Low sodium inotropy is accompanied by diastolic Ca²⁺ gain and systolic loss in isolated guinea-pig ventricular myocytes

Abstract

1. We measured sarcolemmal Ca²⁺ fluxes responsible for the positive inotropic effects of solutions with reduced Na⁺ concentration in voltage-clamped guinea-pig ventricular myocytes; intracellular Ca²⁺ concentration ([Ca²⁺]_i) was measured with Indo-1.

2. Reduction of external Na⁺ concentration by 50 % (to 67 mM) produced an increase in systolic [Ca²⁺]_i accompanied by a decrease in Ca²⁺ entry via the L-type Ca²⁺ current. With reduced Na⁺ concentration, there was an initial decrease in the Na⁺–Ca²⁺ exchange current on repolarization followed by an increase to greater than control. We attribute this initial decrease to a decrease in the Na⁺ gradient and the subsequent increase to a fall in intracellular Na⁺ concentration and increase in systolic [Ca²⁺]_i.

3. The decreased L-type Ca²⁺ current and increased Ca²⁺ efflux on Na⁺–Ca²⁺ exchange resulted in a calculated systolic loss of Ca²⁺.

4. The calculated systolic loss of Ca²⁺ was accompanied by a measured increase in sarcoplasmic reticulum (SR) Ca²⁺ content.

5. Reduction of the external Na⁺ concentration also produced an outward shift of holding current which was blocked by Ni²⁺. This is taken to represent Ca²⁺ influx via Na⁺–Ca²⁺ exchange.

6. When diastolic influx is taken into account, the observed gain in SR Ca²⁺ content can be predicted. The measurements show that, in reduced Na⁺, much of the entry of Ca²⁺ into the cell occurs during diastole (via Na⁺–Ca²⁺ exchange) rather than in systole (via the L-type Ca²⁺ current).

Na⁺–Ca²⁺ exchange plays an important role in the regulation of intracellular Ca²⁺ concentration ([Ca²⁺]_i) in the heart. Under resting conditions in most species, it produces a net Ca²⁺ efflux from the cell that helps balance Ca²⁺ entry (Bers, 1991). In the beating heart, part of the systolic increase of Ca²⁺ is pumped out of the cell via Na⁺–Ca²⁺ exchange, the combined activity of Na⁺–Ca²⁺ exchange and the quantitatively smaller sarcolemmal Ca²⁺-ATPase must produce a Ca²⁺ efflux that balances the entry of Ca²⁺ into the cell via the L-type Ca²⁺ current (Choi & Eisner, 1999a).

Na⁺–Ca²⁺ exchange was first identified from the effects of reduction of external Na⁺ concentration on contraction and cardiac Ca²⁺ regulation. Thus reduction of external Na⁺ concentration increases the force of contraction and can produce an increase in resting force (contracture) (Luttgau & Niedergerke, 1958; Chapman & Tunstall, 1980) and calcium concentration (Sheu & Fozzard, 1982; Allen et al. 1983). The aim of the work in this paper was to elucidate the nature of the Ca²⁺ fluxes which result in increased contraction. There are at least three possibilities. (1) A decreased ability of Na⁺–Ca²⁺ exchange to remove Ca²⁺ ions from the cell will make it less able to compete with SR Ca²⁺-ATPase during systole resulting in a larger fraction of the systolic Ca²⁺ being returned to the SR and thence increasing SR Ca²⁺ content. (2) There may be an increased Ca²⁺ influx via Na⁺–Ca²⁺ exchange during the initial phase of depolarisation of the action potential. (3) Although under normal conditions, most Ca²⁺ influx and efflux occurs during systole, with little Ca²⁺ movement during the diastolic period (see above), it is possible that Na⁺ reduction results in significant Ca²⁺ flux during diastole.

In this paper, we therefore measured the Ca²⁺ content of the SR and the fluxes which occur during both systole and diastole. The results show that when external Na⁺ concentration is reduced the extra Ca²⁺ entry occurs during diastole. In the steady state this diastolic Ca²⁺ influx via Na⁺–Ca²⁺ exchange plus the Ca²⁺ entry during systole via the L-type Ca²⁺ current balance the systolic efflux via Na⁺–Ca²⁺ exchange.

METHODS

General procedures

Experiments were performed on single ventricular myocytes enzymatically isolated from guinea-pig heart. Guinea-pigs were killed by cervical dislocation (in accordance with local and national guidelines). Cells were prepared using a modification of a published method (Mitra & Morad, 1985). Briefly, the heart was removed and retrogradely perfused through the aorta for 5 min at 37°C with a nominally Ca²⁺-free solution containing (mM): NaCl, 134; glucose, 11; Hepes, 10; KCl, 4; MgSO₄, 1.2; Na₂HPO₄, 1.2; adjusted to pH 7.4 with NaOH. All solutions were gassed with 95 % O₂ and 5 % CO₂. Protease (Sigma, type XIV) and collagenase (Worthington, type I) were then added to a final concentration of 0.1 mg ml⁻¹ and 1 mg ml⁻¹, respectively, and the heart was perfused for 7-9 min until flaccid. The perfusate was then changed to a taurine-containing solution for 10 min containing (mM): NaCl, 109; taurine, 25; glucose, 11; Hepes, 10; KCl, 4; MgSO₄, 1.2; Na₂HPO₄, 1.2; CaCl₂, 0.1; adjusted to pH 7.4 with NaOH. The ventricles were then dissected from the heart, suspended in the taurine-containing solution and gently triturated to release single cells. The isolated cells were stored at room temperature (22-23°C) in the taurine-containing solution until use.

Measurement of [Ca²⁺]_i

Cells were loaded with the fluorescent Ca²⁺ indicator Indo-1 by incubation with 5 μM Indo-1 acetoxymethylester (AM) (Molecular Probes, Eugene, OR, USA) for 5 min followed by at least 30 min of de-esterification. Cells were then placed in a superfusion chamber on the stage of an inverted microscope adapted for epifluorescence (Nikon Diaphot 300, Nikon). Indo-1 fluorescence was excited at 340 nm and measured at 400 nm and 500 nm. The ratio of the emitted fluorescence (400/500) was used as an index of [Ca²⁺]_i (O'Neill et al. 1990).

Voltage clamp

Cells were voltage clamped using the perforated patch-clamp technique with amphotericin-B and depolarised with 100 ms duration pulses from -40 to 0 mV at 0.33 Hz. To compensate for the relatively high access resistance of the perforated patch technique, the switch clamp facility (3-5 kHz) of the Axoclamp-2B voltage-clamp amplifier (Axon Instruments Inc., CA, USA) was used. Pipettes (2-3 MΩ) were filled with (mM): KCH₃O₃S, 125; KCl, 22; K₂EGTA, 0.1; NaCl, 10; Hepes, 10; MgCl₂, 5; adjusted to pH 7.2 with NaOH. Amphotericin-B was added to the pipette solution to a final concentration of 240 μg ml⁻¹.

Electrophysiological and fluorescence signals were digitised at 2 kHz and stored using Axoscope 1.1 software amplifier (Axon Instruments Inc.) for later analysis using custom software (Visual Basic 4, Microsoft Corp.).

Measurement of SR Ca²⁺ content and sarcolemmal Ca²⁺ flux

The integral of the inward Na⁺–Ca²⁺ exchange current on application of 10 mM caffeine was used as a measure of the SR Ca²⁺ content (Varro et al. 1993). Baseline was taken as the final level of current reached in the presence of caffeine. A measure of Ca²⁺ influx via the L-type Ca²⁺ current was obtained by integrating the current. Baseline was taken as the mean level of current before the pulse. The Ca²⁺ efflux via Na⁺–Ca²⁺ exchange on repolarization was obtained by integrating the tail current. Baseline was taken as the mean level of current between 1.5 and 2 s after the pulse. In order to calculate the total Ca²⁺ efflux one has to allow for the fact that part of the Ca²⁺ efflux occurs via the electroneutral Ca²⁺-ATPase. We have assumed that this contributes 20 % of the total Ca²⁺ efflux (Bennett et al. 1999). In some experiments (for example see Fig 5), in order to avoid the need for this correction, cells were pre-incubated in carboxyeosin to inhibit the sarcolemmal Ca²⁺-ATPase (Bassani et al. 1995; Choi & Eisner, 1999b). Cells were loaded with 5- (and 6-) carboxyeosin diacetate (succinimidyl ester) from Molecular Probes. A 10 mM stock solution was made in DMSO. This stock solution was added to the normal experimental solution (see below) to make a final concentration of 20 μM in which cells were loaded for 15 min. Afterwards, cells were superfused with carboxyeosin-free solution for a further 10 min to allow carboxyeosin de-esterification and to wash off the residual compound in the bath solution. The SR Ca²⁺ content and sarcolemmal Ca²⁺ fluxes were related to total cell volume assuming a cell capacitance/volume ratio of 5 pF pl⁻¹ (Page, 1978).

Figure 5.

Calculation of contribution of diastolic Ca²⁺ entry

A, time course of experiment. Traces show: top, Indo-1 ratio; middle, current; bottom, Ca²⁺ fluxes. External Na⁺ was reduced to 67 mM as shown above. The pink records were obtained in the presence of 10 mM Ni²⁺. In the Ca²⁺ flux record the filled circles show the calculated entry on the L-type Ca²⁺ current and the open circles show the efflux on Na⁺–Ca²⁺ exchange. B, comparison of systolic and diastolic fluxes. Traces show (from top to bottom): systolic Ca²⁺ flux (obtained from the difference of I_Ca and I_Na-Ca in A); diastolic Ca²⁺ flux calculated from the Ni²⁺-sensitive current in A; net Ca²⁺ flux obtained by summing the systolic and diastolic fluxes; calculated change of cell Ca²⁺. This takes account of the time course of the individual currents. C, specimen records from control (i) and reduced Na⁺ (ii) obtained from the points indicated on the lower trace of B. Note that, to save space, the two traces have been positioned more closely vertically than is really the case (see B).

Solutions

The superfusion solution contained (mM): NaCl, 134; glucose, 11; Hepes, 10; KCl, 4; MgCl₂, 1.2; CaCl₂, 1.5; 4-aminopyridine, 5; BaCl₂, 0.1; adjusted to pH 7.4 with HCl. In low [${\mathrm{Na}}_{\mathrm{o}}^{+}$] solutions, 67 mM ${\mathrm{Na}}_{\mathrm{o}}^{+}$ (50 % of the control [${\mathrm{Na}}_{\mathrm{o}}^{+}$) was replaced with equimolar amounts of either N-methylglucamine (NMDG) or lithium. Caffeine was added to the required solutions to a final concentration of 10 mM. All experiments were performed at room temperature (22-23°C).

Statistics

Values are presented as means ±s.e.m. from n cells. Student's paired t test was used to compare the parameters between groups. Statistical significance was reached when P < 0.05.

RESULTS

In the present experiments, single voltage-clamped guinea-pig ventricular myocytes were exposed to solutions of reduced sodium concentration ([Na⁺]_o). Removal of more than about half the sodium was found to produce a state of Ca²⁺ overload in which there is spontaneous release of Ca²⁺ from the SR (not shown). To avoid such Ca²⁺ oscillations, which complicated the quantitative analysis of sarcolemmal Ca²⁺ flux and SR Ca²⁺ content, [Na⁺]_o was not reduced below 67 mM. Typical effects of [Na⁺]_o reduction on the Ca²⁺ transient are illustrated in Fig 1A. Here, exposure to 67 mM ${\mathrm{Na}}_{\mathrm{o}}^{+}$ increased both diastolic and systolic [Ca²⁺]_i. The increase of the Ca²⁺ transient amplitude was maximal within about 1 min and then subsequently decreased to a new steady-state level over the next 10 min. This pattern of change in [Ca²⁺]_i was seen in 15 of 17 cells.

Figure 1.

The effects of reduced sodium on [Ca²⁺]_i and membrane Ca²⁺ fluxes

A, time course showing Indo-1 ratio. The membrane was held at -40 mV and 100 ms duration depolarizing pulses applied to 0 mV at 0.33 Hz. B, specimen records obtained at the times indicated in A. Traces show: top, [Ca²⁺]_i; middle, membrane current; bottom, net Ca²⁺ fluxes (Ca²⁺ entry is positive). C, expanded records of membrane current from a and b.

Specimen records of [Ca²⁺]_i and membrane current are shown in Fig 1B. They emphasise the increase in amplitude of the systolic Ca²⁺ transient in 67 mM ${\mathrm{Na}}_{\mathrm{o}}^{+}$. Under control conditions, the membrane current record shows an L-type Ca²⁺ current on depolarisation and a Na⁺–Ca²⁺ exchange tail current on repolarisation. The integrated traces show that the Ca²⁺ entry via the L-type Ca²⁺ current is about 4 μmol l⁻¹. On repolarisation there is a similar Ca²⁺ loss on Na⁺–Ca²⁺ exchange. Thus, during steady-state stimulation, Ca²⁺ influx and efflux are in balance. In low Na⁺ solution the L-type Ca²⁺ current was depressed whereas the Na⁺–Ca²⁺ exchange tail current was increased (Fig 1C). Consequently, in Fig 1B record b (obtained after 1 min in 67 mM ${\mathrm{Na}}_{\mathrm{o}}^{+}$) the Ca²⁺ influx was reduced (2 μmol l⁻¹) while the efflux was increased (7 μmol l⁻¹), resulting in a net systolic Ca²⁺ loss of 5 μmol l⁻¹ per pulse. These changes in Ca²⁺ flux were accompanied by a small outward steady-state current, consistent with a reduced ability of Na⁺–Ca²⁺ exchange to extrude Ca²⁺ during diastole in low Na⁺ solutions (see below). Record c in Fig 1B (8 min in reduced [Na⁺]_o) shows a reduced Ca²⁺ transient and Na⁺–Ca²⁺ exchange Ca²⁺ efflux. The integral trace still shows a net calculated efflux but now of only 3.9 μmol l⁻¹ per pulse.

The time course of changes in sarcolemmal Ca²⁺ fluxes during the whole period of this experiment is plotted in Fig 2. The fluxes were calculated for each pulse in the same way as shown in Fig 1B. It is clear that (with the exception of the first few stimuli in reduced [Na⁺]_o, see below) increase in the amplitude of the Ca²⁺ transient was accompanied by increased Ca²⁺ removal via the Na⁺–Ca²⁺ exchange current (I_Na-Ca) and a decrease in Ca²⁺ influx via the L-type current (I_Ca). On average, after 1 min in 67 mM ${\mathrm{Na}}_{\mathrm{o}}^{+}$, the magnitude of the Ca²⁺ efflux via Na⁺–Ca²⁺ exchange was significantly increased from 5.43 ± 1.2 to 10.5 ± 1.2 μmol l⁻¹ (n = 5, P < 0.01) whereas the Ca²⁺ influx was significantly decreased from 5.42 ± 1.2 to 2.8 ± 0.7 μmol l⁻¹ (n = 5, P < 0.01). Over the next 2-3 min in low Na⁺ solution, there was a gradual decrease in both the Ca²⁺ transient and Ca²⁺ efflux while the Ca²⁺ influx remained depressed. As a result, after 10 min, the Ca²⁺ influx was 2.5 ± 0.7 μmol l⁻¹ while the Ca²⁺ efflux was significantly reduced to 5.3 ± 1.1 μmol l⁻¹ (P < 0.05). Summing these fluxes over 10 min resulted in a cumulative Ca²⁺ loss from the cell of approximately 400 μmol l⁻¹ (Fig 2D) which would have been expected to deplete the cell of Ca²⁺.

Figure 2.

Pulse by pulse analysis of systolic Ca²⁺ fluxes during sodium reduction

External Na⁺ concentration was decreased to 50 % of control for the period shown. A, magnitude of the systolic Ca²⁺ transient. B, calculated Ca²⁺ entry on the L-type current (•) and efflux on Na⁺–Ca²⁺ exchange (○). C, net Ca²⁺ movement on each pulse calculated as efflux - entry. D, cumulative calculated change of Ca²⁺ obtained by adding up the data in C.

The experiment in Fig 2 shows the time course of change of both the Na⁺–Ca²⁺ exchange flux and the underlying Ca²⁺ transient. The Na⁺–Ca²⁺ exchange flux is influenced by both the change in systolic Ca²⁺ and by the imposed change in Na⁺ gradient. In order to distinguish between the effects of these two parameters, we have looked at the relationship between the magnitude of the Na⁺–Ca²⁺ exchange tail current and [Ca²⁺]_i. Figure 3 shows recordings of [Ca²⁺]_i and current. In this figure we have focused on the initial events, i.e. the period when the Ca²⁺ transient is increasing in amplitude. The initial effect of Na⁺ removal was an outward shift of current (see later). However, the Na⁺–Ca²⁺ exchange tail current initially decreased in magnitude (compare transients a and b). This was then followed by a gradual increase of the amplitude of the tail current in parallel with that of the systolic Ca²⁺ transient. The tail current is plotted against the Indo-1 ratio for three specimen transients in Fig 3B. The early transient in low Na⁺ (b) has a smaller slope than the control (a). However, transient c (obtained after 40 s in low Na⁺) has a slope which has recovered almost towards the control level. The slopes of the relationships are plotted as a function of time in Fig 3C. It is clear that the slope initially decreases before recovering. It is likely that the initial decrease of slope is due to the reduction of Na⁺ gradient produced by reduction in external Na⁺ concentration. The subsequent recovery of the slope could then be due to a decrease in intracellular Na⁺ concentration tending to restore the Na⁺ gradient.

Figure 3.

Transient effects of Na⁺ removal on Na⁺–Ca²⁺ exchange

A, time course of the experiment. Traces show: top, [Ca²⁺]_i; bottom, current. For clarity, the capacity currents and the currents during depolarization have been removed so that the Na⁺–Ca²⁺ exchange current is the only visible time-dependent current. B, plots of the Na⁺–Ca²⁺ exchange tail current as a function of Indo-1 ratio for the three pulses identified in A. The continuous lines are linear regressions to the data. The dashed lines in b and c show the regression line from control (a) for comparison. C, plot of the slope of currents versus ratio as a function of duration of exposure to reduced Na⁺ solution.

The experiments described above show that in a reduced Na⁺ solution there is a decreased entry and increased efflux of calcium. This would result in a decrease of cell calcium. In contrast to this calculated loss of calcium from the cell, there is an increase in the amplitude of the systolic Ca²⁺ transient. Furthermore, direct measurements of SR Ca²⁺ content showed that, as expected from previous work (Smith et al. 1988), the SR Ca²⁺ content was increased. This is demonstrated in Fig 4 which shows the effects of reduction of external Na⁺ on the response to 10 mM caffeine application. There is an increase in both the amplitude of the caffeine-evoked increase of [Ca²⁺]_i and the accompanying Na⁺–Ca²⁺ exchange current. On average, the SR Ca²⁺ content (calculated from the integral of the current - see Methods) was significantly increased from a control value of 16.9 ± 0.9 to 31.7 ± 2.4 μmol l⁻¹ (n = 5, P < 0.02) 10 min after lowering Na⁺. This effect was reversible when [Na⁺]_o was returned for 10 min to 134 mM (17.3 ± 1.8 μmol l⁻¹, n = 3) (not shown).

Figure 4.

Na⁺ reduction increases the SR Ca²⁺ content

In both panels caffeine (10 mM) was applied to estimate the SR Ca²⁺ content. Traces show (from top to bottom): Indo-1 ratio; membrane current; calculated SR Ca²⁺ content.

The results thus far are consistent with the idea that in reduced external Na⁺ concentration the cell gains Ca²⁺ during diastole. Subsequent experiments were designed to obtain direct evidence for such a diastolic Ca²⁺ gain. In these experiments, we also improved the quantification of Ca²⁺ efflux. The results of Fig 3 showed that there is a transient depression of Na⁺–Ca²⁺ exchange. This presumably means that, during this initial period, the sarcolemmal Ca²⁺-ATPase contributes more to Ca²⁺ removal from the cell than is the case in the steady state. In order to avoid problems produced by this phenomenon, these experiments were performed following pre-incubation in carboxyeosin, an inhibitor of Ca²⁺-ATPase (Bassani et al. 1995; Gatto et al. 1995; Choi & Eisner, 1999b). Results obtained in carboxyeosin were qualitatively similar to those described above. In particular a decrease in L-type Ca²⁺ current and increase in Na⁺–Ca²⁺ exchange current were observed. One difference was that, in the presence of carboxyeosin, the secondary decay of systolic Ca²⁺ following Na⁺ removal was less pronounced than in its absence. On average, in the absence of carboxyeosin, the Ca²⁺ transient measured after 5 min in low Na⁺ was 67.2 ± 3.2 % (n = 9) of the peak value reached whereas it was 92.7 ± 2.9 % (n = 7, P < 0.01) in the presence of carboxyeosin.

The experiment depicted in Fig 5A shows the time course of increase of systolic [Ca²⁺]_i on reduction of external Na⁺ concentration. This is accompanied by an outward shift of the holding current and, following an initial decrease, an increase in the magnitude of the Na⁺–Ca²⁺ exchange current on repolarization. The pink traces show the effects of the same manoeuvre in the presence of Ni²⁺ (10 mM). The Ca²⁺ transient is abolished and the outward shift of current is virtually absent suggesting that much of the current is due to Na⁺–Ca²⁺ exchange. The lower traces show the calculated Ca²⁺ movements during systole via Na⁺–Ca²⁺ exchange and the L-type Ca²⁺ current. The difference between these two currents is shown as the trace labelled systolic in Fig 5B. There is an initial predicted gain of Ca²⁺ followed by a loss. The open circles in Fig 5B show the calculated diastolic Ca²⁺ gain calculated from the Ni²⁺-sensitive shift in holding current. This is expressed as the Ca²⁺ movement per stimulation period (here 3 s). This allows comparison with the systolic fluxes. The data show that the calculated influx is initially zero and rapidly rises to a maintained level following removal of Na⁺. Finally the triangles show the calculated net Ca²⁺ movements. This is initially zero and shows a net influx on reduction of external Na⁺ before decaying again to zero. The bottom trace of Fig 5B shows the time course of the predicted changes of cell Ca²⁺. In this experiment there is a predicted net gain of Ca²⁺ of about 29 μmol l⁻¹. On average, in five experiments, the calculated increase in Ca²⁺ was 29.8 ± 4.2 μmol l⁻¹. In the same cells the increase in SR Ca²⁺ measured directly by application of caffeine (see Fig 4) was 23.9 ± 4 μmol l⁻¹ (n = 5, P > 0.2). These values are not dissimilar.

The lower traces of Fig 5B show the time course of change of total cell Ca²⁺ concentration. It is important to note that this changes significantly during exposure to reduced external Na⁺ concentration. Under control conditions the efflux via Na⁺–Ca²⁺ exchange balances the entry of Ca²⁺ via the L-type current. There is therefore very little flux of Ca²⁺ during the diastolic period. In contrast, in the steady state in low Na⁺ there is a significant entry of calcium during each diastole which is compensated by extra Ca²⁺ efflux on systole. Indeed more Ca²⁺ enters during the diastolic period by Na⁺–Ca²⁺ exchange than leaves during systole by Na⁺–Ca²⁺ exchange. These differences are emphasised in the expanded traces of Fig 5C.

DISCUSSION

The present work shows that the increase in the systolic Ca²⁺ transient produced by reduction of external Na⁺ concentration is associated with an increase in SR Ca²⁺ content. This quantitative demonstration of an increase in SR Ca²⁺ content reinforces the conclusions drawn by previous qualitative estimates (Smith et al. 1988). The major contribution of this paper is a dissection of the contributions of various sarcolemmal Ca²⁺ fluxes to this increase of SR content.

The methods used to measure sarcolemmal fluxes have been described and validated before (Trafford et al. 1997; Bennett et al. 1999). One particular problem does, however, require further discussion. Our electrical measurements record Ca²⁺ efflux only via Na⁺–Ca²⁺ exchange and ignore that via electroneutral Ca²⁺-ATPase. In previous work we have corrected for the contribution of Ca²⁺-ATPase by multiplying the Na⁺–Ca²⁺ exchange flux by a constant factor. In the present experiments, however, it is likely that the ratio of these two fluxes changes during the period of exposure to low Na⁺. This can be seen from experiments such as that in Fig 3 which shows Na⁺–Ca²⁺ exchange plotted as a function of [Ca²⁺]_i. The fact that the slope of this relationship changes means that the activity of Na⁺–Ca²⁺ exchange is affected independent of any change of [Ca²⁺]_i. This slope decreases greatly within 3 s of decreasing external Na⁺, an effect which is probably due to the decrease in the Na⁺ gradient. The slope then recovers almost to control levels over the next minute or so. This recovery is likely to be due to the decrease in intracellular Na⁺ (Ellis, 1977; Sheu & Fozzard, 1982).

If we assume that the sarcolemmal Ca²⁺-ATPase is unaffected by these manoeuvres then the Ca²⁺-ATPase:Na⁺–Ca²⁺ exchange activity ratio is likely to increase transiently during the period of Na⁺ reduction. This will make it impossible to measure the total Ca²⁺ efflux accurately. Therefore the net systolic loss of Ca²⁺ seen in Figs 1 and 2 will have been underestimated as a result of this problem. We overcame these limitations in the experiment shown in Fig 5 by performing the whole experiment in the presence of carboxyeosin so that Ca²⁺-ATPase is inhibited throughout. All Ca²⁺ efflux was, therefore, produced by Na⁺–Ca²⁺ exchange and could be directly measured.

Two measurable sarcolemmal Ca²⁺ fluxes are associated with the systolic rise in [Ca²⁺]_i produced by voltage clamp depolarization. Ca²⁺ enters on depolarization via the L-type Ca²⁺ current and leaves on repolarization via Na⁺–Ca²⁺ exchange. For completeness, it should be noted that there will also be an initial influx of Ca²⁺ through Na⁺–Ca²⁺ exchange but this is likely to be much smaller than the entry through the L-type Ca²⁺ current (Negretti et al. 1995). The results of this paper show that reduction of external Na⁺ decreases the entry of Ca²⁺ via the L-type Ca²⁺ current. It is likely that this is due to increased Ca²⁺-dependent inactivation of the L-type Ca²⁺ current as a result of the larger systolic Ca²⁺ transient. The effects on the Na⁺–Ca²⁺ exchange flux are more complicated: the steady-state effect is an increase in the efflux but this was often preceded by a transient decrease. Plotting the Na⁺–Ca²⁺ exchange current as a function of [Ca²⁺]_i showed that the maintained increase was due to the larger systolic Ca²⁺ transients whereas the initial decrease was a direct effect of reduction of the sarcolemmal Na⁺ gradient.

These fluxes therefore show that reduction of external Na⁺ concentration produces an increase in net Ca²⁺ efflux or a decrease in influx during systole. This measured loss of calcium is associated with an increase in both SR and systolic Ca²⁺. A similar association between systolic Ca²⁺ loss and increased SR Ca²⁺ content has been observed during strophanthidin-induced inotropy (Bennett et al. 1999). In the case of strophanthidin it was suggested that the increased systolic loss of calcium could be compensated for by an increased diastolic entry. However, no measurements of this Ca²⁺ entry could be made. In the present work we have found that reduction of [Na⁺]_o is accompanied by an outward shift of holding current that is largely inhibited by Ni²⁺. Such an outward current has been shown to be due to Na⁺–Ca²⁺ exchange (Chin et al. 1993; Su et al. 1999). It is possible to calculate the Ca²⁺ entry via this route and compare it with the net systolic loss. The results show that in the steady state the fluxes balance. However, during the first few pulses in reduced Na⁺ there is a net gain of Ca²⁺ by the cell that can account for the measured gain of Ca²⁺ by the SR.

It is worth considering the thermodynamics of the increased diastolic gain of Ca²⁺ in reduced Na⁺. The reversal potential of the Na⁺–Ca²⁺ exchange (E_Na-Ca) is given by 3E_Na - 2E_Ca. Under control conditions we have: [Na⁺]_i, ∼10 mM; [Na⁺]_o, 134 mM; [Ca²⁺]_i, 100-200 nM; [Ca²⁺]_o, 1.5. This gives values for E_Na of 67 mV and for E_Ca of 125-116 mV resulting in values for E_Na-Ca of -49 to -31 mV. In 67 mM Na⁺ (before intracellular Na⁺ has fallen) E_Na is 49 mV giving values for E_Na-Ca of -103 to -85 mV. Therefore with the holding potentials used in the present experiments of -40 mV one would expect little net Na⁺–Ca²⁺ exchange flux under control conditions once [Ca²⁺]_i had fallen to diastolic levels and significant Ca²⁺ influx in reduced Na⁺. Of course with a physiological diastolic potential of -80 mV there will be less driving force for Ca²⁺ entry.

The maintenance of Ca²⁺ homeostasis in cardiac cells is of paramount importance; our results show that when external Na⁺ is lowered the means of achieving a balance of Ca²⁺ fluxes changes. Under control conditions the bulk of Ca²⁺ influx during the action potential is balanced by efflux generated during the systolic rise of [Ca²⁺]_i. In low Na⁺ a large influx of Ca²⁺ also takes place during diastole. This is balanced by extra efflux generated during the larger systolic Ca²⁺ transient. The greater than normal SR Ca²⁺ content maintains systolic Ca²⁺ at the required level.