INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

MAMMALIAN HEARTS ARE ABSOLUTELY dependent on extracellular Ca²⁺ to contract. With each beat, a depolarization-induced influx of extracellular Ca²⁺, via voltage-dependent Ca²⁺ channels, triggers release of sufficient intracellular Ca²⁺ (from the sarcoplasmic reticulum, SR) to effect contraction. There is, however, considerable variation across mammalian species in the relative dependence on intra- and extracellular Ca²⁺ to activate contraction. This variation is particularly apparent with respect to the rat and guinea pig (32). We consider that it reflects a species difference in Gibbs free energy of the sarcolemmal Na⁺/Ca²⁺ exchanger (G_Na/Ca).

The sarcolemmal Na⁺/Ca²⁺ exchanger is a membrane-bound molecule that reversibly exchanges Na⁺ and Ca²⁺ across the cardiac cell membrane (for a recent review, see Ref. 6). During the interval between beats, the exchanger utilizes the free energy offered by the large transsarcolemmal Na⁺ gradient to extrude intracellular Ca²⁺, thereby promoting its return to normal diastolic values (3). There is considerable evidence to suggest that this Ca²⁺-buffering role of the exchanger is enhanced in guinea pig vis-à-vis rat myocardium. Thus, in response to activation of voltage-gated Ca²⁺ channels, Na⁺-Ca²⁺ exchange current is fivefold greater in guinea pig than in rat myocytes (39). Caffeine-induced release of Ca²⁺ from the SR fails to elicit a mechanical contracture in guinea pig papillary muscle unless the "forward" activity of the exchanger (i.e., Na⁺_o-induced Ca²⁺ efflux) is impeded (21). Contrariwise, postcaffeine repletion of SR Ca²⁺ content and the accompanying recovery of twitch force are rapid in myocytes of rat, whereas both are slow in the guinea pig (29). These representative electrophysiological and mechanical results are supported by findings from several metabolic studies.

The metabolic rate of the arrested guinea pig heart, relative to that of the rat, is insensitive to interventions designed to increase the flux of Ca²⁺ into the myoplasm. This obtains whether the source of Ca²⁺ is intracellular (i.e., from the SR) or extracellular. Specifically, the resting or basal metabolism of guinea pig myocardium is not affected by caffeine-induced release of Ca²⁺ from the SR (12, 14), in contrast to that of rat myocardium (12). Likewise, K⁺ depolarization-induced influx of Ca²⁺ through L-type Ca²⁺ channels has no effect on the resting cardiac metabolism of the guinea pig, even if extracellular K⁺ concentration ([K⁺]_o) is elevated to 40 mM (14), whereas any increment of K⁺ above its normal value (4-6 mM) elevates the resting rate of oxygen consumption (VO₂) of rat cardiac myocytes to some new value (11).

Hence both mechanical and metabolic responses to agents that release Ca²⁺ from the sarcoplasmic reticulum appear to be blunted in the guinea pig vis-à-vis the rat. We previously argued (14) that this is because Ca²⁺ arising from either intra- or extracellular sources is rapidly extruded from guinea pig cells via Na⁺/Ca²⁺ exchange. We now speculate, in accord with the hypothesis of Shattock and Bers (41), that diastolic Ca²⁺ efflux is comparatively small in the rat heart because the thermodynamic potential of its Na⁺/Ca²⁺ exchanger is relatively close to equilibrium. Hence we predict a difference between rat and guinea pig hearts in their suprabasal metabolic responses to lowering of [Na⁺]_o, the sodium ion concentration of the coronary perfusate.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Isolated perfused whole hearts. Rats and guinea pigs, weighing at least 350 g, were killed by decapitation in accordance with methods approved by the University of Auckland Animal Ethics Committee. As previously described (14), the heart was rapidly excised and its aorta was attached to the perfusion cannula. Constant-flow Langendorff perfusion of the coronary vasculature was achieved using a roller pump adjusted to provide an initial perfusion pressure of 8-11 kPa. All atrial inlets were tied, and a catheter was inserted into the right ventricle via the pulmonary artery to drain coronary venous effluent. Timed collection of this effluent determined coronary flow, which, during the initial period of spontaneous beating, averaged 64 ± 2 ml · min¹ · g dry wt¹ (n = 25) and 50 ± 3 ml · min¹ · g dry wt¹ (n = 21, means ± SE) for rat and guinea pig hearts, respectively. A water-filled latex balloon, at the end of a triple-lumen catheter, was inserted via the left atrium into the left ventricle. Two lumina were open, allowing aspiration of the ventricle. The third lumen connected the balloon to a pressure transducer (Statham P23Db), enabling continuous measurement of left ventricular pressure (P_LV).

Isolated ventricular trabeculae. Trabeculae were dissected from the right ventricles of isolated hearts and mounted between a fixed support and a force transducer (AE-801, SensoNor, Horton, Norway), as previously described (15). Preparations were loaded with the fluorescent Ca²⁺ indicator fura 2 using its acetoxylmethyl form. One of the major factors limiting the use of fura 2 is its rapid extrusion from cells at temperatures >30°C (37). Accordingly, in preliminary experiments we found that fura 2 was poorly retained at 37°C. To alleviate this problem, we used probenecid (1 mM), a blocker of the inorganic anion transporter, which effectively enabled retention of the indicator at body temperature.

Perfusion solutions. The standard solutions for both trabeculae and whole hearts was a modified Krebs-Henseleit bicarbonate buffer containing (in mM) 118 NaCl, 4.8 KCl, 1.18 MgSO₄, 1.18 KH₂PO₄, 24.8 NaHCO₃, 2.5 CaCl₂, and 10 glucose. For whole heart preparations, insulin (10 U/l) and the colloid replacement Haemaccel (20 ml/l) (Hoechst, Auckland, New Zealand) were added. The latter agent contains Na⁺ (145 mM), K⁺ (5.1 mM), and Ca²⁺ (6.25 mM), due account of which was taken in calculating ionic concentrations. Solutions were equilibrated with 95% O₂-5% CO₂ at 37°C. Hearts were arrested either by increasing the KCl concentration of the standard perfusate to a final K⁺ concentration of 20 mM or by adding the Ca²⁺ channel antagonist verapamil (50 µM).

Measurement of VO₂. For most experiments, the oxygen content of the arterial inflow and venous outflow catheters was measured with a fuel cell device (OxyCon, Department of Physiology, University of Tasmania, Hobart, Australia). The VO₂ was subsequently calculated from the arteriovenous difference in oxygen content multiplied by the rate of coronary flow determined volumetrically.

1

1

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Rate of oxygen consumption (VO2) (A and B) and left ventricular pressure (PLV) (C and D) of a perfused rat heart during spontaneous (isovolumic) beating (A and C) and during KCl arrest (B and D) at the values of extracellular Na+ concentration ([Na+]o) indicated (mM). In A and C, cardiac arrest induced at 13.5 min by increase of extracellular K+ concentration ([K+]o). In B: [Na+]o returned (to its standard value of 143 mM) from 3, 31, and 60 mM at 8, 13, and 11 min, respectively. Note different scales on both ordinates and abscissas throughout.

Statistical analyses. Results were analyzed by two-way analyses of variance for unbalanced repeated measures, with procedures available in the SAS statistical software package (SAS Institute, Cary, NC). Differences among means were tested for statistical significance (P < 0.05) using an appropriate set of contrast coefficients. Summary data are expressed as means ± SE.

Curve fitting. Sigmoidal y = y(x) relationships were fitted by a four-parameter version of the Hill equation

(1)

In this expression, x denotes the independent variable (either [Na⁺]_o or Gibbs free energy), y denotes VO₂ (with values between y_min and y_max), and K_m is the value of x that yields the half-maximal increment of y above y_min. The curve-fitting parameter n reflects y'_Km, the slope of the relationship at K_m; explicitly, n = 4 K_m · y'_Km/(y_min  y_max). When n is positive, a negative sigmoid obtains and vice versa. Within the context of the present study, no physical meaning is attributed to this parameter. Data for each species were fitted according to Eq. 1 using nonlinear, weighted, curve-fitting procedures available in the SAS software package. The weighting function was given by the inverse of the standard errors of the means. Goodness of fit is reported both as r² (the square of the correlation coefficient) and as s_{y · x} (the standard error of estimate or square root of the residual variance).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

As can be seen in Fig. 1A, KCl arrest reduces the VO₂ of the heart to ~25% of the value observed during the preceding period of spontaneous, isovolumic contractions. This reduced rate is an index of cardiac basal metabolism. For the remainder of this study, we focus on the increase of metabolic rate, above its "basal" value, of the electrically and mechanically quiescent heart.

Effect of reduced [Na+]_o during KCl arrest. During K⁺ arrest, we reduced [Na⁺]_o below its standard value of 143 mM by equimolar substitution of LiCl for NaCl. [In 2 rat hearts (data not shown), the use of either sucrose or NMDG as substitutes for NaCl (at 31 mM) led to comparable results.] Figure 1 shows representative effects from a rat heart on VO₂ (Fig. 1B) and P_LV (Fig. 1D) as functions of time for selected values of [Na⁺]_o. It can be seen that VO₂ reached a peak within ~1 min of reducing [Na⁺]_o and remained substantially elevated for at least 10-15 min thereafter.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   A: mean (± SE) peak VO2 as a function of [Na+]o, during 20 mM K+ arrest, for 13 rat () and 12 guinea pig () hearts. Data fitted (solid lines) according to Eq. 1 (see MATERIALS AND METHODS); resulting values of regression parameters: (ymin, ymax, Km, n, sy · x, r2) = (6.2 ± 1.1, 32.9 ± 1.1, 66.8 ± 2.1, 8.2 ± 1.6, 1.49, 0.9979) and (3.8 ± 0.3, 30.8 ± 0.4, 25.1 ± 2.7, 5.1 ± 0.3, 0.40, 0.9995) for rat and guinea pig, respectively. B: mean (± SE) values of left ventricular pressure (PLV) recorded at the same time as measurements made in A. In both A and B, standard error bars omitted when less than size of symbols.

Effect of Ca²+-free, low-Na⁺ perfusion. The pronounced effect of reduced [Na⁺]_o on the metabolic rate of K⁺-arrested hearts of either species (Fig. 2A) is consistent with the interpretation of Fiolet and colleagues (1, 2, 10, 11) that [Ca²⁺]_i has been increased subsequent to reversal of the sarcolemmal Na⁺/Ca²⁺ exchanger. To test this interpretation, two K⁺-arrested hearts of each species were subjected to 3 mM [Na⁺]_o in both the presence and absence of extracellular Ca²⁺. As can be seen in Fig. 3, Ca²⁺-free perfusion completely abolished the potentiation of both VO₂ and P_LV previously observed under conditions of low [Na⁺]_o. This result is consistent with the notion that Ca²⁺-free perfusion prevented the intracellular Na⁺-dependent influx of Ca²⁺ that had previously occurred during perfusion with 3 mM Na⁺ (Figs. 1 and 2). This interpretation motivates detailed consideration of the thermodynamics of the Na⁺/Ca²⁺ exchanger (see APPENDIX).

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Mean (± SE) VO2 (A) and PLV development (B) for 2 hearts of each species during spontaneous beating (6 mM [K+]o) and K+ arrest in both presence and absence of 2.5 mM extracellular Ca2+ concentration ([Ca2+]o). Standard error bars commonly less than size of symbols at bottom.

VO₂ as a function of G_Na/Ca. When r, the coupling ratio of the Na⁺/Ca²⁺ exchanger (see Eqs. A1 and A4, APPENDIX), is assigned the value of 3 (20, 31, 36), theG_Na/Ca becomes

(2)

where R is the universal gas constant (8.31 J · mol¹ · K¹), and T is temperature (K). The value of , the diastolic permeability of the cardiac cell membrane to Na⁺ relative to that of K⁺, is assumed to be 0.02 (42) independent of species and of [Na⁺]_o. Its contribution to G_Na/Ca is small and has been included merely for completeness.

(3)

(4)

(5)

View larger version (17K): [in this window] [in a new window]   Fig. 4.   VO2 as a function of Gibbs free energy of Na+/Ca2+ exchange (GNa/Ca) for hearts of rat () and guinea pig (). A: regression parameters for guinea pig curve (see Eq. 1): (ymin, ymax, Km, n, sy · x, r2) = (3.9 ± 0.15, 30.9 ± 1.41, 4.45 ± 0.24, 10.1 ± 0.94, 0.68, 0.9967) achieved with [Na+] = 45 mM (see Eq. 5); latter value also used to calculate GNa/Ca coordinates of rat data. Standard error bars commonly smaller than size of symbols. B: same data as in A but with GNa/Ca coordinates of rat data right shifted according to Eq. 6 with  = 21. Standard error bars (same as in A) omitted for clarity. Regression parameters: (ymin, ymax, Km, n, sy · x, r2) = (4.1 ± 0.27, 33.6 ± 1.31, 4.86 ± 0.27, 9.85 ± 1.15, 1.31, 0.9910). Arrows indicate forward and reverse modes of exchanger: extracellular Na+-dependent Ca2+ efflux and intracellular Na+-dependent Ca2+ influx, respectively.

(6)

Species difference in effect of reduced [Na+]_o on [Ca²⁺]_i. The above thermodynamic model predicts that an increase of [Ca²⁺]_i will not occur in K⁺-arrested guinea pig myocardium until [Na⁺]_o is reduced to the vicinity of 45 mM, the "critical" extracellular Na⁺ concentration that causes G_Na/Ca to change sign. By contrast, in quiescent rat myocardium, an increase of [Ca²⁺]_i is predicted to occur when [Na⁺]_o is reduced only slightly below 143 mM, to the vicinity of 130 mM. [The latter estimate arises by substituting the expression for C (analogous to Eq. 4) into Eq. 6 and solving for [Na⁺] with  = 21.]

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Fura 2 fluorescence ratio (340/380, index of [Ca2+]i), as a function of time, for individual right ventricular trabeculae of rat (A) and guinea pig (B). At left: unfiltered Ca2+ transients in response to electrical stimulation at 1 Hz. At right: records overlaid following digital filtering (second-order Bessel filter with 1-Hz cutoff frequency). Arrows indicate reduction of [Na+]o (from 143 mM) to values specified (mM). Dotted line in A indicates baseline fluorescence before elevation of [K+]o from 6 to 20 mM.

Normokalemic versus hyperkalemic arrest. By using a Ca²⁺ channel antagonist, we avoided the aforementioned possibility of species-dependent influx of Ca²⁺ via voltage-dependent Ca²⁺ channels. Furthermore, because cardiac arrest could now be achieved in 6 mM K⁺, we could simultaneously exploit the voltage sensitivity of the exchanger (Eq. A2) to generate new VO₂-[Na⁺]_o relationships. Our model predicts that these will be left shifted with respect to those previously generated (Fig. 2A) under hyperkalemic arrest. It further predicts that the extent of the left shift will be greater for rat hearts. These predictions were tested using 50 µM verapamil arrest, after which hearts of both species were challenged with various levels of reduced [Na⁺]_o and the peak rates of O₂ measured (as in Fig. 2).

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Cardiac VO2 of rats (circles) and guinea pigs (squares) during normokalemic (solid symbols) and hyperkalemic (open symbols) cardiac arrest. A: VO2 as a function of [Na+]o. Solid lines and open symbols: same data as in Fig 2A. Regression parameters for broken lines and solid symbols (fitted according to Eq. 1): (ymin, ymax, Km, n, sy · x, r2) = (6.7 ± 1.7, 39.5 ± 1.4, 39.5 ± 3.3, 2.47 ± 0.39, 1.96, 0.9977) and (5.1 ± 0.2, 32.6 ± 0.5, 22.7 ± 0.3, 0.23, 0.9944) for rat and guinea pig, respectively. B: VO2 as a function of GNa/Ca. Data for each experimental series normalized between zero (mean value observed at 143 mM [Na+]o) and unity (mean value observed at 3 mM [Na+]o). Regression line fitted (according to Eq. 1), using same values of  and [Na+] (21 and 45 mM, respectively) as in Fig 4B; resulting values of regression parameters: (ymin, ymax, Km, n, sy · x, r2) = (0.017 ± 0.024, 0.989 ± 0.041, 4.94 ± 0.43, 7.03 ± 1.03, 0.081, 0.9732).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

In the present study we documented a species difference in cardiac energetics and developed a simple thermodynamic model to explain it. Our approach capitalizes on the thermodynamic reversibility of the sarcolemmal Na⁺/Ca²⁺ exchanger. That is, when the [Na⁺]_o is sufficiently reduced, the exchanger reverses and facilitates transsarcolemmal Ca²⁺ influx. In turn, exchanger-mediated Ca²⁺ influx releases Ca²⁺ from the SR (1, 2, 26-28). The resulting increase of [Ca²⁺]_i (Fig. 5) stimulates the rate of resting energy expenditure (Fig. 1). Because, under conditions of low [Na⁺]_o, there is a progressive fall of [Na⁺]_i (1), none of the observed increase in metabolic rate can be attributed to the sarcolemmal Na⁺-K⁺-ATPase. Instead, it must reflect increased demands on sarcolemmal and SR Ca²⁺-ATPases, coupled with enhanced actomyosin ATPase activity of cross-bridges (i.e., increased P_LV, as shown in Figs. 1-3). Proof that these mechanical and metabolic responses reflect rising [Ca²⁺]_i due to reversal of the exchanger is provided in Fig. 3. When [Na⁺]_o is reduced in the absence of extracellular Ca²⁺, no increase of VO₂ or P_LV is observed (Fig. 3).

Species difference in metabolic response to reduced [Na+]_o. Our results (Fig. 2A) demonstrate that rat cardiac muscle is much more sensitive than guinea pig cardiac muscle to reduction of [Na⁺]_o. Once again, this cannot be attributed to a species difference in activity of the Na⁺-K⁺ pump. Its metabolic contribution, which is modest in any case (9, 38), must further diminish, reflecting the behavior of [Na⁺]_i that has been documented during low Na⁺ superfusion of rat ventricular myocytes (1). Nor can a species difference in Ca²⁺ influx via voltage-dependent Ca²⁺ channels be invoked as an explanation. We eliminated this possibility by use of the Ca²⁺-channel blocker verapamil to induce cardiac arrest. Whereas this induced a left shift of the VO₂-[Na⁺]_o relationships (Fig. 6A), the extent of displacement was quantitatively attributable (Fig. 6B) to the lower value of [K⁺]_o adopted (reflecting, in turn, the electrogenic nature of Na⁺/Ca²⁺ exchange). Thus the pronounced species difference in metabolic response to hyponatremic perfusion persisted in the face of diminished [Na⁺]_i and reduced [K⁺]_o.

Species difference in G_Na/Ca. It is our contention that the observed species differences, whether metabolic (Figs. 2A and 6A) or ionic (Fig. 5), in response to a reduction of [Na⁺]_o, can be explained by a species difference in G_Na/Ca. To that end, we developed a simple algebraic model that relates measured values of VO₂ to calculated values of G_Na/Ca. The model exploits the existence, in the guinea pig, of a relatively unambiguous "threshold" or "critical" value of [Na⁺]_o that is substantially displaced from the standard perfusate value of 143 mM. This critical value, [Na⁺], represents the [Na⁺]_o that renders the G_Na/Ca zero in the guinea pig heart.

Possible species differences in intracellular ion concentrations. The numeric value of , the stoichiometrically weighted ratio of intracellular ion concentrations in the two species (Eq. 6), warrants further consideration. We do so by defining G_Na/Ca to be the species difference in G_Na/Ca (i.e., the horizontal separation of VO₂-G_Na/Ca relationships, as in Fig. 4A). Then, by substituting Eqs. 3 and 6 into Eq. 2, it follows that

(7)

The numeric value of this expression is 7.84 kJ/mol (at 37°C). That is, the observed difference between the rat and guinea pig, in the response of cardiac metabolism to a reduction of [Na⁺]_o, may be attributed to a 7.84 kJ/mol difference in diastolic G_Na/Ca. To what could this difference be attributed?

Implications of a species difference in G_Na/Ca. Our conclusion that a 7.84 kJ/mol difference of diastolic G_Na/Ca distinguishes rat and guinea pig myocytes may explain a number of previously reported differences between these two species. Our model suggests that, in the quiescent rat heart, the difference between standard and critical [Na⁺]_o is only ~10 mM. Hence Na-dependent Ca²⁺ efflux is expected to be minimal in that species, in line with the hypothesis of Shattock and Bers (41). The resulting reduction in driving force for Ca²⁺ efflux via the exchanger is consistent with 1) the slightly higher resting VO₂ of rat whole hearts shown in Fig. 2, 2) the high frequency of spontaneous, mechanical oscillations in unstimulated cardiac tissues of the rat (22, 23), 3) the fivefold greater loss of exchangeable Ca²⁺ by guinea pig myocardium during prolonged rest (18), 4) the more rapid recovery of caffeine-depleted sarcoplasmic reticular Ca²⁺ stores by rat than by guinea pig cardiac myocytes (29), 5) the relative refractoriness of [Ca²⁺]_i to ryanodine-induced depletion of the SR in rat vis-à-vis guinea pig myocytes (19), 6) the greater extent of sarcoplasmic reticular Ca²⁺-loading during diastole in rat than in guinea pig papillary muscles or myocytes (33), 7) the greater metabolic cost of hyperosmolality-induced futile Ca²⁺ cycling by the SR of rat than of guinea pig whole hearts (13), and 8) the twofold greater magnitude of Na⁺/Ca²⁺ exchange current, for a given [Ca²⁺]_i challenge, in guinea pig than in rat cardiac myocytes (39). It is certainly consistent with the much greater metabolic response to hyponatremic perfusion evinced by rat hearts in the present study.

Perspectives