Sarcoplasmic reticulum calcium mishandling: central tenet in heart failure?

Abstract

Excitation-contraction coupling links excitation of the sarcolemmal surface membrane to mechanical contraction. In the heart this link is established via a Ca²⁺-induced Ca²⁺ release process, which, following sarcolemmal depolarisation, prompts Ca²⁺ release from the sarcoplasmic reticulum (SR) though the ryanodine receptor (RyR2). This substantially raises the cytoplasmic Ca²⁺ concentration to trigger systole. In diastole, Ca²⁺ is removed from the cytoplasm, primarily via the sarcoplasmic-endoplasmic reticulum Ca²⁺-dependent ATPase (SERCA) pump on the SR membrane, returning Ca²⁺ to the SR store. Ca²⁺ movement across the SR is thus fundamental to the systole/diastole cycle and plays an essential role in maintaining cardiac contractile function. Altered SR Ca²⁺ homeostasis (due to disrupted Ca²⁺ release, storage, and reuptake pathways) is a central tenet of heart failure and contributes to depressed contractility, impaired relaxation, and propensity to arrhythmia. This review will focus on the molecular mechanisms that underlie asynchronous Ca²⁺ cycling around the SR in the failing heart. Further, this review will illustrate that the combined effects of expression changes and disruptions to RyR2 and SERCA2a regulatory pathways are critical to the pathogenesis of heart failure.

Cardiac contraction and SR calcium handling

The cyclic release and uptake of Ca²⁺ from the modified endoplasmic reticulum Ca²⁺ store (the sarcoplasmic reticulum; SR) are quintessential events in excitation–contraction (EC) coupling and drive diastole and systole in the heart. Triggered by surface membrane depolarization, L-type Ca²⁺ channels initiate a small inward Ca²⁺ current, which prompts for the larger and sustained release of SR Ca²⁺ through the ligand-gated release channel, the ryanodine receptor (RyR). The resultant increase in cytosolic Ca²⁺ (from 100 nM at rest to 10 μM at the end of systole), via the Ca²⁺-induced Ca²⁺ release mechanism, triggers crossbridge cycling and contraction. L-type Ca²⁺ channels inactivate, RyRs close, and removal of Ca²⁺ from the cytoplasm is achieved via the sarcoplasmic-endoplasmic reticulum Ca²⁺-dependent ATPase pump (SERCA) and the Na⁺-Ca²⁺ exchanger embedded in the surface membrane. SERCA is located on the SR and acts to replenish the SR Ca²⁺ load during diastole, so that Ca²⁺ is stored ready for release on the next beat. The Na⁺-Ca²⁺ exchanger returns Ca²⁺ equalling the initial inward current, back to the extracellular space (Fig. 1).

Fig. 1.

Proteins involved in regulation of SR Ca²⁺ uptake and release in the heart. Upon surface membrane depolarization, a small inward Ca²⁺ transient though the L-type Ca²⁺ channel (LTCC) triggers a larger Ca²⁺ transient through the RyR2, to trigger systole. During diastole, the LTCC inactivates, RyR2 closes, and Ca²⁺ is sequestered back into the SR via SERCA2a (SERCA), lowering cytoplasmic Ca²⁺. The Na⁺-Ca²⁺exchanger (NCX) returns Ca²⁺ equalling the initial inward current, back to the extracellular space. RyR2 is modulated by a number of key proteins and modifications, including cytoplasmic proteins calmodulin (CaM), membrane bound junctophilin (black line connecting the t-tubules to the SR membrane) and the FK binding proteins (FKBP) and by a luminal protein complex including triadin (Tri) and junctin (Jun), Ca²⁺ binding proteins histidine-rich Ca²⁺ binding protein (HRC), calsequestrin 2 (CSQ2; pink circles) and S1001A (S100). That triadin cannot bind to CSQ2 and RyR2 when bound to HRC is depicted by dashed yellow and pink lines. SERCA2a is modulated by the membrane bound regulatory protein phospholamban (PLN) and Ca²⁺ binding proteins HRC and S100A1. Sympathetic activation of the β-adrenergic receptor (β-AR) triggers G-protein (Gs) activation of adenylyl cyclase (Ac) and generation of 3′-5′cyclic adenosine-monophosphate (cAMP), which drives phosphorylation of the LTCC, RyR2 and PLN, mediated by PKA and CaMKII (blue circles). Generation of ROS/RNS by xanthine oxidase, nitric oxide synthase and nicotinamide adenine dinucleotide phosphate oxidase—depicted by red rectangles—which fine-tunes the activity of LTCC, RyR2, SERCA2a, CaMKII and PLN

For healthy heart function, release and uptake of SR Ca²⁺ during systole and diastole is tightly regulated. This is due to the exquisite control of RyR and SERCA function by the integrated effects of accessory proteins, ions, molecules, and post-translational phosphor- and thiol modification. Dysfunction of RyR and SERCA can alter SR Ca²⁺ cycling and deplete the SR Ca²⁺ load (reviewed in (Gorski et al. 2015)), signatures of heart failure and arrhythmia.

Ryanodine receptors

Structure and function

The RyR is a ligand-gated Ca²⁺ release channel located in the SR membrane. Cardiac RyR (RyR2) are high conductance channels with modest ion selectivity (reviewed in (Meissner 2017)) and are largely responsible for the large rise in cytoplasmic Ca²⁺, which triggers systole. In addition to RyR2, two isoforms exist; RyR1, which is the predominantly expressed isoform in skeletal muscle, and RyR3, which is broadly expressed. All isoforms share a > 65% sequence identity and significant functional homology (Hakamata et al. 1992).

RyRs are tetramers composed of four identical ~550 kDa subunits symmetrically arranged in the membranes of the SR. The bulk of the RyR tetramer is located on the cytoplasmic side, occupying much of the junctional space between the SR and sarcolemma (reviewed in (Meissner 2017)). The full structure is yet to be resolved, but cryo-electron microscopy studies have elucidated key features within 11 distinct domains (Peng et al. 2016). The N-terminal domain (a hotspot for mutation), three splA kinase and ryanodine receptor (SPRY) binding domains (including SPRY1 which contains the binding motif for the immunophilin FK506 binding proteins (FKBPs) discussed below), the P1 and P2 domains, and the handle domain (which with SPRY1, forms the FKBP binding motif). These are followed in sequence by the helical domains (HD), HD1 and HD2 (containing residues which are hyperphosphorylation targets in heart failure), the central domain (containing two EF hand motifs which are pivotal in channel gating), and finally the channel domain (which forms the Ca²⁺ pore and contains Ca²⁺ and adenosine triphosphate (ATP) activation sites).

RyR2 regulatory proteins

RyR2 forms a macromolecular complex with numerous cytoplasmic and SR luminal proteins, many of which are implicated in the pathogenesis of heart failure (described below). Numerous cytoplasmic proteins bind to RyR2 and influence its function, including the FKBPs FKPB12.0 and FKBP12.6, and the EF hand Ca²⁺-binding protein calmodulin. FKPB12.0 and FKBP12.6 bind to RyR2 and are postulated to stabilize the RyR2 in its closed conformation during diastole (Brillantes et al. 1994; Richardson et al. 2017). Calmodulin, which binds to RyR2 in both the absence and presence of Ca²⁺, stabilizes interactions between RyR2 N-terminal and central domains, hindering conformational changes which promote RyR2 activation (Huang et al. 2013). Calmodulin also activates Ca²⁺/calmodulin-dependent protein kinase II (CaMKII), an important RyR2 regulator (see below). Emerging evidence illustrates that a second EF hand Ca²⁺ sensing protein, S1001A, has a role in enhancing contractile performance by modulating RyR2 activity (Du et al. 2002) in a Ca²⁺-dependent manner. S1001A enhances the Ca²⁺ transient amplitude during systole and curtails diastolic Ca²⁺ leak (Völkers et al. 2007) in healthy cardiomyocytes. Finally, junctophilin 2 is a membrane-binding protein that forms a structural bridge between the t-tubule and the SR membrane, enabling formation of dyadic junctions (Takeshima et al. 2000). Junctophilin 2 also modulates Ca²⁺ flux by increasing the number of functional L-type Ca²⁺ channels (Poulet et al. 2020) and through interactions with RyR2. Junctophilin 2 stabilizes RyR2 in the closed state (Munro et al. 2016) to enable RyR2 inactivation in diastole (van Oort et al. 2011) and curtail channel activity during diastole (Beavers et al. 2013).

A SR luminal complex is formed between RyR2 and anchoring proteins triadin and junctin, which tether the Ca²⁺ binding proteins calsequestrin 2 (CSQ2) (Wei et al. 2009) and the histidine-rich Ca²⁺ binding protein (HRC) (Rani et al. 2012) to RyR2. Together, this complex serves as both a luminal Ca²⁺ sensor and to facilitate RyR2 luminal Ca²⁺ sensitivity, curtailing RyR2 activity under diastolic conditions (Dulhunty et al. 2012). CSQ2 is a glycoprotein residing in the SR lumen, which has a moderate to high capacity but low affinity for SR Ca²⁺ and whose secondary role is to buffer free SR [Ca²⁺] to ~ 1 mM. HRC has a high affinity and low capacity for Ca²⁺. Bi-functional in nature, HRC binds to SERCA at low SR luminal Ca²⁺ (inhibiting pump function) and indirectly with RyR2 at moderate SR Ca²⁺, curtailing RyR2 activity and preventing Ca²⁺ leak (Rani et al. 2012).

RyR2 post-translational modification

RyR2 activity and gating are also modulated by ions (including Ca²⁺ and Mg²⁺), co-proteins and, of importance for this review, post-translational modification due to β-adrenergic receptor (β-AR) stimulation and redox modification. Two important proteins, which tether to the RyR2 cytoplasmic domain, are the ubiquitous serine/threonine protein kinases, protein kinase A (PKA), and CaMKII. In response to a β-AR signalling cascade or a rise in cytoplasmic Ca²⁺, these proteins post-translationally phosphorylate RyR2, transiently increasing channel activity and SR Ca²⁺ release (Marx and Marks 2013; Camors and Valdivia 2014).

When haemodynamic demand is increased in heart failure, sympathetic nervous system activity is increased, causing an increase in noradrenaline and adrenaline levels. Noradrenaline and adrenaline bind to sarcolemmal β₁-AR and β₂-ARs, inducing a conformational change that leads to stimulatory Gsα-protein coupling, activation of adenylyl cyclase, accumulation of 3′-5′cyclic adenosine-monophosphate (cAMP), and activation of PKA, which drives a phosphorylation cascade ((Kaumann et al. 1999); Fig. 2a). PKA has diverse targets (including RyR2, L-type Ca²⁺ channels, the SERCA-regulator phospholamban, troponin I, and C-protein) (Kaumann et al. 1999). In the healthy heart, the phosphorylation events PKA mediates coalesce to augment cardiac function by enhancing systolic Ca²⁺ transients and boosting diastolic Ca²⁺ uptake. In the human non-failing heart, activation of both β₁-AR and β₂-ARs mediates a positive inotropic response with a hastening of relaxation (Molenaar et al. 2000).

Fig. 2.

Pathway to protein phosphorylation and cysteine modification of sarcoplasmic reticulum Ca²⁺ handling proteins. a In response to an increased sympathetic drive, catecholamines bind to and alter structural conformation of G-protein (Gs) coupled β-AR on the sarcolemma. The resultant activation of adenyl cyclase (Ac) stimulates 3′-5′cyclic adenosine-monophosphate (cAMP) generation, activating PKA. PDE4 acts to modulate cAMP production. CaMKII is reported to be activated by a rise in cellular cAMP; however, the prevailing consensus supports a Ca²⁺-mediated mechanism. PKA and CaMKII target and phosphorylate Ca²⁺ handling proteins including the LTCC, RyR2 and phospholamban (PLN), depicted by blue circles. In healthy muscle, the ensuing increase in systolic SR Ca²⁺ release through RyR2 and SR Ca²⁺ uptake through SERCA2a (SERCA) during diastole drives a larger Ca²⁺ transient and enhances contractility. Only proteins involved in or modified by β-AR signalling are included in this figure, for simplicity. b The major sources of reactive oxygen species/reactive nitrogen species (ROS/RNS) in the heart are xanthine oxidase (XO), nitric oxide synthase (NOS), nicotinamide adenine dinucleotide phosphate oxidase (NADPH oxidase or NOX) and the mitochondria, all of which generate superoxide (O₂^.-), which can S-oxidize cysteine residues. NOS produce nitric oxide (NO), which in the presence of O₂⁻, forms peroxynitrite (ONOO⁻); these processes can both S-nitrosylate cysteine residues and nitrate tyrosine residues. The major targets of ROS/RNS include the LTCC, NCX, RyR2, CaMKII, SERCA and PLN, depicted by red circles. For simplicity, and as there has been no role implicated in failing heart, the process of cysteine S-glutathionylation is not presented in this figure. In healthy muscle, the basal levels of tyrosine and cysteine modification (maintained by the opposing antioxidant system) have a potential ionotropic role to increase Ca²⁺ release under conditions of increased heart rate. Only proteins involved in or modified by ROS/RNS signalling are included in this figure, for simplicity

While CaMKII activation has been linked to mediation of the β-AR response (Grimm and Brown 2010), its activity is considered primarily a reaction to Ca²⁺ fluxes in the contraction-relaxation cycle. During systole (or when cytoplasmic Ca²⁺ levels are high), Ca²⁺ binds to calmodulin; Ca²⁺-calmodulin in turn binds to CaMKII, inducing a conformational change which exposes the catalytic domain, readying it to mediate local phosphorylation (Grimm and Brown 2010). CaMKII may also operate in a Ca²⁺-independent fashion as a result of auto-phosphorylation during β-AR signalling (Erickson et al. 2011) or during oxidative stress (Erickson et al. 2008). CaMKII has diverse targets (including RyR2, L-type Ca²⁺ channels, and phospholamban), whose phosphorylation-dependant modulation has been well-described (reviewed in (Bers and Grandi 2009)).

Arguably just as important in regulating RyR2 function are transient redox modifications, which fine-tune the activity of Ca²⁺ handling proteins and are induced by reactive oxygen species (ROS) and reactive nitrogen species (RNS) (Fig. 2b). ROS and RNS modify the thiol side chains of cysteine residues. ROS, including superoxide, hydrogen peroxide, and hydroxyl radicals, are a by-product of the mitochondrial electron transport chain. The localization of ROS-generating enzymes xanthine oxidase and nicotinamide adenine dinucleotide phosphate oxidase (on the SR and sarcolemmal membranes) (reviewed in (Niggli et al. 2013)) is thought to be responsible for RyR2 oxidative and nitrosative modification. Two cardiac isoforms of nitric oxide synthase produce both nitric oxide and peroxynitrite, which can modify RyR2 (Flores et al. 2016). The three primary modifications induced are S-oxidation (two oxidized thiols react to form disulphide bridges), S-nitrosylation (oxidized thiols react with RNS), and S-glutathionylation (oxidized thiols react with glutathione, forming a mixed disulphide). Generally, ROS-induced modifications increase RyR2 activity (Nikolaienko et al. 2018); however, some observe an initial RyR2 activation followed by a persistent inhibition (Eager et al. 1997; Eager and Dulhunty 1999).

Sarcoplasmic-endoplasmic Ca²⁺-ATPase

Structure and function

SERCA is a 110 kDa Ca²⁺-ATPase (P-type), which spans the SR membrane. The primary role of SERCA is to restore cytosolic Ca²⁺ to diastolic levels and to replenish SR Ca²⁺ load (Misquitta et al. 1999). There are three mammalian SERCA isoforms (SERCA1, SERCA2, and SERCA3), which are implicated in diverse cell functions (reviewed by (Misquitta et al. 1999)) with the prominent cardiac variant being SERCA2a (Periasamy and Kalyanasundaram 2007). Despite variation in size, protein interactions, and tissue localisation (Misquitta et al. 1999), SERCA isoforms display an ~ 84% sequence identity (Periasamy and Kalyanasundaram 2007).

SERCA2a is a monomer, comprised of four regions: a transmembrane region of ten helices in the SR membrane and three cytosolic regions, which include phosphorylation and nucleotide binding domains (reviewed in (Wuytack et al. 2002)). Within the transmembrane regions are two Ca²⁺-binding sites. When these sites face the cytoplasm, they have a high affinity for Ca²⁺, and when cytosolic Ca²⁺ binds (with ATP), a conformational switch occurs to transport Ca²⁺ across the SR (Toyoshima et al. 2000). In the alternate conformation, Ca²⁺ binding sites are low affinity and face the SR lumen (Toyoshima et al. 2000). Ca²⁺ is then released, inducing SERCA2a’s reversion to the high affinity conformation, allowing the transport cycle to continue (Moller et al. 2005).

SERCA regulatory proteins

In the context of this review, prominent influences on SERCA2a activity are interactions with the inhibitory protein phospholamban and activator S100A1, both of which bind to the cytoplasmic/transmembrane domains of SERCA2a. Phospholamban serves as a brake on SERCA2a activity, with the dephosphorylated form inhibiting SERCA2a by reducing its affinity for Ca²⁺ (Ling et al. 2012; Akin et al. 2013). When phospholamban is phosphorylated by PKA and/or CaMKII, this inhibition is lost. The subsequent increase in SERCA2a pump activity (Simmerman and Jones 1998) is associated with shortening of the duration of contraction in ventricle (Kaumann et al. 1999; Molenaar et al. 2000) and atrium (Molenaar et al. 2007). A novel Dwarf Open Reading Frame Micropeptide (DWORF), which binds to SERCA, has recently been characterized (Nelson et al. 2016). DWORF has a higher affinity for SERCA2a than phospholamban. This allows DWORF to displace phospholamban from SERCA2a to relieve phospholamban inhibition of pump activity (Makarewich et al. 2018). DWORF’s potential as a treatment in heart failure is discussed further below. Finally, in addition to regulating Ca²⁺ release, S100A1 can also influence SERCA function, by forming a Ca²⁺-dependent complex with both SERCA2a and phospholamban (Kiewitz et al. 2003). Additionally, S100A1 serves to enhance Ca²⁺ uptake during diastole to support both Ca²⁺ loading and timely relaxation (Kiewitz et al. 2003; Kettlewell et al. 2005). This mechanism appears independent of any effect on phospholamban; that is, it does not alter phospholamban binding to (or regulation of) SERCA2a (Rohde et al. 2010).

SR Ca²⁺ mishandling in heart failure

Balanced SR Ca²⁺ cycling and maintenance of cytoplasmic Ca²⁺ are essential for maintaining contractile function. Reduced systolic Ca²⁺ delivery to the cytoplasm (reduced SR Ca²⁺ release), defective diastolic Ca²⁺ removal from the cytoplasm (reduced SR Ca²⁺ uptake), and aberrant diastolic cytoplasmic Ca²⁺ accrual (diastolic Ca²⁺ leak) underpin contractility declines and can serve as an arrhythmic substrate (Fig. 3). A reduction in Ca²⁺ transmitted to the cytoplasm during systole contributes to a smaller Ca²⁺ transient. The resultant dampened EC coupling gain and hypocontractility are found in heart failure (Hobai and O’Rourke 2001). Reduced uptake of Ca²⁺ into the SR in ventricular tissue is a feature of heart failure, which, in turn, is linked to reductions in the SR Ca²⁺ load, reduced subsequent systolic Ca²⁺ transients, and hypocontractility. Both reduced SR Ca²⁺ uptake and diastolic Ca²⁺ leak from the SR (through RyR2) result in elevated cytoplasmic Ca²⁺ during diastole and together serve to increase Na⁺-Ca²⁺ exchanger forward mode activity. This generates a depolarizing inward current and a delayed-afterdepolarization (DAD), which, if large enough, can trigger arrhythmia.

Fig. 3.

Mechanisms of impaired Ca²⁺ cycling in heart failure. In heart failure, impaired Ca²⁺ cycling can manifest as defective SR Ca²⁺uptake and/or Ca²⁺ buffering capacity, reduced SR Ca²⁺ release and diastolic Ca²⁺ leak. a Decreases in SERCA2a expression, the SERCA2a to phospholamban (PLN) ratio and in small ubiquitin-like modified (SUMO; grey circle) expression, coupled with hypophosphorylation of PLN due to aberrant phosphatase activity, depresses SR Ca²⁺ uptake during diastole. Post-translational modification of SERCA2a (tyrosine nitration) due to enhanced ROS/RNS and protein acetylation (orange circle, labelled a) contributes to the decline in diastolic SR Ca²⁺ uptake SERCA2a and cardiac dysfunction. Altered trafficking of calsequestrin 2 (CSQ2; pink circles) to the junctional SR (reducing total SR calsequestrin 2 content) and mutations within the Ca²⁺ binding protein histidine-rich Ca²⁺ binding protein (HRC; with mutation depicted by a cross) contribute to a reduced Ca²⁺ buffering capacity. b Reduced SR Ca²⁺ release, which manifests as a decreased Ca²⁺ transient, precipitates the reduction in contractility and cardiac output in heart failure. The declining transient can be due to a lowered Ca²⁺ store load or to changes in CICR. Alterations in subcellular architecture and the existence of RyR2 outside coupons result in a decrease in Ca²⁺-induced Ca²⁺ release (CICR), and to the declining ability of the LTCC to stimulate RyR2 Ca²⁺ release. Further, a decreased expression in junctophilin 2 (black line connecting the t-tubule to the SR membrane) in early heart failure contributes to the loss of CICR. c Due to chronic β-AR stimulation during heart failure, upregulation of phosphor pathways hyperphosphorylate RyR2 (blue circle) via PKA and CaMKII, dissociating the FK binding proteins (FKBPs) and inducing aberrant RyR2 activity during diastole (diastolic Ca²⁺ leak). In addition, chronic levels of ROS/RNS found in failing heart redox modify RyR2 (red circle), resulting in aberrant RyR2 activity during diastole (diastolic Ca²⁺ leak), and dissociation of calmodulin (CaM) and FKBP. A downregulation of junctophilin 2 (black line connecting the t-tubule to the SR membrane) in failing heart leads to the destabilization of RyR2 in the closed state and diastolic Ca²⁺ leak. A reduced expression of S100A1 (S100) is linked to end stage heart failure and purported to play a role in inducing diastolic Ca²⁺ leak

Heart failure and associated arrhythmias

Heart failure is one of the leading cases of morbidity and mortality world-wide, affecting over 37 million people globally (Bui et al. 2011). Heart failure is a complex clinical syndrome, which renders the heart unable to pump enough blood to meet the body’s metabolic needs. Hemodynamically, this can manifest as an impairment of cardiac output and organ perfusion. Abnormalities associated with heart failure include left ventricular dysfunction, left ventricular dilatation, heart rhythm and conduction irregularities, and impaired Ca²⁺ cycling (Ponikowski et al. 2016). Altered SR Ca²⁺ homeostasis contributes to the depressed contractility, impaired relaxation, and propensity to arrhythmia, which typifies the failing heart. Such Ca²⁺ handling dysfunction results from the combined effects of alterations in both the expression and activity of key Ca²⁺ handling proteins (reviewed in (Lou et al. 2012)) and leads to a vicious cycle of stress-induced modification (via enhanced β-AR and ROS/RNS signalling). Crucial is the acquired arrhythmia which is responsible for at least 50% of all heart failure deaths (Kjekshus 1990) and can be linked in part to diastolic Ca²⁺ leak and increased DAD incidence.

Reduced SR Ca²⁺ release

Early heart failure studies showed a decreased Ca²⁺ amplitude duration (in the absence of any change in L-type Ca²⁺ current (Beuckelmann et al. 1992; Zhang et al. 1996; Lindner et al. 1998; Pieske et al. 1999)), a decrease in peak stretch amplitude, and a decreased SR Ca²⁺ load (Hobai and O’Rourke 2001). The reduced systolic Ca²⁺ transient appears secondary to the diminished SR Ca²⁺ store load and precipitates the contractility and cardiac output reductions observed in heart failure patients (Hobai and O’Rourke 2001). Further, the decreased Ca²⁺ transients in a rat failing heart model can be attributed to EC uncoupling in the absence of reduced Ca²⁺ store load (Gomez et al. 1997) and are linked to a decreased capacity of the L-type Ca²⁺ channel to stimulate RyR2 Ca²⁺ release and reduced responsiveness to β-AR signalling.

Independent of SR Ca²⁺ load, changes in subcellular architecture and targeting of RyR2 can lead to Ca²⁺ transient dyssynchrony and are a feature of heart failure progression. In failing myocardium, a higher proportion of RyR2 exist outside a couplon, and thus are not under direct L-type Ca²⁺ channel Ca²⁺-induced Ca²⁺ release control, which contributes to a reduced Ca²⁺ transient amplitude (Louch et al. 2004). Functional grouping of RyR2 into clusters is also disrupted during heart failure and is thought to contribute to a slower Ca²⁺ release kinetic and diastolic Ca²⁺ leak (Kolstad et al. 2018). Downregulation of junctophilin 2 occurs as part of cardiac remodelling in early stage heart failure (Minamisawa et al. 2004; Xu et al. 2007; Wei et al. 2010). Mechanistic insight into the associated Ca²⁺ mishandling has come from junctophilin 2 silencing and knockdown models. Partial silencing of junctophilin 2 in HL-1 cells caused a depression in the maximal Ca²⁺ transient amplitude (Landstrom et al. 2011). Junctophilin 2 knockdown disrupted the close localization of the L-type Ca²⁺ channel with RyR2, resulting in a loss of Ca²⁺-induced Ca²⁺ release and development of acute heart failure (van Oort et al. 2011). Together, this illustrates the importance of effective Ca²⁺-induced Ca²⁺ release in maintaining a healthy Ca²⁺ transient and contractility.

Defective SR Ca²⁺ uptake

Reduced uptake of Ca²⁺ in to the SR during diastole is a feature of heart failure (Zima and Blatter 2006). The locus of pathologic control is SERCA2a, with significant evidence for a reduction in SERCA2a mRNA and protein expression in heart failure (Houser et al. 2000), which contributes to reduced diastolic SR Ca²⁺ uptake and cardiac dysfunction. There are, however, a few notable exceptions (reviewed in (Houser et al. 2000)). Mechanistically, the decline in SR Ca²⁺ uptake can diminish the SR Ca²⁺ load and subsequent Ca²⁺ transients, resulting in hypocontractility. More recent work has linked reduced function of SERCA2a and the subsequent decline in SR Ca²⁺ to the development of beat-to-beat variability in the Ca²⁺ transient amplitude, or Ca²⁺ alternans (a recognized risk factor for lethal arrhythmic events) (Xie et al. 2008). The importance of altered SERCA2a expression and function in mediating heart failure is best illustrated in phase two clinical trials for end-stage heart failure, where patients were treated with recombinant adeno-associated viral transfer of SERCA2a gene (Zsebo et al. 2014). Enhancement of SERCA2a expression and function resulted in increased contractility and relaxation/diastole.

Compounding the consequences of declines in SERCA2a expression in heart failure are disturbances in the regulatory balance between SERCA2a and phospholamban. In particular, declining SERCA2a expression leads to an imbalance in the ratio of SERCA2a to phospholamban (Schwinger et al. 1998; Schmitt 2003), as only very modest declines in phospholamban protein expression are found in failing hearts. As phospholamban functions as a brake on SERCA2a activity, the relative increase in the ratio of phospholamban to SERCA2a contributes to the observed decline in SR Ca²⁺ uptake in failing hearts during diastole (reviewed in (Chu and Kranias 2006)). Further, phosphorylation of phospholamban at Ser16 and Thr17 by PKA and CaMKII (respectively) relieves phospholamban’s inhibition of SERCA2a, leading to increases in SR Ca²⁺ uptake (Simmerman and Jones 1998). Hypophosphorylation of phospholamban is consistently reported in heart failure and is linked to attenuation of the β-AR receptor pathway (Molenaar et al. 2007), increased protein phosphatase PP2Ce expression (which dephosphorylates Thr17 on phospholamban (Akaike et al. 2017)), and protein phosphatase 1 activity (again dephosphorylating phospholamban, reviewed in Chu and Kranias (2006)). As hypophosphorylation of phospholamban favours SERCA2a inhibition, this means that re-uptake of Ca²⁺ to the SR by SERCA2a is ultimately depressed, even if SERCA2a expression is unaltered. Overexpression of DWORF represents a promising heart failure treatment strategy and can counteract the effects of any regulatory imbalance between SERCA2a and phospholamban. Overexpression of DWORF in a mouse model of dilated cardiomyopathy prevents the development of the typical heart failure phenotype, by improving SERCA2a and overall cardiac function and preventing the initial cardiac remodelling synonymous with heart failure (Makarewich et al. 2018). Mechanistically, DWORF relieves the inhibitory effect of phospholamban on SERCA2a Ca²⁺ uptake (even in the background of an imbalanced phospholamban to SERCA2a ratio), improving relaxation time and enhancing contractility (Makarewich et al. 2018).

Less well understood—however potentially just as important—is the phospholamban-independent role of SERCA2a post-translational modification in dysfunction in heart failure. Heart failure increases oxidative stress-inducible nitric oxide synthase (Lokuta et al. 2005), substrates for modification of tyrosine, and cysteine residues. A two-fold increase in the levels of tyrosine nitration is found in human dilated cardiomyopathy patients (Knyushko et al. 2005), with tyrosine nitration known to inactivate SERCA2a and lead to slower relaxation (Lokuta et al. 2005). The post-translational covalent addition of a small ubiquitin-like modified (SUMO) protein to lysine residues is known to influence target protein activity, stability, and, in the case of SERCA2a, preserve function during diastole (Kho et al. 2011). Both the expression of SUMO and SUMO conjugation to Lys480 and Lys585 on SERCA2a are significantly reduced in the failing heart, resulting in declining SR Ca²⁺ uptake (Baumeister and Quinn 2016). Finally, SERCA2a acetylation is significantly increased in human animal heart failure models, with acetylation of Lys492 deemed responsible for the decrease in SR Ca²⁺ uptake and contractile abnormality (Gorski et al. 2019).

Whatever the mechanism, it is now established that dysfunctional SERCA2a pump activity and decreased/slower Ca²⁺ extrusion can result not only in a higher cytosolic Ca²⁺ concentration but also in a decreased SR Ca²⁺ content, so can contribute to both systolic and diastolic dysfunction.

Reduced SR Ca²⁺ buffering

In heart failure, while the expression of CSQ2 remains similar to that found in healthy tissue, altered trafficking of CSQ2 to the junctional SR has been reported (Jacob et al. 2013). Up to 40% of all CSQ2 (which is N-glycosylated during biosynthesis) was found to contain altered glycan structures, which were deemed uncharacteristic of healthy heart junctional SR (Jacob et al. 2013). The presence of the altered glycan forms is consistent with altered trafficking and localization within the rough endoplasmic reticulum of the perinuclear cisternae (McFarland et al. 2010). The subsequent reduced junctional SR CSQ2 content, and thus reduced interaction with RyR2, is consistent with a loss of Ca²⁺ buffering capacity, in addition to higher diastolic RyR2 activity (Chopra et al. 2007; Dulhunty et al. 2012).

A Ser96Ala polymorphism in HRC has been associated with higher risk of ventricular arrhythmia and sudden cardiac death in heart failure patients (Arvanitis et al. 2008). An overall increase in SR Ca²⁺ leak and DADs were reported, which can explain the arrhythmia phenotype. Further mechanistic insights were highlighted by (Zhang et al. 2014), where the reduced Ca²⁺ store load in the absence of compromised SERCA2a activity was attributed to a decline in HRC Ca²⁺ buffering capacity (Zhang et al. 2014). Overall, the extent of involvement that disrupted SR Ca²⁺ buffering has in the pathogenesis of Ca²⁺ mishandling in heart failure requires clarity.

Diastolic SR Ca²⁺ leak

A central role for RyR2 dysfunction in the pathogenesis of heart failure was first eluded to in 1987, where higher diastolic Ca²⁺ concentrations were observed in human end stage heart failure patients (Gwathmey et al. 1987). Since then, the diastolic Ca²⁺ leak hypothesis has been defined and constitutes a substrate for DADs and arrhythmia in heart failure. In a compensatory mechanism to enhance contractility and cardiac output in heart failure, sympathetic nervous system hyperactivity results in a chronic catecholamine cascade, which can induce hyperphosphorylation of RyR2 and diastolic Ca²⁺ leak. Mechanistically, RyR2 phosphorylation is consistent with the adoption of a conformation, which is energetically favourable to RyR2 opening and loss of FKBP capacity to stabilize RyR2 in the closed formation (Dhindwal et al. 2017).

PKA hyperphosphorylation of RyR2 at Ser2808 was first identified as the cause of diastolic Ca²⁺ leak in heart failure, inducing increases in cytosolic Ca²⁺ sensitivity and diastolic channel open probability, due to the dissociation of FKBP12.6 (Marx et al. 2000). Similar findings have been made by others (Reiken et al. 2003; Shan et al. 2010; Walweel et al. 2017). Supporting these findings is that inhibition of PKA-mediated phosphorylation retains FKBP12.6 binding to RyR2 and prevents heart failure onset (Wehrens et al. 2006) and that constitutive activation of PKA leads to hyperphosphorylation, hypocontractility, arrhythmia, and a higher likelihood of sudden cardiac death (Antos et al. 2001). Phosphodiesterase 4D3 (PDE4D3) protein expression and activity are reduced by approximately half in the failing heart, which correlates with hyperphosphorylation of RyR2 at Ser2808 and increased diastolic Ca²⁺ channel activity (Lehnart et al. 2005). PDE4D3 serves as a break in the G-protein/adenylyl cyclase/cAMP phosphorylation cascade, by hydrolyzing cAMP to the inactive 5’-AMP. Studies using PDE4D3 ± mice confirm these findings and provide a causative link between loss of PDE4D3 and an accelerated heart failure progression and are associated with enhanced RyR2 phosphorylation, higher channel activity, and hypocontractility (Lehnart et al. 2005).

There is not universal support, however, for the influence of PKA-mediated phosphorylation or FKBP12.6 dissociation in the development of diastolic Ca²⁺ leak. Indeed, there are a number of independent laboratories that have shown that ablating the S2808 phosphorylation site (S2808A) does not alter responses to β-AR signalling (Benkusky et al. 2007; MacDonnell et al. 2008; Zhang et al. 2012; Alvarado et al. 2017), nor does it prevent progression of cardiac dysfunction or heart failure (Zhang et al. 2012; Alvarado et al. 2017), compared with responses in wild type mice. Similarly, studies in two independent laboratories failed to induce any FKBP12.6 dissociation from RyR2 upon PKA phosphorylation (Xiao et al. 2004; Zhang et al. 2012).

There is significant support, however, for a role of CaMKII-mediated phosphorylation of RyR2 in RyR2 dysfunction in failing heart. CaMKII upregulation, triggered by β-AR stimulation or cytosolic Ca²⁺, was first reported to hyperphosphorylate RyR2 Ser2814 in heart failure by Wehrens et al. (2004). CaMKII-mediated phosphorylation of RyR2 has been shown to persist in failing heart (Ai et al. 2005) and to enhance diastolic Ca²⁺ leak and arrhythmogenic propensity in models of heart failure (Guo et al. 2006; Curran et al. 2007). This is further supported by findings in S2814A ablation models, where mice are protected from induced tachyarrhythmia (van Oort et al. 2010) and heart failure development post transverse aortic constriction (Respress et al. 2012). Together, this firmly places Ser2814 phosphorylation as a major pathogenic mechanism in the SR Ca²⁺ leak pathway. That a number of studies report RyR2 hyperphosphorylation at both Ser2814 and Ser2808 (Respress et al. 2012; Walweel et al. 2017) seems to compound the inconsistency in the literature, however may highlight the existence of aetiology-dependant induction of protein kinase activity, leading to different hyperphosphorylation sites on failing RyR2 (Respress et al. 2012; Ather et al. 2013). What is consistent, however, is that phosphorylation of Ser2808 and Ser2814 can induce aberrant diastolic activity (Ca²⁺ leak) and contribute to the development of DADs and arrhythmogenesis.

Phosphorylation of RyR2 during heart failure may also occur independent of enhanced β-AR signalling. The expression (Marx et al. 2000) and localization (Belevych et al. 2011) of the ubiquitous protein phosphatase 2a, which curtails phosphorylation, is lowered in heart failure. The reduced expression at the SR membrane thus contributes to the RyR2 phosphorylation increases and diastolic Ca²⁺ leak observed (Belevych et al. 2011). Moreover, Ser2808 hyperphosphorylation, Ca²⁺ sparks, and diastolic Ca²⁺ leak are each linked to heart failure progression in models of doxorubicin cardiotoxicity (Llach et al. 2019). In the absence of a known β-AR stimulation pathway, doxorubicin-oxidant activation of PKA (Brennan et al. 2006) is the postulated trigger.

Constituting a refinement of the diastolic Ca²⁺ leak hypothesis is evidence for an altered redox balance towards a more oxidized intracellular environment in heart failure (Terentyev et al. 2008). The resultant-enhanced oxidative modifications (primarily S-oxidation) of any of the 89 cysteine residues on RyR2 in heart failure promote RyR2 activity (reviewed in (Nikolaienko et al. 2018)). While traditionally associated with overall gain of function, it is now clear that in failing hearts, oxidative modification of RyR2 manifests as aberrant diastolic RyR2 activity, diastolic Ca²⁺ leak, arrhythmogenesis, and contractile dysfunction (Shannon et al. 2003; Kubalova et al. 2005; Belevych et al. 2007; Blayney and Lai 2009; Bers 2014; Zima et al. 2014; Walweel et al. 2017). It is important to note that the observed higher diastolic RyR2 channel activity is in the background of RyR2 hyperphosphorylation (Shan et al. 2010; Walweel et al. 2017). One mechanism linking conformational change to diastolic Ca²⁺ leak in RyR2 is that oxidation (and the associated loss of calmodulin) unzips the otherwise tight interaction between the N-terminal and central domains of the channel, contributing to the leaky RyR2 phenotype (Yano et al. 2005; Mochizuki et al. 2007; Oda et al. 2015).

Links between RyR2 redox modification and accessory protein interactions have been observed and indicate that disrupted associations may underpin redox-mediated changes in heart failure. Two studies show that association of FKBP12.0/12.6 is reduced in failing hearts with hyperoxidized RyR2 (Shan et al. 2010; Walweel et al. 2017), concurrent with increased Ca²⁺ sensitivity, which manifests as higher diastolic open probability and is reversed by thiol reduction (Walweel et al. 2017). In dogs, higher overall RyR2 S-oxidation in failing ventricles similarly corresponds with decreased FKBP12.6 association (Yano et al. 2005). Zissimopoulos et al. (2007) found some oxidants decrease RyR2/FKBP12.6 binding, with disruptions attributed to redox modification of RyR2. It should be noted that RyR2 aberrant activity upon thiol modification is observed in RyR2 expressed in stable HEK-293 cell lines (Waddell et al. 2016). FKBP12/12.6 is not associated with RyR2 in this cell line, demonstrating that FKBP dissociation alone is not the only mechanism that induces diastolic Ca²⁺ leak.

How redox-induced S-nitrosylation might influence function in heart failure has produced contradictory results. Increased S-nitrosylation is a feature of ischemic heart failure (Shan et al. 2010); however, there are opposing reports illustrating a higher overall oxidation, but hypo-S-nitrosylation of RyR2 in nitric oxide synthase-knockout mice and hypertensive heart failure (Gonzalez et al. 2007, 2010). This is consistent with a cardioprotective mechanism, whereby S-nitrosylation affords a degree of cardio-protection by shielding proteins from oxidative damage during ischemia/reperfusion injury (Kohr et al. 2011). This is suggestive of an interplay between S-oxidation and S-nitrosylation (Hare and Stamler 2005). Reminiscent of the aetiological patterns observed in RyR2 hyperphosphorylation, the underlying cause of heart failure might prompt an up- or down-regulation of the nitrosylation pathway in failing hearts (Gonzalez et al. 2007, 2009, 2010; Shan et al. 2010). Others, however, propose that RyR2 S-nitrosylation disrupts calmodulin binding and that it is this loss of calmodulin that triggers dysfunctional Ca²⁺ release through RyR2 (Ono et al. 2010; Hino et al. 2012; Oda et al. 2015). That calmodulin binding to RyR2 is reduced in heart failure is well established and linked to the spontaneous release of Ca²⁺ via RyR2. An emerging hypothesis is that the synergistic contribution of both phosphor- and ROS/RNS-derived post-translational modification of RyR2 drives RyR2 conformational change, leading to diastolic Ca²⁺ leak and an arrhythmogenic substrate (discussed in (Denniss et al. 2018)).

Finally, a reduction in expression of key regulatory proteins junctophilin 2 and S100A1 in the early and end stages of heart failure (Remppis et al. 1996; Minamisawa et al. 2004; Xu et al. 2007; Wei et al. 2010) can also contribute to diastolic Ca²⁺ leak. In addition to contributing to a reduced Ca²⁺ transient in heart failure, the downregulation of junctophilin 2 in heart failure destabilized RyR2 in the closed state (van Oort et al. 2011). Reduced junctophilin can lead to spontaneous Ca²⁺ oscillations and a higher Ca²⁺ spark frequency (Landstrom et al. 2011; van Oort et al. 2011) and would contribute to diastolic Ca²⁺ leak in the early stages of heart failure. Downregulation of S100A1 in end stage heart failure drives dysregulated SR Ca²⁺ cycling and deterioration in contractile function (Most et al. 2004). The significance of this contribution to diastolic Ca²⁺ leak is unknown; however, restoration of S100A1 levels via gene delivery can rescue much of the heart failure phenotype (reviewed in (Rohde et al. 2011)), improving overall SR Ca²⁺ cycling, and reducing pathological SR Ca²⁺ leak.

Conclusion

Just as Ca²⁺ cycling around the SR assumes critical importance in facilitating normal hemodynamic function, it is clear that aberrant Ca²⁺ SR cycling is a central tenet of pathological dysfunction in the failing heart. Such dysfunction manifests as cardiac insufficiency due to hypocontractility and, crucially, a propensity for arrhythmia, which culminates in an increased incidence of sudden death due to a major cardiac event. Research to date has firmly established that changes around the release and reuptake of SR Ca²⁺ due to detrimental remodelling of RyR2 and SERCA2a function assume particular significance in the progression to an arrhythmogenic phenotype. There is compelling evidence that such remodelling is largely the result of the combined effects of expression changes and major disruptions to RyR2 and SERCA2a regulatory pathways due to aberrant post-translational modification and altered co-protein interactions. While there have been advances in the treatment of detrimental downregulation of SERCA2a expression using gene therapy, there remains much to clarify in terms of our understanding of the relative significance of modification and co-protein disruptions which affect SR Ca²⁺ cycling. Further research will undoubtedly shed light on these complex regulatory relationships and clarify potentially important therapeutic targets.

Authors’ contributions

All authors contributed to the review conception and design. The first draft of the manuscript was written by Nicole Beard and Amanda Denniss. All authors contributed and commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding information

Amanda Denniss is supported by a Research Training Program stipend. Alexander Dashwood is supported by a University of Queensland Graduate School Scholarship. Nicole Beard and Peter Molenaar are supported funded by a National Heart Foundation Vanguard Grant.

Data availability

Not applicable.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest

.

Ethics approval

Not applicable.

Consent to participate

Not applicable.

Consent for publication

Not applicable.