The interaction of Ca²⁺ with sarcomeric proteins: role in function and dysfunction of the heart

Abstract

The hallmarks of the normal heartbeat are both rapid onset of contraction and rapid relaxation as well as an inotropic response to both increased end-diastolic volume and increased heart rate. At the microscopic level, Ca²⁺ plays a crucial role in normal cardiac contraction. This paper reviews the cycle of Ca²⁺ fluxes during the normal heartbeat, which underlie the coupling between excitation and contraction and permit a highly synchronized action of cardiac sarcomeres. Length dependence of the response of the regulatory sarcomeric proteins mediates the Frank-Starling Law of the heart. However, Ca²⁺ transport may go astray in heart disease such as in congestive heart failure, and both jeopardize systole and diastole and triggering arrhythmias. The interaction between weak and strong segments in nonuniform cardiac muscle allows partial preservation of force of contraction but may further lead to mechanoelectric feedback or reverse excitation-contraction coupling mediating an early diastolic Ca²⁺ transient caused by the rapid force decrease during the relaxation phase. These rapid force changes in nonuniform muscle may cause arrhythmogenic Ca²⁺ waves to propagate by the activation of neighboring sarcoplasmic reticulum by diffusing Ca²⁺ ions.

this article is part of a collection on Electrophysiology and Excitation-Contraction Coupling. Other articles appearing in this collection, as well as a full archive of all collections, can be found online at http://ajpheart.physiology.org/.

Function of Normal Myocardium

This review will discuss reverse excitation-contraction coupling (RECC) against the background of the principles of the normally functioning heart. The normal heart contracts at regular intervals driven by the electrical impulses of the sinus node. Rapid conduction via the His and Purkinje system permits a near to synchronous contraction of the main pump chambers (67). To achieve a forceful heartbeat as well as rapid complete relaxation in between heartbeats, Ca²⁺ ions should both appear and disappear rapidly and at the right time in the cardiac myocyte. For the purpose of the current discussion, we will adopt the simplest scenario that allows a heart to contract and put the maximal amount of energy into the circulation during systole while it completely relaxes and receives venous blood during diastole as a superbly compliant structure. How this might work is a useful introduction into the mainstay of the discussion of what might happen when Ca²⁺ transport goes astray.

A brief history of the role of Ca²⁺.

Sidney Ringer spent several years at the end of the 19th century studying the composition of extracellular solutions required to maintain contraction of the heart (112). In the process he discovered, albeit serendipitously, the role of the most important component, i.e., Ca²⁺ ions, when he realized that contraction was maintained when he had prepared his saline solutions unwittingly with Ca²⁺-rich London tap water, whereas contraction disappeared rapidly when the saline solution was properly prepared with distilled water without the addition of Ca²⁺ salts (113). This work has led to the now generally accepted composition of well-buffered saline solutions with ion concentrations that approximate the extracellular fluid (114). Sixty years later it was shown that Ca²⁺ is the only physiologically occurring ion that activates contraction (59). Further research into the role of Ca²⁺ in the activation of contraction in striated muscle led to the concept that striated muscle employs an intracellular store, interposed between the extracellular milieu and the cytosol, which boosts Ca²⁺ entry into the cell during the action potential into full-fledged Ca²⁺ release from the stores required to activate a substantial fraction of the regulatory troponin C (TnC) on the actin filaments. Structural studies led to the discovery of the elements underlying this amplifier system of the Ca²⁺ signal: the transverse tubules (47, 49, 111) that invaginate the myocyte and form a network with transverse and longitudinal elements (74, 100, 103, 120) and sarcoplasmic reticulum (SR) (46, 94, 107, 130). Concurrent with the elucidation of these structures, a role of transverse tubules and SR, details of Ca²⁺ transport emerged. The SR appeared to sequester (26) Ca²⁺, which is quintessential in the relaxation of striated muscle (11, 87 94, 117). In turn, SR-Ca²⁺ release initiates the contractile events (38, 48) and the release appeared to respond to transient increases of the intracellular Ca²⁺ concentration ([Ca²⁺]_i) (38, 47). These observations led Fabiato (40, 42) to his landmark studies that founded the concept of Ca²⁺-induced Ca²⁺ release from the SR in cardiac myocytes.

Current view of the structures involved in Ca²⁺ transport.

Figure 1 shows that on average, a cardiac myocyte (cell volume, 20 pl) is completely packed with myofibrils enveloped in a network of Ca²⁺-storing SR and mitochondria that accompany every sarcomere in the myofibril. {In the following, I will simplify Ca²⁺ transport in a single typical myocyte by an indication of the approximate cell volume (20 pl), fluxes, and currents per cell and the resulting [Ca²⁺] and the corresponding number of Ca²⁺ ions that arrive in a half sarcomere of a typical myofibril.} The cell membrane conveys electrical information in the form of the action potential in a fraction of a millisecond to the latter trio of organelles by its transverse tubules, which invaginate the surface perpendicular to the cell axis at every Z disk. The tubules form a helicoid transverse and longitudinal network (74), which reduces the distance between transmembrane events and the center of the sarcomere to <1 μm (100). The transverse tubules comprise ∼60% of the total cell membrane area (10⁴ μm²) (103). Ca²⁺ transport, illustrated in Fig. 1, allows the heart to cycle between diastole and systole by shuttling Ca²⁺ ions between the extracellular milieu and the interior of the myocytes and even more importantly by shuttling Ca²⁺ ions between the SR and the cytosol.

Fig. 1.

Magnification of a dyad, the contact site of sarcoplasmic reticulum (SR) and transverse tubule (58). The diagram illustrates the sequence of Ca²⁺ events starting with the action potential. The Ca²⁺ channels of the SR dominate the scene because of the magnitude of its Ca²⁺ fluxes (indicated by arrows: red for Ca²⁺ release; green for Ca²⁺ removal; and black, indicating the interaction with the contractile proteins); after (58). The plant Ryania speciosa (inset) defends itself against insects by production of the alkaloid ryanodine, which binds specifically to the SR-Ca²⁺ channel. The dominant ligand for Ca²⁺ in the cardiac myocyte is the troponin-C molecule on the actin filament. [Ca²⁺]_i, intracellular Ca²⁺ concentration; I_Ca, Ca²⁺ current; LCC, L-type Ca²⁺ channel; NCX, Na⁺/Ca²⁺ exchange; RyR2, ryanodine receptor 2; SERCA, sarco(endo)plasmic reticulum Ca²⁺-ATPase; NKP, natrium/kalium pump; CaP, calcium pump. Modified from ter Keurs (139) with permission.

It is known that cell structures turn over constantly and at a surprising rate; for example, myofibrillar proteins are turned over at 10%/day (69, 118), and myofibrils are being rebuilt “as they work” (15, 99), allowing for a constant adjustment of their properties to the hemodynamic load (12). One may expect that the structure of the transverse tubules is equally dynamic and that the integrity of the transverse tubular system varies with the working conditions under which the heart operates. The loss of transverse tubular integrity has indeed been observed in ventricular hypertrophy and even more so in congestive heart failure (CHF) (127, 157). As a result, the high-speed connection between the surface membrane and the SR in the core of the cell is lost and the cell will have to rely on other means to activate its core. We will see in the following why the speed of activation decreases by the loss of integrity of the transverse tubules (20).

Mechanisms underlying excitation-contraction coupling.

An asset of cardiac myocytes that allows rapid activation and relaxation of cardiac muscle is that when the transverse tubules pass the SR, a specialized structure, the dyad or triad, is formed by the terminal cisternae of the SR, which embrace the transverse tubular membrane (58, 86) (Fig. 1). Na⁺ and Ca²⁺ channels, Na^+/Ca²⁺ exchange (NCX) proteins, as well as Na⁺/K⁺ pump proteins aggregate on the transverse tubular membranes, colocalizing (102, 122) with the Ca²⁺ channels that are clustered on the SR membrane [ryanodine receptors (RyRs) named after ryanodine, the insecticidal alkaloid produced by the plant Ryania speciosa] (6, 23, 121). At the interface between the terminal cisterna and transverse tubules, clusters of up to 35 L-type Ca²⁺ channels make close contact with ∼10-fold more SR-Ca²⁺ channels (68, 106, 106) in so-called couplons (131). Still, some of the terminal cisternae remain free (73).

Figure 1 shows the relationship between the membrane action potential, which invades the myocyte along its transverse tubules, and Ca²⁺ release by the SR. In short, Na⁺ current (∼a total of 50 nA/cell) depolarizes the cell to +30 mV in a millisecond mostly and is carried, at least in guinea pig cardiac myocytes, by the cardiac Na⁺ channel isoform on the intercalated discs (85). Several brain channel isoforms on the transverse tubular membrane conduct, where they colocalize with NCX proteins in the couplons (150) a smaller Na⁺ current (85), which raises the [Na⁺] in the dyadic cleft and reverses the gradient for NCX; the increased [Na⁺] dissipates in a few (5–10) milliseconds. The depolarization also opens, again in a millisecond, the membrane L-type Ca²⁺ channels. Rapid partial repolarization to 0 mV by the transient outward K⁺ current drives a maximal Ca²⁺ current spike (∼0.5 nA/cell) through the Ca²⁺ channels (106) into the clefts of the dyads and triads between the transverse tubules and the terminal cisternae of the SR (110, 141). The Ca²⁺ ions in the clefts bind to regulatory calmodulin sites on the cytosolic domains of L-type Ca²⁺ channel itself and rapidly, but partially, inactivate the channels terminating the Ca²⁺ current spike and leaving a tonic Ca²⁺ current (∼0.3 nA/cell) until the cells repolarize to −40 mV (138). The binding of Ca²⁺ to the Ca²⁺-binding domains of NCX rapidly activates (61, 62, 96, 101) the exchanger, which facilitates Ca²⁺ entry through “reverse mode” transport by NCX, adding to Ca²⁺ entry into the cleft (96). Lastly and most importantly, the Ca²⁺ that has entered the cell (∼10 μM) during the transient of Ca²⁺ current and reverse-mode NCX binds to the RyRs and triggers (116) a rapid and 3–5-fold larger flux of Ca²⁺ ions through these channels from the SR into the cytosol (24, 156). The tonic component of Ca²⁺ current loads the SR with Ca²⁺ ions (116). The interaction between membrane Ca²⁺ entry and SR-Ca²⁺ release underlies the mechanism of excitation-contraction coupling (ECC) that has been known since Fabiato and Fabiato's work (43) as Ca²⁺-induced Ca²⁺ release, while the spatially restricted properties of the couplon protect the system from unbridled regenerative Ca²⁺ release (131). The magnitude of the SR-Ca²⁺ release flux depends on the degree of filling of the SR with Ca²⁺ ions, the magnitude of the rapid component of the Ca²⁺ current, and the interval between the action potentials because the RyRs need to recover (∼0.3 s) before they are ready for the next release (7). The average diastolic SR-Ca²⁺ content is kept close to constant by feedback of [Ca²⁺]_i to the surface membrane Ca²⁺ channels and NCX, whereby an increase of SR-Ca²⁺ content increases the systolic Ca²⁺ transient, which results in an increased efflux of Ca²⁺ from and decreased Ca²⁺ entry into the cell (35). Adrenergic mechanisms modulate the magnitude and the dynamics of the Ca²⁺ fluxes in addition to the effects of varied [Ca²⁺] itself on the Ca²⁺/calmodulin-dependent kinase II system that modulates virtually all Ca²⁺ transporters, e.g., in response to changes in heart rate (56, 84).

The released Ca²⁺ diffuses over a very short distance (∼1 μm) into the cytosol of the adjacent myofibril and starts two simultaneous events. The mechanically important event is that the released Ca²⁺ (∼15·10³ ions/half sarcomere) binds to a substantial fraction of the protein TnC (25,000 mol/half sarcomere) of the troponin-tropomyosin complex on the actin filaments in sarcomeres of every myofibril (39, 126). Binding of Ca²⁺ to TnC starts a chain reaction that moves tropomyosin away from the binding sites for myosin cross bridges, thus enabling the cross bridges (90·10³/half sarcomere) to bind to actin and convert the chemical energy of ATP into force development (2 pN/bridge for a maximal force of ∼100 mN/mm²) and filament sliding and thus sarcomere shortening (128). This process is amazingly close to synchronous in the 10¹⁴ sarcomeres of the normal left ventricle and allows rapid pressure development and brisk ejection of blood (66, 67). A singular property of cardiac muscle is that the process of activation of contraction is strongly sarcomere length (SL) dependent (32, 79, 143), far more than fast or slow skeletal muscle (80).

The released Ca²⁺ also rapidly reaches the Ca²⁺ pumps [the sarco(endo)plasmic Ca²⁺-ATPase 2 pump proteins (SERCA2)] of the SR that envelops the same myofibril. SERCA2 responds to a slight rise in the [Ca²⁺]_i by immediately sequestering Ca²⁺ back into the SR; this response starts relaxation. A more modest amount (∼30%) of the released Ca²⁺ is simultaneously extruded from the cell by NCX proteins that borrow the energy from the gradient of Na⁺ ions across the cell membrane. The fractions of NCX-mediated Ca²⁺ extrusion and SERCA2-mediated Ca²⁺ sequestration by the SR vary somewhat from species to species (13). Furthermore, NCX is important because three to four Na⁺ ions enter the cell during transport of a single Ca²⁺ ion. By consequence, the cell interior tends to become less negative during extrusion of Ca²⁺. During the normal action potential, this Ca²⁺ extrusion by NCX only slows the repolarization somewhat, but we will see the consequences of this depolarization when spontaneous Ca²⁺ release occurs later in this review.

Cardiac cells, such as atrial and Purkinje cells, which operate without a transverse tubular system but still have a regular apposition of SR near the myofibrils, activate their contractile apparatus by initial Ca²⁺ release at the surface membrane in response to the action potential, together with the propagation of Ca²⁺ release mediated by Ca²⁺ diffusion to the core of the cell. The nonjunctional SR of these cells responds to the arrival of diffusing Ca²⁺ ions by SR-Ca²⁺ release, which can be modulated by second messengers that control the SR-Ca²⁺ load (16). Purkinje cells have much larger diameters than atrial cells and are also devoid of a transverse tubular system (34). These cells have developed a specialized triplicate layer of inositol 1,4,5-trisphosphate-Ca²⁺ channels directly under the membrane juxtaposed to a ∼5-μm layer of SR-RyR₃ channels that borders the SR throughout the remainder of the core of the cell endowed with RyR₂ channels (132). This triplicate system may serve to carry the Ca²⁺ signal expediently to the core of the cell by a chemical transmission process (17). Cardiac myocytes themselves may lose the integrity of the transverse tubular system in disease, such as hypertrophy heart failure and ischemic disease (157); in this case, ECC also has to rely on the centripetal propagation of Ca²⁺ release, which is likely to slow the onset and relaxation of contraction (19, 20). As the tonic component of Ca²⁺ current via the transverse tubules loads the SR with Ca²⁺ ions (116), “detubulation” rapidly leads to a reduced Ca²⁺ load of the SR with, as a result, reduced force development. This effect is likely to contribute to the smaller inotropic effect of increased heart rate as has been described in cardiac muscle during heart failure.

An important question is whether the Ca²⁺ transport shows any activity during diastole. It is not unexpected that when a L-type Ca²⁺ channel opens in the transverse tubule, the resulting Ca²⁺ influx may trigger Ca²⁺ release by RyRs or clusters thereof in the terminal cisterna: events now known as Ca²⁺ sparklets triggering Ca²⁺ sparks (24, 156). Furthermore, the RyRs respond to Ca²⁺, so some may respond to Ca²⁺ released by neighboring terminal cisternae by Ca²⁺ release. This phenomenon is noticeable in the form of Ca²⁺ waves in cardiac myocytes.

The puzzle regarding spontaneous activity of the Ca²⁺ system during diastole thus reduces to the question, Do Ca²⁺ sparks and Ca²⁺ waves occur during diastole in normal myocardium? Again, we propose a simple scenario: the incidence of these events in normal cardiac fascicles is small. We do so on the basis of many years of work with isolated cardiac fascicles or trabeculae that can be found on the endocardial side of the ventricle (143) in many species, including mouse, ferret, cat, pig, and rabbit, but the main source over the years has been the rat right ventricle. It has been shown by Goo and collaborators (52a) that these trabeculae are virtually identical to the fascicles that comprise the wall of the heart. The attraction of using these thin free running (usually <100 μm) muscles is that one can study ECC at the level of the sarcomeres by using light diffraction techniques, while loading the specimen with Ca²⁺-sensitive fluorescent dyes permits a continuous measurement of the [Ca²⁺]_i (5). Their behavior is strikingly similar. Figure 2, A and B, shows that the action potential-induced [Ca²⁺]_i transient is uniform and shows a rapid rise and a simple monotonic decline throughout the muscle without variations in time or space during diastole.

Fig. 2.

A: trabeculae, endocardially running fascicles of myocardium, respond to an action potential by uniform calcium release and uniform sarcomere length (SL) changes, followed by quiescence during the diastolic interval. B: duration of the cytosolic [Ca²⁺] transient increases with increase of SL. C and D: duration of the twitch at constant SL increases with stretch of the sarcomeres, adapted from ter Keurs (147) and van Heuningen et al. (152) with permisison. F, force, ROI, region of interest; 3-D, three-dimensional.

Thus the action potential brings the muscle uniformly (143) from an exquisitely compliant passive structure to an impressive maximal stress development of ∼1 kPa, which is more than sufficient to generate the end-systolic pressure of the normal LV. Four major factors dictate the active force: [Ca²⁺]_i in the cytosol of the myocytes, the phosphorylation status of troponin [see for review (123)] SL and the velocity of sarcomere shortening (5, 30, 143). The latter two components are important to the maintenance of force in muscle in which the activation is nonuniform (see Sarcomere dynamics in nonuniform muscle). The relationship between SL and force shows little dependence on shortening before peak force of the twitch and forms the basis for the well-known end-systolic force SL relationship (ESFSLR), which underlies the end-systolic pressure-volume relationship (ESPVR) of the pump chambers of the heart (135). [Ca²⁺]_i and the sensitivity of the contractile proteins to Ca²⁺ ions dictate force generated at any length and underlie the inotropic state of the heart. In in vitro experiments, the sarcomeres shorten during contraction because they are able to stretch the connections of the muscle with the equipment. When sarcomere shortening is prevented, it is clear that both the magnitude and the time course of the force of contraction are length dependent (Fig. 2C). Figure 2B also shows that the rate of decline of the [Ca²⁺]_i is decreased when force is increased when the sarcomeres operate at greater length (5, 76). The observed slowing of the relaxation rates of the Ca²⁺ transient and force are also observed when, at a fixed SL, force is increased by raising the [Ca²⁺] in the medium surrounding the muscle (21, 71) or by raising [Ca²⁺]_i through procedures such as frequency potentiation (8). Conversely, the decline of the [Ca²⁺]_i transient accelerates when force development by the cross bridges is inhibited by compounds like 2,3-butanedione monoxime (76). Simultaneous studies of the [Ca²⁺]_i transient and force by us and others have shown that the rate of decline of the [Ca²⁺]_i transient always decreases when force is increased (5, 65, 70, 76), suggesting an interaction between force development and Ca²⁺ binding to TnC on the contractile apparatus (see Contraction of uniform muscle: interaction between SL, [Ca²⁺], and active force development). Furthermore, any rapid decline of force during contraction leads to a transient increase in the [Ca²⁺] surrounding the myofibril, suggesting that Ca²⁺ dissociates rapidly when the muscle is unloaded (1, 2, 65). Such a transient increase of [Ca²⁺]_i upon rapid unloading also occurs in skinned fibers, demonstrating that it is a property of the myofibrillar apparatus (1).

After relaxation, both force and SL as well as [Ca²⁺]_i remain constant without perceptible fluctuations (133) and the microscopic image of the muscle is crystal clear and completely quiescent. Studies using confocal microscopy after loading of the muscle with the ester of the Ca²⁺ probe fluo-4 AM confirm that the cells rarely show Ca²⁺ sparks and no Ca²⁺ waves (G. L. Wang and H. E. D. J. ter Keurs, unpublished data). These observations support the notion that cells in intact normal undamaged myocardium are quiescent during diastole unless the [Ca²⁺] in the medium surrounding the muscle is raised above 2 mM. The consequence of the resulting low [Ca²⁺] in the cytosol is that the passive force in the structure is completely supported by the titin molecules inside the myocytes (see Fig. 1) and the collagen meshwork surrounding myocytes and collagen bundles alongside the fascicles (10, 55, 57). The combined compliance of titin and perimysial collagen allows stretch of the sarcomeres to the usual end-diastolic length of ∼2.1 μm by a stress of 0.06 kPa. Cardiac sarcomere stretch above ∼2.3 μm is rendered impossible when the collagen bundles parallel to the fascicles are made taut by the stretch. The importance of the role of the collagen bundles is that the cardiac sarcomere always operate on the ascending limb of the force length relationship.

Contraction of uniform muscle: interaction between SL, [Ca²⁺], and active force development.

The release of Ca²⁺ by the SR during a normal heartbeat brings the free [Ca²⁺]_i from ∼0.1 to ∼1 μM. TnC has one binding site for Ca²⁺ ions in this range of [Ca²⁺]_i . Binding of Ca²⁺ to TnC enables force development by myosin cross bridges. A characteristic of cardiac muscle is that at a [Ca²⁺]_i of 2 μM, the contractile response saturates (4, 145). At saturating [Ca²⁺]_i, force increases with stretch (144) over the range of SL at which the heart operates (1.5 to 2.3 μm) (115). The shape of the resultant relationship between SL and Ca²⁺-saturated force is quite similar to the relationship that was predicted by Huxley and colleagues (54, 144) for skeletal muscle based on the length-dependent overlap of myosin cross bridges with the actin filament (22).

The response of force to [Ca²⁺]_i follows Hill kinetics [EC₅₀ = 0.65 μM; n = ∼5.2 (4)], but the Hill coefficient is fivefold higher than is predicted for the binding of Ca²⁺ to one low affinity site on TnC (36). Furthermore, the sensitivity of the contractile response at submaximal [Ca²⁺]_i levels is strongly SL dependent: a phenomenon known as length-dependent activation (LDA) (44, 60, 78, 143). This remarkable phenomenon is found in all striated muscle but is stronger in cardiac muscle than in fast skeletal muscle, which again is more length dependent than slow skeletal muscle (80). Both the ESFSLR and the force [Ca²⁺]_i relationship suggest that factors which depend on length facilitate binding of Ca²⁺ to TnC by one or more cooperative mechanisms.

Several of the possible mechanisms contributing to LDA have recently been reviewed by de Tombe et al. (32). We suspect that more than one mechanism may be in operation, which makes the unraveling of the individual contributions clearly quite challenging. The first mechanism that has been proposed is that sarcomere operates at constant volume; hence, length changes are accompanied by changes in the spacing of the actin and myosin lattice, which could mediate LDA by modulating the probability of cross-bridge attachment. Titin could play a role in the maintenance of the constant volume of the sarcomere and, thereby, play a role in LDA (22, 51); alternatively, Titin could influence the binding probability of cross bridges by aligning the cross bridges within the lattice (45). Elegant X-ray diffraction experiments on intact rat cardiac trabeculae by de Tombe and collaborators show little correlation between lattice spacing and the Ca²⁺ sensitivity of the contractile system, suggesting a minor, but still debated (50), role for lattice spacing-dependent mechanisms.

Several other mechanisms of cooperativity at the level of the actin-troponin-tropomyosin and cross-bridge interaction with actin complex are intensely studied: 1) TnC bound to Ca²⁺ enhances binding of Ca²⁺ ions by neighboring TnC (53, 137), 2) length dependence of the effect of TnC-Ca²⁺ activation on the interaction between troponin-tropomyosin units (44), 3) cross-bridge attachment enhances binding of Ca²⁺ ions to adjacent TnC (129), and 4) force exerted by the cross bridges on the actin filament deforms the thin filament; the resulting deformation of the actin-troponin-tropomyosin complex enhances binding of Ca²⁺ ions to TnC.

The responsible feedback mechanisms have not yet been resolved but are clearly of outstanding importance because it permits a highly effective coupling between the Ca²⁺ release process and the mechanical response of the cardiac sarcomere to the released Ca²⁺ ions.

Modeling of ECC, done to explore the effect of feedback of force to the off rate of Ca²⁺ from TnC (97, 144) on contractile mechanics, has shown that this particular mechanism may explain many fundamental properties known for cardiac muscle, including the fact that the muscle behaves during the heartbeat as if it were in a steady state and shows only little effect of shortening against a load, thereby explaining a fixed ESFSLR, as observed experimentally (134). Hence, the model predicts the well-known ESPVR of the left ventricle as well as Starling's Law of the heart (104, 105, 144). The time course of [Ca²⁺]_i and force during contraction predicted by the model matches experimental data on trabeculae (such as shown in Fig. 2). An important consequence of this behavior is that a force decrease accelerates the dissociation of Ca²⁺ from Ca²⁺·TnC. This phenomenon is experimentally well known (65, 81) and has been shown to be arrhythmogenic (153).

The molecular mechanism that dictates the TnC·Ca²⁺ off rate is unknown, but in view of the realistic reproduction of the steady-state force-[Ca²⁺] relationships and ESFSLR at varied activation levels for [Ca²⁺]_i as well as the realistic prediction of the [Ca²⁺]_i transients at varied force levels, the mechanism clearly merits further investigation.

Cardiac Dysfunction

Reduced and sluggish Ca²⁺ release and slow removal.

The most common cause of failure of the heart is dilatation and dysfunction of the pump after myocardial infarction. Myocardial infarction of the left ventricle causes remodeling of the infarct area, which turns into a scar, and of the remaining viable myocardium so that it can take on the increased workload. With time, though, a substantial infarct may be followed by CHF, characterized by a reduced force of the heartbeat, as well as a slower onset of contraction and slowed relaxation. Another hallmark of CHF is that the inotropic effect of increased heart rate is lost (93), so that the heart loses a powerful feedback mechanism which normally allows it to respond to an increased hemodynamic load. Hence, the heart will rely more on other feedback mechanisms, such as increased filling pressure and increased β-adrenergic drive, which may enhance further pathophysiological remodeling.

There is considerable evidence that failure of Ca²⁺ transport in myocytes is a central cause of contractile dysfunction in CHF. The problem starts with action potential generation, as loss of transient outward K⁺ current (14, 95) eliminates the rapid early repolarization and increases the duration of the action potential. The loss of the early repolarization reduces the driving force for the phasic component of L-type Ca²⁺ current. The reduced Ca²⁺ influx both triggers a smaller Ca²⁺ release through the SR-Ca²⁺ channels (25, 106) and delays the activation of NCX so that the contribution of Ca²⁺ entry through early reverse mode NCX will decrease, thus further lowering the trigger for SR-Ca²⁺ release (148). Both mechanisms would lead to a reduction of force, especially at increased heart rate. In addition, reduced SR-Ca²⁺ pump activity and an increased expression and activity of the NCX (64) not only slow down relaxation but also reduce the SR-Ca²⁺ load and force, particularly at an increased heart rate. Gomez et al. (52) showed in models ranging from left ventricular hypertrophy to CHF no changes of sparks compared with control and suggested that abnormalities in ECC in CHF myocytes may be due to changes in the spatial organization of the dyad. Some RyRs in transverse tubuli are still coupled to L-type Ca²⁺ channels, but other RyRs may be “orphaned” in CHF (52); that is, RyRs are still able to function but are “physically” isolated from Ca²⁺ influx channels, predicting reduced global cellular Ca²⁺ transients (119) and reduced force.

The duration of action potential prolongation increases the tonic Ca²⁺ influx (149) increasing SR-Ca²⁺ load (116), which would counter the negative inotropic effects of the aforementioned changes in phasic L-type Ca²⁺ current, SERCA2, and NCX. On the other hand, there is mounting evidence that the maintenance of structurally intact transverse tubules is also at fault in CHF: the tubules are not rigorously connected to the surface membrane anymore, and their regular structure is lost (20). We know that the transverse tubules carry more than half of the NCX and L-type Ca²⁺ channel proteins, thus contributing substantially to the SR-Ca²⁺ load. Loss of this source of Ca²⁺ would contribute to a reduced inotropism of increased heart rate in CHF. Furthermore, cell regions that lose their tight coupling to the action potential owing to the loss of the transverse tubular integrity will now rely more on Ca²⁺ diffusion and propagated Ca²⁺ release, and the rate of development of force will decrease. The loss of the proximity of NCX also predicts a shift of Ca²⁺ removal toward the downregulated SERCA2, slowing relaxation in the CHF muscle even more.

Ca²⁺ at the wrong time: Ca²⁺ waves in CHF.

Nonfailing trabeculae typically are quiescent during diastole at physiological extracellular Ca²⁺ concentration ([Ca²⁺]_o). Figures 2A and 3A illustrate both the quiescence during diastole and reproducible shortening during the twitch as well as rapid uniform lengthening during relaxation (31). Figure 3 shows dramatically that this uniform behavior is lost in trabeculae from the heart of rats with CHF, where spontaneous sarcomere motion occurs at low [Ca²⁺] and steeply increased with further increase of [Ca²⁺]_o. This increase was sevenfold larger in the muscles from CHF compared with controls. A microscopic inspection of these muscles confirmed that the spontaneous SL variations were in fact caused by Ca²⁺ waves causing contractions that propagate within individual cells. The spontaneous diastolic activity that caused that shortening of sarcomeres following an action potential became highly variable and reduced the force of the twitch by as much as 30% (98). Spontaneous diastolic activity also increased following stimulation at a high rate and during catecholamine stimulation in cardiac muscle from CHF, contributing to the loss of inotropic effects of increased heart rate and sympathetic stimulation.

Fig. 3.

Sarcomere uniformity is lost in congestive heart failure (CHF). A: sarcomere shortening during twitches of hypertrophied but still compensated cardiac muscle. B: sarcomere shortening of failing cardiac muscle from rats 4 mo following coronary artery ligation. Right ventricular hypertrophy (RVH) after a myocardial infarction does not affect quiescence of the Ca²⁺ transport system (in A). However, development of frank CHF by the animal is associated with dramatic spontaneous diastolic Ca²⁺ release, leading to spontaneous diastolic sarcomere contractions and loss of active force. Inset: peak of the power spectrum (PSD) of the SL fluctuations occurs at ∼3 Hz, which is similar to that of the unloaded twitch.

The spontaneous diastolic contractile activity in CHF appeared to be due to a greater sensitivity of SR-Ca²⁺ channels to the SR-Ca²⁺ load. Fabiato and colleague (40, 41) described in the early 70s that the SR starts to release its Ca²⁺ when the amount of Ca²⁺ in the SR exceeds a threshold level. Indeed, reduction of the SR-Ca²⁺ load in muscles from animals with CHF, for example, by a small amount of ryanodine, eliminated the spontaneous contractions and restored force development. The failure of membrane L-type Ca²⁺ channel blockers to do the same reinforces the notion that the malfunction underlying spontaneous diastolic SR-Ca²⁺ release in CHF resides in the SR-Ca²⁺ channel. These spontaneous Ca²⁺ transients are known to lead to arrhythmogenic membrane depolarizations (144).

The molecular mechanism underlying the acquired abnormality of SR-Ca²⁺ channels in CHF is still elusive, although several mechanisms have been proposed, including channel oxidation or nitrosylation (31) and excessive phosphorylation of the channel by activated protein kinase A, causing FK506-binding protein 12.6 to dissociate from the channel and increase its open probability (88). Intuitively, this is an attractive proposition because β-blockers, especially carvedilol, form the backbone of treatment of CHF. However, both the questions as to whether the SR-Ca²⁺ channel is excessively phosphorylated in CHF and the effect of increased phosphorylation of the channel on dissociation of FK506-binding protein 12.6 are still debated (75, 158, 159). It has been known for a decade that mutations of the SR-Ca²⁺ channel protein are involved in malignant arrhythmias (108). Hence, it is fascinating to learn that common RyR₂ variants associate with ventricular arrhythmias and sudden cardiac death: the A allele of the rs3766871 variant associates with the incidence of lethal cardiac arrhythmias in acquired CHF, whereas the A allele of the rs790896 variant may act as a protective factor against sudden cardiac death (109).

Sarcomere dynamics in nonuniform muscle.

Nonuniformity of muscle properties is a scenario that is expected in diseased myocardium, for example during bouts of ischemia. We will see in the paragraphs below that this scenario may lead to arrhythmias because of the derailment of ECC. Here, we will summarize the effects of transient nonuniformity on sarcomere mechanics in weak and strong segments of such muscle. The reference scenario is uniform sarcomere behavior in the muscle when superfused by one solution at a fixed physiological [Ca²⁺]_o. Nonuniform muscle could be studied by mounting two inflow ports for solution in the bath: the main buffer solution with 1.2 mM [Ca²⁺]_o ran longitudinally through the bath; the other inflow of buffered saline entered from a side port in the bath and covered exactly half the muscle. Exposure of the muscle to a buffer solution with low [Ca²⁺]_o (0.4 mM) from the side port created in a reversible manner nonuniform ECC, consisting of a weak force-generating segment in low [Ca²⁺]_o and a strong segment in [Ca²⁺]_o (1.2 mM).

This paradigm caused sarcomere stretch in the jet-exposed region during the stimulated twitches (153) and led to the generation of arrhythmogenic Ca²⁺ waves starting in the exposed area.

Figure 4 illustrates the effect of nonuniformity on force development and shortening in a cardiac trabecula. Force was modulated by lowering [Ca²⁺]_o along the whole muscle or only along a segment of (∼50% of the length of) of the muscle. Reducing [Ca²⁺]_o uniformly along the muscle from 1.2 to 0.4 mM depressed force at all SL and caused a reduction of force of the twitch by 90% as well as reduced sarcomere shortening (Fig. 4, A and B). The force SL relationship (FSLR) of uniform muscles shows the well-known shapes at normal and low [Ca²⁺]_o (Fig. 4D) (143). In contrast, when only half of the length of the muscle was exposed to 0.4 mM [Ca²⁺]_o, force decreases by only 60% (Fig. 4, A and C), showing that the segment that was weakened by exposure to a low [Ca²⁺]_o is able to support the higher force. This raises the question as to whether the force generation by the weak segment can fully be explained by the mechanics of contraction and specifically by the FSLR and force-sarcomere shortening velocity relationship (FSVR) of the sarcomeres. Indeed, the FSLR at lowered [Ca²⁺]_o shows that maintaining a longer SL in the weak segment is responsible for a 2.5-fold increase of force in the weak segment (Fig. 4D). The remainder of the force enhancement is due to the act of stretching the weak segment during the contraction, as is shown by analysis of the effect of stretch on the FSVR (Fig. 4E) (30). The FSVR and maximal shortening velocity shown in Fig. 4D match closely the well-known FSVR in cardiac muscle previously reported (30, 33, 37), but the data extend the FSVR to the force range where the muscle is stretched (30, 37, 63). Figure 4E shows that the shortening velocity in the strong muscle segment matched the lengthening velocity in the weak segment as one would expect when overall muscle length is constant. Along the same line of reasoning, sarcomere shortening in the strong segments appears to be responsible for a force deficit of 30% by the FSVR (Fig. 4) and another 30% by the FSLR. These data show that the force generating capacity in nonuniform muscle is completely explained by the FSLR and FSVR of the weak and strong segments of the muscle. It is clear that the FSLR and FSVR indeed explain that force in the nonuniform contraction is fourfold higher than the force when the whole muscle is exposed to 0.4 mM [Ca²⁺]_o. Force in nonuniform myocardium is sustained by stretch of the weak segments. It is noteworthy that the weak segment exhibits rapid shortening during the decline of force in the relaxation phase. We will see that this phenomenon may play a role in the arrhythmogenic nature of contraction by nonuniform muscle.

Fig. 4.

Force development in a uniform and nonuniform trabecula. A and B: twitch force and sarcomere shortening during the twitch at normal and low [Ca²⁺] during nonuniform contraction (blue and green traces); the red traces show force and shortening in the strong segment of the same trabecula that is exposed to 1.2 mM [Ca²⁺], whereas the weak segment is exposed to 0.4 mM [Ca²⁺]. C: simultaneous SL records in the strong and weak segment of a nonuniform muscle as well as SL in the border zone (see text). D: the force SL relationship (FSLR) of the unstimulated muscle (red curve) and at peak force at 1.2 mM [Ca²⁺] (blue curve) and 0.4 mM [Ca²⁺] (green curve). Arrows indicate effect of SL change in the strong and the weak segments in the nonuniform muscle. Stretch of the weak segment as in C enhances its force development 2.5-fold. E: force-sarcomere shortening velocity relationship (FSVR) of the same muscle as in A. The passive and active FSLR are shown for 0.4 mM extracellular Ca²⁺ concentration ([Ca²⁺]_o) and 1.2 mM [Ca²⁺ ]_o. Force in the weak segment is enhanced 2.5-fold by the fact that the sarcomeres of the weak segment are forced to operated at a higher SL (black arrow). D: FSVR shows that the shortening velocity in the strong segment is offset by the lengthening velocity from weak segment at peak force of the nonuniform muscle (dashed line), thereby enhancing force another 1.5-fold. The dashed line shows a balance of velocity of sarcomere shortening in the strong and weak segments (For further explanation, see text). %F_max, percent maximum force.

Reverse Excitation-Contraction Coupling

We discovered arrhythmogenic contractile waves initially in damaged muscle (142). When cardiac muscle is damaged locally, in and near the damaged region Ca²⁺ waves are initiated, which then propagate in a coordinated fashion into adjacent tissue (27, 142). Several observations suggest that aftercontractions in multicellular preparations occur as the combined result of the mechanical effects and elevated cellular Ca²⁺ levels owing to regional damage and may give rise to premature beats as well as triggered arrhythmias. Unique aspects of the aftercontractions that arise in damaged regions of cardiac muscle are that they appear to be initiated by the stretch and release of the damaged, and apparently weaker, region during the regular twitch and that they propagate into neighboring myocardium; hence, the term triggered propagated contractions (TPCs) (151).

We hypothesized that the initiation of the TPCs was due to effects of the arrangement of a strong normal muscle segment in series with a segment that is weaker owing to the effect of damage. The initial studies on sarcomere mechanics of isolated rat ventricular and human atrial trabeculae (28) suggested that the TPCs may be caused by Ca²⁺ release, first by the myofilaments and then by the SR in the cells near the damaged region. Local release of Ca²⁺ in the damaged region causes a Ca²⁺ transient that is conducted into adjacent muscle by the combination of Ca²⁺ diffusion and Ca²⁺-induced Ca²⁺ release, causing the TPCs. Figure 5 shows a typical example of both a TPC and of a propagating intracellular Ca²⁺ wave (89). The latter causes a depolarization of the cell membrane, delayed afterdepolarization (DAD), which may reach threshold and become responsible for action potential generation (27, 136). The sequence of events generating the TPCs is as follows: Ca²⁺ release by the myofilaments leading to a propagated wave of SR-Ca²⁺ release into adjacent muscle that causes a DAD is clearly the reverse of the normal ECC process and was coined by us as RECC (18, 140). RECC may be the mechanism that couples regional damage, for example, in ischemic heart disease with the initiation of premature beats and arrhythmias in the adjacent myocardium. TPCs appeared to arise in the ends of the trabeculae (151), which are easily damaged by dissection and mounting of the muscle. The difficulty to control such damage of a small segment of the muscle could be overcome in studies in which a segment of the muscle was again superfused, now with a small jet of bathing solution that differs from the main solution in the muscle bath.

Fig. 5.

A: examples of propagated contractions triggered in a damaged region (left attachment side to as force transducer) [adapted from ter Keurs et al. (146) with permission]. TPC, triggered propagated contraction. The corresponding intracellular Ca²⁺ waves are illustrated in B [adapted from Miura et al. (90) with permission] in which the relative [Ca²⁺ ]_i is color coded.

With this approach it was possible to weaken the segment in the jet reversibly by either reducing the superfusate [Ca²⁺] or by emptying the SR by adding caffeine to the jet solution or by paralyzing the contractile apparatus using 2,3-butanedione monoxime. These studies have shown that the force of the twitch of the muscle with the weak and the strong segment in series is reduced indeed (Fig. 5A) and that an intracellular Ca²⁺ transient (the Ca²⁺ surge) is triggered during the rapid decline of the electrically driven twitch. The Ca²⁺ surge appeared to give rise to a propagating Ca²⁺ wave that invaded the regions of muscle with a functional SR (153). The trigger site for the Ca²⁺ surge is always the region with weakened but preserved contractile activity (144, 153); triggering is enhanced by increasing the rate of decline of force during relaxation (155). The propagation velocity of the Ca²⁺ waves is dictated by the magnitude of the original Ca²⁺ surge and by the Ca²⁺ load in the cells involved in the propagation process (136). The propagating Ca²⁺ wave is invariably accompanied by an electrotonically spreading afterdepolarization (DAD) (27, 28).

The occurrence of a DAD during propagated Ca²⁺ waves and TPCs predicts that RECC is arrhythmogenic, because every action potential should lead to a contraction and every contraction would lead, by RECC, to a DAD which again could trigger an action potential. This prediction was borne out both in experiments on muscles with small damaged region (27, 92) and in the experiments with the jet-exposed muscles: when the muscle segment was weakened, stimulation of the muscle steadfastly led to triggered arrhythmias (see Fig. 6). Restoration of the uniformity of the muscle by superfusion with a single superfusate rapidly stopped the arrhythmias (Fig. 6) (153).

Fig. 6.

A: effects of a local jet with 2,3-butanedione monoxime (BDM): A jet containing BDM inhibits the activation of sarcomeres in the jet, resulting in regional stretch of the muscle in the jet. The spatiotemporal Ca²⁺ distribution (righthand panels) shows that Ca²⁺ waves arise from the border between regions with and without the BDM jet following a repeated stimulation. Arrowhead indicates moment of electrical stimulation. At low levels of activation, this region generates only a localized Ca²⁺ surge during the late part of the relaxation phase (middle). F/F_max, ratio of force to maximum force. B: nonuniform excitation-contraction coupling caused by the jet containing BDM (20 mmol/l) is arrhythmogenic. Recording of force showing that stimulus trains during local exposure to BDM (gray bars below the tracings) consistently induced arrhythmias. An expanded force tracing shows that spontaneous contractions were both preceded and followed by aftercontractions induced by the stimulus train. OFF (arrow) indicates when the jet was turned off; this rapidly eliminated the contractile nonuniformity and its arrhythmogenic effects. S indicates stimulus trains [reproduced from Wakayama et al. (153) with permission].

The observation that TPCs and the underlying Ca²⁺ waves propagate at speeds of several millimeters per second (92) is surprising since many studies have shown and predicted from models of ECC speeds <1 mm/s in isolated myocytes (74), although modeling of the propagation process in the presence of a high occupancy of the ligands (such after a stimulus train that induces the TPCs in these experiments) suggested the possibility of velocity of propagation of the Ca²⁺ wave in excess of 5 mm/s (3). It does not seem likely that Gd³⁺-sensitive membrane stretch-activated channels play an important role in enhancing the velocity of propagation (161). Further work is needed to test the possible mechanisms underlying a high velocity of propagation, including the effect of the depolarization that accompanies the Ca²⁺ wave. Figure 7 illustrates that RECC incorporates all mechanisms involved in the normal forward ECC, with one specific difference; i.e., the effect of a rapid release of Ca²⁺-activated sarcomeres leads to a surge of Ca²⁺ ions released into the cytosol. The coupling mechanism between the quick mechanical release and the Ca²⁺ surge is hitherto hypothetical. We chose a mechanism that makes use of the aforementioned feedback of force to the off-rate of Ca²⁺ from TnC, whereby a quick reduction of force would enhance the dissociation rate of Ca²⁺ and create a robust cytosolic Ca²⁺ transient following relaxation [see for detailed discussion (144)]. This Ca²⁺ surge, at that moment of the cardiac cycle, meets a SR in which the RyR Ca²⁺ channels have sufficiently recovered (9) to respond by vigorous SR-Ca²⁺ release. The released Ca²⁺ diffuses through the cytosol and triggers SR-Ca²⁺ release in the adjacent sarcomeres, generating a wave of Ca²⁺ release and the resulting TPC. The velocity at which the Ca²⁺ wave propagates depends on the magnitude of the initial Ca²⁺ surge (155) as well as on the SR-Ca²⁺ load, which dictates the rate of SR-Ca²⁺ release, the amount of Ca²⁺ released, the threshold of the RyR channels (75). In addition, Ca²⁺ diffusion accelerates when the cytosolic ligands are still partially occupied by Ca²⁺ (3). Gap junctions between cells form an additional barrier, which reduces the rate of Ca²⁺ diffusion from cell to cell and may stop propagation at low SR-Ca levels (82), whereas increasing gap junctional resistance blocks Ca²⁺ waves even at high SR-Ca²⁺ load (160).

Fig. 7.

A model for Ca²⁺-mediated arrhythmias: Rapid reduction of force (e.g., quick release of a weak segment by relaxation of the strong segment) induces a surge of Ca²⁺ ions coming off troponin C. The SR is sufficiently recovered to respond by local Ca²⁺ release. Ca²⁺ diffuses along the cell and triggers further Ca²⁺ release in the form of Ca²⁺ waves. The rise of [Ca²⁺] owing to these events leads to NCX-mediated Ca²⁺ extrusion. This process depolarizes the cell [delayed afterdepolarization (DAD)] and may initiate a new action potential. Modified from ter Keurs et al. (144) with permission.

It follows that the electrogenic Ca²⁺ removal by means of NCX at the membrane and the amplitude of the resultant DAD (77) is proportional to the amplitude of the Ca²⁺ wave and the propagation velocity, as has indeed been found (27, 91). The magnitude of the DAD itself is the main parameter that controls the probability of action potential generation that would lead to a premature beat or even an arrhythmia.

Summary and Outlook

In summary, a near molecular model of ECC is gracefully evolving based on the spectacular development of imaging techniques with mesoscopic resolution and Ca²⁺-sensitive intracellular probes. The current state of the ECC model explains many features of the synchronized action of the billions of sarcomeres in the cardiac pump chambers. The most pressing, lacking insight is still the last step of the ECC model: the mechanisms that render the response of the regulatory troponin-tropomyosin complex in the cardiac sarcomere to Ca²⁺ exquisitely length dependent. I am in no doubt that this century's old enigma underlying the Frank-Starling Law of the heart will be clarified in the next decennium.

It is also clear that the model of ECC can be used to predict the cardiac cycle only if one assumes uniformity of the underlying structure. For understanding of the normal heart, the model should be refined by incorporating detailed properties of the endocardial, midmyocardial, and subepicardial layers. In addition, the heart in disease is commonly nonuniform; further studies of microscopic behavior of ECC and RECC are needed to generate targets for molecular therapies in these conditions.

This review did not touch on the effect of second messenger systems that modulate function of the normal heart in response to the requirement of the circulation, e.g., by accelerating all phases of the cardiac cycle during tachycardia (72). Neither did we touch on the rapidly burgeoning field of posttranslational modification of the sarcomeric proteins-involved control of its function in health and disease (124, 125), a topic that clearly merits a separate review.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants HL-58860-O6A2 and HL-66140 and grants from the Canadian Institutes for Health Research and the Heart and Stroke Foundation of Alberta the North West Territories and Nunavut.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author (s).

AUTHOR CONTRIBUTIONS

Author contributions: H.E.D.J.T.K. prepared figures; H.E.D.J.T.K. drafted manuscript; H.E.D.J.T.K. edited and revised manuscript; H.E.D.J.T.K. approved final version of manuscript.