Intracellular Na+ Concentration Is Elevated in Heart Failure But Na/K Pump Function Is Unchanged
Background— Intracellular sodium concentration ([Na+]i) modulates cardiac contractile and electrical activity through Na/Ca exchange (NCX). Upregulation of NCX in heart failure (HF) may magnify the functional impact of altered [Na+]i.
Methods and Results— We measured [Na+]i by using sodium binding benzofuran isophthalate in control and HF rabbit ventricular myocytes (HF induced by aortic insufficiency and constriction). Resting [Na+]i was 9.7±0.7 versus 6.6±0.5 mmol/L in HF versus control. In both cases, [Na+]i increased by ≈2 mmol/L when myocytes were stimulated (0.5 to 3 Hz). To identify the mechanisms responsible for [Na+]i elevation in HF, we measured the [Na+]i dependence of Na/K pump–mediated Na+ extrusion. There was no difference in Vmax (8.3±0.7 versus 8.0±0.8 mmol/L/min) or Km (9.2±1.0 versus 9.9±0.8 mmol/L in HF and control, respectively). Therefore, at measured [Na+]i levels, the Na/K pump rate is actually higher in HF. However, resting Na+ influx was twice as high in HF versus control (2.3±0.3 versus 1.1±0.2 mmol/L/min), primarily the result of a tetrodotoxin-sensitive pathway.
Conclusions— Myocyte [Na+]i is elevated in HF as a result of higher diastolic Na+ influx (with unaltered Na/K-ATPase characteristics). In HF, the combined increased [Na+]i, decreased Ca2+ transient, and prolonged action potential all profoundly affect cellular Ca2+ regulation, promoting greater Ca2+ influx through NCX during action potentials. Notably, the elevated [Na+]i may be critical in limiting the contractile dysfunction observed in HF.
Received January 17, 2002; revision received March 6, 2002; accepted March 6, 2002.
Cardiac contractile dysfunction in heart failure (HF) is due largely to altered Ca2+ transport. Specifically, HF cardiac myocytes exhibit depressed and prolonged Ca2+ transients.1 These changes can be explained by reductions in sarcoplasmic reticulum (SR) Ca2+-ATPase2 and increased Na/Ca exchange (NCX)3–6⇓⇓⇓ function.
Intracellular sodium concentration ([Na+]i) affects excitation-contraction coupling by modulating pH and [Ca2+]i through Na/H exchange and NCX, respectively.6 The upregulated NCX in HF may increase the functional impact of altered [Na+]i. Recent reports7–10⇓⇓⇓ indicate that [Na+]i is increased in hypertrophy; however, HF data are limited. Preliminary data in human HF suggest some increase in [Na+]i,11 but a slight decrease in [Na+]i was found in pacing-induced HF rabbits.12
Higher [Na+]i could be explained by lower Na/K pump activity, consistent with reports of decreased Na/K pump expression and isoform shifts in some HF models.13–17⇓⇓⇓⇓ However, functional studies in HF ventricular myocytes are sparse and contradictory.8,15⇓ Enhanced Na+ influx could also raise [Na+]i, and this explains the higher [Na+]i in rat versus rabbit ventricular myocytes.18
Our first aim was to determine whether [Na+]i is altered in a nonischemic rabbit HF model that we have extensively characterized.4,19⇓ HF induced by aortic insufficiency and constriction resulted in 90% of animals having nonsustained ventricular tachycardias (10% incidence of sudden death). NCX expression is 2-fold increased, and contractions, Ca2+ transients, and SR Ca2+ load are reduced.4,19⇓ We measured [Na+]i in ventricular myocytes at 37°C, using the fluorescent indicator sodium binding benzofuran isophthalate (SBFI) and validated calibration methods.18 We found that [Na+]i is higher in HF versus control myocytes at rest and during stimulation. A second aim was to determine why [Na+]i is higher in HF (altered extrusion or influx). Using Na+ loading/recovery protocols18 to measure Na+ extrusion by the Na/K pump, we found that the maximal Na+ transport rate (Vmax) and [Na+]i for half-maximal stimulation (Km) are comparable in control and HF myocytes. On the other hand, resting Na+ influx (measured as the initial rate of [Na+]i rise on abrupt Na/K pump inhibition) was significantly higher in HF than control. Therefore, higher [Na+]i is due to elevated diastolic Na+ influx rather than altered Na/K pump characteristics in this HF model. We also show how the elevated [Na+]i in HF has major functional consequences for calcium flux during the action potential (AP).
Rabbit HF Model and Myocyte Isolation
HF was induced in rabbits by aortic insufficiency, followed 2 to 4 weeks later by aortic constriction as described.19 Rabbits were studied ≈5 months later, when the left ventricular end-systolic dimension exceeded 1.20 cm (University of Illinois Animal Studies Committee–approved protocols). Myocytes were isolated as described,4,19⇓ with back-flow across the incompetent aortic valve in HF rabbits blocked by an inflated balloon-tipped catheter. Data were obtained from 14 control and 11 HF rabbits.
Myocytes were plated on laminin-coated coverslips and incubated with 10 μmol/L SBFI-AM and Pluronic F-127 (0.05% wt/vol) for 90 minutes at room temperature. After washout, SBFI-AM was allowed to deesterify for 20 minutes. The normal Tyrode’s solution contained (in mmol/L): 140 NaCl, 4 KCl, 1 MgCl2, 2 CaCl2, 10 HEPES, and 10 glucose (pH=7.4). Fluorescence excitation at 340 and 380 nm (F340 and F380, alternating at 100 Hz) was by a 75-W xenon lamp, and emission was recorded at 535±20 nm. F340/F380 was calculated after background subtraction and converted to [Na+]i by calibration at the end of each experiment (using divalent-free solutions with 0, 10, or 20 mmol/L extracellular [Na+], [Na+]o) in the presence of 10 μmol/L gramicidin and 100 μmol/L strophanthidin.18 Measurements were at 35 to 37°C.
Na+ Efflux Through the Na/K Pump
Na/K pump flux was determined as the rate of pump-mediated [Na+]i decline.18 Myocytes were Na+-loaded by inhibiting the Na/K pump in a K+-free solution containing (in mmol/L): 145 NaCl, 2 EGTA, 10 HEPES, and 10 glucose (pH=7.4). [Na+]i decline was measured on pump reactivation in solution containing (mmol/L): 140 TEA-Cl, 4 KCl, 2 EGTA, 10 HEPES, and 10 glucose (pH=7.4). Since cell volume did not change with this protocol (n=4), [Na+]i decline reflects Na+ efflux. The rate of [Na+]i decline (−d[Na+]i/dt) was plotted versus [Na+]i and fitted with Jpump=Vmax/(1+(Km/[Na+]i)nHill).
In some experiments, this protocol was repeated with the use of whole-cell voltage clamp or current clamp; 5 to 10 MΩ pipettes were filled with (in mmol/L): 30 KCl, 110 K-aspartate, 5 NaCl, 10 HEPES, 5 MgATP, 0.72 MgCl2 (1 free [Mg2+]), 3 BAPTA, 1.15 CaCl2 (100 nmol/L free [Ca2+]), and 0.2 SBFI, pH=7.2. In some experiments, further isolation of the Na/K pump current (Ip) used isoosmotic replacement of (mmol/L) 20 internal KCl with TEA-Cl, and 7 external NaCl or TEA-Cl with 5 NiCl2 and 2 BaCl2, with Em=−30 mV to inactivate sodium channels.
Resting Na+ Influx
Resting Na+ influx was taken as the initial rate of [Na+]i rise after abrupt Na/K pump inhibition with strophanthidin (200 μmol/L), with physiological [Na+]o and [Ca2+]o. Some measurements were made with tetrodotoxin, HOE 642 (provided by Dr J. Punter, Aventis Pharma, Frankfurt, Germany), and/or Ni2+ present.
Data are expressed as mean±SEM, and the Student’s unpaired t test was used.
[Na+]i at Rest and During Stimulation
[Na+]i was measured in control and HF myocytes at rest and during stimulation at 0.5 to 3 Hz. Figure 1, A and B, shows a control cell in which [Na+]i was measured in normal Tyrode’s solution, first at rest and then stimulated at 3 Hz for 10 minutes, reaching a new steady state in 3 to 5 minutes. When stimulation was stopped, [Na+]i returned to baseline (Figure 1A, inset shows SBFI calibration).18 Resting [Na+]i was significantly higher in HF (9.7±0.7 mmol/L, n=20) versus control (6.6±0.5 mmol/L, n=24; Figure 1D and the Table). On stimulation, [Na+]i rose by similar amounts in control and HF at all frequencies (Figure 1C). Thus, [Na+]i remained significantly higher in HF during stimulation. Because there was little difference in [Na+]i between 0.5 to 3 Hz, we pooled all the stimulation data for Figure 1D and the Table. Average [Na+]i in contracting myocytes was 11.3±0.9 mmol/L (n=15) in HF and 8.0±1.1 mmol/L (n=12) in control.
Na/K Pump–Mediated Na+ Efflux
Higher [Na+]i in HF versus control might be explained by a reduced ability of the Na/K pump to extrude Na+. To test this, we determined [Na+]i dependence of Na+ efflux through the Na/K pump. Myocytes were Na+-loaded by incubating in K+-free solution to block the Na/K pump18 (Figure 2A). Extracellular Na+ was then removed, and we measured the time course of [Na+]i decline with the Na/K pump active (4 mmol/L [K+]o). Additionally, [Na+]i decline was measured with 100 μmol/L strophanthidin to isolate passive sodium efflux.
To determine Na/K pump–mediated flux, [Na+]i decline in the presence and in the absence of strophanthidin was numerically differentiated, and d[Na+]i/dt was plotted as a function of [Na+]i (Figure 2B). The maximal pump-independent Na+ efflux rate (passive) was higher in HF (5.2±0.2 mmol/L/min, n=5, versus 3.2±0.1 mmol/L, n=6, in control, P<0.001). Na/K pump–mediated Na+ efflux (total minus passive) is plotted versus [Na+]i (Figure 2C). Data were fit with a Hill expression to derive Vmax, Km, and nHill (Table). These data show that Na/K pump characteristics are similar in control and HF, so elevated [Na+]i in HF is not due to a lower Na/K pump rate. Moreover, Figure 2C shows that at resting [Na+]i (6.6 and 9.7 mmol/L for control and HF), Na+ efflux mediated by the Na/K pump is ≈2 times higher in HF versus control (Table). Since Na+ influx and efflux must be equal and opposite at steady state, this suggests that resting Na+ influx is also higher in HF; this could cause the higher [Na+]i in HF myocytes.
Membrane potential (Em) was not controlled in Figure 2. Although Na/K pump activity is nearly voltage-insensitive in Na+-free solution,20 non–pump-mediated Na+ efflux could vary with Em, complicating our analysis. To test this, we repeated Na+ efflux experiments under voltage clamp and current clamp conditions. Figure 3A shows a typical experiment (n=4) in which a myocyte was first clamped at constant Em (−80 mV), whereas [Na+]i decline was measured in the presence and absence of strophanthidin. We then switched to current clamp and repeated the protocol. The cell depolarized to −25.7 mV during Na+ loading in K+-free solution (Figure 3A, lower trace). However, with 4 mmol/L K+ during Na+ efflux measurements, Em repolarized to −85.3 and −86.5 mV (with and without strophanthidin, respectively). This is consistent with a very small Na/K pump contribution to resting Em in ventricular myocytes.21,22⇓ Figure 3B shows that the rate of [Na+]i decline is similar under voltage-clamp and current-clamp conditions, both for passive (with strophanthidin) and total Na+ efflux (without strophanthidin). Thus, results shown in Figure 2 are not affected by lack of voltage control. In voltage-clamp experiments in which ionic conditions were chosen to isolate Ip with simultaneous [Na+]i measurement, we compared the [Na+]i dependence of Ip and pump d[Na+]i/dt. Since these curves are comparable (Figure 3C), Na+ efflux rate measured by d[Na+]i/dt reports Na/K pump function much like Ip (although converting Ip to d[Na]i/dt requires an assumed surface-to-volume ratio).
Resting Na+ Influx
To test the hypothesis that resting Na+ influx is higher in HF, we measured resting Na+ influx as the initial rate of [Na+]i increase on abrupt Na/K pump inhibition (Figure 4A). As expected, Na+ influx was twice as high in HF (2.26±0.31 mmol/L/min, n=9) versus control (1.13±0.15 mmol/L/min, n=10) (Figure 4B).
Next we investigated various Na+ entry pathways. First, we measured Na+ influx with Na+ channels blocked (30 μmol/L tetrodotoxin, TTX). Figure 4B shows that TTX abolishes the HF versus control difference and that TTX-sensitive Na+ entry is significantly higher in resting HF versus control cells. This higher TTX-sensitive Na+ influx accounts for ≈85% of the difference in the total Na+ entry. In another series of experiments, we measured Na+ influx in the presence of TTX (30 μmol/L), HOE 642 (2 μmol/L, to block the Na/H exchange), and Ni2+ (5 mmol/L, to block NCX). The presence of all these blockers did not completely abolish Na+ influx, suggesting that other mechanisms may contribute to Na+ entry (eg, TTX-insensitive background Na+ leak channel or Na/K/2Cl cotransport). Subtracting the influx measured in the presence of all three blockers from that in the presence of TTX gives the Ni2+- and HOE 642-sensitive pathways, which are not significantly altered in HF (Figure 4B).
We studied Na+ regulation in myocytes from a rabbit model of HF. Using SBFI, we determined [Na+]i, the [Na+]i dependence of Na+ transport by the Na/K pump, and the rate of Na+ influx. We found that (1) [Na+]i is higher in HF versus control, both at rest and during stimulation, (2) [Na+]i dependence of the Na/K pump is unchanged in HF, (3) resting Na+ influx and efflux are greater in HF, and (4) increased Na+ influx in HF is largely TTX-sensitive.
[Na+]i Is higher in HF Myocytes: Physiological Consequences
[Na+]i is ≈3 mmol/L higher in resting as well as contracting HF myocytes. Higher [Na+]i has been reported in cardiac hypertrophy,7–10⇓⇓⇓ but the only published [Na+]i measurements in HF12 showed a slight decrease (by 0.8 mmol/L) in rabbits with pacing-induced HF. However, this HF model differs significantly from the model used here because the density of the NCX current was significantly reduced, whereas in our model4,19⇓ and in human HF,3,5⇓ NCX expression is increased.
Higher [Na+]i in HF has very important implications for NCX function and consequently on Ca2+ transients and contractility. Figure 5 shows simulations of how NCX current (INCX) and net Ca2+ transport by NCX (∫INCX) may be expected to vary during a steady-state AP in HF and control, with the use of values of [Na+]i measured here (8.0 and 11.3 mmol/L in control and HF myocytes, respectively). INCX was calculated by means of the equation described by Weber et al,24 and we accounted for enhanced NCX expression, AP prolongation by 42 ms (Figure 5A), and reduced [Ca2+]i transients, as previously recorded in HF and control rabbit myocytes.4,19⇓ Because NCX actually senses the submembrane [Ca2+]i ([Ca2+]sm), which can differ from the bulk [Ca2+]i,23 we approximated [Ca2+]sm by using the procedure described by Weber et al24 (Figure 5B). We assumed that submembrane [Na+]i ([Na+]sm) is equal to bulk [Na+]i. However, this may not be the case, because [Na+]sm may be elevated transiently as the result of Na+ influx through Na+ channels or NCX, favoring more outward INCX.
In control, outward INCX is only expected for a brief period (Figure 5, C and D). The rise in [Ca2+]sm quickly favors inward INCX. This would also be true in HF if [Na+]i were unaltered (Figure 5C, dotted line). However, the measured [Na+]i increase in HF shifts INCX in the outward direction. Outward INCX produces calcium entry for ≈150 ms, reducing the time for NCX-mediated Ca2+ efflux. Indeed, in human HF, it was suggested that Ca2+ influx through outward INCX may occur during the AP.25 Three factors contribute to the increase in Ca2+ influx through NCX in HF: higher [Na+]i, prolonged AP, and smaller Ca2+ transients.19 Figure 6 shows the individual effect of each of these factors on the total Ca2+ efflux through NCX. That is, the INCX integral is less inward as [Na+]i rises, twitch Δ[Ca]i declines, and AP duration increases. For the changes observed in this HF model (reported here and by Pogwizd et al),4,19⇓ increased [Na+]i has the largest impact on NCX function. Indeed, the greater Ca2+ influx and lower Ca2+ efflux would tend to load the cell (and SR) with Ca2+, making more available for contractile activation. Therefore, by favoring Ca2+ influx through NCX, higher [Na+]i may minimize contractile dysfunction in HF. Moreover, we speculate that NCX upregulation is compensatory in allowing NCX to still extrude the greater amount of calcium that enters during the AP.
[Na+]i Dependence of the Na/K Pump Is Unchanged in HF Myocytes
We found that Vmax and Km for the Na/K pump are similar in control and HF. This is somewhat surprising, considering that several biochemical studies revealed decreased expression and/or isoform shifts of the Na/K pump in failing or hypertrophied hearts.13–17⇓⇓⇓⇓ However, most of these studies were performed in tissue homogenates and might reflect changes in nonmyocytes. Moreover, such measurements cannot differentiate between the internalized versus the sarcolemmal Na/K pumps26 nor between functional and inactive pumps. It is therefore possible that whereas the total pool size of immunoreactive subunits decreases, the density of the Na/K pump in the sarcolemma is relatively unchanged. There are few functional Na/K pump studies in HF myocytes, with somewhat contradictory results. Decreased Vmax and unchanged [Na+]i affinity have been reported in rats with HF after myocardial infarction.15 On the contrary, unaltered maximal Ip and lower [Na+]i affinity have been found in myocytes from dogs with chronic atrioventricular block and hypertrophy.8 Reduced Ip has been reported in a hypertrophic rat model with increased SR Ca2+ content.10
The Vmax values here at 37°C are approximately twice what we measured at 23°C in rabbit ventricular myocytes.18 The Km found here compares well with values derived from Na/K pump current measurements in cardiac cells in the presence of intracellular K+ or Cs+.20,27,28⇓⇓ The most important result regarding the Na/K pump is that under physiological conditions (ie, appropriate resting [Na+]i), the rate of Na+ extrusion by the pump is higher in HF (Table and Figure 2C). Higher resting Na/K pump current (≈2-fold) has also been reported for dogs with chronic atrioventricular block.8
Higher [Na+]i in HF Is Due to Enhanced Na+ Influx
Higher Na+ efflux in HF must be balanced by enhanced Na+ influx to maintain steady-state [Na+]i balance. Our results confirm that resting Na+ influx in HF is twice as high as in control (Table and Figure 4B). Most of the excess resting Na+ influx in HF myocytes occurs through a TTX-sensitive pathway. This suggests that a larger number of Na+ channels are open in quiescent HF myocytes versus control (perhaps like a window current). Resting HF cells were not more depolarized (not shown), ruling out one simple explanation. Another explanation might be functional alteration or expression of Na+ channels in HF. There are indications that the density of a slow inactivating, persistent, TTX-sensitive Na+ current is increased in failing ventricular myocytes,29 with this current being partially responsible for AP prolongation in HF.
In summary, [Na+]i is higher in HF as the result of an elevation of diastolic Na+ influx with unaltered Na/K pump characteristics. Along with decreased Ca2+ transients and longer AP duration, the increased [Na+]i contributes to greater Ca2+ influx through NCX during the AP in HF. This would increase cellular and SR calcium content. This increased Ca2+ influx in HF depends especially on the elevated [Na+]i and may be functionally important in limiting the extent of contractile dysfunction in HF.
This work was supported by National Institutes of Health grants HL-30077 (Dr Bers), HL-64724 (Dr Bers), and HL-46929 ( Dr Pogwizd) and American Heart Association Fellowship 0010180Z (Dr Weber). We thank Lu Leach, Jodi Jeanes, and Jorge Acevedo for technical support.
- ↵Bailey B, Houser SR. Sarcoplasmic reticulum-related changes in cytosolic calcium in pressure-overload-induced feline left ventricular hypertrophy. Am J Physiol. 1993; 265: H2009–H2026.
- ↵Mercadier JJ, Lompre AM, Duc P, et al. Altered sarcoplasmic reticulum Ca2+-ATPase gene expression in the human ventricle during end-stage heart failure. J Clin Invest. 1990; 85: 305–309.
- ↵Studer R, Reinecke H, Bilger J, et al. Gene expression of the cardiac Na+-Ca2+ exchanger in end-stage human heart failure. Circ Res. 1994; 75: 443–453.
- ↵Pogwizd SM, Qi M, Yuan W, et al. Upregulation of Na+/Ca2+ exchanger expression in an arrhythmogenic rabbit model of heart failure. Circ Res. 1999; 85: 1009–1019.
- ↵Hasenfuss G, Schillinger W, Preuss M, et al. Relationship between Na+-Ca2+ exchanger protein levels and diastolic function of failing human myocardium. Circulation. 1999; 99: 641–648.
- ↵Bers DM. Excitation-Contraction Coupling and Cardiac Contractile Force. 2nd ed. Dordrecht, The Netherlands: Kluwer Academic Press; 2001: 427.
- ↵Jelicks LA, Siri FM. Effects of hypertrophy and heart failure on [Na+]i in pressure-overloaded guinea pig heart. Am J Hypertens. 1995; 8: 934–943.
- ↵Verdonk F, Volders PGA, Vos MA, et al. Cardiac hypertrophy is associated with an increase in subsarcolemal Na+. Biophys J. 2001; 80: 598a.Abstract.
- ↵Mészáros J, Khananshvili D, Hart G. Mechanisms underlying delayed afterdepolarizations in hypertrophied left ventricular myocytes of rats. Am J Physiol. 2001; 281: H903–H914.
- ↵Maier LS, Hasenfuss G, Pieske B. Frequency-dependent changes in intracellular Na+ concentration in isolated human myocardium. Circulation. 1997; 96 (suppl I): I-178. Abstract.
- ↵Yao A, Su Z, Nonaka A, et al. Abnormal myocyte Ca2+ homeostasis in rabbits with pacing-induced heart failure. Am J Physiol. 1998; 275: H1441–H1448.
- ↵Schwinger RH, Wang J, Frank K, et al. Reduced sodium pump α1, α3, and β1-isoform protein levels and Na+,K+-ATPase activity but unchanged Na+-Ca2+ exchanger protein levels in human heart failure. Circulation. 1999; 99: 2105–2112.
- ↵Kim CH, Fan TH, Kelly PF, et al. Isoform-specific regulation of myocardial Na, K-ATPase alpha-subunit in congestive heart failure: role of norepinephrine. Circulation. 1994; 89: 313–320.
- ↵Ellingsen O, Holthe MR, Svindland A, et al. Na, K-pump concentration in hypertrophied human hearts. Eur Heart J. 1994; 15: 1184–1190.
- ↵Shamraj OI, Grupp IL, Grupp G, et al. Characterisation of Na/K-ATPase, its isoforms, and the inotropic response to ouabain in isolated failing human hearts. Cardiovasc Res. 1993; 27: 2229–2237.
- ↵Pogwizd SM, Schlotthauer K, Li L, et al. Arrhythmogenesis and contractile dysfunction in heart failure: roles of sodium-calcium exchange, inward rectifier potassium current and residual β-adrenergic responsiveness. Circ Res. 2001; 88: 1159–1167.
- ↵Nakao M, Gadsby DC. [Na] and [K] dependence of the Na/K pump current-voltage relationship in guinea pig ventricular myocytes. J Gen Physiol. 1989; 94: 539–565.
- ↵Stimers JR, Shigeto N, Lieberman M. Na/K pump current in aggregates of cultured chick cardiac myocytes. J Gen Physiol. 1990; 95: 61–76.
- ↵Levi AJ. The electrogenic sodium/potassium pump and passive sodium influx of isolated guinea pig ventricular myocytes. J Cardiovasc Electrophysiol. 1992; 3: 225–238.
- ↵Weber CR, Piacentino V III, Ginsburg KS, et al. Na/Ca exchange current and submembrane [Ca] during the cardiac action potential. Circ Res. 2002; 90: 182–189.
- ↵Dipla K, Mattiello JA, Margulies KB, et al. The sarcoplasmic reticulum and the Na+/Ca2+ exchanger both contribute to the Ca2+ transient of failing human ventricular myocytes. Circ Res. 1999; 84: 435–444.
- ↵Gao J, Mathias RT, Cohen IS, et al. Two functionally different Na/K pumps in cardiac ventricular myocytes. J Gen Physiol. 1995; 106: 995–1030.