Ups and downs of calcium in the heart

Abstract

Contraction and relaxation of the heart result from cyclical changes of intracellular Ca²⁺ concentration ([Ca²⁺]_i). The entry of Ca²⁺ into the cell via the L‐type Ca²⁺ current leads to the release of more from the sarcoplasmic reticulum (SR). Compared to other regulatory mechanisms such as phosphorylation, Ca²⁺ signalling is very rapid. However, since Ca²⁺ cannot be destroyed, Ca²⁺ signalling can only be controlled by pumping across membranes. In the steady state, on each beat, the amount of Ca²⁺ released from the SR must equal that taken back and influx and efflux across the sarcolemma must be equal. Any imbalance in these fluxes will result in a change of SR Ca²⁺ content and this provides a mechanism for regulation of SR Ca²⁺ content. These flux balance considerations also explain why simply potentiating Ca²⁺ release from the SR has no maintained effect on the amplitude of the Ca²⁺ transient. A low diastolic [Ca²⁺]_i is essential for cardiac relaxation, but the factors that control diastolic [Ca²⁺]_i are poorly understood. Recent work suggests that flux balance is also important here. In particular, decreasing SR function decreases the amplitude of the systolic Ca²⁺ transient and the resulting decrease of Ca²⁺ efflux results in an increase of diastolic [Ca²⁺]_i to maintain total efflux.

Introduction

It is well established that contraction of the heart is controlled by changes of cytoplasmic Ca²⁺ concentration ([Ca²⁺]_i). [Ca²⁺]_i must be high enough in systole to activate the contractile proteins in order to pump blood out of the heart. Equally importantly, it must fall during diastole to low enough levels that the muscle of the heart relaxes so that the chambers can refill with blood. The aim of this article is to discuss factors responsible for both the rise and fall of [Ca²⁺]_i.

The importance of Ca²⁺ in contraction of the heart was originally demonstrated by Sydney Ringer (Ringer, 1883). I have recently described both the serendipitous nature of his discovery and the manner in which it was published (Eisner, 2014). In the subsequent 130 years it has become clear that changes of [Ca²⁺]_i regulate the function of virtually all organs. It took, however, almost 100 years after Ringer's discovery to measure directly the transient elevation of [Ca²⁺]_i (the Ca²⁺ transient) which underlies cardiac contraction (Allen & Blinks, 1978; Allen & Kurihara, 1980).

Calcium versus other signalling modalities

Calcium has its actions by binding to proteins and thereby changing their shape and properties (Fig. 1 A). In the case of striated muscle, Ca²⁺ binds to troponin and the resulting movement of tropomyosin allows actin and myosin to interact. There are, of course, many other ways to alter the properties of proteins including phosphorylation and nitrosylation. This immediately raises the question, what are the advantages and disadvantages of using Ca²⁺ to modify protein structure? One great advantage of Ca²⁺ is its speed of action. [Ca²⁺]_i rises and Ca²⁺ ions bind to troponin very quickly. In cardiac muscle contraction can be 50% complete within 200 ms (Fig. 1 A). In contrast regulation by phosphorylation is much slower. Figure 1 B shows the effects of β‐adrenergic stimulation on the Ca²⁺ current. This takes about 20 s to be 50% complete (Frace et al. 1993). This slower time scale is due to the many steps involved; agonist binding to the receptor, generation of cyclic AMP, and, finally, the activation of protein kinase A which then phosphorylates the L‐type channel. Although signalling by phosphorylation is slower than that by Ca²⁺, it has the advantage that it can be regulated in a much more subtle way. Specific phosphodiesterases can break down cAMP and various regulators affect the activities of the kinases and phosphatases. Another significant advantage of control by phosphorylation is that it can be reversed simply by dephoshorylation. In contrast, Ca²⁺ ions cannot be destroyed. The only way to reverse Ca²⁺ binding is to pump the Ca²⁺ across a membrane, either out of the cell or into an intracellular store. Put simply, Ca²⁺ must be recycled and many of the results and conclusions of the remainder of this review follow from this.

Figure 1. Comparison of control by Ca²⁺ and phosphorylation.

A, Ca²⁺. The upper part shows a schematic diagram of a protein changing shape when it binds Ca²⁺. The lower part shows time course of (from top to bottom): action potential, [Ca²⁺]_i and cell shortening recorded from a sheep ventricular myocyte. (Record kindly provided by K. M. Dibb.) B, phosphorylation. The upper part shows the change of shape produced by phosphorylation and its reversal by dephosphorylaton with a phosphatase. The steps between binding to the receptor and phosphorylation are shown: agonist binds to the β‐adrenergic receptor (β‐AR) which, via a G‐protein, activates adenylate cyclase (AC) producing cyclic AMP (cAMP) which activates protein kinase A (PKA). (Illustrations by Jessica Caldwell.) The lower part shows the time course of the effects of isoprenaline (ISO) on the L‐type Ca²⁺ current in a frog ventricular myocyte. Trace reproduced with permission from Frace et al. (1993).

The above argument needs some qualification. Although it describes well the situation in striated muscle, where Ca²⁺ binding is followed rapidly by contraction, in other tissues the steps subsequent to Ca²⁺ binding may involve slower, enzymatic processes. This is, for example, the case in smooth muscle where Ca²⁺ binds to calmodulin, activating the enzyme myosin light chain kinase which, in turn, phosphorylates myosin. There is therefore a longer delay between the rise of [Ca²⁺]_i and that of contraction (Shabir et al. 2004). It should also be noted that factors other than speed are also involved in the evolution of Ca²⁺ signalling (Case et al. 2007).

Flux balance

The events underlying Ca²⁺ cycling during the heartbeat are illustrated in Fig. 2 (see Bers (2001, 2008) and Eisner et al. (2017) for reviews). Ca²⁺ enters the cell through the L‐type Ca²⁺ channel resulting in a large increase of [Ca²⁺]_i in the small dyadic space between the surface membrane/transverse tubule and the sarcoplasmic reticulum (SR). This Ca²⁺ binds to the SR Ca²⁺ release channel (ryanodine receptor; RyR) making it open and releasing a much greater amount of Ca²⁺ into the cytoplasm from the SR, a process known as Ca²⁺‐induced Ca²⁺ release (Fabiato, 1985). Relaxation requires that [Ca²⁺]_i be lowered to resting levels. This occurs in two ways. (1) Ca²⁺ is pumped back into the SR by the sarco/endoplasmic reticulum Ca²⁺‐ATPase (SERCA). The activity of SERCA is regulated in part by [Ca²⁺]_i and also by the inhibitory accessory protein phospholamban. Phosphorylation of phospholamban, as occurs during β‐adrenergic stimulation, relieves the inhibition and accelerates SERCA (Kirchberber et al. 1975). (2) Ca²⁺ is also removed from the cell, largely by the Na⁺/Ca²⁺ exchanger (NCX), which uses the entry of three Na⁺ to power the exit of one Ca²⁺ from the cell (Kimura et al. 1986). The resulting electrogenic current can be used to measure the activity of NCX in the intact cell. It is important to note that the various Ca²⁺ fluxes must be balanced on each cycle. In other words, in the steady state, the Ca²⁺ influx into the cell must exactly equal that pumped out. Likewise, the amount of Ca²⁺ released from the SR must equal that pumped back. This Ca²⁺ flux balance is an example of the need to recycle Ca²⁺ mentioned above.

Figure 2. Control points in cellular Ca²⁺ cycling.

A, schematic diagram showing Ca²⁺ fluxes. (1) Ca²⁺ enters the cell via the L‐type Ca²⁺ current. (2) Ca²⁺ is released via the RyR from the SR. (The amplitude of the Ca²⁺ transient depends, therefore, on the sum of Ca²⁺ entry and release from the SR.) (3) Ca²⁺ is pumped back into the SR by SERCA and pumped out of the cell largely via NCX. B, the dependence of the amplitude of the Ca²⁺ transient on SR Ca content. (Redrawn from Trafford et al. (1997).)

Control of the amplitude of the Ca²⁺ transient

The size of the systolic Ca²⁺ transient is not fixed. It increases, for example, during exercise due to β‐adrenergic stimulation (Endoh & Blinks, 1988) and decreases in heart failure (Beuckelmann et al. 1992). Three of the steps indicated in Fig. 2 A are potential control points. (1) The size of the L‐type Ca²⁺ current; the larger the current, the greater the triggering of RyRs to open. (2) The ease with which the RyRs open. I will show later that this is probably not an important control point. (3) The amount of Ca²⁺ in the SR. Figure 2 B shows that, as SR Ca content increases, there is a steep increase of the amplitude of the Ca²⁺ transient (Bassani et al. 1995). On average, the amplitude of the Ca²⁺ transient is proportional to the cube of SR Ca content (Trafford et al. 2000). Several factors account for the steepness. An increase of SR luminal Ca²⁺ increases the driving force for Ca²⁺ to leave the SR. It will also increase the open probability of the RyR. This can occur either by Ca²⁺ binding directly to sites on the luminal part of the RyR or by the Ca²⁺ which flows out through the RyR binding to sites on the cytoplasmic surface (Guo et al. 2012). Whatever the origin of the steep relationship, it is of great significance. For example, a small reduction of SR Ca²⁺ content in cardiac hypertrophy will have a much larger fractional effect on the amplitude of the systolic Ca²⁺ transient (Díaz et al. 2004). Furthermore, the steep dependence means that regulation of the amplitude of the systolic Ca²⁺ transient requires exquisitely precise control of SR Ca content. This therefore raises the question of how SR Ca content is controlled.

Control of SR Ca content

Much work has shown that SR Ca content in the heart is controlled by a simple, negative feedback mechanism (Trafford et al. 1997). This depends on the fact that (i) the Ca²⁺ transient amplitude is a steep function of SR Ca content and (ii) fluxes of Ca²⁺ across the sarcolemma are sensitive to the amplitude of the Ca²⁺ transient. An example of this mechanism is shown in Fig. 3 A. A high (10 mm) concentration of caffeine had been applied to empty the SR. Following removal of caffeine, the cell was electrically stimulated. The Ca²⁺ transient was initially very small as a result of the low SR Ca content but then increased over the course of about 10 beats as the SR refilled. The underlying changes of sarcolemmal fluxes are illustrated in Fig. 3 B. When the Ca²⁺ transient is small (a), there is less Ca²⁺‐dependent inactivation of the L‐type current (Sipido et al. 1995) and therefore increased influx. There is also less efflux on NCX. The cell is therefore not in Ca²⁺ flux balance and gains Ca²⁺. The resulting increase of SR Ca content increases the amplitude of the Ca²⁺ transient making the L‐type current inactivate more quickly, decreasing Ca²⁺ influx and there is also more efflux on NCX such that a new steady state is reached.

Figure 3. Negative feedback control of SR Ca²⁺ content.

A, time course. The trace shows [Ca²⁺]_i from a ferret ventricular myocyte stimulated with voltage‐clamp pulses. Before the record began, the cell had been exposed (in the absence of stimulation) to caffeine (10 mm) to empty the SR. Caffeine was then removed. The trace shows the effects of recommencing stimulation. B, expanded traces of L‐type current (left) and NCX (right). In both panels the traces correspond to the points in A; i.e. a is the first record after commencing stimulation and b the steady state. (Traces in A and B redrawn from Trafford et al. (1997).) C, schematic diagram of negative feedback loop. An increase of SR Ca²⁺ increases the amplitude of the Ca²⁺ transient. This decreases Ca²⁺ influx via the L‐type Ca²⁺ current (1) and increases efflux on NCX (2) leading to a decrease of cell and therefore SR Ca²⁺ content. (A and B redrawn from Trafford et al. (1997).)

These effects of the Ca²⁺ transient on sarcolemmal Ca²⁺ fluxes underpin a simple yet powerful negative feedback loop which controls SR Ca²⁺ content (Fig. 3 C). This can be illustrated by considering the consequence of an increase of SR Ca content. This will increase the amplitude of the Ca²⁺ transient which, in turn, will increase Ca²⁺ efflux and decrease influx, thereby decreasing cell and SR Ca content. Not only does this mechanism regulate SR Ca content but it is fundamental to ensuring that the fluxes of Ca²⁺ into and out of the cell and across the SR membrane are balanced on each beat. Any imbalance of Ca²⁺ flux across the surface membrane will result in a change of cell and SR Ca leading to an alteration of the amplitude of the Ca²⁺ transient and thence to changes of Ca²⁺ fluxes that will return the cell to Ca²⁺ flux balance. Ca²⁺ fluxes can, of course, be out of balance but only for short periods. One example has already been given: the effect of stimulating when the SR is empty. Under these conditions, the cell is initially not in Ca²⁺ flux balance; influx is greater than efflux until the SR Ca²⁺ content and Ca²⁺ efflux increase sufficiently that efflux equals influx and the cell is back in flux balance. Another, well‐known example is when the frequency of stimulation is changed or stimulation is recommenced after a pause. For example, in species other than rodents, when stimulation is stopped Ca²⁺ leaks out of the SR (Bridge, 1986). The SR refills during stimulation resulting in a gradual increase of the amplitude of the Ca²⁺ transient and contraction (Allen et al. 1976).

The effects of increasing RyR opening

A particularly striking demonstration of the consequences of the maintenance of Ca²⁺ flux balance is provided by the effects of increasing the open probability of the RyR. Figure 4 demonstrates that the application of 500 μm caffeine to increase RyR opening produces an immediate increase of the amplitude of the systolic Ca²⁺ transient. This increase is, however, short‐lived and the amplitude of the Ca²⁺ transient decays to basal levels within a few beats (Trafford et al. 2000). The explanation of this result is as follows. Under control conditions, Ca²⁺ efflux and influx are equal. The increase in the size of the systolic Ca²⁺ transient produced by caffeine increases Ca²⁺ efflux until it is greater than influx resulting in a decrease of SR Ca content. (See Shannon et al. (2005) for detailed mathematical modelling of this phenomenon.) The decrease of SR Ca²⁺ content in Fig. 4 was calculated from the change of NCX flux. More direct evidence for it comes from showing that the amount of SR Ca²⁺ that can be released by a high concentration (10 mm) of caffeine is also decreased (Trafford et al. 1998) and measuring the decrease of SR Ca²⁺ directly with a low affinity Ca²⁺ indicator (Greensmith et al. 2014). The decrease of SR Ca²⁺, in turn, decreases the amplitude of the Ca²⁺ transient until a new steady state is reached at which Ca²⁺ influx and efflux are again equal. Similar transient effects are seen for other manoeuvres which affect the opening of the RyR (Choi et al. 2000; for review see Eisner et al. 2009). This transient response of Ca²⁺ transient amplitude to an increase of RyR opening can be thought of as an emergent property of the system. It results because the various, independent Ca²⁺ transporters are effectively coupled by their dependence on cytoplasmic and SR [Ca²⁺].

Figure 4. The effects of potentiating RyR opening.

The effects of adding caffeine (500 μm) to increase RyR opening in a rat ventricular myocyte. Traces show (from top to bottom): [Ca²⁺]_i; sarcolemmal Ca²⁺ fluxes (influx on L‐type Ca²⁺ current, efflux on NCX; predicted change of SR Ca content. (Redrawn from Trafford et al. (2000).)

The regulation of resting and diastolic [Ca²⁺]_i

For the heart to work properly as a pump, diastolic [Ca²⁺]_i must be sufficiently low that the ventricle is relaxed and can therefore refill with blood between beats. An increase of diastolic force, accompanied by impaired relaxation, is seen in many patients with heart failure. Furthermore, heart failure with preserved ejection fraction (HFpEF) is increasingly being regarded as an important clinical problem (Borlaug & Paulus, 2011). Such patients have symptoms of heart failure but normal ejection fractions and their major clinical problem is diastolic. While other factors may well also contribute to diastolic dysfunction, there is evidence that diastolic [Ca²⁺]_i is increased in patients with heart failure (Gwathmey et al. 1987, 1991; Beuckelmann et al. 1992; Sipido et al. 1998; Runte et al. 2017). It is therefore important to understand how diastolic [Ca²⁺]_i is regulated.

Control of resting [Ca²⁺]_i

Many studies have investigated the regulation of [Ca²⁺]_i in unstimulated (resting) cardiac preparations. Early work, performed before the introduction of Ca²⁺‐sensitive indicators, used resting force as a bio‐assay of [Ca²⁺]_i. The role of the NCX, for example, was originally demonstrated by examining the effects of changing extracellular Na⁺ concentration (Niedergerke, 1963; Chapman & Tunstall, 1980). Subsequent work measuring intracellular Na⁺ concentration showed that the voltage‐dependent NCX played a major role in controlling resting force and therefore [Ca²⁺]_i (Eisner et al. 1983). The advent of fluorescent Ca²⁺ indicators allowed more direct measurements and also identified a role for the plasma membrane Ca²⁺‐ATPase in the regulation of resting [Ca²⁺]_i (Lamont & Eisner, 1996). What is less well established is the route by which Ca²⁺ enters a resting cell. One possibility might be the random opening of L‐type Ca²⁺ channels but the likely flux would be low at the normal resting potential where the open probability is low. Work on rat ventricular myocytes found evidence for three voltage‐sensitive mechanisms involved in the control of resting [Ca²⁺]_i. (1) NCX; the greater the depolarization, the more the Ca²⁺ influx. (2) L‐type Ca²⁺ current; depolarization to about 0 mV increased influx but the magnitude decreased on further depolarization as the driving force for Ca²⁺ entry was reduced. (3) Finally, there was an increase of [Ca²⁺]_i on hyperpolarization, attributed to increased driving force (Kupittayanant et al. 2006) on an un‐gated Ca²⁺ entry pathway. Gadolinium at a concentration of 250 μm was the only substance found to completely inhibit this Ca²⁺ entry leaving its identity unresolved. One of the classes of channels which are sensitive to gadolinium is the so called transient receptor potential (TRP). There has been considerable interest in the properties of such channels in the heart with their being implicated in the development of hypertrophy (Gao et al. 2012) and the cell damage in muscular dystrophy (Lorin et al. 2015). Another channel which may contribute to the background Ca²⁺ entry mechanism is that responsible for store‐operated Ca²⁺ entry since its inhibition has been reported to decrease resting [Ca²⁺]_i in HL‐1 cells (Touchberry et al. 2011). The importance of store‐operated channels in normal adult cardiac myocytes is, however, still controversial (for review see Eisner et al. 2017).

Diastolic [Ca²⁺]_i

At first sight one might think that the diastolic level of resting [Ca²⁺]_i would be similar to the resting level in an unstimulated cell. In point of fact this is not the case and diastolic is higher (Fig. 5 and Dibb et al. 2007). This is because, in the steady state, in a resting cell, [Ca²⁺]_i is controlled solely by the surface membrane with the sarcolemmal Ca²⁺ influx being exactly balanced by efflux (Eisner et al. 1984; Friel & Tsien, 1992; Rios, 2010). The question then is what does control diastolic [Ca²⁺]_i? In recent work we have investigated the effects of interfering with SR function. This work was prompted by the fact that heart failure affects the SR by both making the RyR leaky and decreasing SERCA activity (for review see Bers, 2014).

Figure 5. Effects of caffeine and isoprenaline on diastolic and systolic [Ca²⁺]_i .

A, time course (rat ventricular myocyte). The grey trace shows [Ca²⁺]_i during rest and stimulation at 2 Hz. The blue is a low‐pass filtered version. Caffeine (Caf, 1 mm) and isoprenaline (ISO, 1 μm) were applied as shown. B, mean data showing normalized values for the amplitude of the Ca²⁺ transient, diastolic and average [Ca²⁺]_i. In each group the unlabelled bar is control; C, caffeine; I, ISO; I+C, ISO + caffeine. (Reproduced from Sankaranarayanan et al. (2017).)

The experiment illustrated in Fig. 5 A shows the effects on diastolic [Ca²⁺]_i of making the RyR leaky with caffeine (1 mm) (Sankaranarayanan et al. 2017). This is a higher concentration than was used in Fig. 4 and results in a release of Ca²⁺ from the SR even in the absence of stimulation indicating a marked decrease of SR Ca²⁺. The grey trace shows that, under basal conditions, the diastolic [Ca²⁺]_i at 2 Hz stimulation is only slightly greater than resting [Ca²⁺]_i. The subsequent application of caffeine decreased systolic [Ca²⁺]_i and increased diastolic [Ca²⁺]_i. After removing caffeine, application of the β‐adrenergic agonist isoprenaline (ISO) increased systolic [Ca²⁺]_i. The subsequent addition of caffeine again decreased systolic [Ca²⁺]_i and produced a more marked increase of diastolic [Ca²⁺]_i than was seen in the absence of ISO. The average data (Fig. 5 B) confirm the effects of caffeine on diastolic and systolic [Ca²⁺]_i. It is also noteworthy that, in the absence of caffeine, ISO produces a large increase of the amplitude of the Ca²⁺ transient with little or no effect on diastolic [Ca²⁺]_i. In contrast, in the presence of caffeine, the effect of ISO on the amplitude is greatly diminished and that on diastolic is increased.

The maintained decrease of systolic [Ca²⁺]_i contrasts with the transient potentiation with no steady state effect seen in Fig. 4. The explanation is that the higher concentration of caffeine results in a large release of Ca²⁺ from the SR thereby greatly decreasing SR Ca. If SR Ca²⁺ falls sufficiently then, even if RyR opening is potentiated enormously and all the Ca²⁺ is released from the SR, the Ca²⁺ transient will be smaller than in control (Negretti et al. 1993; Sankaranarayanan et al. 2016). A similar argument has been used to explain why RyR potentiation decreases the amplitude of the Ca²⁺ transient in heart failure (Belevych et al. 2007). An explanation of the associated increase of diastolic [Ca²⁺]_i is that it is a direct result of the decrease of systolic [Ca²⁺]_i and the need to preserve flux balance. Before caffeine is added, influx and efflux will be equal. In the presence of caffeine the smaller Ca²⁺ transient will decrease the systolic Ca²⁺ efflux to less than the influx leading to the cell gaining Ca²⁺. This is compensated by the increase of diastolic [Ca²⁺]_i, which will increase efflux. Assuming that Ca²⁺ influx is unaffected by caffeine, then the total (systolic plus diastolic) efflux must be the same in the absence and presence of caffeine. If we assume that NCX activity is proportional to [Ca²⁺]_i (Barcenas‐Ruiz et al. 1987), then one would expect the time‐averaged [Ca²⁺]_i concentration to be constant. That this is indeed the case is demonstrated by the low‐pass filtered trace of Fig. 5 A and the mean data of Fig. 5 B, which show that average [Ca²⁺]_i is unaffected by making the RyR leaky with caffeine (Sankaranarayanan et al. 2017). Figure 6 demonstrates the results of stimulating over a wider range of frequencies. The average level of [Ca²⁺]_i is unaffected by addition of caffeine (Fig. 6 A – see Fig. 6 B for mean data). Increasing frequency increases average [Ca²⁺]_i but in a manner that tends to saturation. The average level of [Ca²⁺]_i can usefully be compared with the Ca²⁺ influx into the cell. As the frequency of stimulation is increased, the influx per pulse decreases due to increased inactivation (Fig. 6 C) (Sipido et al. 1998; Dibb et al. 2007). The influx per unit time, however, increases in a saturating manner with frequency. The linear relationships between influx per unit time and average [Ca²⁺]_i are shown in Fig. 6 D. The lines are statistically identical in the absence and presence of caffeine. One noteworthy feature of Fig. 6 D is that, even when there is no Ca²⁺ influx, there is still a finite level of [Ca²⁺]_i. The extrapolation of the fitted lines provides horizontal intercepts of the order of 4 μmol l⁻¹ s⁻¹ suggesting that, in addition to the L‐type Ca²⁺ current, there is a continuous influx of Ca²⁺ into the cell of this magnitude. This is a similar magnitude to the value obtained with a different method referred to above. It therefore appears that, at least under the conditions of these experiments, Ca²⁺ influx consists of two components: (i) background and (ii) L‐type Ca²⁺ current. There will doubtless be other routes including possible entry through reverse mode NCX but these are not resolved here. The tight correlation between Ca²⁺ entry and average [Ca²⁺]_i demonstrates that it is the influx per unit time which sets the average level of [Ca²⁺]_i; the properties of the SR simply determine the relative contributions of diastolic and systolic [Ca²⁺]_i to the average. A good example of this is the effect of β‐adrenergic stimulation on [Ca²⁺]_i. Under normal conditions (Fig. 5 B) this results in an increase of systolic [Ca²⁺]_i with little or no effect on diastolic. In contrast, when the RyR is made leaky with ryanodine, the bulk of the increase is diastolic.

Figure 6. Interactions between RyR leak and stimulation frequency on diastolic, systolic and average [Ca²⁺]_i .

A, time course (rat ventricular myocyte). Stimulation frequency was altered and 1 mm caffeine (Caf) applied as shown above. B, mean data showing (from top to bottom): amplitude, diastolic, average [Ca²⁺]_i, Ca²⁺ influx per pulse, Ca²⁺ influx per second. C, top, specimen Ca²⁺ currents; bottom, integrated Ca²⁺ influx at the frequencies indicated. D, average [Ca²⁺]_i as a function of Ca²⁺ influx per second. In all panels the red symbols indicate the presence of caffeine. (Reproduced from Sankaranarayanan et al. (2017).)

In addition to diastolic and systolic [Ca²⁺]_i, Ca²⁺ waves also contribute to Ca²⁺ efflux and average [Ca²⁺]_i. An increase of SR Ca²⁺ content increases the frequency of Ca²⁺ sparks resulting in the occurrence of waves of Ca²⁺‐induced Ca²⁺ release which propagate through the cell (Cheng et al. 1996). Some of the Ca²⁺ in these waves is pumped out of the cell by NCX and the resulting current has been shown to produce delayed afterdepolarizations and thereby initiate various arrhythmias (Ferrier et al. 1973; Mechmann & Pott, 1986). Despite their arrhythmogenic nature, these Ca²⁺ waves may serve a useful function in providing a route for Ca²⁺ efflux, in addition to elevated diastolic [Ca²⁺]_i under conditions of Ca²⁺ overload (Díaz et al. 1997). More generally, at a given Ca²⁺ influx, the sum of the effluxes produced by the three components of efflux, systolic, diastolic and Ca²⁺ waves, must be constant. One example of the consequences of this is shown in Fig. 7. Under basal conditions (left hand panel), the systolic efflux (SE) balances Ca²⁺ influx. The application of a high concentration of isoprenaline (1 μm, middle panels), increases Ca²⁺ influx on the L‐type current (I) and results in Ca²⁺ waves. It is largely the Ca²⁺ efflux associated with the wave (DE) which balances the increased influx due to isoprenaline. Finally (right hand panel), tetracaine, which decreases RyR opening, abolishes the waves and the associated efflux. Ca²⁺ flux balance is now maintained by an increase of systolic efflux due to the increased Ca²⁺ transient (Venetucci et al. 2006).

Figure 7. Ca²⁺ efflux associated with Ca²⁺ waves.

In each panel traces show (from top to bottom): [Ca²⁺]_i; membrane current (both low gain and amplified); Ca²⁺ balance (the calculated change of total intracellular Ca²⁺ concentration). Panels show (from left to right): control; addition of isoprenaline (ISO,1 μm); addition of isoprenaline (1 μm) and tetracaine (50 μm). Experiment performed on a rat ventricular myocyte. I: influx; SE: Ca²⁺ efflux associated with the systolic Ca²⁺ transient; DE: Ca²⁺ efflux mediated by diastolic Ca²⁺ release. (Reproduced from Venetucci et al. (2006).)

The previous section has investigated the effects of disabling SR function by making the RyR leaky. We have also studied the effects of decreasing Ca²⁺ uptake into the SR by decreasing SERCA activity with thapsigargin. As has been shown previously (Janczewski & Lakatta, 1993; Negretti et al. 1993), the resulting decrease of SR Ca content decreases the amplitude of the Ca²⁺ transient. Figure 8 also shows that there is an increase of diastolic [Ca²⁺]_i. This, in turn, will increase the diastolic efflux thereby compensating for the decrease of systolic efflux. Therefore, no matter whether the amplitude of the Ca²⁺ transient is decreased by making the RyR leaky or decreasing SERCA activity, the result is a compensatory increase of diastolic [Ca²⁺]_i. Figure 8 B shows, again, that increasing stimulation frequency increases average [Ca²⁺]_i. Thapsigargin has a much greater effect on diastolic [Ca²⁺]_i at the higher stimulation rate. This is presumably because, at lower rates, there is time for [Ca²⁺]_i to fall to close to the resting (unstimulated) level. Indeed, as shown in Fig. 8 A, in the absence of stimulation thapsigargin has no effect on resting [Ca²⁺]_i.

Figure 8. The effects of SERCA inhibition on [Ca²⁺]_i .

A, original record (rat ventricular myocyte). The cell was stimulated at the frequencies indicated and thapsigargin (1 μm) applied as shown. B, mean data showing the effects of stimulation frequency on (from top to bottom): amplitude, diastolic and average [Ca²⁺]_i. (Reproduced from Sankaranarayanan et al. (2017).)

From these experiments, we therefore conclude that systolic [Ca²⁺]_i is a major factor regulating diastolic [Ca²⁺]_i. The overall regulation of diastolic [Ca²⁺]_i is yet another example of the importance of considering Ca²⁺ flux balance. A given Ca²⁺ influx must be balanced by an equal efflux. Assuming that the properties of NCX are constant and the rate of NCX pumping is proportional to [Ca²⁺]_i, then the level of Ca²⁺ influx sets the average level of [Ca²⁺]_i. The relative diastolic and systolic levels that determine the average are determined by factors such as SR function. Making the RyR leaky or inhibiting SERCA will result in a situation with elevated diastolic and decreased systolic [Ca²⁺]_i compared to the situation with a normal SR. All other things being equal, a decrease of systolic [Ca²⁺]_i as occurs in heart failure would be expected to be associated with elevated diastolic [Ca²⁺]_i. What actually happens will depend, also, on the Ca²⁺ influx. Although many studies in heart failure find that the L‐type current is unaffected, some work has reported a decrease (for review see Benitah et al. 2010). The latter case should be associated with a decrease of average [Ca²⁺]_i thereby mitigating the expected rise of diastolic [Ca²⁺]_i. It should, however, also be noted that the situation is potentially more complex. Firstly, the decrease of the Ca²⁺ transient may slow Ca²⁺‐dependent inactivation of the L‐type current and thereby maintain or even increase Ca²⁺ influx. Secondly, the prolongation of the action potential which occurs in heart failure will increase Ca²⁺ entry through the L‐type current.

Conclusions

I have referred extensively to the fact that the ventricular myocyte is in Ca²⁺ flux balance: on each beat, in the steady state, the amount of Ca²⁺ that enters the cell equals that which leaves. As I have reviewed, this has important consequences for cellular Ca²⁺ cycling and cardiac function. It underlies the regulation of SR Ca content. It explains why an increase in RyR open probability has no effect, in the steady state, on the amplitude of the Ca²⁺ transient. Finally, it explains the inverse relationship between the levels of diastolic and systolic [Ca²⁺]_i as well as the importance of average [Ca²⁺]_i.

Additional information

Competing interests

None.

Funding

Work from the author's laboratory was supported by a grant from the British Heart Foundation (grant number: CH/2000004/12801).

Biography

David Eisner is The British Heart Foundation Professor of Cardiac Physiology at The University of Manchester. He studied Natural Sciences in Cambridge and then obtained his DPhil in Oxford in the laboratory of Denis Noble. He has held academic positions at University College London and Liverpool University before moving to Manchester. His research focuses on the regulation of intracellular calcium both in normal conditions and in the alterations that lead to cardiac arrhythmias and heart failure.