INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE NA⁺/ca²⁺ exchanger (NCX1) is an important plasma membrane transport protein that regulates cellular Ca²⁺ homeostasis in many tissues. In cardiac muscle, by exchanging three extracellular Na⁺ for one intracellular Ca²⁺ (forward mode), NCX1 is universally accepted as the dominant Ca²⁺ efflux mechanism during diastole. The reverse-mode operation of NCX1 (3 Na⁺ out:1 Ca²⁺ in) has been proposed to load the sarcoplasmic reticulum (SR) with Ca²⁺ (17), trigger SR Ca²⁺ release (16), and directly activate the contractile elements (17).

The cytosolic free Ca²⁺ concentration ([Ca²⁺]_i) occupies a key role in cardiac excitation-contraction coupling. During an action potential, the influx of extracellular Ca²⁺ via L-type Ca²⁺ channels and possibly reverse NCX1 (16, 17) trigger SR Ca²⁺ release, resulting in increase in [Ca²⁺]_i and myofilament activation. During diastole, most of the myoplasmic Ca²⁺ (70-92%) is resequestered in the SR by sarco(endo)plasmic recticulum Ca²⁺-ATPase (SERCA2), with a small amount of Ca²⁺ extruded by forward NXC1 (7-28%) and, to a minor extent, by sarcolemmal Ca²⁺-ATPase (2, 3, 26). Thus the principal cardiac Ca²⁺ extrusion mechanism is mediated by NCX1. In addition, forward Na⁺/Ca²⁺ exchange contributed importantly to the rate of [Ca²⁺]_i decline from the midportion of the [Ca²⁺]_i transient and was necessary to reach end-diastolic [Ca²⁺]_i levels (26). Whereas there is unanimous agreement on the important role of forward Na⁺/Ca²⁺ exchange on Ca²⁺ extrusion during diastole, the physiological role of reverse Na⁺/Ca²⁺ exchange in mediating SR Ca²⁺ release and loading the SR with Ca²⁺ is highly controversial (1, 16, 17, 20, 23, 24, 27).

The recent recognition that adult cardiac myocytes, when placed in culture and subjected to continual electrical field stimulation, retained normal contractile function (4), and the advent of high-efficiency gene transfer into adult cardiac myocytes by recombinant adenovirus (13, 15) together provide the investigator with useful tools to introduce exogenous genes into cultured adult cardiac myocytes and examine the functional consequences. This approach compliments the transgenic mouse model with the additional theoretical advantage that compensatory responses to overexpression of an exogenous gene are less likely to occur in a short-term cell culture model than in a transgenic animal. The present study was undertaken to test the hypothesis that, in adult rat cardiac myocytes, NCX1 functions in both the forward (Ca²⁺ efflux) and reverse (Ca²⁺ influx) modes of operation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Myocyte isolation and culture. Cardiac myocytes were isolated from the septum and left ventricular free wall of male Sprague-Dawley rats (~300 g) by successive perfusion with collagenase and hyaluronidase (6). Portions of freshly isolated myocytes were seeded on laminin-coated coverslips (7) and used within 2 h of isolation for contractility measurements. For myocyte culture, myocytes were either seeded on laminin-coated coverslips, which were subsequently placed in four-well trays (Nuclone), or plated directly on laminin-coated four-well trays. Culture medium consisted of serum-free medium 199 (Earle's balanced salt solution without L-glutamine and NaHCO₃) supplemented with (in mM) 25 NaHCO₃, 5 creatine, 2 carnitine, 5 taurine, and 0.1 ascorbic acid. Insulin (100 U/ml), 5-bromo-2-deoxyuridine (31 µg/ml), bovine serum albumin (0.2%), penicillin (500 U/ml), and gentamicin (4 µg/ml) were also added to the culture medium. After 2 h, media were then changed to remove nonadherent myocytes. The myocytes were incubated for an additional 3-4 h before initiation of electrical stimulation.

Construction of recombinant replication-deficient adenovirus. The basic protocol is described by He et al. (11). Briefly, the coding sequence of rat heart NCX1 was released from pcDNA3.1(+) by sequential digestion with HindIII and XbaI. The NCX1 sequence (3,067 bp) was inserted into the shuttle vector pAdTrack-CMV, using the same HindIII and XbaI restriction sites on the shuttle vector. The resulting shuttle vector plasmid was linearized with PmeI, mixed with supercoiled pAdEasy-1, and used to electroporate BJ-5183 cells. The recombinant plasmids usually generated smaller colonies, which were selected and screened by restriction endonuclease mapping (HindIII/XbaI and BglII/ClaI). Once recombination was confirmed, supercoiled plasmid DNA (isolated from minipreps) was transformed into DH10B cells for large-scale amplification. The recombinant construct was linearized with PacI and used to transfect HEK-293 cells. Recombinant adenovirus containing both green fluorescent protein (GFP) and NCX1 (each under a separate cytomegalovirus promoter) (Adv-GFP-NCX1) was harvested in 7 days and stored at 20°C in 5 mM Tris (pH 8.0), 50 mM NaCl, 0.05% bovine serum albumin, and 25% glycerol.

Adenoviral infection of cardiac myocytes. Two hours after isolation, myocytes seeded in four-well trays were infected with either Adv-GFP-NCX1 or adenovirus expressing GFP alone (Adv-GFP) at a multiplicity of infection of 1 for 3 h. Media were then changed and myocytes were studied after 6, 24, and 48 h in continued pacing culture. Over 95% of myocytes fluoresced green (excitation 478 nm, emission 535 nm) within 6 h, indicating successful adenoviral infection and GFP expression. For brevity, myocytes infected with Adv-GFP and Adv-GFP-NCX1 are referred to as GFP and NCX1 myocytes, respectively.

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 either 0.6, 1.8, or 5.0 mM [Ca²⁺]_o and were placed on a temperature-controlled stage (37°C) of a Zeiss IM35 microscope (29). Fields of myocytes were chosen at random, and myocytes were field stimulated to contract (1 Hz) between platinum electrodes spaced 2 mm apart, as previously described (29-31). Myocytes were viewed through an Olympus [DApo UV ×40, 1.30 numerical aperture (NA)] oil objective and were imaged by a charge-coupled device video camera (Ionoptix; Milton, MA). Myocyte motion measurements were acquired with the use of a personal computer with interface and software that were purchased from Ionoptix. Data were permanently stored on a zip drive (Iomega; Roy, UT) and analyzed off-line by Ionoptix software. For calibration of pixels versus micrometer, a high-resolution test target (model 22-8635, Ealing Electro-Optics; Natick, MA) was used.

[Ca²+]_i transient measurements. Myocytes were exposed to 0.67 µM fura 2-AM for 15 min at 37°C (7). Fura 2-loaded myocytes mounted in a Dvorak-Stotler chamber situated in a temperature-controlled stage (37°C) of a Zeiss IM35 inverted microscope were field stimulated to contract at 1 Hz between platinum wire electrodes, as previously described (31). [Ca²⁺]_o was varied between 0.6 and 5.0 mM. Excitation light (360 and 380 nm, ±10 nm band pass, Ionoptix) was directed to individual myocytes only during data acquisition to minimize photobleaching. Epiflourescence (510 ± 18 nm) collected by an Olympus DApo UV ×40/1.30 NA oil objective was passed through a pinhole (1.6 mm) and captured by a photomultiplier (model R928-07; Hamamatsu). Output from the photomultiplier was routed through an amplifier/discriminator (model C609, Thorn EMI; Middlesex, UK) before arrival at a counter/timer board (model C660, Thorn EMI).

Measurement of SR Ca²+ content. SR-releasable Ca²⁺ content was estimated by measuring the time integral of forward Na⁺/Ca²⁺ exchange current induced by caffeine exposure, as described by us previously (32, 33). Briefly, whole cell patch-clamp recordings were performed at 30°C. Pipette solution consisted of (in mM) 100 Cs⁺ glutamate, 1 MgCl₂, 30 HEPES, and 2.5 MgATP; pH 7.2. External solution was (in mM) 130 NaCl, 5 CsCl, 1.2 MgSO₄, 1.2 NaH₂PO₄, 20 HEPES, and 10 glucose; pH 7.4. [Ca²⁺]_o was either 0.6 or 5.0 mM. Holding potential was 80 mV. At 200 ms after the 11th conditioning pulse (from 80 to 0 mV, 300 ms, 1 Hz), with membrane potential held at 80 mV, caffeine (5 mM; 2.4 s) was applied by puffer superfusion (30, 33). Currents were digitized at 0.5 kHz and collected for ~4 s.

NCX1, SERCA2, and calsequestrin immunoblotting. Cultured myocytes from a four-well tray were rinsed three times with ice-cold phosphate-buffered saline and then scraped with a rubber policeman into a 1 ml of ice-cold lysis buffer containing 20 mM Tris (pH 8.1), 137 mM NaCl, 2 mM EDTA, 10% glycerol, 0.3% sodium dodecyl sulfate (SDS), 0.4% deoxycholate, 100 µg/ml phenylmethylsulfonyl fluoride, and 1 µl/ml protease inhibitor cocktail (Sigma). The cell lysate was snap-frozen with dry ice-ethanol and stored at 80°C. The amount of myocyte protein recovered from a 4-well culture tray was 887.6 ± 117.8 µg (n = 19).

Statistics. All results are expressed as means ± SE. In experiments in which maximal contraction amplitudes were measured as functions of experimental group (Adv-GFP vs. +Adv-GFP), [Ca²⁺]_o, and days in culture, three-way ANOVA was performed to determine significance of difference. A linear model fitted standard-least squares (JMP version 4, SAS Institutes; Cary, NC) was used. For analysis of a parameter (e.g., systolic [Ca²⁺]_i and maximal contraction amplitude) as a function of group (GFP vs. NCX1) and [Ca²⁺]_o, two-way analysis of variance (ANOVA) was used to determine statistical significance. In all analyses, P  0.05 was taken to be statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effects of continual pacing on cultured myocyte contractility. We have previously demonstrated that myocytes cultured under quiescent conditions for as little as 18 h exhibited significantly less cell shortening compared with freshly isolated myocytes (28). When placed under continual electrical stimulation conditions, myocytes retained similar contraction amplitudes at 24 and 48 h in culture when compared with freshly isolated cells (Table 1) whether cells were studied at 0.6, 1.8, or 5.0 mM [Ca²⁺]_o. This conclusion is supported by two-way ANOVA ([Ca²⁺]_o, day in culture), which indicated significant [Ca²⁺]_o effect (P < 0.0001) but insignificant day in culture (P = 0.6331) and [Ca²⁺]_o × day interaction (P = 0.5070) effects.

Effects of adenoviral infection on cultured myocyte contractility. Infection with recombinant, replication-deficient adenovirus has been shown to be highly efficient for gene delivery to adult cardiac myocytes (13, 15). Compared with uninfected myocytes, adenoviral-infected myocytes at 6 (day 0), 24, and 48 h of continual pacing in culture had similar cell shortening amplitudes regardless of whether cell shortening was measured at 0.6, 1.8, or 5.0 mM [Ca²⁺]_o (Table 1). Three-way ANOVA (group, [Ca²⁺]_o, and day) indicated insignificant group (Adv-GFP vs. +Adv-GFP; P = 0.1341) and day (P = 0.4577) but significant [Ca²⁺]_o (P < 0.0001) effects. In addition, the group × [Ca²⁺]_o × day interaction effect was also insignificant (P = 0.6844).

Effects of Adv-GFP-NCX1 infection on NCX1 abundance and cell size. Forty-eight hours after infection with either Adv-GFP or Adv-GFP-NCX1, cardiac myocyte lysates were collected to analyze for NCX1 abundance by immunoblotting. Under nonreducing gel conditions, NCX1 was detected as a band of apparent molecular weight of 160 kDa (18, 32). Compared with GFP myocytes, NCX1 myocytes had significantly (P < 0.02) more NCX1 protein (558 ± 97 vs. 199 ± 29 arbitrary units) (Fig. 1). In contrast, there were no (P < 0.24) differences in SERCA2 amounts between NCX1 (111 ± 4) and GFP (102 ± 6 arbitrary units) myocytes (Fig. 1). As an additional control for protein loading, we measured calsequestrin in each of the myocyte lysates (Fig. 1). Calsequestrin expression has been shown to be unchanged during ontogenic development, aging, cardiac hypertrophy, and failing human myocardium (10) and this can be used as an internal control. There were no significant (P = 0.9884) differences in calsequestrin protein amounts between GFP and NCX1 (91.3 ± 0.7 vs. 91.3 ± 0.6 arbitrary units, respectively) myocytes (Fig. 1). In another series of experiments, in which whole cell capacitance was measured, it was found that there were no significant differences (P = 0.2744) in whole cell capacitance between GFP (181.2 ± 2.5 pF; n = 12) and NCX1 (184.5 ± 1.7 pF; n = 14) myocytes, indicating that NCX1 overexpression had no effect on cell sizes.

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Immunoblots of Na+/Ca2+ exchanger (NCX1), sarco(endo) plasmic reticulum Ca2+-ATPase (SERCA2), and calsequestrin. Proteins in myocyte homogenates (25 µg for NCX1 lane and 50 µg for SERCA2 and calsequestrin lanes) were separated by gel electrophoresis and transferred to Trans-Blot membranes, and NCX1, SERCA2, and calsequestrin were identified by immunoblotting, as described in METHODS. Note that when compared with green fluorescent protein (GFP) myocytes, NCX1 myocytes had significant increases in NCX1. There were no differences in either SERCA2 or calsequestrin expression between the 2 groups. Composite results are presented in RESULTS. Numbers on left refer to apparent molecular mass.

Effects of NCX1 overexpression on myocyte contractile function. At 0.6 mM [Ca²⁺]_o, GFP myocytes shortened 456% more than NCX1 myocytes (Table 2). This is because of the 21 NCX1 myocytes examined, 15 exhibited contraction amplitudes that were <1% of resting cell length (Fig. 2). Indeed, 11 of the 21 NCX1 myocytes had no discernible contractions when paced in media containing 0.6 mM [Ca²⁺]_o. In contrast, at 5.0 mM [Ca²⁺]_o, GFP myocytes shortened 30% less than the NCX1 myocytes did (Fig. 2 and Table 2). The differences in twitch amplitudes were no longer apparent at intermediate [Ca²⁺]_o levels (Fig. 2, Table 2). These conclusions are supported by highly significant (P < 0.0001) [Ca²⁺]_o and group × [Ca²⁺]_o interaction effects, indicating that the magnitude and/or direction of the effects of [Ca²⁺]_o on cell shortening was different across experimental groups (GFP vs. NCX1).

View larger version (38K): [in this window] [in a new window]   Fig. 2.   Overexpression of NCX1 alters contractile function in rat myocytes. Isolated myocytes were infected with recombinant adenovirus that expresses either GFP or GFP and NCX1 and then cultured for 48 h under continuous electrical stimulation (1 Hz) conditions. For contraction studies, cultured myocytes were paced (1 Hz) to contract at 37°C and extracellular Ca2+ concentration ([Ca2+]o) of 0.6, 1.8, or 5.0 mM. Shown are steady-state paced twitches from myocyte overexpressing either GFP (A, C, and E) or NCX1 (B, D, and F). Note that at 0.6 mM [Ca2+]o (A and B), contraction amplitude was larger in GFP myocytes, whereas at 5.0 mM [Ca2+]o (E and F), NCX1 myocytes contracted more than GFP myocytes. At 1.8 mM [Ca2+]o (C and D), the contraction amplitudes were similar between GFP and NCX1 myocytes. Results are summarized in Table 2.

Effects of NCX1 overexpression on [Ca²+]_i transients. [Ca²⁺]_i occupies a central role in cardiac myocyte excitation-contraction coupling. Thus the differences in contractile behavior between GFP and NCX1 myocytes may be related to differences in [Ca²⁺]_i homeostasis brought about by NCX1 overexpression in cardiac myocytes. Indeed, resting [Ca²⁺]_i was significantly lower in NCX1 myocytes (Table 3; group effect, P < 0.0001). Changing [Ca²⁺]_o apparently had no significant effect on resting [Ca²⁺]_i (Table 3; [Ca²⁺]_o effect, P = 0.0596). Diastolic [Ca²⁺]_i comparisons between GFP and NCX1 myocytes essentially mirrored the results of resting [Ca²⁺]_i comparisons in Table 3; significant group (P < 0.0001) but insignificant [Ca²⁺]_o (P = 0.4338) effects.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   [Ca2+]i transients in myocytes overexpresing NCX1. Fura 2-loaded myocytes were paced (1 Hz) to contract at 37°C and [Ca2+]o of 0.6 (A and B), 1.8 (C and D), and 5.0 mM (E and F). Note that at 0.6 mM [Ca2+]o, amplitudes of [Ca2+]i transients were lower in NCX1 (B) than GFP (A) myocytes. At 5.0 mM [Ca2+]o, intracellular Ca2+ concentration ([Ca2+]i) transient amplitudes were larger in NCX1 (F) than GFP (E) myocytes. At 1.8 mM [Ca2+]o, the [Ca2+]i transient amplitudes were similar between GFP (C) and NCX1 (D) myocytes. Note also the rates of [Ca2+]i decline during the first half of the [Ca2+]i transient were similar between the two groups. In another 2 groups of myocytes expressing either GFP or GFP and NCX1, but not loaded with fura 2, it was determined that autofluoresence from GFP accounted for <10% of the fura 2 signal obtained either at 360- or 380-nm excitation. Composite results are summarized in Table 3.

Effects of NCX1 overexpression on SR-releasable Ca²+ contents. Application of 5 mM caffeine for 2.4 s on cardiac myocytes at 80 mV caused a large inward current that rapidly returned to baseline (Fig. 4). This inward current represented forward Na⁺/Ca²⁺ exchange (3 Na⁺ in:1 Ca²⁺ out) current (I_NaCa) operating to extrude the SR Ca²⁺ released by caffeine. Under our experimental conditions, the time integral of the caffeine-induced inward current was an estimate of SR-releasable Ca²⁺ of the myocyte (32, 33). At 0.6 mM [Ca²⁺]_o, both I_NaCa time integral (pC/cell) and SR-releasable Ca²⁺ (normalized to cell size; fmol/fF) were smaller in NCX1 myocytes (Fig. 4B) than GFP myocytes (Fig. 4A) (Table 4). In contrast, at 5.0 mM [Ca²⁺]_o, SR-releasable Ca²⁺ was larger in NCX1 (Fig. 4D) than GFP (Fig. 4C) myocytes (Table 4). For both I_NaCa time integral and SR-releasable Ca²⁺, two-way ANOVA indicated significant group (P < 0.03), [Ca²⁺]_o (P < 0.0001), and group × [Ca²⁺]_o interaction (P < 0.001) effects, indicating that changing [Ca²⁺]_o affected the inherent differences in SR Ca²⁺ contents between GFP and NCX1 myocytes.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Estimation of sarcoplasmic reticulum (SR)-releasable Ca2+ in myocytes overexpressing NCX1. Myocytes were incubated with either 0.6 (A and B) or 5.0 mM (C and D) [Ca2+]o at 30°C and were voltage clamped at 80 mV. At 200 ms after the 11th conditioning pulse (from 80 to 0 mV, 300 ms, and 1 Hz), caffeine (5 mM) was puffed onto the myocyte for 2.4 s. Large transient inward current caused by caffeine-induced SR Ca2+ release was observed. This current represents Na+ entry caused by Ca2+ extrusion by the NCX1, and the time integral of this current provides an estimate of SR-releasable Ca2+ (32, 33). At 0.6 mM [Ca2+]o, SR-releasable Ca2+ was higher in GFP (A) than NCX1 (B) myocytes. At 5.0 mM [Ca2+]o, NXC1 current time integral was higher in NCX1 (D) than GFP (C) myocytes. Composite results are summarized in Table 4.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Adult cardiac myocytes placed in primary culture may provide a useful in vitro model to study gene expression in the heart. Unlike neonatal cardiomyocytes, healthy isolated adult cardiac myocytes are rod-shaped, striated, and do not exhibit spontaneous contractions. Early studies on primary cultures of adult cardiac myocytes demonstrated that the myocytes rapidly lost their rod-shaped morphology, spread on the culture substratum, and acquired spontaneous contractile activity (22). The loss of the normal rod-shaped morphology was prevented by exclusion of fetal calf serum and addition of creatine, carnitine, and taurine (CCT medium) to the culture media (25). Despite maintenance of rod-shaped morphology, quiescent cultured adult cardiac myocytes suffered a rapid decline in contractile function such that maximal contraction amplitude decreased by 20% within 18 h (28) and 30-50% within 6-24 h of plating (8). In addition, quiescent cultured myocytes gradually atrophied as indicated by loss of one-third of cellular protein content after 15 days of culture (25). These shortcomings were mitigated by the adoption of continual electrical stimulation of adult cardiac myocytes in culture, which resulted in preservation of normal contractile activity for at least 72 h (4) and maintenance of cellular protein content after 7 days (14). Thus the first major finding of our study is that we confirm and extend Berger et al.'s (4) observation that continuous pacing of adult cardiac myocytes while in culture maintained normal contractile function, whether measured at low, intermediate, or high [Ca²⁺]_o (Table 1). This important result allows the use of cultured adult cardiac myocytes in studies of the effects of various gene manipulations on excitation-contraction coupling.

A second technical difficulty encountered when using adult cardiac myocytes in gene manipulation studies relates to the relative inefficiency (compared with neonatal cardiomyocytes) of introduction of foreign genes by conventional, plasmid-based procedures for gene transfer. With the advent of recombinant adenovirus-mediated gene delivery methodology, high-efficiency gene transfer into adult cardiac myocytes can be accomplished (13, 15). Thus a second major finding of our present study is that there were no differences in contractile function between uninfected and adenovirus-infected adult cardiac myocytes, whether examined at low, intermediate, and high [Ca²⁺]_o, or at 6-, 24-, and 48-h postadenovirus exposure (Table 1). Our observations provide the foundation for using continuously paced cultured adult cardiac myocytes in studies of effects of adenovirus-mediated gene transfer on cardiac excitation-contraction coupling.

To determine the physiological role of the cardiac NCX1, we studied adult rat cardiac myocytes, in which overexpression of the rat heart NCX1 was achieved by recombinant adenovirus-mediated gene delivery (Fig. 1). At "physiological" [Ca²⁺]_o of 1.8 mM, there were no differences in twitch amplitudes between NCX1 and GFP myocytes, although maximal shortening velocity was greater and t_1/2 was shorter in NCX1 myocytes (Table 2). Our observations are similar to the findings in ventricular myocytes from transgenic mice overexpressing NCX1, in which twitches were of similar size between transgenic and wild-type myocytes but the rate of contraction and relaxation were significantly faster in transgenic than in wild-type myocytes (24). This subtle change in contractile behavior in myocytes overexpressing the NCX1 can be made more apparent by manipulating the experimental conditions. For example, under conditions favoring Ca²⁺ efflux (0.6 mM [Ca²⁺]_o) (29, 30), NCX1 myocytes contracted significantly less than GFP myocytes (Table 2). Conversely, under conditions favoring Ca²⁺ influx (5.0 mM [Ca²⁺]_o), NCX1 myocytes shortened more than GFP myocytes (Table 2). The differences in twitch amplitudes between GFP and NCX1 myocytes were no longer apparent at intermediate [Ca²⁺]_o. Assuming similar myofilament Ca²⁺ sensitivity between GFP and NCX1 myocytes, this complex pattern of contractile responses (Fig. 2) is consistent with the hypothesis that alterations in both Ca²⁺ influx and efflux pathways contributed to the different contractile behavior in NCX1 myocytes. Because NCX1 can affect both Ca²⁺ influx and efflux pathways, and because there is no a priori reason to suspect L-type Ca²⁺ channel expression in cardiac myocytes is affected by NCX1 overexpression (especially in view of the observation in transgenic myocytes overexpressing NCX1) (27), our findings are supportive of our hypothesis that the overexpressed NCX1 was functional in both Ca²⁺ influx and efflux modes during stimulated twitches.

If the differences in contractile behavior between GFP and NCX1 myocytes were brought about by the overexpressed NCX1 perturbing Ca²⁺ fluxes, then one would expect corresponding changes in [Ca²⁺]_i transient during a twitch. At 0.6 mM [Ca²⁺]_o, conditions which favored Ca²⁺ efflux, the magnitude of [Ca²⁺]_i transient was higher in the control GFP myocytes (Table 3). At 5.0 mM [Ca²⁺]_o, conditions that favored Ca²⁺ influx, [Ca²⁺]_i transient amplitude was higher in NCX1 myocytes. At intermediate [Ca²⁺]_o levels of 1.8 mM, differences in [Ca²⁺]_i transient amplitude or systolic [Ca²⁺]_i were narrowed (Table 3). This pattern of [Ca²⁺]_i transient amplitude differences between NCX1 and GFP myocytes mirrors the pattern of contractile amplitude differences and indicates that the overexpressed NCX1 was functional in both modes of operation. It should be noted that the lack of significant differences in [Ca²⁺]_i transient amplitudes measured in GFP and NCX1 myocytes stimulated at physiological [Ca²⁺]_o of 1.8 mM agrees with the observation in transgenic murine myocytes overexpressing NCX1 (1, 24, 27). It is also interesting to note that both resting and diastolic [Ca²⁺]_i were significantly lower in NCX1 myocytes (Table 3), consistent with the notion that because forward NCX1 was necessary in maintaining diastolic [Ca²⁺]_i levels (26), enhanced forward NCX1 function in NCX1 myocytes resulted in a lower resting and diastolic [Ca²⁺]_i levels.

Despite no changes in SERCA2 expression in transgenic murine myocytes overexpressing NCX1 (24), SR Ca²⁺ uptake activity was reported to be increased (24) or unchanged (27). In our rat myocytes overexpressing NCX1, the t_1/2 of [Ca²⁺]_i decline (a measure of in situ SR Ca²⁺ uptake activity; Ref. 31), was not different compared with control GFP myocytes (Table 3). Thus SR Ca²⁺ uptake activity was not affected by overexpression of NCX1. This is also consistent with the observation that SERCA2 expression was unaffected by NCX1 overexpression in cultured adult rat myocytes (Fig. 1).

At high [Ca²⁺]_o, SR Ca²⁺ content was significantly higher in NCX1 myocytes (Fig. 4D and Table 4). Because neither SERCA2 activity (Table 3) nor abundance (Fig. 1) was affected by NCX1 overexpression, and in the absence of evidence suggesting L-type Ca²⁺ channel function or abundance and SR Ca²⁺ leak were altered by NCX1 overexpression, increased SR Ca²⁺ content in NCX1 myocytes was likely due to enhanced reverse Na⁺/Ca²⁺ exchange during depolarization, with the resultant increase in [Ca²⁺]_i transient (Fig. 3F) and twitch amplitudes (Fig. 2F). Higher SR Ca²⁺ content was also observed in transgenic myocytes overexpressing NCX1 (24), although this finding is inconsistent among investigators (1, 27). Increased SR Ca²⁺ loading during diastole, rather than increased Ca²⁺ influx on the exchanger during a twitch, was proposed by Terracciano et al. (24) as the cause of enhanced contraction in transgenic myocytes overexpressing NCX1. This appears unlikely, as both resting and diastolic [Ca²⁺]_i were significantly lower in NCX1 myocytes (Table 3), indicating increased extrusion of Ca²⁺ during rest and in diastole by the exchanger. Yao et al. (27) also reported a somewhat lower diastolic [Ca²⁺]_i in transgenic myocytes, although this difference did not reach statistical significance.

Under conditions that promote Ca²⁺ efflux ([Ca²⁺]_o = 0.6 mM), SR Ca²⁺ content was significantly lower in NCX1 myocytes (Fig. 4B and Table 4). The lower SR Ca²⁺ content in NCX1 myocytes at 0.6 mM [Ca²⁺]_o explained very well the lower [Ca²⁺]_i transient and twitch amplitudes. Thus overexpressing NCX1 in adult rat cardiac myocytes can either load or deplete the SR of Ca²⁺, depending on the experimental conditions. When one considers in total the patterns of changes in twitch amplitudes, [Ca²⁺]_i transients, and SR Ca²⁺ contents observed in NCX1 myocytes, the most logical explanation is that the overexpressed NCX1 is functional in both forward and reverse modes. Enhanced Ca²⁺ influx during depolarization associated with the action potential was a major effect of NCX1 overexpression, and increased Ca²⁺ extrusion during diastole and at rest by the overexpressed exchanger resulted in a lower resting and diastolic [Ca²⁺]_i. Our conclusions are in agreement with the results obtained from transgenic myocytes overexpressing NCX1 (24, 27), although Adachi-Akahane et al. (1) reported that the overexpressed NCX1 in transgenic murine myocytes only produced functional consequences in the Ca²⁺ efflux mode. The discrepancy is most likely due to artifactual lowering of cytosolic Na⁺ by patch pipette dialysis in the study of Adachi-Akahane et al. (1).

A recent study (21) employing adenovirus-mediated gene transfer in cultured adult rabbit myocytes suggested that contractile performance was impaired, rather than enhanced, by NCX1 overexpression (21). In addition, SR Ca²⁺ content, as measured by caffeine-induced contracture, was decreased in rabbit myocytes overexpressing NCX1. The apparent discrepancy between our data and those of Schillinger et al. (21) can be reconciled if the following factors are taken into account. First, there is the well-known species difference in NCX1 function between adult rat and rabbit myocytes (2). For example, during a twitch, NCX1 provided ~28% of Ca²⁺ transport in rabbit myocytes compared with ~7% in rat myocytes (2). Second, in contrast to our study, adult rabbit myocytes were not continually paced while in culture for 48 h (21) and thus would be expected to suffer deterioration in contractile function (8, 28). Indeed, comparison of twitch amplitudes of freshly isolated adult rabbit myocytes (7.6 ± 0.8% at 2 mM [Ca²⁺]_o, 0.5 Hz and 35°C; Ref. 19) with those of cultured and quiescent rabbit myocytes (2.7 ± 0.2% at 1.75 mM [Ca²⁺]_o, 0.5 Hz and 37°C; Ref. 21) strongly suggests this to be the case. Thus the effects of NCX1 overexpression in adult rabbit myocytes with normal contractile function is at present unclear. Third, decreased contractile performance in cultured rabbit myocytes overexpressing NCX1 was observed only at higher pacing frequencies (21). At lower pacing frequency (0.5 Hz), differences in twitch amplitude between NCX1-overexpressing and control rabbit myocytes were no longer significant (21), similar to the observations in transgenic mouse myocytes overexpressing NCX1 (24) and our own contraction data collected at 1.8 mM [Ca²⁺]_o and 1 Hz (Fig. 2 and Table 2). In contrast to normal rabbit myocytes, which exhibited positive staircase phenomenon, in rabbit myocytes overexpressing NCX1, increasing pacing frequency resulted in no increases in twitch amplitudes (21), suggesting that Ca²⁺ efflux was enhanced at high beating rates. This conclusion is consistent with our observations that contractions in NCX1 rat myocytes were "impaired" compared with GFP myocytes when studied under conditions that favored Ca²⁺ efflux pathways (0.6 mM [Ca²⁺]_o; Fig. 2 and Table 2) (29, 30). Thus the apparently contradictory results of NCX1 overexpression on myocyte contractile function can be explained by the different experimental conditions that may promote either Ca²⁺ efflux or influx. Fourth, it is well known that intracellular Na⁺ is lower in rabbit (3-5 mM) than rat (12 mM) myocytes (2). On the basis of energetics considerations, species differences in intracellular Na⁺ would bias the rat NCX1 more toward the reverse mode of operation. This would partly account for our observation that at high [Ca²⁺]_o, rat NCX1 myocytes contracted more vigorously than GFP myocytes due to enhanced SR Ca²⁺ loading via reverse Na⁺/Ca²⁺ exchange during depolarization. Fifth, canine NCX1 was overexpressed in rabbit myocytes in the study of Schillinger et al. (21), whereas we overexpressed rat NCX1 in myocytes of the same species. Whether there are functional differences when an important membrane transport protein is overexpressed in cells from a different species is at present unknown. Finally, the lower caffeine contracture amplitude in rabbit myocytes overexpressing NCX1 (21), in contrast to the findings of Terracciano et al. (24), can be explained by SR Ca²⁺ content depletion due to enhanced Ca²⁺ efflux in rabbit myocytes overexpressing NCX1 during 5 min of electrical pacing at 2 Hz before caffeine application.

Finally, there are several important limitations to this study. The first is that we have measured some, but not all, of the known Ca²⁺ regulatory pathways that may be altered by continuous pacing culture and adenoviral infection. For example, Ellingsen et al. (8) reported that L-type Ca²⁺ current (I_Ca) increased ~55% at 24-48 h of quiescent rat myocyte culture, returning to levels observed in freshly isolated myocytes by 72 h. Continuous pacing for 48 h increased I_Ca density by another 33% compared with quiescent myocyte culture (4). Despite these caveats, it has to be emphasized that our present culture conditions were effective in maintaining normal contractile activity compared with freshly isolated myocytes (Table 1). In addition, comparisons between GFP and NCX1 myocytes in the present study were performed under similar conditions, i.e., 48 h after adenoviral infection of continual electrically stimulated myocytes in culture, and thus should provide identical background for our measurements. A second limitation relates to the concern that NCX1 overexpression may affect other Ca²⁺ regulatory pathways in rat cardiac myocytes. We have provided direct (SERCA2 and calsequestrin expression; Fig. 1) and indirect (SR Ca²⁺ uptake activity; Table 3) evidence that some of the Ca²⁺ homeostatic pathways were unaffected by NCX1 overexpression in our short-term (48 h) culture. In addition, in transgenic myocytes, in which compensatory responses are more likely to occur, NCX1 overexpression had no affect on peak I_Ca density (27), phospholamban, SERCA2, and calsequestrin levels (24). The weight of current evidence suggests that NCX1 overexpression was unlikely to significantly affect other important Ca²⁺ homeostatic (I_Ca and SERCA2) pathways. The third limitation is that we have not measured action potentials in either GFP or NCX1 myocytes. Potential differences in action potential waveforms between GFP and NCX1 myocytes may have major effects on the energetics of the exchanger. For example, if the action potential duration were prolonged in NCX1 myocytes, as was the case in transgenic murine myocytes overexpressing NCX1 (27), thermodynamic considerations would argue for enhanced reverse NCX1 during depolarization in NCX1 myocytes. The fourth limitation is that the overexpressed NCX1 may occupy a membrane domain that enables it to load the SR with Ca²⁺ or trigger SR Ca²⁺ release more readily than the endogenous native exchanger. This may account for the higher SR Ca²⁺ content (Table 4) and [Ca²⁺]_i transient magnitude (Table 2) observed in NCX1 myocytes. Viewed in this context, it can be argued that Ca²⁺ influx via reverse NCX1 to trigger SR Ca²⁺ release may not occur at all in normal myocytes. The present technology, however, does not allow one to target the NCX1 to a specific membrane domain (e.g., the Ca²⁺ microdomains of the Ca²⁺ channel/ryanodine receptor complex) to definitively settle this question.

In summary, we have established an in vitro myocyte culture model system in which normal contractile function was maintained with continual electrical stimulation. Infection with adenovirus to deliver exogenous genes had no effect on myocyte contractile function. By using our cultured myocyte system, which is fundamentally different than transgenic approach, we found that overexpression of NCX1 resulted in altered patterns of contraction and [Ca²⁺]_i transients, changed SR Ca²⁺ contents, and lower resting and diastolic [Ca²⁺]_i. Calsequestrin and SERCA2 expression, SR Ca²⁺ uptake activity, and cell size were unaffected by overexpression of NCX1. We conclude that the most consistent explanation of our observations is that the overexpressed NCX1 was functional in both Ca²⁺ influx and efflux modes. We speculate that changes in NCX1 expression or activity in pathological conditions may have profound effects on cardiac contractile function.