Dynamic Regulation of Sodium/Calcium Exchange Function in Human Heart Failure
Background— Sarcolemmal Na/Ca exchange (NCX) regulates cardiac Ca and contractility. NCX function during the cardiac cycle is determined by intracellular [Ca] and [Na] ([Ca]i, and [Na]i) and membrane potential (Em), which all change in human heart failure (HF). Therefore, changes in NCX function may contribute to abnormal Ca regulation in human HF.
Methods and Results— We assessed the cellular bases of differences in NCX function in ventricular myocytes from failing (F) and nonfailing (NF) human hearts. Allosteric activation of NCX by [Ca]i was comparable in F and NF myocytes (K1/2=150±31 nmol/L, n=7). The steady-state relation between [Ca]i and NCX current (INCX) was used to infer the local submembrane [Ca]i ([Ca]sm) that is sensed by NCX dynamically during the action potential (AP) and Ca transient (37°C). This involved “tail” INCX measurement during abrupt repolarization of APs and Ca transients, where peak inward INCX indicates [Ca]sm. This allows inference of the direction of Ca transport by the NCX during the AP. In NF myocytes, NCX extrudes Ca for most of the AP. Three factors shift the direction of NCX-mediated Ca transport (to favor more Ca influx) in F versus NF myocytes, as follows: (1) reduced [Ca]sm, (2) prolonged AP duration, and (3) elevated [Na]i.
Conclusions— These results show that Ca entry through NCX may limit systolic dysfunction due to reduced sarcoplasmic reticulum Ca stores in HF but could contribute to slow decay of the [Ca]i transient and to diastolic dysfunction.
Received April 2, 2003; de novo received June 4, 2003; revision received August 8, 2003; accepted August 13, 2003.
Sarcolemmal Na/Ca exchange (NCX) competes with sarcoplasmic reticulum (SR) Ca-ATPase in causing systolic Ca transient decline and relaxation.1 In many hypertrophy and heart failure (HF) models, NCX expression is increased,2–6 and this could enhance the contribution of NCX to twitch relaxation. Because NCX and SR Ca-ATPase compete during [Ca]i decline, reduced SR Ca-ATPase function indirectly favors greater NCX contribution to relaxation (even without altered NCX expression or properties).1,7 In addition, NCX function is strongly influenced by [Ca]i, [Na]i, and membrane potential (Em), all of which change in HF. Typically in HF (versus control), Ca transients are smaller in amplitude and decline more slowly,8,9 [Na]i is higher,10,11 and action potential duration (APD) is prolonged.4,12,13 These would all reduce Ca extrusion through NCX during the cardiac cycle and bias NCX toward Ca influx. In addition, local submembrane [Ca]i ([Ca]sm) sensed by NCX during normal Ca transients is higher than global [Ca]i measured by fluorescent Ca indicators.14,15 Thus, quantitative analysis of NCX function is imperative to understand how NCX functions during action potentials (AP) in human HF.
When Em is negative to the NCX reversal potential (ENCX=3ENa−2ECa), Ca extrusion is favored (inward NCX current, INCX), whereas Em above ENCX favors Ca entry (outward INCX). As [Ca]i, [Ca]sm, and Em change during the AP, net INCX changes dynamically. Low [Ca]i, high [Na]i, and long APD favor Ca entry through NCX. Our goal here is to evaluate how NCX functions during the AP in ventricular myocytes from failing (F) and nonfailing (NF) human hearts.
We show that Ca entry through NCX is readily demonstrable, characterize allosteric NCX activation by [Ca]i, determine [Ca]sm during normal AP, and infer the actual Ca flux through NCX during the AP. In NF myocytes, NCX primarily extrudes Ca throughout the cardiac cycle. However, in typical F myocytes, there is substantial Ca entry through NCX during the AP (explaining the slowly rising [Ca]i seen in many F myocytes). Additionally, we demonstrate how important thermodynamic determinants are (versus number of NCX molecules) to NCX behavior.
Cell Isolation, Electrophysiology, and [Ca]i Measurement
Myocytes were isolated from F and NF human hearts as described previously and were from the same 11 F and 7 NF patients (mean ejection fractions were 17.5±4.9% and 54.1±4.6% by echocardiography).7 Em and current were controlled and recorded using discontinuous, single-electrode voltage clamp at 37°C. Ca signals were recorded with the use of fluo-3.
Solutions Used for Electrophysiological Recording
Pipettes (1 to 3 MΩ) were filled with the following (in mmol/L): 130 cesium aspartate, 20 tetraethylammonium chloride, 1 MgCl2, 10 HEPES, 2.5 NaCl, 5 Na2ATP, and 0.1 potassium–fluo-3, pH 7.2, using CsOH at 37°C. Total [Na] (12.5 mmol/L) was based on human myocyte estimates.10 Seals were attained in solution (normal Tyrode) containing the following (in mmol/L): 150 NaCl, 5.4 KCl, 10 glucose, 5 HEPES, and 1 MgCl2, 1 CaCl2, with pH 7.4 at 37°C using NaOH. After patch rupture, 30 μmol/L niflumate and 3 mmol/L 4-aminopyridine were added (to block Ca-activated chloride and potassium currents), and CsCl replaced KCl.
Allosteric Ca Activation of INCX
Cells were pretreated with 1 μmol/L thapsigargin to block SR Ca-ATPase and 20 μmol/L nifedipine-blocked L-type Ca channels. Em was initially −100 mV to keep [Ca]i low but then oscillated between +100 mV and −100 mV to drive Ca flux through INCX.16
Time Course of [Ca]sm During AP
Nifedipine and thapsigargin were omitted, allowing normal excitation contraction coupling (ECC). A typical AP, recorded with physiological solutions at 1 Hz (37°C) from a F human myocyte, was used as an AP-clamp waveform. AP clamps were interrupted at 10 to 500 ms by repolarization to −70 mV, yielding tail INCX. Capacitative current and ICa were subtracted,15 leaving only INCX, which was used to infer [Ca]i sensed by NCX ([Ca]sm).
Steady-State INCX Versus [Ca]i
The steady-state relation between [Ca]i and INCX is described in the Equation.15,16
The three factors are allosteric regulation by [Ca]i, which is not transported (Allo), competitive binding of Na and Ca as substrates (Bind) and the effects of Em (Elect). The latter two combined give an electrochemical term. Km values are intracellular (i) and extracellular (o) Na and Ca dissociation constants. KmCaAct=150 nmol was determined here; other constants are as in Reference 15. Maximal INCX flux (Vmax) was determined for each cell by using the measured [Ca]i dependence of INCX (Vmax=8.8±3.6 and 6.8±1.8 A/F for NF and F), consistent with unaltered NCX mRNA and [Ca]i dependence in these F hearts.7
Ca Fluxes Through NCX
Slow depolarizing voltage ramps from −100 mV (Figure 1A) show Ca entry through ICa and INCX (with SR function blocked). [Ca]i increased biphasically with depolarization (see also rate of [Ca]i rise, d[Ca]i/dt). Figure 1B shows that d[Ca]i/dt rises, with a hump between −25 and +30 mV reflecting Ca entry through ICa, and rises again at more positive Em (as Ca entry through ICa declines and that through INCX rises). Moreover, d[Ca]i/dt attributable to INCX reaches levels comparable to that through ICa and drives [Ca]i considerably higher. These Ca fluxes emphasize that substantial Ca entry can occur through INCX (and was similar in F or NF human myocytes).
Repolarization promotes Ca removal by NCX, shown three ways in Figure 1C. One is based on d[Ca]i/dt (after correcting for cytosolic Ca buffering).7 The solid red curve is measured INCX (corrected for surface-to-volume ratio=13 pF/pL).7 The dashed curve is based on the Equation (using Vmax=5.1 A/F, pipette [Na]i=12.5 mmol/L, and the measured [Ca]i and Em). This demonstrates that [Ca]i decline correlates well with measured INCX and also that the Equation describes well the [Ca]i dependence of INCX in human ventricular myocytes.
During the cardiac cycle, Em, [Ca]i, and [Na]i all influence INCX activity electrochemically (and [Ca]i activates NCX allosterically), even at constant NCX expression. Figure 2 illustrates the independent effects of Em, [Na]i, and [Ca]i on outward INCX, using the Equation. As Em increases from 0 mV to +100 mV, outward INCX increases dramatically (Figure 2A), corresponding with the progressive rise in d[Ca]i/dt at positive Em in Figure 1A. Figure 2B shows how increasing [Na]i promotes outward INCX. Alterations in [Ca]i have complex effects on outward INCX because the electrochemical and allosteric effects are directionally opposite (Figure 2C). As [Ca]i rises, outward INCX at +20 mV (Ca influx) first increases as a result of allosteric Ca activation, but at higher [Ca]i the electrochemical effect becomes more dominant and outward INCX decreases. For inward INCX (Em=−80 mV), higher [Ca]i increases both allosteric and electrochemical factors.
Allosteric Activation of INCX
We measured allosteric activation of INCX in intact human myocytes (with other transporters and channels blocked),16 using Em to control INCX direction (Figure 3B and 3C). Outward INCX at +100 mV drives Ca entry, whereas inward INCX extrudes Ca at −100 mV, creating [Ca]i oscillations (Figure 3A). During each pulse, INCX and [Ca]i values were taken to create Figure 3D. Inward INCX at −100 mV increases as [Ca]i rises, as a result of both allosteric Ca activation and increased thermodynamic driving force. Outward INCX at +100 mV decreases with rising [Ca]i, (reduced driving force), but the increase in outward INCX with [Ca]i must be attributed to allosteric activation. Figure 3D shows mean fits of the Equation to both the inward and outward INCX. There was no major difference detected between F and NF human myocytes, and the mean allosteric KmCaAct was 150±31 nmol/L.
Submembrane [Ca] and INCX During AP
Fluorescent indicators do not sense local [Ca]sm near the NCX. This is a minor complication for Figure 3 but more problematic when ICa and SR release are functional, and [Ca]sm may far exceed bulk [Ca]i.15 High [Ca]sm biases INCX in the inward direction versus that expected from [Ca]i. Using our novel approach developed in rabbit myocytes,15 we assessed [Ca]sm in human F and NF myocytes. This allows more direct evaluation of NCX function during the AP.
Peak twitch [Ca]i is lower in F versus NF human myocytes.7 In Figure 4A, the AP was interrupted at different times and INCX tails were measured at −70 mV (Figure 4B). Here, ICa, SR Ca transport, and ECC are fully functional. Conditioning pulses before each test AP ensured steady-state SR Ca loading. We plot measured INCX versus [Ca]i (Figure 4D) and compared initial INCX with the steady-state relation (fitting the Equation to the late [Ca]i decline; thick line in Figure 4D). Initial INCX is higher than expected for [Ca]i, and we extrapolate [Ca]sm from INCX (at the moment of repolarization) to the steady-state curve. Repeating this procedure at different times allows reconstruction of the [Ca]sm versus global [Ca]i transient in this HF cell (Figure 4E).
In NF human myocytes, INCX is inward for almost the entire AP, except very briefly early in the AP. The first inward current hump is driven largely by high [Ca]sm, whereas the later inward hump is driven mainly by repolarization. Integrated Ca flux by INCX (Figure 5E) reaches >30 μmol Ca/L cytosol. We calculate that to reach the observed peak [Ca]i requires 120 μmol Ca/L cytosol (based on our cytosolic Ca buffering measurement).7 This implies that NCX extrudes 25% of activator Ca during the cardiac cycle in NF human myocytes (matching values from simpler analysis of [Ca]i decline).7
In HF, there is substantial outward INCX during the AP (Figure 5D). This is largely due to lower peak [Ca]i, since the AP-clamp was identical and 12.5 mmol/L [Na]i was used for F and NF in both the pipette and calculations. Thus, the low [Ca]i and [Ca]sm cause sufficient outward shift in INCX to bring in substantial Ca during the AP. The integral in Figure 5E peaks at +12 μmol Ca/L cytosol before repolarization causes INCX to switch directions and extrude Ca. This suggests significant Ca entry during the AP in HF, which can contribute to the Ca transient and contraction (thereby limiting the negative inotropic impact of reduced SR Ca content).7 The mean rate of Ca influx through INCX during 100 to 300 ms of the HF AP (Figure 5D) is 36 μmol/L cytosol/s. This is almost identical to the gain in Ca based on the slow rise in [Ca]i during this time in Figure 5B (35 μmol/L cytosol/s; accounting for cytosolic buffering). We conclude that this slowly rising [Ca]i during the HF AP is largely due to Ca entry through NCX. Figure 5E shows that by end-diastole in HF (1 second at 1 Hz), NCX has extruded a net 20 μmol/L cytosol. Thus, in the HF case, NCX can remove the Ca that entered through NCX and ICa. At a heart rate of 1.5 Hz, Ca would tend to accumulate in F myocytes, but APD also shortens, reducing Ca entry and allowing more time for diastolic Ca extrusion through INCX.
Figure 5 is appropriate for myocytes studied here (with some cellular dialysis), but in NF versus F myocytes, the APD is shorter and [Na]i is lower.7,10,11 Figure 6A extends our analysis to simulate NF conditions, with shorter AP duration (by 100 ms), different [Na]i values, Ca transient as in Figure 5 (but with 10% faster [Ca]i decline), and at [Ca]o=1.25 mmol/L (physiological free [Ca]o).17 The rate and integral of SR Ca pumping is also shown. For NF cells ([Na]i=8 to 10 mmol/L), Ca transport through NCX is similar to that shown in Figure 5, consistent with [Na]i in NF cells.10 In F myocytes (versus NF), there is more Ca influx through NCX at any [Na]i. This effect is more pronounced at the elevated [Na]i (≈12 mmol/L) in HF.10,11
In the F simulations at 12 mmol/L [Na]i, NCX barely extrudes the Ca that entered during the AP and at 1 second has only extruded 5 μmol/L of the Ca that enters through ICa (net extrusion does not reach 13 μmol/L cytosol until 1.9 seconds). At 1 Hz, this would cause cellular Ca gain, larger Ca transients, and consequently a greater Ca extrusion, until a new steady state is attained (where Ca influx and efflux are balanced during the cardiac cycle). Ca extrusion rate through NCX is relatively low during the AP versus SR Ca pump, but on repolarization the Ca flux rates are more comparable.
We quantitatively characterized NCX function during APs in human ventricular myocytes (F and NF) for the first time. Our major findings are the following. (1) Allosteric NCX activation by [Ca]i in F and NF myocytes occurs at physiological [Ca]i (Km=150 nmol/L). (2) [Ca]sm sensed by NCX during normal Ca transients exceeds bulk [Ca]i. (3) In NF myocytes, NCX extrudes Ca nearly throughout the cardiac cycle. (4) In F myocytes, substantial Ca enters the cell through NCX during the AP, as a result of lower [Ca]i, prolonged APD and higher [Na]i in HF.
Allosteric Regulation of Human NCX by [Ca]
Estimates of Ca affinity of the allosteric NCX site vary from KmCaAct of 22 to 125 nmol/L in intact myocytes18,16 to 300 to 600 nmol/L in giant excised patches.19 The higher excised patch values might reflect loss of some unknown NCX regulator. Although our protocol measures physiologically relevant NCX activation by Ca, KmCaAct could be slightly underestimated because we use [Ca]i versus [Ca]sm (but this is a minor issue here versus during SR Ca release). Nevertheless, we conclude that as [Ca]i increases in human myocytes, NCX activates much like ferret, canine,16 and rabbit NCX (Weber and Bers, unpublished observation, 2002), but unlike mouse NCX, which displays weak allosteric Ca regulation at physiological [Ca]i.16 Furthermore, we did not detect a difference in allosteric Ca activation between F and NF myocytes, consistent with unaltered [Ca]i dependence of INCX in Reference 7.
Allosteric Ca activation augments inward and outward INCX. As [Ca]i rises during an AP, Ca activation will activate NCX molecules regardless of the direction favored thermodynamically. This could enhance Ca influx early in the AP (and later in HF), but as [Ca]i (and [Ca]sm) increase, NCX switches direction to Ca efflux. At this time, inward INCX will be stimulated by allosteric activation, thermodynamic driving force, and higher substrate concentration. This accelerates Ca extrusion through NCX. As [Ca]i approaches diastolic levels, NCX partially deactivates, helping to prevent [Ca]i from going too low. This KmCaAct (≈150 nmol/L) may help to set diastolic [Ca]i near to but below the threshold for contractile and ryanodine receptor activation and allow NCX to optimally stabilize resting [Ca]i. This KmCaAct may also keep NCX partially active for the next beat.
[Ca]sm Sensed by NCX
During SR Ca release, inward INCX rises more rapidly than bulk [Ca]i.14,20,21 However, one cannot simply block INCX (or other currents) to measure physiologically relevant INCX because it is interdependent on ICa and SR Ca release. In our quantitative approach developed in rabbit ventricular myocytes,15 peak [Ca]sm was ≈3-fold higher than peak [Ca]i and occurred much earlier with respect to the AP upstroke (<30 versus 81 ms). From this [Ca]sm transient (and Em dependence of INCX), we inferred that INCX is inward for almost the entire normal rabbit cardiac cycle. This analysis reflects how average NCX molecules function, but individual NCX molecules that are closer to Ca entry or release sites would sense higher [Ca]sm. During SR Ca release, [Ca]cleft near ryanodine receptors may be >10 times higher than the average [Ca]sm sensed by NCX. NCX molecules nearest the cleft would sense higher [Ca]sm (and earlier), causing those NCX molecules to shift to Ca extrusion even earlier during the AP. Thus, the submembrane space here is not anatomically defined but is functionally defined by average NCX behavior.
In F and NF human ventricular myocytes (37°C), we found that even with unaltered NCX expression or properties,7 changes in [Ca]i, [Na]i, and Em characteristic of HF alter INCX during the AP (Figure 5). Our results with NF human and rabbit ventricular myocytes15 are similar. That is, the rapid [Ca]sm rise causes INCX to be almost entirely inward (extruding Ca) throughout the cardiac cycle. In contrast, INCX differs in F human myocytes, which exhibit a small early peak [Ca]i and slowly rising [Ca]i during the AP. The lower [Ca]sm during the AP plateau causes substantial Ca entry through INCX. This functional difference between F and NF human myocytes is further magnified when we account for the longer AP duration and elevated [Na]i seen in F versus NF myocytes (Figure 6). Previous simulations of INCX in HF (human and rabbit) suggest such an outward shift in INCX,11,22 but in this study we demonstrate this directly. Indeed, Despa et al11 compared how altered [Ca]i, [Na]i, and APD in rabbit HF are expected to influence Ca entry through INCX. Higher [Na]i was the most effective, but all would contribute in HF.
Not all F human myocytes show the prominent and typical phenotype in Figure 5B. About 50% of F cells exhibit various intermediate phenotypes between that in Figure 5B and the NF phenotype (although overall mean Ca transients were clearly smaller and slower in F).7 Thus, there is cellular heterogeneity, but in many F human ventricular myocytes there is clearly sufficient Ca entry through NCX during the AP to boost [Ca]i. This Ca entry may contribute directly to support contraction by enhancing Ca binding to myofilaments and limiting the reduction in SR Ca load. If this shift in NCX fluxes did not occur in HF, the degree of systolic dysfunction would be even worse.
Similar qualitative conclusions were made for a nonischemic rabbit HF model, but [Ca]sm was not measured.4,11 That is, reduced Ca transients, elevated [Na]i, and prolonged APD caused similar shifts in INCX during the AP. Thus, these conclusions have broad impact with respect to excitation-contraction coupling in HF (although rat and mouse models that lack AP plateau may differ).
We previously considered possible sources of error in this method,15,23 but these effects are small and occur very early during the AP and so do not affect our conclusions.
Balance of Ca Fluxes During the Cardiac Cycle
Many studies report elevated NCX expression in HF,2–5,24 which could partly compensate for lowered SR Ca pump activity5 in preserving relaxation rate and diastolic function. However, either reduced SR Ca pump function or increased Ca extrusion through NCX shift Ca from the SR out of the cell, unloading the SR.2–5,25,26 In the human hearts studied here, the [Ca]i dependence of INCX was unaltered in F versus NF, so the reduced SR Ca-pump function is more important in reducing SR Ca content.7 Notably, SR Ca content is reduced despite the shift toward outward INCX during the AP. Indeed, unaltered INCX properties do not preclude major alterations in NCX function. In F versus NF, lower [Ca]sm allows NCX-mediated Ca influx during the AP, but slowed [Ca]sm decline drives an increase in NCX-mediated Ca efflux during diastole (Figure 5D).
In F, there is less SR Ca release (due to lower SR Ca content) and greater Ca influx (due to less Ca-dependent ICa inactivation, longer APD, and Ca entry through INCX). Thus, there must also be less SR Ca uptake and greater Ca extrusion through NCX during the cardiac cycle. This must be true, regardless of details. Although most aspects of our quantitative analysis agree with former work, net extrusion of >30 μmol/L Ca during a cardiac cycle implies an equal amount of Ca entry, presumably through ICa. This is somewhat higher than expected for integrated ICa influx during the AP1 but is consistent with 25% of the 120 μmol/L cytosol of activator Ca crossing the sarcolemma. It is possible that we underestimate how completely NCX shuts off during diastole. Thus, although the present quantitative analysis is valuable, further refinement should be sought.
In conclusion, INCX reverses physiologically, making knowledge of [Ca]i, [Na]i, and Em essential in understanding INCX. Thermodynamic factors are as important as protein expression levels. In NF human myocytes, Ca extrusion through INCX occurs almost throughout the cardiac cycle, whereas in F myocytes, substantial Ca entry occurs through INCX during the AP. This Ca entry depends on lower peak [Ca]i, higher [Na]i, and prolonged APD in F myocytes. This may limit the systolic dysfunction observed in HF but could also slow relaxation and worsen diastolic dysfunction.
This study was supported by National Institutes of Health grants HL-30077 and HL-64098 (Dr Bers), HL-33921 and HL-61495 (Dr Houser), and American Heart Association predoctoral fellowships (Drs Weber and Piacentino). We thank Drs K.S. Ginsburg and K.B. Margulies for their help.
Drs Weber and Piacentino contributed equally to this work.
Bers DM. Excitation-Contraction Coupling and Cardiac Contractile Force. 2nd ed. Dordrecht/Boston: Kluwer Academic; 2001.
O’Rourke B, Kass DA, Tomaselli GF, et al. Mechanisms of altered excitation-contraction coupling in canine tachycardia-induced heart failure, I: experimental studies. Circ Res. 1999; 84: 562–570.
Pogwizd SM, Qi M, Yuan W, et al. Upregulation of Na+/Ca2+ exchanger expression and function in an arrhythmogenic rabbit model of heart failure. Circ Res. 1999; 85: 1009–1019.
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 beta-adrenergic responsiveness. Circ Res. 2001; 88: 1159–1167.
Hasenfuss G, Schillinger W, Lehnart SE, et al. Relationship between Na+-Ca2+-exchanger protein levels and diastolic function of failing human myocardium. Circulation. 1999; 99: 641–648.
Piacentino V III, Weber CR, Chen X, et al. Cellular basis of abnormal calcium transients of failing human ventricular myocytes. Circ Res. 2003; 92: 651–658.
Davies CH, Davia K, Bennett JG, et al. Reduced contraction and altered frequency response of isolated ventricular myocytes from patients with heart failure. Circulation. 1995; 92: 2540–2549.
Kääb S, Nuss HB, Chiamvimonvat N, et al. Ionic mechanism of action potential prolongation in ventricular myocytes from dogs with pacing-induced heart failure. Circ Res. 1996; 78: 262–273.
Pieske B, Maier LS, Piacentino V III, et al. Rate dependence of [Na+]i and contractility in nonfailing and failing human myocardium. Circulation. 2002; 106: 447–453.
Despa S, Islam MA, Weber CR, et al. Intracellular Na+ concentration is elevated in heart failure but Na/K pump function is unchanged. Circulation. 2002; 105: 2543–2548.
Beuckelmann DJ, Nèabauer M, Erdmann E, et al. Alterations of K+ currents in isolated human ventricular myocytes from patients with terminal heart failure. Circ Res. 1993; 73: 379–385.
Weber CR, Piacentino V III, Ginsburg KS, et al. Na+-Ca2+ exchange current and submembrane [Ca2+] during the cardiac action potential. Circ Res. 2002; 90: 182–189.
Weber CR, Ginsburg KS, Philipson KD, et al. Allosteric regulation of Na/Ca exchange current by cytosolic Ca in intact cardiac myocytes. J Gen Physiol. 2001; 117: 119–131.
Miura Y, Kimura J. Sodium-calcium exchange current: dependence on internal Ca and Na and competitive binding of external Na and Ca. J Gen Physiol. 1989; 93: 1129–1145.
Hilgemann DW, Collins A, Matsuoka S. Steady-state and dynamic properties of cardiac sodium-calcium exchange: secondary modulation by cytoplasmic calcium and ATP. J Gen Physiol. 1992; 100: 933–961.
Priebe L, Beuckelmann DJ. Simulation study of cellular electric properties in heart failure. Circ Res. 1998; 82: 1206–1223.
Weber CR, Ginsburg KS, Bers DM. Cardiac submembrane [Na] transients sensed by Na-Ca exchange current. Circ Res. 2003; 92: 950–952.
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.
Hobai IA, O’Rourke B. Decreased sarcoplasmic reticulum calcium content is responsible for defective excitation-contraction coupling in canine heart failure. Circulation. 2001; 103: 1577–1584.