ca²⁺ sparks occur at a very low rate (100 s⁻¹ cell⁻¹). When the AP occurs, Ca²⁺ influx causes a large increase in cytosolic [Ca²⁺]_i within the tight junctional subspace (Fig. 1a). A raise of junctional [Ca²⁺]_i above nanomolar concentrations increases the spark rate 10³ to 10⁶ fold. Therefore, the I _Ca influx current becomes amplified by a dramatic increase in Ca²⁺ spark rate during CICR. This implies that modulation of the Ca²⁺ spark rate may control the Ca²⁺ transient amplitude and cardiac force development (inotropy). Indeed, imaging experiments of subcellular Ca²⁺ signaling have provided evidence that β-adrenergic stimulation increases local Ca²⁺ release possibly due to RyR2 phosphorylation [33].

An important advance why changes in the intracellular Ca²⁺ transient occur in HF come from combined voltage-clamp and Ca²⁺ spark imaging experiments to assess if changes in EC coupling are responsible for depressed cardiac function. A rat model of hypertension induced HF with preserved Ca²⁺ influx current (I _Ca) density showed that the ability of any given I _Ca to activate SR Ca²⁺ sparks (Δ[Ca²⁺]_i) was significantly decreased [24]. Thus, HF is accompanied by a decrease in the gain of EC coupling (Δ[Ca²⁺]_i/I_Ca) which has been confirmed in rats with myocardial infarction [27], dogs with pacing-induced HF [34], mice with viral myocarditis [29], and muscle LIM protein knockout mice [35]. Additionally, structural changes of the T-tubules, SR storage organelles, and/or the architecture of the junctional microdomain cleft are all likely contributors to defective EC coupling by altering the geometry of the Ca²⁺ release unit potentially resulting in orphaned RyR2 release channels and/or abnormal CICR [36, 37].

Sudden unexpected death accounts for up to 50% of all deaths in HF patients and is most often due to ventricular tachyarrhythmias [38]. Both reentry and focal mechanisms have been documented in patients with ischemic cardiomyopathy [39]. At the cellular level, arrhythmias have been associated with Ca²⁺ induced electrical abnormalities including SR Ca²⁺ overload leading to intracellular Ca²⁺ waves and Ca²⁺ activated transient inward current (I _TI) [40‐42]. In digitalis treated cells, I _TI was shown to initiate delayed after depolarizations (DADs) [43]. Association of SR Ca²⁺ leak with activation of a depolarizing Na⁺/Ca²⁺ exchange current is a likely mechanism of arrhythmogenic I _TI in HF [44‐46].

Genetic linkage and translational studies have significantly advanced our understanding about specific Ca²⁺ dependent arrhythmia mechanisms (Table 1). For instance, RyR2 missense mutations may cause stress-induced syncope and sudden death in a syndrome called Catecholaminergic Polymorphic Ventricular Tachycardia (CPVT) [47, 48]. Consistent with a β-AR and cAMP mediated stress mechanism, PKA phosphorylated RyR2 channels containing CPVT mutations showed a significant gain-of-function defect and a Ca²⁺ leak mediated arrhythmia trigger mechanism [49, 50]. Indeed, cardiomyocytes from knockin mice with the RyR2-R2474S mutation identified earlier in CPVT patients [47] showed Ca²⁺ leak mediated I _TI currents and DADs during stimulation with catecholamines (Fig. 1b) [51]. While we have to anticipate additional mechanisms of Ca²⁺ leak in HF, mechanistic linkage of CPVT mutant RyR2 dependent Ca²⁺ leak to arrhythmia initiation [51] provides a specific mechanism of Ca²⁺ dependent arrhythmia initiation which can be targeted and tested by a novel class of RyR2 stabilizing compounds [51] (Table 2).

Ca²⁺ is the central intracellular messenger which links an incoming action potential to myofilament activation and cardiac contraction. Apart from EC coupling, [Ca²⁺]_i is subject to physiological stress adaptation and a higher systolic ∆[Ca²⁺]_i increases force development and cardiac output (inotropy) [33]. Following the realization that SR Ca²⁺ uptake might be depressed in failing cardiomyocytes, the intracellular storage organelles and SR Ca²⁺ content became a focus of intense research interest.

HF has been associated with decreased function of different Ca²⁺ transport mechanisms (Table 1) including the sarcoendoplasmic reticulum (SR) Ca²⁺ ATPase (SERCA2) [52, 53] and with increased function of the Na⁺/Ca²⁺ exchanger (NCX) [54]. The ratio of PLN:SERCA2 is a critical determinant of cardiac Ca²⁺ homeostasis, and increases in this ratio have been suggested to contribute to increased diastolic Ca²⁺ levels and cardiac dysfunction [53, 55, 56]. On the other hand, due to the changes in the relative expression and function of NCX and SERCA2, a net shift toward increased Ca²⁺ extrusion to the extracellular space and a net decrease in SR Ca²⁺ uptake occurs (Fig. 1c). Therefore, intracellular Ca²⁺ stores may develop a relatively Ca²⁺ depleted state which has been confirmed experimentally in cardiomyocytes from dogs and humans with HF [26, 34]. Decreased SR Ca²⁺ load may result in a decreased amplitude and slower upstroke of the intracellular Ca²⁺ transient. Since ∆[Ca²⁺]_i is an important mediator of cardiac force development, a decreased ∆[Ca²⁺]_i due to decreased SR load may result in decreased force development. However, SR Ca²⁺ store depletion in HF does not necessarily prevent an increase in Ca²⁺ spark frequency [31] likely due to chronically increased RyR2 phosphorylation [57] and consistent with an increased propensity for Ca²⁺ induced arrhythmias in HF.

Decreased SERCA2 function has been associated with changes in the regulatory subunit PLN (Fig. 1c). PLN is a phosphoprotein which inhibits SERCA2 in its dephosphorylated state, while phosphorylation of PLN during β-AR stimulation relieves this inhibitory effect. The regulatory role of PLN in myocardial contractility has been established through the generation and characterization of genetically altered mouse models, which revealed a correlation between PLN expression and contractile function [58‐60]. PLN phosphorylation in HF is chronically decreased (PLN hypophosphorylation) which may directly contribute to depressed SR Ca²⁺ uptake [61]. Chronic PLN hypophosphorylation can be predicted to compromise cardiac stress adaptation mediated by cAMP and PKA. Overexpression of the cardiac SERCA2a isoform or the constitutively PKA phosphorylated PLN-Ser¹⁶ improved function in rats following aortic banding-induced HF [62] and decreased arrhythmia susceptibility following ischemia-reperfusion cardiomyopathy in pigs [63].

Phospholamban is potently inhibited by phosphatase 1 (PP1), which in turn is inhibited by phosphatase inhibitor I-1 following PKA phosphorylation and I-1 activation. Consistent with this mechanism, overexpression of an activated I-1 form in mice protects the animals from HF development [64]. Moreover, the identification of PLN mutations in patients with dilated cardiomyopathy has strengthened its critical role in cardiac function. Thus, inhibition of PLN activity and restoration of SR Ca²⁺ cycling were suggested to hold promise for treating heart failure [65].

In samples from patients and animals with HF, RyR2 is chronically PKA hyperphosphorylated contributing to intracellular Ca²⁺ leak and remodeling of the macromolecular channel complex [7, 66, 67]. If the hyperadrenergic state in HF leads to secondary downregulation of β-adrenergic signaling and globally reduced intracellular cAMP synthesis, which molecular mechanism maintains RyR2 PKA hyperphosphorylation and intracellular Ca²⁺ leak? During HF, phosphatases (PP1 and PP2A) are depleted from the RyR2 complex which may contribute to a reduced rate of RyR2 dephosphorylation [61, 66]. In human HF, the cAMP-specific phosphodiesterase isoform PDE4D3 was found to be decreased in the RyR2 complex paralleled by a decrease in cAMP hydrolyzing activity [57]. Thus, PKA hyperphosphorylation and calstabin2 depletion may both contribute to chronic RyR2 activation and intracellular Ca²⁺ leak (Fig. 1c). Indeed, single RyR2 channels from human failing hearts showed an increased open probability consistent with intracellular Ca²⁺ leak [66].

Using site-directed mutagenesis of RyR2, a specific Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) phosphorylation site (Ser²⁸¹⁴) distinct from the PKA phosphorylation site (Ser²⁸⁰⁸) has been identified [6]. CaMKII phosphorylation increased RyR2 Ca²⁺ sensitivity and open probability. Since CaMKII was activated by higher cardiac pacing rates, RyR2 CaMKII phosphorylation may contribute to enhanced CICR and enhanced contractility at higher heart rates. However, the rate-dependent increase in RyR2 phosphorylation by CaMKII as seen in sham-operated, healthy rat hearts was compromised in rat hearts with HF [6]. Transgenic mice overexpressing the cytosolic CaMKIIδ_c isoform develop cardiac hypertrophy, HF, and intracellular Ca²⁺ leak [68]. Acute overexpression of CaMKIIδ_c in cardiomyocytes resulted in significant SR Ca²⁺ leak without changes in contractile function [69]. Adenoviral CaMKII overexpression was not associated with calstabin2 dissociation from RyR2 [69] confirming earlier results [6]. On the other hand, PKA phosphorylation of RyR2 resulted in depletion of the stabilizing calstabin2 subunit [6], and decreased calstabin2 levels have been linked to Ca²⁺ triggered arrhythmias in the structurally normal heart [49]. Disruption of RyR2 PKA phosphorylation in Ser2808Ala knockin mice with myocardial infarct induced HF has provided evidence that PKA hyperphosphorylation induced SR Ca²⁺ leak may directly contribute to HF progression [70]. However, the significance of concomitant PKA and/or CaMKII RyR2 phosphorylation in Ca²⁺ dysregulation during HF has not been addressed conclusively.

Intracellular Na⁺ and Ca²⁺ concentrations are intricately coupled through the NCX current. In addition, Na⁺ influx through sarcolemmal Na_V1.5 channels contributes to intracellular [Na⁺]_i homeostasis (Fig. 1a). Studies about [Na⁺]_i in human myocardium showed a stimulation rate dependent increase in [Na⁺]_i [71, 72]. For any given stimulation rate, human end-stage failing myocardium showed a shift toward higher intracellular [Na⁺]_i [71]. In a rabbit HF model with elevated [Na⁺]_i, the function of the Na⁺/K⁺-ATPase function was found normal [73]. A potential role of altered Na⁺ channel inactivation and a significantly increased late I _Na,L in HF as a cause of intracellular [Na⁺]_i overload has been documented in animal models of HF and human failing myocardium [74, 75]. Slow pacing of failing cardiomyocytes with increased [Na⁺]_i may enhance Ca²⁺ influx from the extracellular side through reverse mode NCX contributing to increased SR Ca²⁺ load and force development. Indeed, at higher pacing rates failing cardiomyocytes with high [Na⁺]_i are prone to diastolic Ca²⁺ overload and contractile dysfunction. Interestingly, pharmacological inhibition of I _Na,L was found to improve diastolic dysfunction in failing human myocardium [76]. Additionally, increased late I _Na,L may result in Ca²⁺ induced electrical cardiomyocyte dysfunction as shown in SCN5A-∆KPQ knockin mice resembling the Long-QT3 syndrome [77].

Pathophysiological consequences of altered EC coupling have an immediate impact on cardiac stress adaptation during higher heart rates. The force–frequency relation (FFR) or staircase phenomenon of healthy human myocardium and within physiological limits describes a heart rate dependent increase in contractile force and cardiac output. However, in failing human hearts or isolated muscle preparations, the frequency-dependent potentiation of contractile force is blunted or, even worse, inversed [78]. Alteration of the FFR in the failing human heart has been accepted as a functional milestone which may partially explain the decreased exercise capacity of symptomatic HF patients.

Using aequorin Ca²⁺ measurements it has been shown that a positive FFR in normal healthy myocardium is associated with a positive ∆[Ca²⁺]_i transient amplitude–frequency relation, and vice versa, a negative FFR in the failing myocardium is associated with an inverted ∆[Ca²⁺]_i–frequency relation [18]. Alterations of the frequency response of contractile performance have been confirmed in vivo in HF patients [79]. A different experimental protocol which is thought to enable increased Ca²⁺ accumulation into the SR stores in diastole, uses post-rest pacing which typically produces an increase in developed force in healthy myocardium also referred to as ‘post-rest potentiation’. However, alterations in post-rest potentiation that may underlie reduced stress adaptation and the FFR inversion have been associated with a blunted post-rest ∆[Ca²⁺]_i increase of SR Ca²⁺ content (Fig. 2) [80]. While in healthy human myocardium a pronounced upregulation of SR Ca²⁺ content occurs at higher heart rates, failing human myocardium shows a blunted regulation of SR Ca²⁺ content [26]. Therefore, the chronic hyperadrenergic state and associated higher heart rates in HF patients may create an increased risk for Ca²⁺ induced cardiac dysfunction and arrhythmias. Accordingly, novel therapeutic strategies have successfully used pharmacological heart rate reduction in HF patients to improve prognostic outcome [81].

Several therapeutic rationales exist which aim to correct known molecular Ca²⁺ signaling abnormalities in HF as summarized in Table 2. Among these strategies, NCX blockers may have therapeutic utility if mode-selective block can be established and/or risks for disturbing intracellular Ca²⁺ metabolism can be avoided. Partial inhibition of NCX by SEA-0400 in MLP^−/− knockout cardiomyocytes with heart failure showed a net gain of [Ca²⁺]_i and SR load but no improvement in contractility whereas cardiomyocytes from mice with aortic-banding induced hypertrophy and HF showed improved contractile function [82]. Further development of NCX blockers for HF will depend on the critical assessment of the potential benefits of NCX reduction versus effects on [Ca²⁺]_i by refining mode dependence and/or including additional targeting strategies.

Some traditional ion channel blockers like tetracaine inhibit ion permeation through RyR2 channels, however, lack of specificity and potential side effects on contractile function indicate significant limitations toward therapeutic applicability [83, 84]. Recently, the efficacy of novel RyR2 channel stabilizing drugs of the 1,4-benzothiazepine class (JTV519 or K201) which stabilize the RyR2 closed state but do not block ion permeation has been established in animal models of heart failure where they inhibit progression of cardiac remodeling and dysfunction [57, 85, 86]. RyR2 stabilizing compounds with high specificity and cardiac activity have been developed [51] and their efficacy as potential HF treatment is under investigation. Additionally, β-blockers and angiotensin-II receptor blockers have been associated with beneficial effects on RyR2 channel dysfunction in HF through mechanisms which indirectly prevent excess post-translational modification of RyR2, e.g. by PKA hyperphosphorylation or nitrosylation [87‐89].

Adeno-associated viruses (AAVs) are currently best suited for myocardial gene delivery due to minimal pathogenicity and several serotypes exhibit tropism for the heart. Following identification and comprehensive characterization of SERCA2a as a potential therapeutic target in HF, cardiac gene therapy trials using replication-deficient viral vectors (AAV-SERCA2a) have been approved by the US Food and Drug Administration and are under review with the UK regulatory authorities for clinical trials [90].

HF is characterized by chronic maladaptive changes and a poor prognosis. Chronic activation of intracellular signaling pathways mediates structural and functional remodeling of the failing heart. Changes in EC coupling at the level of local and global Ca²⁺ signals represent a key mechanism of contractile depression and arrhythmia propensity. It has become increasingly clear that cardiac remodeling in HF occurs within cytosolic Ca²⁺ signaling microdomains. Translational approaches about the local and global Ca²⁺ signaling mechanisms, remodeling mechanisms, and related disease processes will be of key importance to develop novel and specific therapeutic rationales. We anticipate that Ca²⁺ imaging techniques will significantly increase our understanding of the cardiac pathophysiology underlying HF and boost development of novel therapeutic strategies in the future.