Ca²⁺ Cycling in Heart Failure

Abstract

Ca²⁺ plays a crucial role in connecting membrane excitability with contraction in myocardium. The hallmark features of heart failure are mechanical dysfunction and arrhythmias; defective intracellular Ca²⁺ homeostasis is a central cause of contractile dysfunction and arrhythmias in failing myocardium. Defective Ca²⁺ homeostasis in heart failure can result from pathological alteration in the expression and activity of an increasingly understood collection of Ca²⁺ homeostatic binding proteins, ion channels and enzymes. This review focuses on the molecular mechanisms of defective Ca²⁺ cycling in heart failure and consider how fundamental understanding of these pathways may translate into novel and innovative therapies.

Among the many causes of myocardial injury that can lead to CHF, myocardial infarction is the most common in the developed world ¹. The hallmark features of heart failure include reduced contractile function manifested as blunted, slowed, dysynchronous contraction and impaired relaxation. The physiological positive force-frequency relationship and increased myocardial contractile response to increased preload is compromised in heart failure ². The failing heart attempts to compensate for injury by various mechanisms, such as myocardial hypertrophy, increasing filling pressure and enhanced neurohumoral signals, which together drive a feed forward pathophysiological spiral leading to adverse ventricular remodeling and electrical instability ³. Each of these maladaptive events is associated with loss of myocardial Ca²⁺ homeostasis.

I. Ca²⁺ homeostasis and mechanisms underlying excitation-contraction coupling

Ca²⁺ plays a crucial role in coupling cell membrane excitation and contraction, so-called excitation-contraction coupling (ECC) (Figure 1). Cardiac contraction depends on a transient increase in the cytosolic Ca²⁺ concentration ([Ca²⁺]_i) to activate cross bridge formation between myofilament proteins that ultimately elicits pressure development in the cardiac chambers and provides energy for ejection of blood. Cardiomyocytes are packed with myofibrils enveloped in a network of Ca²⁺ storing sarcoplasmic reticulum (SR)⁴ and mitochondria. ECC in ventricular myocytes is built around dyads, specialized membrane ultrastructure formed by the terminal cisternae of the SR and invaginations of the cell membrane called transverse tubules. Voltage-gated ion channels, exchangers and Na⁺/K⁺ ATPase pump proteins are enriched on the transverse tubular membranes and colocalize with the intracellular ryanodine receptor (RyR2) Ca²⁺-release channels, which are clustered on the SR membrane. ECC is initiated when the cell membrane action potential invades the myocyte along its transverse tubules. The flow of inward current depolarizes the cell membrane and rapidly (in 1–2 ms) opens voltage-gated Na⁺ channels (mostly Na_V1.5) that are responsible for a large inward Na+ current (I_Na). I_Na rapidly inactivates (1–2 ms) and Na_V1.5 channels remain inactive until the action potential is complete and the cell membrane returns to a negative resting potential (~−90 mV). The inward I_Na depolarizes the cell membrane, reaching a cell membrane potential that is permissive for opening voltage-gated Ca²⁺ channels (mostly Ca_V1.2 in ventricular myocardium). Inward Ca²⁺ current (I_Ca) 5 triggers opening of RyR2 channels by a Ca²⁺-induced Ca²⁺ release process⁶, resulting in coordinated release of SR Ca²⁺ that contributes the major portion of myofilament-activating Ca²⁺. The I_Ca contributes to the long action potential plateau (200–400 ms) characteristic of ventricular myocytes in humans ⁷. The Ca²⁺ released from the SR diffuses over a very short distance to engage the adjacent myofibrils, binding to troponin C of the troponin-tropomyosin complex on the actin filaments in sarcomeres, which moves tropomyosin away from the binding sites, facilitating formation of cross bridges between actin and myosin to enable myocardial contraction. I_Ca inactivates by voltage and [Ca²⁺]_i-dependent mechanisms ⁸ at the same time that voltage-gated K⁺ channels open to allow an outward current that orchestrates action potential repolarization, establishing conditions required for relaxation ⁷.

Fig 1. Ca²⁺ homeostasis and Excitation Coupling (ECC).

The ECC process is initiated when an action potential (AP) excites the myocyte cell membrane (sarcolemma) along its transverse tubules. This depolarization rapidly opens voltage-gated Na⁺ channels (mostly Na_V1.5) that further depolarize the cell membrane, allowing opening of voltage-gated Ca²⁺ channels (mostly Ca_V1.2). Inward Ca²⁺ current triggers opening of ryanodine receptor (RyR2) channels by a Ca²⁺-induced Ca²⁺ release process, resulting in coordinated release of sarcoplasmic reticulum (SR) Ca²⁺ that contributes the major portion of the myofilament-activating increase in [Ca²⁺]_i. The Ca²⁺ released from the SR binds to troponin C of the troponin-tropomyosin complex on the actin filaments in sarcomeres, facilitating formation of cross bridges between actin and myosin and myocardial contraction. Voltage-gated K⁺ channels open to allow an outward current that favors action potential repolarization, establishing conditions required for relaxation. Relaxation occurs when Ca²⁺ is taken back up into the SR through the action of the SR Ca²⁺ adenosine triphosphatase SERCA2a and is extruded from the cell by the sarcolemmal Na⁺ and Ca²⁺ exchanger (NCX). SERCA2a is constrained by phospholamban (PLN) under resting conditions.

Cardiac relaxation depends on a decrease in [Ca²⁺]_I that is permissive for unbinding of myofilament crossbridges. Sequestration of cytoplasmic Ca²⁺ occurs mainly through active Ca²⁺ uptake by the SR, through the sarcoplasmic-endoplasmic reticulum Ca²⁺ ATPase (SERCA2a) ⁹, and to a lesser extent by extrusion to the extracellular space by the Na⁺/Ca²⁺ exchanger (NCX) ¹⁰, the sarcolemmal Ca²⁺ ATPase ¹¹ and mitochondria ¹². The binding of Ca²⁺ rapidly activates NCX, which facilitates Ca²⁺ efflux into the extracellular milieu using the energy from the cell membrane Na⁺ gradient established by the Na⁺/K⁺ ATPase. NCX generates a current because it exchanges 3Na⁺ for 1Ca²⁺, a net positive charge. Depending upon the electrochemical gradient, NCX current may be inward (forward mode), extruding cytoplasmic Ca²⁺ to the extracellular space, or outward (reverse mode), importing extracellular Ca²⁺ to the cytoplasm. Thus, Ca²⁺ cycling between the extracellular space, cytosol and SR allows rapid contraction and relaxation of the heart.

II. Defective ECC and alterations of Ca²⁺ handling proteins in heart failure

Consistently, cardiomyocytes from failing heart show defective ECC characterized by decreased [Ca²⁺]_i transients, enhanced diastolic SR Ca²⁺ ‘leak’ and diminished SR Ca²⁺ sequestration, events that contribute to impaired contractility and relaxation ¹³. These abnormalities are due to alterations of a collection of key Ca²⁺ handling proteins.

Impaired SR Ca²⁺ release contributes to systolic heart failure

Ca_V1.2/Na_V1.5

Voltage-dependent opening of L-type calcium channels (LTCC) enables cellular Ca²⁺ entry that triggers Ca²⁺-induced Ca²⁺ release from the SR by promoting RyR2 opening, leading to myofilament cross-bridge formation and mechanical force development. The cardiac action potential plateau in ventricular myocytes is optimized for grading Ca_V1.2 openings to initiate Ca²⁺-induced Ca²⁺ release and ECC. Similar to all known voltage-gated ion channels, Ca_V1.2 consists of a pore forming α subunit, auxiliary subunits and connections to various cytoskeletal proteins ¹⁴. Protein kinase A (PKA), protein kinase C (PKC) and the multifunctional Ca²⁺ and calmodulin-dependent protein kinase II (CaMKII) are serine-threonine kinases that catalyze ATP-dependent phosphorylation of Ca_V1.2 proteins ¹⁵ (Figure 2). CaMKII ¹⁶ and PKA ¹⁷ increase the frequency of prolonged Ca_V1.2 openings, while the functional significance of PKC actions at Ca_V1.2 are less clear ¹⁵. These prolonged and frequent Ca_V1.2 channel openings are due to mode 2 Ca_V1.2 gating, a biophysical response shared with β adrenergic receptor (β–AR) agonists, CaMKII and the dihydropyridine agonist BayK 8644 ^16,17,18. Phosphorylation by CaMKII or by PKA, the principal kinase activated by β AR (β–AR) agonists, collaborates with cell membrane potential to enhance the probability of Ca_V1.2 opening. Mode 2 gating appears to underlie I_Ca facilitation, a dynamic pattern of increasing peak I_Ca and slowed I_Ca inactivation ¹⁹. Mode 2 gating and I_Ca facilitation are proarrhythmic, in part, by favoring early afterdepolariazations (EADs)^{20, 21,16}.

Figure 2. Regulation of [Ca²⁺]_i homeostasis by Ca²⁺ binding proteins and kinases.

Regulation of Ca²⁺ homeostasis involves a multitude of Ca²⁺ binding proteins and enzymes, including CaMKII, PKC, PKA and S100A1: (1). CaMKII catalyzes phosphorylation of voltage-gated Ca²⁺ channels (mostly Ca_V1.2 in ventricle) to increase Ca²⁺ entry, RyR2 to increase Ca²⁺ release, voltage-gated Na⁺ channels (mostly Na_V1.5 in ventricle) to increase subsarcolemmal [Na+]_i,, which decreases the driving force for Ca²⁺ extrusion by the Na⁺/Ca²⁺ exchanger (NCX), and PLN to reduce the inhibitory activity of PLN on SERCA2a. In general, the increased phosphorylation of these proteins by CaMKII increases Ca²⁺ influx, and storage by the SR, which leads to increased systolic [Ca²⁺]_i and increased rate and magnitude of force (pressure) generation and improved lusitropy. (2) PKA is activated by β–AR agonists and catalyzes phosphorylation of the same Ca²⁺ regulatory proteins modified by CaMKII, but at different amino acids. (3) Classical PKC isoforms are activated downstream to a variety of G protein coupled receptors and are activated by increased [Ca²⁺]_i, leading to decreased activity SERCA2 by phosphorylating inhibitor 1 (I-1) resulting in PLN dephosphorylation, reducing SR Ca²⁺ load and Ca²⁺ release, causing reduced contractility. (4) S100A1 interacts with the SERCA2a/PLN complex in a Ca²⁺-dependent manner to augment SR Ca²⁺ uptake and increase SR Ca²⁺ content. S100A1 also directly regulates RyR2 function, stimulates ATP synthase activity and promotes the adenosine nucleotide translocator (ANT) function to increase ATP synthesis and mitochondrial ATP efflux in cardiomyocytes.

Elevated [Na⁺]_i is present in failing myocardium from humans ^{22,23,24. 25} Changes in [Na⁺]_i may have a large impact on [Ca²⁺]i homeostasis²⁶. Small increases in [Na⁺]_i may increase Ca²⁺ influx via reverse mode NCX during systole and limit Ca²⁺ extrusion via forward mode NCX during diastole, leading to increased subsarcolemmal [Ca²⁺]_i ^27,28. Hence, Increased [Na⁺]_I levels lead to Ca²⁺ overload, contributing to arrhythmias and impaired diastolic function ²². The major pathway for Na⁺ influx in cardiomyocytoes is through voltage-gated Na⁺ channels, primarily Na_V1.5, which open and close rapidly (1–10 ms) to trigger the upstroke of action potential depolarization in working myocardium. CaMKII associates with and phosphorylates the Na_V1.5 α subunit at a ‘hot spot’ in the cytoplasmic I–II linker domain, an event that promotes a non-inactivating, long-lasting component of I_Na (I_NaL) and arrhythmia-triggering EADs and delayed afterdepolarizations (DADs)^{29, 30}. CaMKII inhibition reverses the increase of I_NaL in heart failure ³¹, suggesting that Na_V1.5 is an important target for the antiarrhythmic effect of CaMKII inhibition ³². [Na⁺]_i is also maintained by the Na+/K+ ATPase pump. It was reported that in failing human hearts the tissue concentration of the Na⁺/K⁺ ATPase pumps are reduced ³³. Whether the functional capacity of the Na⁺/K⁺ ATPase pump in heart failure is altered remains inconclusive, as some studies show unaltered maximum transport rate and affinity for Na⁺ in a rabbit heart failure model ³⁴ whereas the Na⁺/K⁺ ATPase pump was reduced in a rat heart failure model ³⁵.

Reduced SR Ca²⁺ release and increased RyR2 opening probability

RyR2, the largest ion channel protein (560 kDa), exists as a homotetramer (~2.2 MDa). The predominant isoform expressed in cardiac muscle is RyR2 ³⁶. RyR2 works as a multi-protein Ca²⁺ release unit where the RyR2 Ca²⁺ channel is composed of four membrane spanning subunits ³⁷ coupled to various regulatory proteins. Calsequestrin, triadin 1, and junctin bind to RyR2 at the luminal SR membrane face where they transmit information about SR Ca²⁺ content to RyR2 ³⁸. It is known that congenital mutations in RyR2, calsequestrin and triadin can cause increased SR Ca²⁺ leak, disorganized diastolic Ca²⁺ release, arrhythmias and sudden death ^39,40.

Under physiological conditions RyR2 opening probability is increased by the cytoplasmic Ca²⁺ trigger from I_Ca ⁴¹. RyR2 activity is also regulated by multiple factors, including PKA, CaMKII, protein phosphatases 1 and 2A, calmodulin, and FKBP12.6, which are associated with the cytoplasmic face of RyR2. The Marks group demonstrated that PKA phosphorylates RyR2 ⁴² which enables the ‘fight or flight’ response by increasing RyR2 opening probability and [Ca²⁺]_i⁴³. They also showed that hyperphosphorylation of RyR2 by PKA (at serine 2808) causes an FKBP12.6-RyR2 dissociation, increased RyR2 opening probability and SR Ca²⁺ leak in human ^42,44 and animal models of CHF ^{45, 46,47, 48}. In addition their results also suggest that improved cardiac function by β-AR antagonist drugs in failing human heart is associated with restoration of FKBP12.6 levels and repair of RyR2 channel leak ⁴⁴. However, other groups reported conflicting results that PKA does not increase RyR2 phosphorylation ⁴⁹ and that phosphorylation at the S2808 site does not mediate β-AR agonist induced cardiac response ^50,51 or dysfunction after myocardial infarction⁵². These highly controversial results⁵³ indicate that alternative mechanisms may also be important for RyR2 dysfunction in heart failure.

CaMKII is activated by β-AR agonist stimulation ⁵⁴ and increased ROS ⁵⁵ and can phosphorylate RyR2 at two sites: serines 2809 and 2814 ⁵⁶, although the 2814 site appears to be preferred ⁵⁷. CaMKII-dependent RyR2 phosphorylation increases diastolic SR Ca²⁺ release ⁵⁸. Mice genetically lacking serine 2814 (S2814A) have an impaired force-frequency relationship ⁵⁹ and are resistant to MI-induced heart failure and arrhythmias^60,61. It was also shown that oxidative stress generated in the failing heart could directly alter RyR2 function by post-translational modification causing its increased sensitivity to activation by luminal Ca^{2+ 62}. A growing body of evidence suggests that reduced Ca²⁺ release in failing cardiomyocytes is a result of increased and improperly regulated activity of multiple Ca²⁺ handling proteins including Ca_V1.2, Na_V1.5 and RyR2, all of which appeared to be targets of CaMKII.

Impaired Ca²⁺ sequestration during diastole

To achieve relaxation, cytosolic Ca²⁺ must be sequestered, mainly to the SR by SERCA2a ⁹. Diastolic [Ca²⁺]_i is increased in human heart failure, a condition that is likely related, at least in part, to defects in cytosolic Ca²⁺ removal ⁶³. Taken together with loss of physiological SR Ca²⁺ release, elevated diastolic [Ca²⁺]_i results in reduced contractile force, impaired relaxation, and abnormal force–frequency relationship in human heart failure. The sarcomere is the primary functional unit of cardiac muscle that is responsible for contraction and force generation. During diastole sarcomeres are typically quiescent and show uniform lengthening. However, in the failing heart sarcomere uniformity is lost ⁶⁴. Failing myocardium is marked by spontaneous diastolic SR Ca²⁺ release, leading to spontaneous and highly variable diastolic sarcomere contractions ⁶⁴, which significantly reduces contractile force ⁶⁵ and contributes to the loss of inotropic effects in CHF ⁶⁶.

SR Ca²⁺ uptake is impaired in the failing human heart ^67,68, an outcome that is due to several mechanisms. First, there are reduced expression and activity of SERCA2a in failing human heart ^{69, 70}. However, in some human failing hearts SERCA2a expression or activity is normal ^71,72. Over-expression of SERCA2a can restore the Ca²⁺ handling and the contractile function in animal models ⁷³ and in human heart failure ^{74, 75}, suggesting that repairing SERCA2a expression may be a viable therapy for CHF. Defects in SR Ca²⁺ release may be due to loss of normal ‘gain’ of ECC, a condition where a given I_Ca trigger elicits a lesser amount of SR Ca²⁺ release ⁷⁶. Comparisons of ECC gain require experimental conditions that control for SR Ca²⁺ content. Nevertheless, failing human cardiomyocytes may have preserved fractional SR Ca release¹³ despite reduced SR Ca²⁺ pump activity, SR Ca²⁺ content and systolic [Ca²⁺]_i transients, suggesting that defects in ECC gain are not an obligate aspect of failing myocardiocytes.

Second, reduced SR Ca²⁺ uptake could be due to increased inhibitory activity of PLN ^{77, 78}. PLN inhibits SERCA2a in its dephosphorylated form whereas in its phosphorylated form (by PKA at serine-16 and CaMKII at threonine-17)⁷⁹ PLN assembles into a pentamer that lacks SERCA2a inhibitory activity.

Multiple studies suggest that phosphorylation of PLN is decreased in the failing human heart, accounting for increased inhibition of SERCA2a ^{78, 80}. For example, phosphorylation of PLN at threonine 17 is decreased in ventricular myocardium due to increased dephosphorylation by protein phosphatase 2B (PP2B), also called calcineurin, despite increased activity of CaMKII in failing myocardium ⁸¹. PLN phosphorylation at serine 16 is decreased due to increased activity of Type 1 protein phosphatase (PP1) in the failing human heart ⁷⁸. Several mutations in the human PLN gene (such as R9L, R9H, L39stop)⁸² have been identified that provide important insights into PLN regulation of SERCA2a. Two mutations (R9C and R14del) result in enhanced inhibition of SERCA2 by PLN, partly due to decreased PKA-mediated phosphorylation ^{83, 84}. The phenotypes of R9C or R14del carriers include dilated cardiomyopathy and premature death ^{83, 84}.

Another human mutation causing loss of function of PLN (Leu 39 stop) and uninhibited SERCA2a activity also results in dilated cardiomyopathy and premature death ⁸⁵. Genetic manipulation of PLN in mouse models yielded similar and contrasting results compared to human mutations. PLN-KO mice showed enhanced cardiac contractile function with increased affinity of SERCA2a for Ca²⁺, consistent with the concept that PLN down-regulates myocardial contractility by suppressing SERCA activity ⁸⁶. PLN knockout prevented heart failure in a mouse model of dilated cardiomyopathy caused by deficiency of the muscle-specific LIM protein (MLP) ^87,88. Gene therapy with antisense against PLN improved contractile and diastolic function in isolated failing human cardiomyocytes ⁸⁹. However, PLN knockout in mice with severe cardiomyopathy due to transgenic over-expression of CaMKII improved SR Ca²⁺ content and myocardial contraction but nevertheless increased mortality, mitochondrial Ca²⁺ and myocardial cell death ⁹⁰. Taken together, these studies in mice and humans suggest that SERCA2a/PLN activity needs to be maintained within certain boundaries to support physiological function and prevent cardiomyopathy.

Another emerging regulator of SERCA activity is the Histidine-Rich Ca²⁺ Binding Protein (HRC), a low-affinity, high- capacity Ca²⁺-binding protein located in the SR lumen ⁹¹. HRC also affect RyR function through its binding to triadin and it was suggested that HRC may mediate a cross-talk between SR Ca²⁺ -uptake and release. A human HRC variant (S96A) with substitution of Ala in position 96 is associated with life-threatening ventricular arrhythmias in dilated cardiomyopathy patients accompanied by a reduced [Ca²⁺]_i transient and a prolonged decay time ⁹². Transgenic overexpression of HRC in the heart decreases SR Ca²⁺ uptake rates, suggesting that HRC inhibit SERCA2a and intracellular Ca²⁺ cycling and promote progression to heart failure⁹³. These studies suggest an important role of HRC in maintaining Ca²⁺ homeostasis in the SR.

The relative contribution of NCX to cytoplasmic Ca²⁺ sequestration is increased in failing myocardium, probably due to the depressed SR Ca²⁺ uptake ⁹⁴. Expression of NCX in human CHF has been reported to increase ¹⁰ or be unchanged ⁹⁵. Because subsarcolemmal [Na⁺]_i is increased in failing ventricular myocytes, NCX current (I_NCX) current shifts from inward to outward 96, which contributes to prolonged cytoplasmic [Ca²⁺]_i transients, Ca²⁺ overload and diastolic dysfunction ^22,96,97. Thus, enhanced I_NCX may be adaptive to defects in SERCA2a/PLN in CHF, while also contributing to subsarcolemmal [Na⁺]_i and [Ca²⁺]_i overload in CHF.

ATP, mitochondrial Ca²⁺ uptake and retention

Adenosine triphosphate (ATP) is the predominant form of readily available energy in myocardium, which consumes about 6 kg of ATP daily ⁹⁸. The Ca²⁺ concentration gradient between the extracellular and intracellular environments is massive, with approximately 10,000 fold higher extracellular than bulk cytoplasmic (^~100 nM)⁹⁹ [Ca²⁺]_i; Maintaining Ca²⁺ homeostasis constitutes a major ATP cost for cardiomyocytes. SERCA2a and the Na⁺-K⁺ ATPase are amongst the largest energy consuming proteins ¹⁰⁰. A proper equilibrium between Ca²⁺ cycling and ATP production must be maintained to ensure proper intracellular Ca²⁺ handling and a physiological range of myocardial performance ^101,102. Mathematical modeling^{103, 104} and experiments in excised myocardial cell membrane patches using the ATP sensitive K⁺ current (I_KATP) as a read out for subsarcolemmal ATP^{103, 104} support a view that ATP availability can be rate limiting under stress conditions, due to high local ATP consumption and compartmentalization. Thus, it is plausible that subcellular domains of ATP deficiency contribute to myocardial dysfunction in CHF.

CHF is associated with abnormal energy metabolism including decreased energy production and impaired energy utilization ^105–107, which appear to adversely affect [Ca²⁺]_i homeostasis ^107,101. On one hand, reduced ATP/ADP ratio, due to mitochondrial dysfunction, caused impaired function of SERCA2a in animal models of CHF ¹⁰⁸. On the other hand, Ca²⁺ transport regulates ATP production in mitochondria ^{109, 110}. Some validated clinical therapies for CHF improve myocardial energetics and normalize [Ca²⁺]_i homeostasis. For example, β-AR antagonists were designed by Sir James Black, in part, to reduce myocardial O₂ consumption with a goal of preventing myocardial infarction ¹¹¹. β-blockers, which decrease energy consumption, have been shown to normalize the contractile function and Ca²⁺ handling in failing human hearts ^{112, 113}. Left ventricular assist devices, which decrease the workload of the heart, improved Ca²⁺ handling in CHF patients ^{14, 114}. Restoration of mitochondrial Ca²⁺ homeostasis by unloading mitochondrial Ca²⁺ restored cardiac energetics including ATP synthesis ¹¹⁵. Thus, CHF appears to be a condition that arises, at least in part, by interrelated defects in [Ca²⁺]_i homeostasis and metabolism and successful CHF therapies often restore physiological [Ca²⁺]_i homeostasis and metabolism.

Mitochondrial Ca²⁺ regulates cell metabolism and cell death

Mitochondria comprise about 20–30%¹¹⁶ of cardiac mass where they are essential for providing ATP to meet the heightened energy demand for cardiac function. Ca²⁺ appears to be a critical second messenger for communicating cellular energy demands to mitochondria for the purpose of matching ATP production by oxidative phosphorylation with metabolic requirements ¹¹⁰. Oxidative phosphorylation is a Ca²⁺ regulated process, as Ca²⁺ increases the activity of key tricarboxylic acid dehydrogenases involved in producing reducing equivalents (NADH/NADPH) for electron transport ¹¹⁷. Metabolic regulation by mitochondrial Ca²⁺ uptake, however, is not limited to the effects on dehydrogenases. The aspartate/glutamate exchangers located at the inner mitochondrial membrane have Ca²⁺ binding domains, which support increased ATP production in response to local and temporal Ca²⁺ signals ^{118, 119}. Furthermore, the close physical association between mitochondria, SR and plasma membrane Ca²⁺ channels ensures prompt Ca²⁺ transfer to the mitochondrial matrix, which stimulates oxidative phosphorylation in response to activation of ATP-consuming processes in the cytosol ^{120, 121}.

Compared to the SR, mitochondria have a lower affinity but a higher capacity for taking up Ca²⁺. Mitochondria constitute an important buffer for cytoplasmic Ca^2+122,120, but excessive accumulation of mitochondrial Ca²⁺ causes mitochondrial damage and myocardial death ¹²³ (Figure 3). Excessive mitochondrial [Ca²⁺] ([Ca²⁺]_m) and ROS ¹²⁴ trigger mitochondrial permeability transition pore (mPTP) opening and subsequent dissipation of inner mitochondrial membrane potential (ΔΨm) ¹²⁵ and release of apoptotic mediators such as cytochrome C ¹²⁶, leading to cell death ^{127, 128}. The mPTP appears to be an important but incompletely understood target for CaMKII ¹²⁹. Our group recently reported that cardiomyocytes from mice with transgenic expression of a mitochondrial-targeted CaMKII inhibitory protein (CaMKIIN) ¹³⁰ were able to sustain higher mitochondrial Ca²⁺ entry prior to mPTP opening and were resistant to programmed cell death from ischemia/reperfusion-, catecholamine- and myocardial infarction-related injury, suggesting that CaMKII promotes mPTP opening and myocardial death ¹³¹(Figure 3).

Figure 3. A scenario for mitochondrial Ca²⁺ overload, impaired metabolism and cell death in heart failure.

The mitochondrial Ca²⁺ uniporter ¹³⁵ is a Ca²⁺ selective channel residing in the inner mitochondrial membrane. MCU is a phosphorylation substrate for CaMKII. Mitochondrial CaMKII inhibition reduces MCU current, increases mitochondrial Ca²⁺ retention capacity and is protective against myocardial death in response to ischemia-reperfusion injury, myocardial infarction and toxic doses of isoproterenol. Excessive mitochondrial Ca²⁺ and ROS trigger mitochondrial permeability transition pore (mPTP) opening, leading to cell death. Mitochondria Ca²⁺ overload also promotes ROS generation, which could oxidize CaMKII (ox-CaMKII) and cause sustained activation of CaMKII. ox-CaMKII could enhance MCU activity and further increase mitochondrial Ca²⁺ overload, promoting mPTP opening and impairing energy metabolism in heart failure. At the same time, myocardial energy deficiency could adversely affect [Ca²⁺]_i homeostasis.

Mitochondria are considered a key source for pathological increases in ROS, mainly as a result of electron transport chain uncoupling at the level of complexes I and III^{124, 132}. On one hand, oxidative stress could damage mitochondrial DNA and proteins by forming oxidative adducts, leading to mitochondrial dysfunction, impairing myocardial energetics in heart failure. On the other hand, in heart failure impaired mitochondrial bioenergetic function with depressed electron transport systems could cause increased oxidative stress ^{133, 134}. Thus, mitochondrial dysfunction and ROS are tightly linked elements of an interdependent, feed forward circuit that promotes the pathogenesis of heart failure.

Mitochondrial Ca²⁺ uniporter

The mitochondrial Ca²⁺ uniporter ¹³⁵ is a Ca²⁺ selective channel residing in the inner mitochondrial membrane and the major mitochondrial Ca²⁺ entry pathway ^136–138. MCU can be located in close proximity to the SR ¹³⁹ and thus exposed to high [Ca²⁺] (~20–50 μM) ¹⁴⁰. Although the existence of the MCU was established over 50 years ago ¹⁴¹, it was not until very recently that the molecular identity of MCU was discovered. MCU consists of 2 predicted membrane-spanning domains with a linker/pore loop to form a functional channel^{137, 138}. Over-expression of MCU increases cell death in response to challenge by pro-apoptotic stimuli,¹³⁸ whereas suppressing MCU with Ru360, a pharmacological antagonist related to ruthenium red, protects against ischemia-reperfusion injury ¹⁴². We recently found that MCU is a phosphorylation substrate for CaMKII and that CaMKII mediated increases in MCU current (I_MCU) required serines 57 and 92 when MCU was expressed heterologously, while mitochondrial-targeted CaMKII inhibition reduced I_MCU in myocardium ¹³¹. The role of CaMKII signaling to MCU in heart failure is uncertain at this time, but mitochondrial CaMKII inhibition is protective against myocardial death in response to ischemia-reperfusion injury, myocardial infarction and toxic doses of isoproterenol ¹³¹, suggesting protective effects of mitochondrial CaMKII inhibition may be mediated, at least in part, by reducing I_MCU.

The MICU1 is a MCU binding partner that has a single membrane spanning domain and 2 Ca²⁺ binding EF hand domains ^{137, 143}. Some recent data suggest that MICU1 is essential for setting the Ca²⁺ dependence of I_MCU ^{138, 143} and preserving normal [Ca²⁺]_m by acting as a gatekeeper for Ca²⁺ uptake and preventing mitochondrial Ca²⁺ overload and excessive oxidative stress. ¹⁴⁴ In addition, MCUR1 (mitochondrial calcium uniporter regulator 1) was also recently shown to be required for MCU-dependent mitochondrial Ca²⁺ uptake and maintenance of normal cellular bioenergetics ¹⁴⁵. Thus, MCU appears to be a Ca²⁺ and CaMKII-regulated ion channel associated with various accessory protein subunits.

Very few studies have investigated whether or how mitochondrial Ca²⁺ uptake, transport and homeostasis are altered in heart failure. Limited indirect evidence suggests that mitochondrial Ca²⁺ uptake is reduced in failing cardiac myocytes because there is reduced open probability of Ca²⁺ conductance pathways in mitoplasts isolated from failing myocardium and decreased ΔΨm¹⁴⁶, the electrical driving force for mitochondrial Ca²⁺ uptake ¹⁰⁷. There is an emerging view that defective cytosolic Na⁺ and Ca²⁺ homeostasis affects mitochondrial Ca²⁺ transport in heart failure. Mitochondrial Ca²⁺ efflux is mainly enabled by the mitochondrial Na⁺/Ca²⁺ exchanger (mNCE) ¹⁰². Elevated [Na⁺]_i stimulates mNCE and mitochondrial Ca²⁺ efflux and reduces steady-state [Ca²⁺]_m ¹⁴⁷. Thus, mitochondria are a critical interface between Ca²⁺, metabolism and are key determinants of myocardial survival in response to clinically relevant forms of pathological stress. A growing body of evidence suggests that mitochondria play a central role in heart failure.

Transverse tubules

Transverse tubules are deep invaginations of the ventricular myocyte cell membrane (sarcolemma) where voltage-gated Ca²⁺ channels are richly expressed and tightly coupled with SR RyR2, forming dyads to enable Ca²⁺-induced Ca²⁺ release. There is emerging evidence that normal transverse tubular ultrastructure is disrupted in heart failure ^148,149. Transverse tubules can become spatially dispersed, leaving RyRs “orphaned” from their dyadic association with Ca_V1.2 ¹⁵⁰, which impairs Ca²⁺-induced Ca²⁺ release. In addition, Ca²⁺ transients in these regions will depend on Ca²⁺ diffusion and propagated Ca²⁺ release, thus contributing to dysynchronous Ca²⁺ sparks, inefficient ECC and a propensity toward arrhythmias. Recent studies suggest that junctophilin 2 may play a crucial role in maintenance of normal transverse tubular ultrastructure^{151, 148} and association of Ca_V1.2 with RyR2^{151, 152}, while targeted suppression of microRNA, which inhibits junctophilin, prevents disruption of T tubule structure and transition to heart failure from hypertrophy ¹⁵³. β-AR antagonists ¹⁵⁴ and sildenafil ¹⁵⁵ can defend against transverse tubular disruption in animal models of heart failure. Thus, improved understanding of the interface between membrane and regulatory cytoskeletal proteins may lead to new therapeutic targets to preserve cellular architecture that is required for physiological Ca²⁺ homeostasis.

Myofilament and cytoskeletal proteins

Abnormal Ca²⁺ homeostasis and myofilament function impairs cardiac contractile function and triggers ventricular arrhythmias in heart failure ¹⁵⁶. Ankyrins are adapter proteins that attach membrane proteins to the spectrin-actin based membrane skeleton and thus intimately involved in ion channel and transporter signaling complexes in the cardiovascular system ¹⁵⁷. Ankyrin dysfunction has been linked with abnormal ion channel and transporter membrane organization and human arrhythmias ^{158, 159}. Genetic defects in ankyrins cause altered Na⁺ and Ca²⁺ transport and enhanced RyR2 openings contributing to loss of [Ca²⁺]_i homeostasis ¹⁶⁰, activation of CaMKII and arrhythmias¹⁶¹. It was recently reported that ankyrin B plays a cardioprotective role against ischemia induced cardiac dysfunction and ankyrin-B levels are decreased in human heart failure ¹⁶².

Titin is a large myofilament protein that spans half of the sarcomere and functions as a molecular spring that provides passive stiffness to cardiac myocytes ¹⁶³. Titin isoform composition and phosphorylation regulates myocardial diastolic function ¹⁶³. Titin expression was reported to be increased in pressure-overload hypertrophy but decreased in decompensated CHF ^{164, 165}, suggesting that titin could contribute to the loss of compliance and decreased contractile function featured in heart failure. Titin knockout mice demonstrated reduced SR Ca²⁺ uptake accompanied by reduced levels of PLN and SERCA2a and developed cardiac hypertrophy and heart failure ¹⁶⁶. CaMKII phosphorylates titin and modulates passive force generation in normal and failing myocardium ¹⁶⁷. Deranged CaMKII-dependent titin phosphorylation occurs in heart failure and contributes to altered diastolic stress ¹⁶⁷. These findings suggest that titin is a participant in Ca²⁺-related defects in heart failure and suggest that titin could emerge as a target for future heart failure therapies.

Dystrophin is a cytoplasmic protein and a crucial part of the dystroglycan complex, which consists of tightly associated transmembrane and cytoskeletal proteins that serve to connect the cytoskeleton to the extracellular matrix ¹⁶⁸. Mutation of the dystrophin gene and absence of dystrophin causes Duchnne muscular dystrophy (DMD), a fatal X-linked disease¹⁶⁹, which results in a skeletal as well as a dilated cardiomyopathy. Heart failure accounts for 30% of the mortality in DMD patients¹⁷⁰. An MDX mouse, which is a model of DMD and lacks the protein dystrophin, has decreased levels of SR luminal Ca²⁺ -binding proteins ¹⁷¹, decreased SERCA2a expression ¹⁷², and an increase in resting [Ca²⁺]_i¹⁷³. Patients with DMD are at increased risk for fatal cardiac arrhythmias ^{170, 174}. MDX mice were shown to have “leaky” RyR2 due to S-nitrosylation of the channel and calstabin2 depletion ¹⁷⁵. Suppressing the RyR2-mediated diastolic SR Ca²⁺ leak by inhibiting calstabin2 depletion prevented and fatal sudden cardiac arrhythmias in DMD mice, suggesting that leaky RyR2 trigger ventricular arrhythmia in DMD ¹⁷⁵. Recent studies show that CaMKII inhibition or interbreeding in to a genetic background with a knock in RyR2 S2814A mutation that is resistant to CaMKII prevents arrhythmogenic Ca²⁺ waves and ventricular tachycardia in MDX mice ¹⁷⁶, suggesting that CaMKII phosphorylation at S2814A of RyR2 contributes to the arrhythmia in MDX mice and possibly in DMD patients. Combined together, these studies suggest that myofilament and cytoskeletal proteins are intimately involved in Ca²⁺ homeostasis and contribute to pathogenesis of heart failure and arrhythmias.

III: Alterations in regulatory mechanisms in heart failure

CaMKII

CaMKII is a multifunctional serine-threonine protein kinase that is abundant in nerve and muscle. There are 4 different CaMKII encoding genes with each encoding a distinct CaMKII isoform (α, β, γ, δ). CaMKIIδ appears to be the main isoform expressed in the heart but CaMKIIγ is also present ¹⁷⁷. Whether these two main isoforms have selective roles in cardiac pathophysiology is unclear at this point, as there are very few studies investigating the role of CaMKIIγ. Transaortic banding induced increased expression of both CaMKIIδ and CaMKIIγ isoforms¹⁷⁸ and conditional double knockout of CaMKIIδ and CaMKIIγ in caused decreased phosphorylation of target proteins¹⁶⁷. A recent study suggests that CaMKIIγ is enriched in mitochondria¹⁷⁹. CaMKII connects intracellular Ca²⁺ signaling to ECC and regulates both SR Ca²⁺ uptake and release (Figure 2). CaMKII acts on multiple Ca²⁺ homeostatic proteins involved in ECC ³² including voltage-gated Ca²⁺ channels¹⁶, RyR2 ¹⁸⁰ and PLN ¹⁸¹. In general, CaMKII-mediated phosphorylation of Ca²⁺ homeostatic proteins enhances their activity and promotes performance of physiological events such as ECC and fight/flight mechanical and heart rate responses.

CaMKII consists of stacked hexamers and each monomer consists of an N-terminus catalytic domain and a C-terminus association domain that flank a core regulatory domain ¹⁸². The “hypervariable” region located between the association and regulatory domains is likely responsible for tuning the Ca²⁺ sensitivity of CaMKII activation. ¹⁸² CaMKII is activated when [Ca²⁺]_i binds to calmodulin (CaM) causing conformational changes that release the catalytic domain from the negative regulation by the autoinhibitory region of the regulatory domain ¹⁸³.

Under diastolic, resting [Ca²⁺]_i in the presence of low ROS, CaMKII is enzymatically inactive due to the binding of catalytic domain to an autoinhibitory region. Sustained activation of CaMKII by binding to calcified calmodulin (Ca²⁺/CaM) leads to threonine 287 autophosphorylation (the numbering varies slightly between isoforms), CaM trapping and CaMKII activation that is autonomous from Ca²⁺/CaM (Figure 4) ¹⁸⁴. Ca²⁺/CaM autonomous (constitutively active) CaMKII is also generated by oxidation of paired regulatory domain methionines (281/282) ⁵⁵. In this setting, oxidized CaMKII resets its Ca²⁺ sensitivity so that lower levels of intracellular Ca²⁺ are required for initial activation ¹⁸⁵. Thus, both threonine 287 autophosphoryation and methionine 281/282 oxidation can convert CaMKII into a constitutively active enzyme. The constitutively active forms of CaMKII appear to be particularly effective at driving myocardial disease phenotypes ^{21,186,187,21, 188}; Thus, CaMKII is a highly regulated signal but under pathological stress CaMKII undergoes post-translational modifications that convert it into a Ca²⁺/CaM autonomous enzyme with the potential to promote heart failure and arrhythmias..

Figure 4. Structure and activation of CaMKII.

CaMKII consists of stacked hexamers and each monomer consists of an N-terminus catalytic domain and a C-terminus association domain that flank a core regulatory domain. CaMKII is activated when [Ca²⁺]_i binds to calmodulin causing CaMKII to assume an active, extended conformation. Sustained binding to calcified calmodulin (Ca²⁺/CaM) leads to threonine 287 autophosphorylation and sustained CaMKII activation. Oxidation of paired regulatory domain methionines (281/282) also causes sustained activation of CaMKII as oxidized CaMKII resets its Ca²⁺ sensitivity so that lower levels of intracellular Ca²⁺ are required for initial activation. Thus, both threonine 287 autophosphoryation and methionine 281/282 oxidation can convert CaMKII into a constitutively active enzyme to drive myocardial disease phenotypes.

Chronic and excessive neurohormonal activation contributing to the progression of CHF causes increased [Ca²⁺]_i and ROS ^{189, 190}, which causes sustained activation of CaMKII. Increased myocardial CaMKII activity and expression have been found in various animal models ^{191, 192} and in patients with heart failure ¹⁹³. Mice with myocardial transgenic CaMKII over-expression develop heart failure and premature sudden death ¹⁹⁴; CaMKII activation by β–AR stimulation causes fetal gene induction, pathological hypertrophy ^{54, 195}, myocardial apoptosis ¹⁹⁶, arrhythmia ¹⁹⁷ and worsening heart failure after myocardial infarction (MI) ⁵⁵. Angiotensin II activates CaMKII by methionine oxidation and promotes cardiomyocyte death ^185,55 that contributes to sinus node dysfunction ¹⁸⁷, a frequent counterpart to heart failure ¹⁹⁸. Aldosterone activates CaMKII by methionine oxidation and CaMKII activation by aldosterone leads to increased death after MI by increasing the propensity to myocardial rupture. Intriguingly, excessive oxidized CaMKII activates a myocyte enhancer factor 2 transcriptional signaling pathway to increase myocardial expression of matrix metalloproteinase 9 that contributes to myocardial matrix instability and sudden death due to post-myocardial infarction cardiac rupture ¹⁸⁶.

We recently found that hyperglycemia also leads to increased methionine 281/282 oxidized CaMKII in diabetic patients and in mice and increased oxidized CaMKII is a necessary signal for diabetes-associated excess mortality in a mouse model of MI¹⁸⁸. We found that mitochondrial ROS was increased in cardiac myocytes exposed to hyperglycemia and that mitochondrial-targeted antioxidant therapy or a knockin mutation of CaMKIIδ to prevent oxidative activation (M281/281V) were both effective at preventing excess, diabetes-attributable mortality after MI. Importantly, CaMKII inhibitors significantly improved the force frequency relationship in failing human cardiomyocytes ¹⁹⁹. CaMKIIδ^−/minus; knockout mice are resistant to myocardial hypertrophy and pressure overload-induced heart failure^{200, 201} and mice with transgenic myocardial CaMKII inhibition are resistant to heart failure from MI. ⁵⁷ Taken together, this evidence indicates that CaMKII plays an important role in connecting upstream signals, such as neurohumoral activation, hyperglycemia, ischemic injury and infarction with defective Ca²⁺ signaling and downstream pathological outcomes important for CHF.

PKA

PKA is the principal upstream kinase activated by β-AR agonists. There are multiple β-AR subtypes, including β₁-, β₂-, and β₃-ARs ^{202, 203}. β-ARs belong to the large family of G protein-coupled receptors with seven transmembrane domains ²⁰⁴ and contain phosphorylation sites ²⁰⁵, which serve as targets for protein kinases including PKA and PKC ²⁰⁶. The binding of circulating adrenergic amine agonists to β-ARs activates adenylate cyclase and simulates cyclic adenosine monophosphate (cAMP) production to release the catalytically active subunit of PKA.

PKA, in turn, catalyzes phosphorylation of multiple Ca²⁺ regulatory proteins including PLN, L-type Ca²⁺ channels, and RYR2. Under physiological conditions activation of the β–AR signaling pathway through PKA stimulates Ca²⁺ influx and increases SR Ca²⁺ uptake and storage by the SR, leading to increased systolic [Ca²⁺]_i transients and thus increased contractile function and lusitropy ⁴. However, in the failing heart, chronically elevated adrenergic agonist activity leads to down-regulation of β₁-AR signaling with decreased β₂-AR density ^207,208 and uncoupling of β₂ AR from downstream effector molecules, including Ca²⁺ regulatory target proteins such as PLN ²⁰⁹, leading to inefficient ECC and decreased contractile function. These changes impair the ability of the failing heart to increase contractility to meet hemodynamic demands.

Widely established benefits of β-AR antagonist drugs in treating heart failure ⁴⁴ strongly support that altered β-AR signaling is maladaptive and promotes heart failure progression. However, the mechanisms of therapeutic benefit for β-AR antagonist drugs are likely to be diverse. β-AR antagonists preserve transverse tubular ultrastructure ¹⁵⁴, reverse RyR2 hyperphosphorylation ^{44, 210}, and decrease SR Ca²⁺ leak ^{44, 210}, leading to increased contractility in heart failure. In addition, excessive β-AR agonist stimulation causes apoptosis via activation of a mitochondrial death pathway ²¹¹ while β-AR antagonists such as carvedilol can protect mitochondria from oxidative stress-induced mitochondrial permeability transition pore (mPTP) opening ^212,213.

PKA-dependent β-AR signaling desensitizes after sustained β₁-AR agonist stimulation ²¹⁴. In contrast, CaMKII signaling in ECC is persistent and may be necessary to sustain positive inotropic actions of prolonged catecholamine signaling ²¹⁵. Epac is a guanine nucleotide exchange protein that directly binds to and is activated by cAMP in parallel to the classical PKA signaling pathway. Epac was shown to mediate β-AR induced cardiomyocyte hypertrophy ^{216, 217} and arrhythmias²¹⁸, modulate cardiac nuclear Ca²⁺ signaling by increasing nuclear Ca²⁺ through phospholipase C, inositol trisphosphate and CaMKII, and activate the transcription factor MEF2 ²¹⁹. A recent study demonstrated that Epac may mediate cardioprotection from cell death induced by β-AR activation ²²⁰. Thus, β-AR stimulation activates multiple signaling pathways including cAMP/PKA, cAMP/Epac and the CaMKII pathway. In our view, it is not yet clear how much of the therapeutic benefit of β-AR antagonist drugs is due to reduced PKA activity and what portion is attributable to reduction in the activity of other downstream signals, such as CaMKII.

PKC

Protein kinase C is a family of serine-threonine protein kinases that are present in a wide variety of tissues, including myocardium. PKCα is the most abundantly expressed isoform of the myocardial PKC family. RACKs (receptor for activated c kinase) are isoform selective anchoring proteins for PKCs ²²¹. RACKs are important for determining the subcellular localization of PKC isoenzymes upon their activation and modulate their function ²²¹. PKCα plays an important role in regulating myocardial contractility. For example, mice with PKCα deletion demonstrate an increase in [Ca²⁺]_i transients and contractility, while overexpression of PKCα diminishes contractility ²²². PKCα knockout mice are protected from pressure overload induced heart failure and from dilated cardiomyopathy induced by deleting the gene encoding muscle LIM protein (Csrp3), and from cardiomyopathy associated with overexpression of PP1²²². One experimentally validated pathway for PKCα action to decrease [Ca²⁺]_i transients is that PKCα suppresses SERCA2a activity by phosphorylating inhibitor 1 (I-1) resulting in increased PP1 activity and dephosphorylation of PLN ²²². Decreased SERCA2a activity thus reduces SR Ca²⁺ load leading to reduced Ca²⁺ release during systole, hence reducing contractility. Other PKC isoforms (delta) and (epsilon) may play a significant role in promoting hypertrophy ^{223, 224}. Taken together, these results from animal models support a potential role for PKC in promoting heart failure progression.

S100A1

S100A1 belongs to the S100 protein family, a group of EF-hand containing Ca²⁺-binding proteins. S100A1 shows highest expression in human cardiac muscle and is preferentially expressed in the left ventricle. S100A1 has a molecular weight of 10.4 kDa and contains two functional EF-hand Ca²⁺-binding motifs. Upon Ca²⁺ binding S100A1 undergoes a conformational change to expose a hydrophobic pocket for binding to target proteins ²²⁵. The Ca²⁺ binding affinity of S100A1 is tightly regulated by post-translational modifications, including S-nitrosylation and S-glutathionylation of a cysteine residue in the C-terminal region ^{226,227, 228}. Either modification enhances Ca²⁺ affinity by several orders of magnitude, which augments the ability of S100A1 to sense Ca²⁺ oscillations over a wide dynamic range ^{226,227, 228}. S100A1 has emerged as a key regulator of Ca²⁺ cycling and cardiac contractile function ^226,229. S100A1 enhances SR Ca²⁺ uptake and increases SR Ca²⁺ content ^110,229. S100A1 also directly regulates RyR2 function ^{229, 230}. More recently, S100A1 was found to reside in mitochondria where it stimulates ATP synthase (complex V) activity and promotes the adenosine nucleotide translocator function to increase ATP synthesis and mitochondrial ATP efflux in cardiomyocytes ^{110, 231}.

S100A1 knockout mice have impaired contractility and show enhanced proarrhythmogenic susceptibility to acute β-AR agonist stimulation and pressure overload induced by chronic transaortic constriction ^232,233. There is impaired SR Ca²⁺ uptake, increased SR Ca²⁺ leakage and a reduced SR Ca²⁺ load in heart tissues from the S100A1 knockout mice ^234,235. The S100A1 knockout mice also demonstrated excessive mortality and accelerated CHF after MI as well as increased post-MI cardiac remodeling ^234,235. In contrast, mice with myocardial S100A1 over-expression showed enhanced contractile responses to β-AR stimulation, improved [Ca²⁺]_i homeostasis, survival and preserved left ventricular function after MI ²³⁵. In human heart samples with dilated and ischemic cardiomyopathy, S100A1 mRNA and protein expression was found to be down regulated ^236,237. Decreased S100A1 expression levels were also shown in experimental HF animal models and correlate with the severity of heart failure and mortality ^238,235. These results suggest that S100A1 plays an important role in regulating Ca²⁺ cycling and contractile function while loss of S100A1 may contribute to heart failure in the setting of pathological stress.

Calcineurin

Calcineurin, also known as protein phosphatase 2B (PP2B), is a Ca²⁺/CaM-activated, serine-threonine phosphatase and the first Ca²⁺ dependent signaling molecule explicitly linked to myocardial hypertrophy and heart failure^239,240. Calcineurin signaling stimulates cardiac hypertrophy ^241,242,242 and remodeling through activation of the nuclear factor of activated T cells (NFAT) transcription factor. Upon calcineurin-mediated dephosphorylation NFAT translocates to the nucleus and activates cardiac transcription ²⁴³. The calcineurin- NFAT signaling pathway in myocardium appears to be activated only when there are pathological increases in [Ca²⁺]_I, whereas it is not activated during physiologic hypertrophy induced by exercise or pregnancy ²⁴⁴, suggesting that calcineurin signaling is tightly coupled with pathological defects in Ca²⁺ homeostasis.

There is increased calcineurin activity and/or expression in animal models ²⁴¹ and patients with myocardial hypertrophy and heart failure ^{245,239, 246}. Over-expression of calcineurin causes myocardial hypertrophy, heart failure and premature death ^{240, 244}. Calcineurin inhibition by cyclosporin prevented hypertrophy in mice genetically predisposed to develop hypertrophic cardiomyopathy and in a rat model of pressure overload-induced hypertrophy ²⁴⁷. Calcineurin Aβ-knockout mice, with a 80% decrease in calcineurin enzymatic activity in the heart, show decreased hypertrophic responses induced by pressure overload or agonists infusion including angiotensin II and isoproterenol ²⁴⁸. Intriguingly, CaMKII expression and activity are increased in calcineurin transgenic mice ¹⁹⁷. CaMKII inhibition improved contractile function, reduced arrhythmias and decreased mortality in mice with myocardial transgenic over-expression of a constitutively active form of calcineurin without substantially reducing calcineurin-evoked myocardial hypertrophy ^{197, 244}. We interpret these findings to suggest that myocardial dysfunction and high mortality in calcineurin transgenic mice are at least in part attributable to downstream activation of CaMKII and independent of myocardial hypertrophy. The interactions between calcineurin and CaMKII are complex, as highlighted by the finding that CaMKII catalyzed phosphorylation of calcineurin prevents full activation of calcineurin by inhibiting Ca²⁺/CaM binding. Thus, CaMKII may act as an antihypertrophic agent in the context of the calcineurin/NFAT pathway ²⁴⁹. Overall, these findings support a view that calcineurin is an important regulator of cardiac hypertrophy and heart failure but leave open the question of which downstream events are critical for the cardiomyopathic actions of calcineurin.

IV. Arrhythmias as a common cause of death in heart failure

Heart failure, especially in patients with left ventricular ejection fractions less than 30%, is associated with a high rate of arrhythmia- induced sudden death ²⁵⁰. Various factors appear to enhance the probability of arrhythmias, including defective [Ca²⁺]_i homeostasis. Many ion channels respond to loss of normal [Ca²⁺]_i homeostasis by contributing to cell membrane hyperexcitability. However, as exemplified by the Cardiac Arrhythmia Suppression Trial (CAST) ²⁵¹ and The Survival With Oral d-Sotalol (SWORD) ²⁵², ion channel antagonist therapies are not effective in preventing sudden death in high risk patients. In contrast, neurohumoral antagonist drugs that serve as mainstay therapeutics for heart failure, such as β-adrenergic ²⁵³, angiotensin II ²⁵⁴, and mineralocorticoid receptor antagonists ²⁵⁵, are effective in reducing sudden death. These findings suggest that signals that modulate ionic currents are better therapeutic targets than ion channels.

Electrical remodeling

Proarrhythmic electrical remodeling is a term used to describe multiple changes in ionic currents that collectively lead to action potential and QT interval prolongation and favor arrhythmias in failing ventricular myocardium. Prolongation of the action potential plateau, in particular, contributes to a proarrhythmic substrate for non-inactivating components of Na_V1.5 current ^30,256 and Ca_V1.2 channels in a high activity gating mode ¹⁶. A comprehensive review of electrical remodeling in heart failure is beyond the scope of this review but has been recently published elsewhere ²⁵⁷. Voltage-gated K currents (I_K) are the major driving force for myocardial membrane repolarization ²⁵⁸ and failing myocardium is consistently reported to show reduced repolarizing I_K that contributes to proarrhythmic action potential and QT interval prolongation ²⁵⁹. Interestingly, excessive CaMKII activity also contributes to reduced I_K in failing myocardium by phosphorylation of the pore-forming α-subunit of the voltage-dependent K⁺ channel 4.3 (Kv4.3) at Ser⁵⁵⁰, which encodes a class of rapidly inactivating I_K including the transient outward current in the heart ²⁶⁰.

Cardiac ATP-sensitive K⁺ (K_ATP) channels are metabolic sensors activated in response to various forms of cardiac stress, including ischemia and neurohormonal activation, leading to membrane hyperpolarization, decreased action potential duration and contractility²⁶¹. Hence K_ATP channels play an important role in improving cellular energy efficiency and stress resistance. Association of K_ATP with Ankyrin B via the C-terminus of Kir6.2, the pore forming unit, was shown to be important for K_ATP channel trafficking and membrane metabolic regulation ²⁶². One recent study suggests that CaMKII couples the surface expression of cardiac K_ATP channels with Ca²⁺ signaling to regulate energy efficiency and stress resistance, as Ca²⁺- dependent activation of CaMKII results in phosphorylation of Kir6.2, the pore forming subunit and promotes internalization of K_ATP channels ²⁶³. CaMKII also affects trafficking of a variety of voltage-gated K⁺ currents with the net effect of reducing repolarizing K⁺ current and prolonging the action potential ²⁶⁴. These findings suggest that [Ca²⁺]_i may feedback to control multiple ionic currents through activation of CaMKII and that excessive CaMKII activity in CHF contributes to the proarrhythmic substrate and enhanced risk for sudden death in structural heart disease by altering ion channel function and membrane expression.

CaMKII and arrhythmia

Heart failure is a condition of increased oxidant stress, loss of [Ca²⁺]_i homeostasis and activation of CaMKII. CaMKII exerts proarrhythmic effects through actions at multiple protein targets that are key components of Ca²⁺ homeostasis including CaV1.2 ^265,16, NaV1.5 ^{31, 256}, and RyRs ⁵⁷ (Figure 5). CaMKII increases phosphorylation of a CaV1.2 β subunit (β_2a) at Thr498 ²⁶⁵ leading to high activity mode 2 gating, intracellular Ca²⁺ overload and EADs ¹⁶. Phosphorylation of RyR2 at Ser2814 by CaMKII increases diastolic SR Ca²⁺ leak ⁵⁷, which is proarrhythmic ²⁶⁶ by triggering DADs. CaMKII acts on Nav1.5, the predominant cardiac voltage-gated Na⁺ channel, and increases I_NaL ^{256,30, 31}, which prolongs action potential and triggers early EADs ^256,31. CaMKII inhibition has been shown to prevent or suppress ventricular arrhythmias in myocardial tissues and animal models ^{266, 267}. This evidence consistently suggests that CaMKII can promote arrhythmias and sudden death and that CaMKII inhibition can reduce or prevent arrhythmias.

Fig 5. CaMKII and Mechanisms of arrhythmia.

Sustained activation of CaMKII by oxidative stress and elevated [Ca²⁺]_i contributes to arrhythmia in heart failure by several mechanisms: 1) CaMKII phosphorylates L-type Ca channels (Ca_V1.2) to increase its open probability, causing early afterdepolarizations (EADs). Increased I_Ca also contributes to action potential prolongation, augmented [Ca²⁺]_i and DADs. 2) CaMKII phosphorylates Na⁺ channels (Na_V1.5) and enhances the long-lasting late I_Na (gain of function) promoting EADs and increasing subsarcolemmal [Na⁺]_i to favor delayed afterdepolarizations (DADs). 3) CaMKII favors phosphorylation of RyR2 to increase SR Ca²⁺ leak, which shifts Na⁺/Ca²⁺ exchanger (NCX) to a forward mode, causing DADs. CaMKII contributes to arrhythmogenic structural features of injured myocardium by promoting myocyte death and collagen deposition.

Reverse Excitation-Contraction Coupling

Diseased myocardium is non-uniform in ECC with damaged and non-damaged regions as well as inhomogeneous border zone areas bridging damaged and healthy tissue. Arrhythmogenic contractile waves were observed in non-uniform failing myocardium ²⁶⁸. A potential mechanism underlying this phenomenon is reverse ECC ²⁶⁹, a process during which abnormal contractions of damaged regions causes regional rise of [Ca²⁺]_i leading to arrhythmogenic contractile waves. Aftercontractions appear to be initiated by the weak and damaged region during regular contractions and propagate into neighboring myocardium ²⁷⁰. These contractile waves are likely due to mechanical effects of damaged myocardium, such as stretching and release, and regional elevation of [Ca²⁺]_i as a result of damage ²⁷¹. When cardiac muscle is damaged, intracellular Ca²⁺ waves are initiated locally, but propagate into adjacent tissues ²⁷². Diffusing Ca²⁺ ions activate neighboring SR, which in turn triggers further Ca²⁺ release from SR. These Ca²⁺ waves may give rise to premature contractions and triggered arrhythmias ²⁷³. Purkinje fibers are particularly prone to proarrhythmic [Ca²⁺]_i waves and may serve as an arrhythmia focus for injured myocardium ²⁷⁴. Another potential mechanism underling arrhythmogenic Ca²⁺ waves are the activation of stretch-activated channels (SACs), which are nonselective cation channels activated by mechanical stress ²⁷⁵. In the MDX mouse, lack of dystrophin results in increased activity of SACs and increased resting intracellular [Ca²⁺]_i in skeletal muscles ²⁷⁶. SACs have also been reported in ventricular cardiomyocytes ²⁷⁷ and are proposed to play a role in tachycardia-induced chronic heart failure ²⁷⁸. Thus, the role of Ca²⁺ in maladaptive contractions may be proarrhythmic.

V. Therapeutic targets for heart failure

Current drug therapies for CHF are mainly designed to counteract over-activation of the sympathetic and renin angiotensin–aldosterone systems, which is known to prolong survival ^253,254,255. Advanced CHF associated with increased risk of fatal arrhythmias can also be managed by surgically implantable cardioverter defibrillator, cardiac resynchronization therapy (CRT) and mechanical ventricular assist devices. However, currently available pharmacological and device therapies are far from ideal as they fail to fully correct underlying molecular abnormalities involved in systolic and diastolic dysfunction as well as adverse structural and proarrhythmic electrical remodeling. Given the central role of Ca²⁺ signaling in the progression of CHF, restoration of normal [Ca²⁺]_i homeostasis is a promising strategy to forestall progression and improve function of failing cardiomyocytes.

RyR2

CHF is a condition of leaky RyR2, decreased SR Ca²⁺ content and reduced [Ca²⁺]_i transients. Thus, leaky RyR2 can contribute to myocardial dysfunction and arrhythmias ^{58, 244}. Over-expression of the RyR2 regulatory protein FKBP12.6 caused increased SR Ca²⁺ content and improved myocyte shortening in isolated cardiomyocytes ²⁴⁴. RyR2 leak can also potentially be directly targeted by pharmacologic agents shown to improve cardiac function ²⁴⁴ and prevent arrhythmias ²⁸³. For example, K201, a benzothiazepine derivative and inhibitor of RyR2 was shown to stabilize RyR2s and decrease SR Ca²⁺ leak ²⁸⁴. So-called Rycals, K201-congeners, have emerged as promising agents for targeting RyR2 and reducing arrhythmias and heart failure ³⁶. Another Rycal compound, ARM036, also a benzothiazepine derivative, is in Phase II trials for Heart Failure and catecholaminergic polymorphic ventricular tachycardia. It is anticipated that information on the potential clinical benefits of pharmacologic therapy aiming to modulate RyR2 function will soon become available.

CaMKII

CaMKII links Ca²⁺ homeostasis and cardiac function in myocardium under physiological conditions. Under pathological conditions such as heart failure characterized by excessive neurohormonal activation and oxidative stress, CaMKII activation is sustained, which promotes diastolic Ca²⁺ leak and arrhythmias. Animal studies consistently demonstrate that CaMKII inhibition reduces heart failure and arrhythmias, reducing or preventing sudden death. In our view, CaMKII is a highly validated target that connects to most or all aspects of defective [Ca²⁺]_i homeostasis in heart failure. However, to determine whether the experimentally observed benefits of CaMKII inhibition are applicable to human heart failure, CaMKII inhibitory drugs with drug-like properties and adequate specificity and safety will need to be developed.

PKC

PKCα has been identified to have critical roles in the pathogenesis of heart failure. Deletion of the PKCα gene ^222,285 or inhibition with drugs ^286,287,135 have shown dramatic protective effects against the development of heart failure of various etiologies including ischemia, pressure overload or dilated cardiomyopathy induced by deleting LIM protein in animal models. However, clinical trials with PKC inhibitors or RACK inhibitor peptides were largely disappointing for improving heart failure ²⁸⁸ or reducing myocardial injury in MI patients ^289,290.

Transfer of genes encoding S100A1 and SERCA2a are discussed elsewhere in this compendium (MOST)

OVERALL CONCLUSION

It is now clear that impaired [Ca²⁺]_i homeostasis is a key feature of heart failure that contributes to contractile dysfunction and arrhythmias. Defective Ca²⁺ homeostasis in heart failure is most often the result of altered expression and function of a group of [Ca²⁺]_i handling proteins, ion channels and enzymes.. Numerous laboratories have contributed to the improved understanding of these pathways and this new knowledge has bolstered the quest to develop novel and improved therapeutics. We expect that the next several years will witness the initial results of several promising heart failure therapies designed to correct defects in myocardial [Ca²⁺]_i homeostasis.

Nonstandard Abbreviations and Acronyms

CHF

Congestive heart failure

ECC

Excitation-contraction coupling

SR

Sarcoplasmic reticulum

RyR2

Ryanodine receptor

I_Na

Inward Na⁺ current

SERCA2a

Sarcoplasmic-endoplasmic reticulum Ca²⁺ ATPase

NCX

Na⁺/Ca²⁺ exchanger

LTCC

L-type calcium channels

PKA

Protein kinase A

PKC

Protein kinase C

CaMKII

Ca²⁺ and calmodulin-dependent protein kinase II

βAR

B adrenergic receptor

EADs

Early afterdepolariazations

DADs

Delayed afterdepolarizations

PP2B

Protein phosphatase 2B

PP1

Type 1 protein phosphatase

MLP

Muscle-specific LIM protein

HRC

Histidine-Rich Ca²⁺ Binding Protein

I_NCX

NCX current

ATP

Adenosine triphosphate

I_KATP

ATP sensitive K⁺ current

NADH/NADPH

Nicotinamide Adenine Dinucleotide/Nicotinamide Adenine Dinucleotide Phosphate Hydrogen

mPTP

Mitochondrial permeability transition pore

CaMKIIN

CaMKII inhibitory protein

MCU

Mitochondrial Ca²⁺ uniporter

mNCE

Mitochondrial Na⁺/Ca²⁺ exchanger

DMD

Duchnne muscular dystrophy

cAMP

Cyclic adenosine monophosphate

NFAT

Nuclear factor of activated T cells

I_K

Voltage-gated K currents

Ito

Transient outward current in the heart

K_ATP

Cardiac ATP-sensitive K⁺

AAV

Adeno-associated virus

ANT

Adenosine nucleotide translocator