Induced overexpression of Na⁺/Ca²⁺ exchanger transgene: altered myocyte contractility, [Ca²⁺]_i transients, SR Ca²⁺ contents, and action potential duration

Abstract

We have produced mice in which expression of the rat cardiac Na⁺/Ca²⁺ exchanger (NCX1) transgene was switched on when doxycycline was removed from the feed at 5 wk. At 8 to 10 wk, NCX1 expression in induced (Ind) mouse hearts was 2.5-fold higher but protein levels of sarco(endo)plasmic reticulum Ca²⁺-ATPase, α₁- and α₂-subunits of Na⁺-K⁺-ATPase, phospholamban, ryanodine receptor, calsequestrin, and unphosphorylated and phosphorylated phospholemman were unchanged compared with wild-type (WT) or noninduced (non-Ind) hearts. There was no cellular hypertrophy since WT, non-Ind, and Ind myocytes had similar whole cell membrane capacitance. In Ind myocytes, NCX1 current amplitude was ∼42% higher, L-type Ca²⁺ current amplitude was unchanged, and action potential duration was prolonged compared with WT or non-Ind myocytes. Contraction and intracellular Ca²⁺ concentration ([Ca²⁺]_i) transient amplitudes in Ind myocytes were lower at 0.6, not different at 1.8, and higher at 5.0 mM extracellular Ca²⁺ concentration ([Ca²⁺]_o) compared with WT or non-Ind myocytes. Despite similar Ca²⁺ current amplitude and sarcoplasmic reticulum (SR) Ca²⁺ uptake, SR Ca²⁺ content at 5.0 mM [Ca²⁺]_o was significantly higher in Ind compared with non-Ind myocytes, indicating that NCX1 directly contributed to SR Ca²⁺ loading. Echocardiography demonstrated that heart rate, left ventricular mass, ejection fraction, stroke volume, and cardiac output were similar among the three groups of animals. In vivo close-chest catheterization demonstrated similar contractility and relaxation among the three groups of mice, both at baseline and after stimulation with isoproterenol. We conclude that induced expression of NCX1 transgene resulted in altered [Ca²⁺]_i homeostasis, myocyte contractility, and action potential morphology. In addition, heart failure did not occur 3 to 5 wk after NCX1 transgene was induced to be expressed at levels found in diseased hearts.

ca²⁺ ions are centrally involved in excitation-contraction (EC) in cardiac muscle. Among the many transporters and ion channels involved in Ca²⁺ fluxes in cardiac myocytes, Na⁺/Ca²⁺ exchanger (NCX1) is unique in that it mediates both Ca²⁺ efflux (forward mode, 3 Na⁺ in and 1 Ca²⁺ out) and Ca²⁺ influx (reverse mode, 3 Na⁺ out and 1 Ca²⁺ in) during an EC cycle (4). At physiological extracellular Ca²⁺ concentration ([Ca²⁺]_o) and at rest, it is generally accepted that NCX1 primarily functions in the Ca²⁺ efflux mode. During systole, when the membrane potential (E_m) exceeds the equilibrium potential of NCX1 (E_NaCa), Ca²⁺ influx is favored (2, 38). The extent and duration of Ca²⁺ influx mediated by NCX1 is species dependent, likely due to differences in intracellular Na⁺ concentration ([Na⁺]_i) and action potential morphology among species (31, 38). For example, in rodent myocytes, the relatively high [Na⁺]_i (∼11 to 12 mM) compared with that present in rabbit myocytes (∼4.5 mM) (10, 31) would bias NCX1 to Ca²⁺ influx mode during the action potential most of the time (13, 38). Indeed, although NCX1 overexpression in isolated adult rabbit myocytes (by adenovirus-mediated gene transfer) results primarily in impaired myocyte contractility compared with control myocytes (25, 30), in adult rat myocytes NCX1 overexpression can result in either decreased, no change, or increased contraction and intracellular Ca²⁺ concentration ([Ca²⁺]_i) transient amplitudes, depending on prevailing [Ca²⁺]_o (45).

Over the last decade, Philipson and his colleagues have made major contributions toward our understanding of NCX1 function in vivo and in vitro by the generation of constitutively expressed NCX1 transgenic (1, 27, 28, 36, 40) and the cardiac-specific NCX1 knock-out (17, 22–24) mice. During the same period, it became increasingly clear that alterations in NCX1 expression and/or activity are associated with many models of cardiac hypertrophy and heart failure (for review, see Ref. 32). It remains controversial, however, whether the change in NCX1 expression and/or activity is a beneficial compensatory mechanism in response to contractile dysfunction or detrimental, leading to further deterioration of heart failure. One approach to differentiate between cause and compensatory response is the ability to switch on the expression of the NCX1 transgene concomitantly with the onset of the disease state. The purpose of the present study was to characterize a novel transgenic mouse model in which expression of the rat NCX1 transgene was under the control of a cardiac-specific promoter driving the expression of a tetracycline transactivator (tTA). When doxycycline (Dox) was removed from the feed, expression of NCX1 transgene was induced. Using this novel mouse model, we tested the hypotheses that 1) myocytes with induced NCX1 expression exhibited changes in Ca²⁺ homeostasis and contractility according to the thermodynamic driving force for the exchanger and 2) moderate NCX1 overexpression, by itself, did not cause contractile failure in vivo.

METHODS

Generation of inducible NCX1 transgenic mouse.

A cardiac-specific and inducible controlled vector composed of a modified mouse α-myosin heavy chain (MHC) minimal promoter fused with nucleotide binding sites for tTA (TREMHC) was used to provide robust expression when turned on in the absence of Dox and minimal leakage when turned off in the presence of Dox (29). Sequence-verified rat NCX1 gene (45) was cloned into TREMHC vector and microinjected into the nuclei of FVB mice for transgenic mice production. Eleven transgenic founders (7 males and 4 females) were generated. To quantify the number of transgenes inserted into the genome, genomic DNA from mouse-tail was isolated using the Qiagen DNAeasy kit. Genomic DNA (100 ng) was used for real-time PCR using NCX1-specific primer set (forward, CCCAATGTTTCAATGGGATT; and reverse, AGATGGGTCTTGGGGTTC) and GAPDH set (forward, AACGACCCCTTCATTGAC; and reverse, TCCACGACATACTCAGCAC). NCX1-specific primers were conserved between mouse and rat NCX1 and detected rat and mouse NCX1 gene with equal efficiency (data not shown). Real-time PCR was performed in a 20-μl reaction (100 ng of genomic DNA; 500 nM each primer; 1X SYBRE Green Master Mix). Each experimental sample was performed in triplicate. The ΔCT method was used to quantify the results, which are presented as relative fold changes to the GAPDH gene. Each primer set was designed to assume a melting temperature of 60°C and 50% GC content.

Since expression of the transgene required the presence of Dox-responsive transcription factor tTA, all founder mice were healthy and fertile. To determine phenotype, transgenic founders were crossed to the cardiac tTA transgenic mice in FVB background (MHC-tTA). Littermates that were heterozygous for tTA but negative for NCX1 transgene were used as wild-type (WT) controls. To prevent NCX1 expression, Dox (300 mg/kg mouse diet; Bio-Serv) was administered to all pregnant mothers and offsprings of WT and NCX1 transgenic groups. To induce NCX1 expression in adult mice, Dox was removed from the feed at 5 wk of age and mice were studied 3 to 5 wk later. The MHC-tTA mouse line expressed tTA at very low levels. Expression of tTA in WT mice did not affect mouse heart weight and function up to 12 wk. Similarly, Dox alone did not affect WT mouse heart size or function (data not shown).

Mice were housed and fed on a 12-h:12-h light-dark cycle at the Thomas Jefferson University Animal Facility and were supervised by veterinary staff members. Standard care was provided to all mice used for experiments. All protocols applied to the mice in this study were approved and supervised by the Institutional Animal Care and Use Committee at Thomas Jefferson University.

Echocardiographic and hemodynamic analysis of cardiac function.

Transthoracic two-dimensional echocardiography (TTE) was performed in anesthetized (2% inhaled isoflurane) WT, induced (Ind), and noninduced (non-Ind) mice with a 12-MHz probe (20). TTE in M mode was carried out in the parasternal short-axis to assess left ventricular (LV) diameter and function. For in vivo hemodynamic measurements, a 1.4 French micromanometer-tipped catheter (SPR-671; Millar Instruments) was inserted into the right carotid artery and then advanced into the LV of lightly anesthetized (tribromoethanol-amylene hydrate, Avertin; 2.5% wt/vol, 8 μl/g ip) mice with spontaneous respirations. Hemodynamic parameters including heart rate (beats per min⁻¹), LV end-diastolic pressure, and maximal first time derivative of LV pressure rise (+dP/dt) and fall (−dP/dt) were recorded in closed-chest mode, both at baseline and in response to progressive doses of isoproterenol (0.1, 0.5, 1, 5, and 10 ng).

Isolation of adult murine cardiac myocytes.

Cardiac myocytes were isolated from the septum and LV free wall of WT, Ind, and non-Ind mice (∼27 g) according to the protocol of Zhou et al. (47) and modified by our laboratory (35, 37). Briefly, mice were heparinized (1,500 u/kg ip) and anesthetized (pentobarbital sodium, 50 mg/kg ip). The heart was excised, mounted on a steel cannula, and retrograde perfused (100 cmH₂O, 37°C) with Ca²⁺-free bicarbonate buffer followed by enzymatic digestion (collagenases B and D, protease XIV). Isolated myocytes were plated on laminin-coated glass coverslips in a petri dish, and the Ca²⁺ concentration of the buffer was progressively increased from 0.05 to 0.125 to 0.25 to 0.5 mM in three steps (10-min interval each). The 0.5 mM Ca²⁺ buffer was then aspirated and replaced with minimal essential medium (MEM; Sigma M1018) containing 1.2 mM Ca²⁺, 2.5% fetal bovine serum (FBS), and antibiotics (1% penicillin-streptomycin). The pH was adjusted to 7.0 in 4% CO₂ by the addition of NaHCO₃ (0.57 g/l). After 1 h (4% CO₂, 37°), media was replaced with FBS-free MEM containing 0.1 mg/ml bovine serum albumin and antibiotics. Myocytes were used within 2 to 8 h of isolation.

Myocyte shortening measurements.

Myocytes adherent to coverslips were bathed in 0.6 ml of air- and temperature-equilibrated (37°C), HEPES-buffered (20 mM, pH 7.4) medium 199 containing 0.6, 1.8, or 5.0 mM [Ca²⁺]_o. Measurements of myocyte contraction (1 Hz) were performed as previously described (35, 37).

[Ca²⁺]_i transient measurements.

Myocytes were exposed to 0.67 μM of fura-2 AM for 15 min at 37°C. Fura-2-loaded myocytes were field stimulated to contract (1 Hz, 37°C) in medium 199 containing 0.6, 1.8, or 5.0 mM [Ca²⁺]_o. [Ca²⁺]_i transient measurements, daily calibration of fura-2 fluorescent signals, and [Ca²⁺]_i transient analyses were performed as previously described (35, 37).

Na⁺/Ca²⁺ exchange current measurements.

Whole cell patch-clamp recordings were performed at 30°C as previously described (35, 37, 42, 43). Briefly, fire-polished pipettes with resistances of 1–1.5 MΩ when filled with standard internal solution were used. For Na⁺/Ca²⁺ exchange current (I_NaCa) measurements, pipette solution contained (in mM) 100 Cs⁺-glutamate, 7.25 Na⁺-HEPES, 1 MgCl₂, 12.75 HEPES, 2.5 Na₂ATP, 10 EGTA, and 6 CaCl₂ (pH 7.2). Free Ca²⁺ in the pipette solution was 205 nM. Myocytes were bathed in an external solution containing (in mM) 130 NaCl, 5 CsCl, 1.2 MgSO₄, 1.2 NaH₂PO₄, 5 CaCl₂, 10 HEPES, 10 Na⁺-HEPES, and 10 glucose (pH 7.4) (35, 37, 42). Verapamil (1 μM), ouabain (1 mM), and niflumic acid (10 μM) were used to block L-type Ca²⁺ currents, Na⁺-K⁺-ATPase currents, and Cl⁻ currents, respectively. The myocyte was held at the calculated E_NaCa of −73 mV for at least 5 min before current was elicited with a descending-ascending voltage ramp (from +100 to −120 and back to +100 mV, 500 mV/s). I_NaCa, defined as the difference current in the absence and presence of NiCl₂ (1 mM), was normalized to whole cell capacitance (C_m) before comparisons. Our conditions for measuring I_NaCa were carefully chosen to minimize contamination by Na⁺-K⁺-ATPase activity (K⁺ free and the presence of ouabain) and ion fluxes through the NCX1 before the onset of voltage ramp (by holding the cell at the calculated E_NaCa), thereby allowing [Na⁺]_i and [Ca²⁺]_i to equilibrate with those present in the pipette solution.

L-type Ca²⁺ current measurements.

Pipette solution consisted of (in mM) 100 CsCl, 10 NaCl, 20 triethylammonium chloride, 10 HEPES, 5 MgATP, and 10 EGTA (pH 7.2 with CsOH). External solution contained (in mM) 137 NaCl, 5.4 CsCl, 1.8 CaCl₂, 1.3 MgSO₄, 1.2 NaH₂PO₄, 20 HEPES, 4 4-aminopyridine, and 15 glucose (pH 7.4 with NaOH). Before myocyte stimulation was started, holding potential was changed from −70 to −40 mV to inactivate fast inward Na⁺ current. To ensure steady-state sarcoplasmic reticulum (SR) Ca²⁺ loading, six conditioning pulses (from −40 to 0 mV, 300 ms, 1 Hz) were delivered before arrival of each test pulse (from −30 to +40 mV, 10-mV increments, 400 ms). After the last test pulse at +40 mV, the myocyte was held at −40 mV for 1 s before being returned to holding potential of −70 mV. Leak-subtracted inward currents were used in analysis for L-type Ca²⁺ current (I_Ca) amplitudes and inactivation kinetics. I_Ca was normalized to C_m before comparison among WT, Ind, and non-Ind myocytes.

I_Ca inactivation kinetics were determined by least-square fitting of a two-exponential function [A₁exp(-t/τ₁) + A₂exp(-t/τ₂) + C], where t is time; τ₁ and τ₂ are the fast and slow inactivation time constants, respectively; and A₁, A₂, and C are arbitrary constants. I_Ca data for voltage dependence of steady-state activation were fitted to Boltzmann distribution of the form I/I_max = 1/[1 + exp(V_h − V)/k] where I and I_max are current and maximal current, respectively; V was the test activating potential; V_h was the test potential eliciting one-half I_max; and k was the slope factor. Similarly, voltage dependence of steady-state inactivation of I_Ca was determined by a voltage prepulse protocol (43), and I_Ca data were fitted to a Boltzmann function of the form I/I_max = 1/[1 + exp(V − V_h)/k], where V is the prepulse potential and I_max, V_h, and k have their usual meanings.

Action potential measurements.

Action potentials were recorded using current-clamp configuration at 1.5× threshold stimulus and 4-ms duration (37, 46). Pipette solution consisted of (in mM) 125 KCl, 4 MgCl₂, 0.06 CaCl₂, 10 HEPES, 5 K⁺-EGTA, 3 Na₂ATP, and 5 Na₂-creatine phosphate (pH 7.2). External solution consisted of (in mM) 132 NaCl, 5.4 KCl, 1.8 CaCl₂, 1.8 MgCl₂, 0.6 NaH₂PO₄, 7.5 HEPES, 7.5 Na⁺-HEPES, and 5 glucose (pH 7.4).

SR Ca²⁺ content measurements.

SR Ca²⁺ content was estimated by integrating forward I_NaCa induced by caffeine exposure as described previously (35, 37, 41). The pipette solution consisted of (in mM) 100 Cs⁺-glutamate, 1 MgCl₂, 30 HEPES, and 2.5 MgATP (pH 7.2). The external solution contained (in mM) 130 NaCl, 5 CsCl, 1.2 MgSO₄, 1.2 NaH₂PO₄, 5 CaCl₂, 20 HEPES, and 10 glucose (pH 7.4; 30°C). Holding potential was −70 mV. At 200 ms after the 11th conditioning pulse (from −70 to 0 mV, 300 ms, 1 Hz), with E_m held at −70 mV, caffeine (5 mM, 2.4 s) was applied by puffer superfusion. The resulting inward current was digitized at 1 kHz and collected for 5 s. To convert I_NaCa time integral (coulombs) to moles, charge was divided by Faraday's constant of 96,487 coulombs/equivalent, based on 3 Na⁺ being exchanged for each Ca²⁺. SR Ca²⁺ content was normalized to cell size and given as femtomole per femtofared.

Immunoblotting.

Crude membranes from mouse left ventricles were prepared using a two-step centrifugation protocol described previously (34). For detection of NCX1, sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA2), calsequestrin, and α₁- and α₂-subunits of Na⁺-K⁺-ATPase, proteins in crude membranes were subjected to 7.5% SDS-PAGE under nonreducing (10 mM N-ethylmaleimide for NCX1) or reducing (5% β-mercaptoethanol for SERCA2, calsequestrin, and Na⁺-K⁺-ATPase) conditions. For detection of cardiac ryanodine receptor (RyR2) and phospholamban (PLB), proteins were subjected to 3–8% (Tris-acetate; reducing conditions) and 12% (Tris-glycine; both reducing and nonreducing conditions) SDS-PAGE, respectively. Primary antibodies used were as follows: for NCX1, π11-13 polyclonal antibody (1:500; Swant, Bellinzona, Switzerland); for SERCA2, MA3-919 antibody (1:2,500; Affinity BioReagents, Golden, CO); for RyR2, MA3-916 antibody (1:1,000; Affinity BioReagents); for PLB, MA3-922 antibody (1:500; Affinity BioReagents); for calsequestrin, rabbit anti-calsequestrin antibody (1:2,500; Swant); and polyclonal antibodies against either α₁- (1:1,000; Upstate USA, Charlottesville, VA) or α₂-subunit of Na⁺-K⁺-ATPase (1:1,200; Upstate USA). Secondary antibodies were donkey anti-rabbit or sheep anti-mouse IgG (Amersham). For phospholemman (PLM) immunoblotting, proteins in crude membranes were subjected to 12% SDS-PAGE under reducing conditions. Polyclonal C2 antibody (1:10,000) was used to detect the COOH terminus of endogenous mouse PLM (predominantly unphosphorylated form)(33). Polyclonal CP68 antibody (1:2,000) was used to detect PLM phosphorylated at serine68 (26). Immunoreactive proteins were detected with enhanced chemiluminescence Western blotting system. Protein band signal intensities were quantitated by scanning autoradiograms of the blots with a phosphorimager.

Statistics.

All results are expressed as means ± SE. For analysis of maximal contraction and [Ca²⁺]_i transient amplitudes as a function of group and [Ca²⁺]_o, +dP/dt and −dP/dt as a function of group and isoproterenol, I_NaCa, and I_Ca as a function of group and voltage, two-way ANOVA was used. For analysis of protein abundance, action potential parameters, C_m, echocardiographic data, I_Ca activation and inactivation parameters, one-way ANOVA was used. A commercial software package (JMP version 7; SAS Institute, Cary, NC) was used. In all analyses, P < 0.05 was taken to be statistically significant.

RESULTS

Inducible NCX1 transgenic mouse.

Eleven founders (7 males, 4 females) harboring the rat NCX1 transgene were generated. The transgenic mice were all heterozygous. Analysis of mouse genomic DNA showed that founder lines contained transgene copies that are 2 to 37 times above the endogenous mouse NCX1 gene copies (data not shown). We selected a transgenic line with 11 times more NCX1 transgene copies (line B) for expression studies. To control the timing of overexpression, we crossed the NCX1 transgenic founder with homozygous MHC-tTA mice in the presence of Dox. Mice that were heterozygous for both tTA and NCX1 transgenes were used as the experimental (Ind and non-Ind) groups, whereas littermates that were heterozygous for tTA only were used as WT controls.

We chose founder line B because preliminary experiments showed that 3 wk after induction, NCX1 protein levels were ∼2× that of WT hearts, assuming equal sensitivity of the polyclonal π11-13 antibody to detect rat and mouse NCX1. This magnitude of NCX1 increase in Ind hearts was similar to that found in models of heart failure (32). Another reason to choose to study a founder line with modest levels of NCX1 transgene expression is that high levels of transgenic expression of even seemingly innocuous proteins such as green fluorescent protein (18) or Cre-recombinase (6) can cause dilated cardiomyopathy.

Both Ind and non-Ind mice had no readily observable altered phenotype in that both body weights and LV mass were similar to those measured in WT mice (Table 1). Indexes of cardiac performance, as determined by TTE and in vivo catheterization, were also similar among the three groups of animals (Table 1). Contractility responses (+dP/dt) to isoproterenol stimulation were similar among the three groups of animals (Fig. 1; group × isoproterenol interaction effect, P < 0.42). For example, in response to maximal doses of isoproterenol, +dP/dt increased by 86.6 ± 18.0%, 107.9 ± 7.9%, and 114.4 ± 20.3% in WT (n = 6), Ind (n = 7), and non-Ind (n = 3) hearts, respectively. Likewise, there were no significant differences in relaxation (−dP/dt) in response to maximal doses of isoproterenol in WT (77.7 ± 24.5% increase), Ind (86.9 ± 19.7% increase), and non-Ind (84.7 ± 4.2% increase) mice (group × isoproterenol interaction effect, P < 0.29).

Table 1.

Characteristics of WT and mice with induced NCX1 transgene

	WT	Non-Ind	Induced
Body weight, g	27.4±0.6 (5)	26.2±0.4 (6)	26.6±0.7 (3)
LV mass, mg	53.0±2.6	53.5±2.0	56.7±2.0
Heart rate, beats/min	409±26	393±8	411±18
Ejection fraction, %	75.2±1.7	77.1±2.2	74.3±1.6
Fractional shortening, %	43.2±1.6	45.1±2.1	42.3±1.5
Stroke volume, μl	37.8±3.3	38.3±1.7	38.2±2.4
Cardiac output, ml/min	15.3±1.3	15.1±0.9	15.6±0.5
+dP/dt, mmHg/s	6,277±567 (7)	6,356±228 (3)	5,489±472 (7)
−dP/dt, mmHg/s	6,314±633	5,708±248	5,133±442

Values are means ± SE; numbers in parentheses are numbers of mice. NCX1, cardiac Na⁺/Ca²⁺ exchanger; WT, wild-type; Non-Ind, noninduced; +dP/dt and −dP/dt, maximal first time derivative of left ventricular (LV) pressure rise and fall, respectively.

Fig. 1.

Induced expression of Na⁺/Ca²⁺ exchanger transgene does not affect contractility response to isoproterenol in vivo. Doxycycline (Dox) was administered to all mice from gestation to day of experiment with the exception of induced (Ind) mice in which Dox was removed from the diet at 5 wk of age. Mice were studied at 8 to 10 wk of age. In vivo catheterization was performed in anesthetized mice, and maximal first time derivative of left ventricular pressure rise (+dP/dt) and fall (−dP/dt) and heart rate were continuously monitored, both at baseline and increasing doses of isoproterenol. To construct dose-response curve for isoproterenol, +dP/dt was normalized to the maximal value measured in each mouse. There are 7 wild-type (□), 7 induced (•), and 3 noninduced (◊) mice. Error bars are not shown if they fall within the boundaries of the symbol. Composite results are shown in Table 1.

To quantify the fold increase in NCX1 expression in Ind compared with WT or non-Ind hearts, we evaluated the reactivity of π11-13 antibody (raised against canine NCX1 but epitope is unknown) against the rat NCX1 transgene and native mouse NCX1. Crude LV membranes were prepared from Spraque-Dawley rats and FVB mice (n = 4 each) and subjected to Western blotting. With the assumption of an equal number of NCX1 molecules per milligram of LV membrane between rat and mouse, the signals from π11-13 antibody were 1.70× higher for mouse (2,949 ± 110) than rat (1,732 ± 51 arbitrary units) NCX1 (Fig. 2; P < 0.0001). With the use of the monoclonal antibody 6H2 raised against the NH₂ terminus of rabbit NCX1, the NCX1 signals from rat and mouse LV membranes were similar (Fig. 2), providing support for our assumption that NCX1 density was similar between mouse and rat LV membranes. Applying the correction factor of 1.7 suggests that protein levels of NCX1 in Ind hearts were ∼2.5-fold higher compared with WT or non-Ind hearts (Fig. 2 and Table 2). By contrast, expression of SERCA2, α₁- and α₂-subunits of Na⁺-K⁺-ATPase, RyR2, PLB, calsequestrin, and both unphosphorylated and phosphorylated forms of PLM was not altered by the induction of NCX1 transgene (Fig. 2 and Table 2). Ind myocytes did not undergo hypertrophy since C_m, a measure of cell surface membrane area, was similar among WT (154.8 ± 5.3 pF; n = 24), non-Ind (149.9 ± 7.5 pF; n = 17), and Ind (163.5 ± 9.0 pF; n = 18) myocytes.

Fig. 2.

Induced expression of Na⁺/Ca²⁺ exchanger transgene does not change the expression of selected proteins involved in excitation-contraction coupling. Top: crude membranes were prepared from wild-type (WT), induced, and noninduced (non-Ind) mouse left ventricle (LV) and subjected to SDS-PAGE followed by Western blot analysis. Primary antibodies (see methods) were used to detect cardiac Na⁺/Ca²⁺ exchanger (NCX1), α₁ (alpha1)- and α₂ (alpha2)- subunits of Na⁺-K⁺-ATPase, sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA2), unphosphorylated phospholemman (C2), phospholemman phosphorylated at serine68 (Cp68), cardiac ryanodine receptor (RyR2), high and low molecular weight phospholamban (PLB_H and PLB_L, respectively), and calsequestrin (CLSQ). PLB_H was detected under nonreducing conditions (only 1 band at ∼25 kDa), and PLB_L was detected under reducing conditions (2 major bands at ∼25 and ∼5 KDa; only the low molecular weight PLB is shown for PLB_L lanes). Composite results are shown in Table 2. Bottom: crude membranes prepared from rat and mouse LV were used to evaluate reactivity of polyclonal π11-13 antibody (1:500) against rat and mouse NCX1. To evaluate whether NCX1 density was similar between rat and mouse LV membranes, a monoclonal antibody (6H2; 1:1,000) raised against the NH₂ terminus of NCX1 was used.

Table 2.

Effects of induced NCX1 transgene expression on levels of selected proteins

	WT	Non-Ind	Induced
NCX1	28.4±0.8 (4)	27.1±2.5 (5)	54.2±3.7*(5)
SERCA2	89.2±1.8	85.3±2.2	81.4±2.4
α1, Na+-K+-ATPase	61.3±6.7	74.8±5.4	57.8±3.0
α2, Na+-K+-ATPase	42.8±2.7	49.9±3.0	49.7±0.8
PLM, unphosphorylated	16.8±2.5	13.9±3.8	11.0±1.4
PLM, phosphorylated	46.4±5.9	45.1±5.6	44.4±5.8
Calsequestrin	106.9±4.8	120.2±4.5	112.9±4.5
RyR2	14.9±2.3	11.0±2.8	14.9±0.4
PLBH	43.9±8.8	56.4±11.9	40.4±4.6
PLBL	31.8±4.6	44.6±3.3	35.0±7.0

Values are means ± SE; numbers in parentheses are numbers of hearts used in crude membrane preparations. SERCA2, sarco(endo)plasmic reticulum Ca²⁺-ATPase; PLM, phospholemman; RyR2, cardiac ryanodine receptor; PLB_H, high molecular weight phospholamban; PLB_L, low molecular weight phospholamban.

^*P < 0.0005 compared with WT or Non-Ind. Since the π11-13 signal is 1.7× stronger for mouse than rat NCX1 (Fig. 2), the corrected NCX1 protein level is ∼2.54× in Induced when compared with WT or Non-Ind myocytes.

Effects of induced NCX1 transgene expression on Na⁺/Ca²⁺ exchange activity.

Na⁺/Ca²⁺ exchange activity measured as NCX1 current (I_NaCa) in all three groups of myocytes increased as voltage became more positive (Fig. 3; voltage effect, P < 0.0001). I_NaCa was significantly (group effect, P < 0.0001) increased in Ind myocytes compared with WT or non-Ind myocytes (Fig. 3). Our conditions were biased toward measuring outward (3 Na⁺ out and 1 Ca²⁺ in) rather than inward (3 Na⁺ in and 1 Ca²⁺ out) I_NaCa. This accounts for the significant group × voltage interaction effects (P < 0.0001), indicating that differences in I_NaCa magnitude between Ind and WT or non-Ind myocytes increased with positive voltages. At +100 mV, I_NaCa was ∼40% higher in Ind compared with WT or non-Ind myocytes. There were no differences in I_NaCa between WT and non-Ind myocytes (group effect, P < 0.48; group × voltage interaction effect, P < 0.23).

Fig. 3.

Induced expression of Na⁺/Ca²⁺ exchanger transgene increases Na⁺/Ca²⁺ exchange current (I_NaCa). I_NaCa was measured in WT, induced (Ind), and non-Ind myocytes at 30°C and 5.0 mM extracellular Ca²⁺ concentration ([Ca²⁺]_o; see methods). Current-voltage relationships of I_NaCa (means ± SE) from WT (◊; n = 10), Ind (□; n = 7), and non-Ind (○; n = 8) myocytes are shown. The reversal potential of I_NaCa in all myocytes examined was ∼−60 mV. Error bars are not shown if they fall within the boundaries of the symbol.

Effects of induced NCX1 transgene expression on myocyte contraction.

In all three groups of myocytes, elevating [Ca²⁺]_o resulted in the expected increase in contraction amplitudes (Fig. 4 and Table 3; [Ca²⁺]_o effect, P < 0.0001). However, the dynamic range in response to increasing [Ca²⁺]_o was significantly increased in Ind myocytes. Specifically, at 0.6 mM [Ca²⁺]_o, Ind myocytes shortened less than WT or non-Ind myocytes. By contrast, at 5.0 mM [Ca²⁺]_o, Ind myocytes contracted more than WT or non-Ind myocytes. At 1.8 mM [Ca²⁺]_o, contraction amplitudes were similar among WT, Ind, and non-Ind myocytes. These conclusions are supported by highly significant (P < 0.0001) group × [Ca²⁺]_o interaction effects, indicating that the magnitude and/or direction of the effects of changing [Ca²⁺]_o on cell shortening were different across the experimental groups.

Fig. 4.

Induced expression of Na⁺/Ca²⁺ exchanger transgene alters myocyte contractility. Myocytes isolated from LV of WT, noninduced, and induced mice were paced (1 Hz) to contract at 37°C and [Ca²⁺]_o of 0.6 (top), 1.8 (middle), and 5.0 mM (bottom). Steady-state twitches are shown. Results are summarized in Table 3.

Table 3.

Effects of induced NCX1 transgene expression on myocyte shortening

[Ca2+]o	WT	Non-Ind	Induced
	Maximal contraction amplitude, % resting cell length
0.6	3.18±0.28 (16)	3.28±0.38 (16)	2.12±0.18*(16)
1.8	5.42±0.32 (20)	5.69±0.55 (11)	5.71±0.28 (18)
5.0	9.60±0.37 (21)	9.95±0.43 (11)	11.98±0.54*(20)
	Maximal shortening velocity, cell length/s
0.6	0.44±0.05	0.60±0.07	0.32±0.04*
1.8	0.82±0.05	0.93±0.10	0.98±0.04
5.0	1.41±0.08	1.54±0.10	1.78±0.10*
	Maximal relengthening velocity, cell length/s
0.6	0.27±0.04	0.47±0.06	0.15±0.02*
1.8	0.65±0.06	0.76±0.11	0.72±0.05
5.0	1.16±0.08	1.26±0.11	1.44±0.06*

Values are means ± SE; numbers in parentheses are numbers of myocytes, without regard to the number of cells contributed by each heart (n = 5, 4, and 5 hearts for WT, Non-Ind, and Induced, respectively). [Ca²⁺]_o, extracellular Ca²⁺ concentration.

^*P < 0.007 (group × [Ca²⁺]_o interaction effects), Induced compared with WT or Non-Ind.

Both maximal shortening and relengthening velocities in Ind myocytes were significantly slower at 0.6 and faster at 5.0 mM [Ca²⁺]_o, when compared with those measured in WT and non-Ind myocytes (Table 3; group × [Ca²⁺]_o interaction effects, P < 0.007). There were no group × [Ca²⁺]_o interaction effects in contraction amplitude, maximal shortening, and relaxation velocities when comparing WT and non-Ind myocytes.

Effects of induced NCX1 transgene expression on [Ca²⁺]_i transients and SR Ca²⁺ uptake.

Differences in myocyte contractility between Ind and WT or non-Ind myocytes may be due to changes in [Ca²⁺]_i associated with induced NCX1 overexpression. When compared with WT or non-Ind myocytes, systolic [Ca²⁺]_i and [Ca²⁺]_i transient amplitudes in Ind myocytes were lower at 0.6, similar at 1.8, and higher at 5.0 mM [Ca²⁺]_o (Fig. 5 and Table 4; group × [Ca²⁺]_o interaction effects, P < 0.004). There were no differences in diastolic [Ca²⁺]_i among the three groups of myocytes.

Fig. 5.

Induced expression of Na⁺/Ca²⁺ exchanger transgene modulates intracellular Ca²⁺ concentration ([Ca²⁺]_i) transients. Myocytes isolated from WT, noninduced, and induced murine LV were loaded with the Ca²⁺ indicator fura-2 and paced (1 Hz) to contract at 37°C and [Ca²⁺]_o of 0.6 (top), 1.8 (middle), and 5.0 mM (bottom). Results are summarized in Table 4.

Table 4.

Effects of induced NCX1 transgene expression on [Ca²⁺]_i transients, SR Ca²⁺ uptake, and SR Ca²⁺ contents

[Ca2+]o	WT	Non-Ind	Induced
	Systolic [Ca2+]i, nM
0.6	143±5 (8)	144±7 (11)	124±8*(7)
1.8	165±7 (16)	161±8 (13)	157±7 (21)
5.0	231±9 (19)	233±10 (13)	285±13*(21)
	Diastolic [Ca2+]i, nM
0.6	105±4	101±6	96±6
1.8	91±5	87±4	83±4
5.0	98±5	100±4	100±3
	[Ca2+]i transient amplitude, % increase in fura-2 signal
0.6	8.5±0.8	7.8±0.6	5.7±0.6*
1.8	17.0±1.0	17.3±1.1	17.1±1.0
5.0	28.0±1.2	28.2±1.2	37.2±2.0*
	t1/2 of [Ca2+]i transient decline, ms
1.8	213±13	215±9	206±10
	SR Ca2+ content, fmol/fF
5.0	ND	12.5±1.4 (12)	16.9±1.2† (10)

Values are means ± SE; numbers in parentheses are numbers of myocytes, without regard to the number of cells contributed by each heart [for intracellular Ca²⁺ concentration ([Ca²⁺]_i) transient measurements, n = 4 hearts each for WT, Non-Ind, and Induced; for sarcoplasmic reticulum (SR) Ca²⁺ content measurements, n = 2 and 4 hearts for Non-Ind and Induced, respectively]. ND, not done; t_1/2, half-time.

^*P < 0.004 (group × [Ca²⁺]_o interaction effects), Induced vs. WT or Non-Ind;

^†P < 0.03, Induced vs. Non-Ind.

SR Ca²⁺ uptake activity can be estimated from the half-time (t_1/2) of [Ca²⁺]_i transient decline (44). Since the kinetics of [Ca²⁺]_i transient decline is dependent on peak [Ca²⁺]_i (5), we focused on t_1/2 of [Ca²⁺]_i transient decline measured at 1.8 mM [Ca²⁺]_o at which there were no differences in either systolic [Ca²⁺]_i or [Ca²⁺]_i transient amplitudes among the three groups of myocytes. There were no differences in t_1/2 of [Ca²⁺]_i transient decline among WT, non-Ind, and Ind myocytes (Table 4), indicating that induced expression of NCX1 transgene had no discernible effects on SR Ca²⁺ uptake.

Effects of induced NCX1 transgene expression on I_Ca.

Altered contractility and [Ca²⁺]_i transients in Ind myocytes may partly be due to changes in I_Ca and its kinetics. In addition, in homozygous mice constitutively overexpressing NCX1 transgene, I_Ca was significantly increased (27), whereas in cardiac-specific NCX1 knockout mice, I_Ca was substantially decreased compared with that of control WT mice (24). Therefore, we measured I_Ca in all three groups of myocytes. Our initial attempts to measure I_Ca in FVB mouse myocytes using pipette and extracellular solutions previously designed for adult rat LV myocytes (43) and successfully applied to C57BL/6 mouse myocytes (37) were failures. For FVB mouse myocytes, we obtained consistent I_Ca results after we adopted the pipette and bathing solutions published by Zhou et al. (47) and detailed in methods.

Maximal I_Ca density and current-voltage relationships were not different among WT (−7.53 ± 0.88 pA/pF; n = 14), non-Ind (−7.38 ± 0.91 pA/pF; n = 9), and Ind (−7.25 ± 0.32 pA/pF; n = 11) myocytes (Fig. 6). The test potential at which maximal I_Ca occurred was ∼0 mV for all three groups of myocytes (Fig. 6). Neither the fast (τ₁) nor the slow (τ₂) inactivation time constant measured at 0 mV was significantly (P < 0.07) different in Ind (τ₁, 13.6 ± 1.9 ms; and τ₂, 49.3 ± 6.6 ms) compared with WT (τ₁, 15.4 ± 1.5 ms; and τ₂, 72.7 ± 7.7 ms) or non-Ind myocytes (τ₁, 13.7 ± 1.5 ms; and τ₂, 76.2 ± 10.9 ms).

Fig. 6.

Induced expression of Na⁺/Ca²⁺ exchanger transgene has no effect on L-type Ca²⁺ current (I_Ca) density. I_Ca was measured in WT (○; n = 14), noninduced (□; n = 9), and induced (•; n = 11) myocytes at 30°C and 1.8 mM [Ca²⁺]_o. Current density of peak I_Ca (means ± SE) at each test potential (−40 to + 40 mV) is shown. Error bars are not shown if they fall within the boundaries of a symbol.

To further evaluate voltage-dependent properties of I_Ca, steady-state inactivation of I_Ca was determined using the classical two-pulse protocol (Fig. 7). The sigmoidal relationships between the extent of inactivation (f_∞ = I/I_max) and prepulse potential were superimposable for all three groups of myocytes (Fig. 7). The test potential eliciting one-half maximal I_Ca, V_h, was not different among WT (−23.8 ± 0.5 mV; n = 8), non-Ind (−24.2 ± 0.5 mV; n = 6), and Ind (−23.9 ± 0.5 mV; n = 6) myocytes. Likewise, the slope factor k was similar among WT (5.1 ± 0.4 mV), non-Ind (5.0 ± 0.4 mV), and Ind (5.0 ± 0.4 mV) myocytes.

Fig. 7.

Induced expression of Na⁺/Ca²⁺ exchanger transgene does not affect voltage dependence of activation and inactivation of I_Ca. Top left: family of leak-subtracted I_Ca from a WT myocyte at 1.8 mM [Ca²⁺]_o and 30°C (see methods). Bottom left: voltage dependence of activation (d∞) of I_Ca in WT (○, red line; n = 14), induced (◊, green line; n = 11), and noninduced (▵, blue line; n = 9) myocytes. Top right: family of leak-subtracted I_Ca from another WT myocyte at 1.8 mM [Ca²⁺]_o and 30°C. Myocyte was held at −70 mV and switched to −40 mV immediately before prepulse (−30 to +20 mV, 10-mV increments, 300 ms), after which myocyte was returned to −40 mV for 3 ms before test pulse (−40 to 0 mV, 300 ms). Bottom right: voltage dependence of inactivation (f∞) of I_Ca in WT (○, red line; n = 8), induced (◊, green line; n = 6), and noninduced (▵, blue line; n = 6) myocytes.

Steady-state voltage dependence of I_Ca activation (d_∞ = I/I_max) was also similar among the three groups of myocytes (Fig. 7). V_h was −20.2 ± 1.8 mV, −18.2 ± 2.1 mV, and −20.2 ± 2.0 mV in WT (n = 14), non-Ind (n = 9), and Ind (n = 11) myocytes, respectively. The slope factor k was 5.9 ± 1.8 mV, 7.2 ± 1.8 mV, and 7.0 ± 1.8 mV in WT, non-Ind, and Ind myocytes, respectively.

Effects of induced NCX1 transgene expression on SR Ca²⁺ content.

Altered twitch and [Ca²⁺]_i transient amplitudes are likely related to changes in SR Ca²⁺ contents. Despite no significant differences in SERCA2, PLB and RyR2 expression (Fig. 2 and Table 2), and SR Ca²⁺ uptake activity (Table 4), SR Ca²⁺ content determined at 5.0 mM [Ca²⁺]_o was significantly (P < 0.03) higher in Ind compared with non-Ind myocytes (Table 4). Given that I_Ca is not affected by induced NCX1 overexpression (Figs. 6 and 7), our findings suggest that NCX1 contributed directly to SR Ca²⁺ loading.

Effects of induced NCX1 transgene expression on action potential.

Increased NCX1 activity may alter action potential morphology (2, 31). We therefore measured action potential in WT, non-Ind, and Ind myocytes (Fig. 8). When compared with that of WT myocytes, resting E_m was significantly (P < 0.01) higher in non-Ind but not in Ind myocytes (Table 5). Action potential amplitude was lower in WT compared with either non-Ind (P < 0.008) or Ind (P < 0.035) myocytes. The most dramatic finding is the prolongation of action potential duration (APD) at 50% (APD₅₀) and 90% (APD₉₀) repolarization (P < 0.005) in Ind myocytes compared with either WT or non-Ind myocytes.

Fig. 8.

Induced expression of Na⁺/Ca²⁺ exchanger transgene prolongs action potential duration. Action potentials were measured in WT (solid line), noninduced (dashed line), and induced (dotted line) myocytes at 30°C and 1.8 mM [Ca²⁺]_o (see methods). Composite data are presented in Table 5.

Table 5.

Effects of induced NCX1 transgene expression on AP

	WT	Non-Ind	Induced
Resting Em, mV	−69.3±1.3 (15)	−79.2±1.5† (14)	−72.5±1.9 (17)
AP amplitude, mV	110.5±2.4	120.0±2.1†	120.5±3.6
APD50, ms	4.53±0.76	3.29±0.22	8.85±0.78*
APD90, ms	24.31±1.80	22.73±1.47	55.93±9.52*

Values are means ± SE; numbers in parentheses are numbers of myocytes. E_m, membrane potential; AP, action potential; APD₅₀ and APD₉₀, AP duration at 50% and 90% repolarization, respectively. Cells were paced at 1 Hz.

^†P < 0.01, WT vs. Non-Ind;

^*P < 0.003, Induced vs. WT or Non-Ind.

DISCUSSION

The first major finding is that we have engineered mice with inducible cardiac-specific expression of the rat cardiac NCX1. There are many similarities and some notable differences when comparing our current inducible mouse model with previously published mouse models that constitutively overexpress the canine NCX1 transgene (1, 27, 28, 36, 40). For comparison purposes, we would focus mainly on the heterozygous (Het) mice constitutively overexpressing NCX1 transgene (1, 28, 36, 40) since our inducible mouse model is also heterozygous for the NCX1 transgene. We would also discuss relevant features of the homozygous (Hom) mice constitutively overexpressing the canine NCX1 transgene (27, 28).

To achieve controlled expression of a transgene, investigators utilized a cardiac-specific promoter driving the expression of a tTA or of a reverse tTA in which four point mutations converted the system from a Tet-off (Dox is needed to keep transgene expression silent) to a Tet-on (transgene expression is active only in the presence of the drug) system (14). Unfortunately, early experience with these systems was not favorable because of heterogeneous expression throughout the ventricular and atrial cardiomyocyte populations, the potential for context-dependent expression because of the use of a virus-derived sequence, and leakiness or an inability to activate the gene in the presence or absence of Dox, respectively. The TREMHC vector we used allows robust expression when turned on in the absence of Dox and minimal leakage when turned off in the presence of Dox (29). Minimal leakage is supported by the observation of similar I_NaCa and NCX1 protein levels between WT and non-Ind myocytes, and inducible transgene expression is evidenced by increases in both I_NaCa amplitude and NCX1 protein levels in Ind myocytes compared with WT or non-Ind myocytes.

Similar to Het mice constitutively overexpressing NCX1 examined at 4 mo (28), mice with induced NCX1 overexpression (at 2 to 2.5 mo) had no differences in body weights, LV mass, heart rate, ejection fraction, fractional shortening, and baseline +dP/dt and −dP/dt compared with WT or non-Ind mice. The major difference between constitutive and induced NCX1 expression is that unlike HET mice constitutively overexpressing NCX1 that exhibited reduced contractility (+dP/dt) and relaxation (−dP/dt) in response to increasing doses of dobutamine compared with WT littermates (28), mice with induced NCX1 overexpression demonstrated similar inotropic and lusitropic responses to increasing doses of isoproterenol compared with WT or non-Ind controls. It is noteworthy that Hom mice constitutively overexpressing NCX1 developed cardiac hypertrophy, impaired LV fractional shortening, and markedly reduced contractility and relaxation in response to dobutamine compared with WT littermates (28).

Before comparing myocyte characteristics between Ind and Het mice, it is pertinent to point out that although myocytes were isolated from Ind mice at 8 to 10 wk of age, the ages of Het mice from which myocytes were derived were not specified (1, 36, 40). At the myocyte level, Na⁺-dependent Ca²⁺ uptake in sarcolemmal vesicles isolated from adult Het mice constitutively overexpressing NCX1 is 148% (1) or 2.34× (28) higher and I_NaCa (at +70 mV) is ∼80% larger (36) compared with WT controls. Our results in myocytes with induced NCX1 expression are comparable in that the NCX1 protein was 2.5-fold higher and I_NaCa was ∼40% larger compared with WT or non-Ind myocytes. In an attempt to account for the differences between NCX1 protein expression and activity increase in Ind myocytes, we measured expression of PLM, which is a known endogenous inhibitor of NCX1 (7). Neither the expression level nor fractional phosphorylation of PLM was altered by induced NCX1 overexpression. The slight variability in increases in NCX1 activity between Ind and Het myocytes is most likely due to different techniques to estimate NCX1 activity and the conditions employed in measuring I_NaCa. Similar to Het myocytes constitutively overexpressing NCX1 (28, 36), there were no alterations in the expression of SERCA2, PLB, calsequestrin, and Na⁺-K⁺-ATPase in myocytes with induced NCX1 overexpression. Similar to Ind myocytes overexpressing rat NCX1, Het myocytes constitutively overexpressing canine NCX1 also had no differences in I_Ca (1, 40) compared with WT myocytes.

In contrast with myocytes constitutively overexpressing NCX1 transgene in which absolute cell shortening magnitude (1.0 mM [Ca²⁺]_o and 0.5 Hz) was reported to be similar to that observed in WT myocytes (36), we found altered contractility in Ind myocytes depending on the prevailing [Ca²⁺]_o, which is known to be a relatively strong determinant of the exchanger equilibrium. At low [Ca²⁺]_o (conditions that promote Ca²⁺ efflux), myocytes with induced NCX1 overexpression contracted less, whereas at high [Ca²⁺]_o (conditions that promote Ca²⁺ influx), myocytes with induced NCX1 overexpression contracted more vigorously compared with their WT or non-Ind controls. This increased dynamic range in response to increased [Ca²⁺]_o is characteristic of adult rat myocytes in which NCX1 was overexpressed by adenovirus-mediated gene transfer (45) and mouse myocytes in which NCX1 activity was increased by genetically eliminating PLM (37). The results from these three fundamentally different approaches [in vitro NCX1 gene transfer (45), induced NCX1 transgene expression in vivo (present study), and relief of NCX1 inhibition by PLM knockout (37)] support our conclusion that alterations in NCX1 expression and/or activity can significantly impact myocyte contractile behavior.

Changes in NCX1 activity in NCX1 overexpressed myocytes would be expected to alter Ca²⁺ fluxes during EC coupling. At intermediate [Ca²⁺]_o (1 to 1.8 mM), systolic and diastolic [Ca²⁺]_i, as well as [Ca²⁺]_i transient amplitude in Het myocytes with either constitutive (1, 36, 40) or induced NCX1 overexpression, were not different than those measured in WT or non-Ind myocytes. When [Ca²⁺]_o was increased to promote Ca²⁺ influx via NCX1, Ind myocytes exhibited higher [Ca²⁺]_i transient amplitudes. Larger [Ca²⁺]_i transient amplitudes in Ind myocytes studied under elevated [Ca²⁺]_o conditions are most likely the result of increased SR Ca²⁺ contents due to enhanced NCX1 activity (35, 37, 45). Indeed, this is confirmed by increased SR Ca²⁺ contents in Ind myocytes compared with non-Ind myocytes. Conversely, when myocytes are studied under conditions that promote Ca²⁺ efflux (0.6 mM [Ca²⁺]_o), increasing NCX1 expression (45) or enhancing NCX1 activity by genetically eliminating its endogenous inhibitor PLM (37, 42) results in lower [Ca²⁺]_i transient and contraction amplitudes, as well as SR Ca²⁺ contents. It is the unique ability of NCX1 to drive Ca²⁺ in and pump Ca²⁺ out during an EC cycle, depending on the prevailing thermodynamic driving force, that explains the differential effects of low versus high [Ca²⁺]_o on [Ca²⁺]_i transient and contraction amplitudes and SR Ca²⁺ contents in myocytes in which either NCX1 expression or activity is altered.

In cardiac myocytes, NCX1 competes with SERCA2 for Ca²⁺ during decline of the [Ca²⁺]_i transient. Therefore, theoretically, rate of [Ca²⁺]_i transient decline would be expected to increase with NCX1 overexpression. In Het myocytes, Terracciano et al. (36) reported shorter t₅₀ of [Ca²⁺]_i transient decline, but this occurs even in the absence of Na⁺ and Ca²⁺, suggesting faster SR Ca²⁺ uptake. Yao et al. (40) reported faster time constant (τ) of [Ca²⁺]_i transient decline in Het myocytes; the most prominent difference is in the terminal phase (t_50–10%) of [Ca²⁺]_i transient decline. In Ind myocytes, we did not detect differences in the early phase (t_1/2) of [Ca²⁺]_i transient decline compared with WT or non-Ind myocytes. The slight discrepancy in results may be due to differences in temperature (37° vs. 22° or 25°C) at which experiments were performed and the phase of [Ca²⁺]_i transient decline (early vs. late) that was evaluated. It is also useful to recall that unlike rabbit myocytes in which NCX1 contributes some 28% of Ca²⁺ transport, the fraction of Ca²⁺ transported by NCX1 is only 7% in rodent myocytes (3). Indeed, NCX1 has been demonstrated to play only a minor role in relaxation in mouse myocytes compared with other species (36). Even in rabbit myocytes in which NCX1 plays a much bigger role in [Ca²⁺]_i transient decline (3), Yao et al. (39) showed that abrupt inhibition of NCX1 does not affect the initial rate of decline of the [Ca²⁺]_i transient. The weight of current evidence suggests that in rodent myocytes, the effects of NCX1 on [Ca²⁺]_i transient decline are primarily manifest during the later rather than the early phase.

SR Ca²⁺ content is determined by Ca²⁺ influx via L-type Ca²⁺ channels and NCX1 operating in the reverse mode, SR Ca²⁺ uptake activity, SR Ca²⁺ release, SR volume, calsequestrin content, and competition between SR Ca²⁺ uptake and NCX1 operating in the forward mode. There were no differences in cell sizes (and presumably SR volume), I_Ca parameters, SERCA2, PLB, calsequestrin, and RyR2 protein levels, and SR Ca²⁺ uptake activity among WT, non-Ind, and Ind myocytes. Therefore, increased SR Ca²⁺ content in Ind myocytes cannot be ascribed to changes in SR transport proteins or SR Ca²⁺ uptake activity. Theoretically, prolonged APD in Ind myocytes may allow more Ca²⁺ to enter via L-type Ca²⁺ channels, thereby partly explaining the increased SR Ca²⁺ content. We think that this explanation is less likely. First, increased Ca²⁺ flux through I_Ca due to prolonged APD cannot account for the decreased SR Ca²⁺ content at 0.6 mM [Ca²⁺]_o in myocytes in which NCX1 is either overexpressed (45) or its activity increased (37). Second, APD prolongation is most dramatic at the tail-end of the action potential (APD₉₀), at which E_m (−40 mV) I_Ca would be minimally activated. Finally, even with prolongation of early APD (APD₅₀), Ca²⁺-induced inactivation of I_Ca would be expected to limit the amount of Ca²⁺ that can enter. Therefore, the most consistent and unifying hypothesis for altered SR Ca²⁺ contents in NCX1 overexpressed myocytes is due to enhanced reverse or forward NCX1 activity, depending on prevailing [Ca²⁺]_o and hence the thermodynamic driving force of the exchanger.

The thermodynamic driving force for NCX1 is given by E_m − E_NaCa; the latter is determined by [Na⁺]_i, [Ca²⁺]_i, [Ca²⁺]_o, and [Na⁺]_o. In our experiments with intact myocytes, [Ca²⁺]_o and [Na⁺]_o were controlled while [Ca²⁺]_i and action potential were measured. Theoretically, changes in [Na⁺]_i brought on by NCX1 overexpression in Ind myocytes can alter NCX1 activity. Although we did not measure [Na⁺]_i, some insights can be gleaned from [Na⁺]_i measurements in resting and field-stimulated PLM knockout myocytes. In PLM knockout myocytes in which I_NaCa is ∼22% higher than WT myocytes (42), resting [Na⁺]_i (12.0 ± 1.5 mM) is not different than that measured in WT myocytes (12.5 ± 1.8 mM)(9). When field-stimulated to contract at 2 Hz, steady-state [Na⁺]_i in PLM knockout myocytes is 17.0 ± 1.5 mM and not significantly different than that measured in WT myocytes (15.2 ± 1.5 mM)(11). Since the increase in I_NaCa is ∼40% in Ind myocytes, comparable with the ∼22% increase in PLM knockout myocytes (as compared with their respective controls), changes in [Na⁺]_i due to moderately increased NCX1 activity, if any, would be small.

Another major finding is that when compared with non-Ind myocytes, both APD₅₀ and APD₉₀ were dramatically prolonged in Ind myocytes, despite no changes in resting E_m and action potential amplitude. Prolongation of action potential was not due to changes in I_Ca since neither the current density nor the inactivation time constants were affected by induced NCX1 transgene expression. Altered NCX1 activity may theoretically affect APD (2). In rat and mouse cardiac myocytes with short APD, there is a net Ca²⁺ efflux via NCX1 at the shoulder of the action potential. This is because the action potential is already repolarizing before the [Ca²⁺]_i transient reaches its peak, and I_NaCa is predominantly inward during this phase of the action potential (31). Increase in NCX1 activity in Ind myocytes would therefore be expected to prolong APD, which is what we observed. In Het myocytes constitutively overexpressing NCX1, Yao et al. (40) reported decreased action potential amplitude and abbreviated APD₅₀ but prolonged APD₉₀ compared with WT myocytes. These authors hypothesized that decreased action potential amplitude and shortening of APD₅₀ were due to increased outward current secondary to enhanced reverse NCX1 (3 Na⁺ out and 1 Ca²⁺ in) activity during early depolarization, whereas prolonged APD₉₀ was due to more marked forward NCX1 (3 Na⁺ in and 1 Ca²⁺ out) activity during the [Ca²⁺]_i transient. The differences between our action potential findings in Ind myocytes and Het myocytes constitutively overexpressing NCX1 are not clear but may be due to different genetic backgrounds (FVB vs. C57Bl/6xC3H1), animal models (induced vs. constitutive), pipette solutions (Ca²⁺ buffered vs. not buffered), temperature (30° vs. 25°C), and potential differences in other repolarizing currents such as transient outward current or other K⁺ currents. We do not know the reason for the observed differences in resting E_m and action potential amplitude between WT and non-Ind myocytes. It is not due to Dox since it was administered to both WT and non-Ind mice from gestation up to the day of experiment.

In some, but not all, models of heart failure and hypertrophy, NCX1 expression has been reported to be increased (32) and hypothesized to be a major contributor to contractile dysfunction (19, 21). In heart failure, expression of many proteins involved in cardiac Ca²⁺ homeostasis such as SERCA2, PLB, RyR2, NCX1, etc., is known to be altered (8, 16). Our current observations clearly indicated that induction of NCX1 overexpression in adult mice (to simulate switching on NCX1 expression with onset of disease state), without concomitant changes in other proteins involved in Ca²⁺ transport, did not result in contractile dysfunction in vivo. In this light, it is interesting to note that NCX1 has been shown to significantly contribute to the maintenance of contractile function and [Ca²⁺]_i transients in failing human ventricular myocytes (12, 15). Our novel model of inducible NCX1 transgenic expression is eminently suited for studies to dissect whether overexpression of NCX1 found in heart failure models is beneficial or harmful to the animal.

In summary, we have generated a novel mouse model in which cardiac-specific expression of rat NCX1 can be controlled. Despite no detectable differences in cardiac performance in vivo, LV myocytes with induced expression of NCX1 transgene exhibited prolonged APD, enhanced contractility, [Ca²⁺]_i transient amplitude, and SR Ca²⁺ content at high [Ca²⁺]_o, but no changes in I_Ca, SERCA2, RyR2, PLB, Na⁺-K⁺-ATPase, phospholemman, calsequestrin, and SR Ca²⁺ uptake. We conclude that NCX1 operates in both forward and reverse modes during an action potential, directly contributes to SR Ca²⁺ filling, and regulates myocyte contractility. In addition, induced NCX1 transgene expression at levels found in diseased hearts, by itself, was not associated with cardiac dysfunction in young adult mice.

GRANTS

This work was supported in part by National Heart, Lung, and Blood Institute Grants RO1-HL-58672 and RO1-HL-74854 (to J. Y. Cheung); RO1-HL-56205, RO1-HL-61690, RO1-HL85503, PO1-HL-75443, and PO1-HL-91799 (to W. J. Koch); and PO1-HL-91799 (Project 2; to A. M. Feldman); by the Pennsylvania Research Formulary Fund (to A. M. Feldman); and by American Heart Association Scientist Development Grant F64702 (to T. O. Chan).