INTRODUCTION

The Na⁺/Ca²⁺ exchanger is a major regulator of [Ca²⁺]_i¹ in excitable as well as in many nonexcitable cells (1, 2). The exchanger catalyzes bidirectional electrogenic exchange of Na⁺ for Ca²⁺ across the plasma membrane, its direction being determined by the magnitude and orientation of electrical and chemical ion gradients. The exchanger works in concert with other cellular Ca²⁺ transporters including the sarcolemmal Ca²⁺ pump and Ca²⁺ channels and intracellular Ca²⁺ sequestration and release systems. Thus the function of the exchanger under physiological or pathological conditions is often difficult to define, because the membrane potential or intracellular concentrations of Na⁺ and Ca²⁺ may vary in different cell types and change in response to agonist or electrical stimulation.

Recent molecular cloning studies have revealed that the Na⁺/Ca²⁺ exchanger isoforms expressed in various cell types are highly homologous to the cardiac clone and are the product of the same gene (NCX1) (3, 4, 5). These isoforms, however, differ in a small region near the carboxyl end of the large central loop, which is due to alternative splicing (4, 5). In brain and skeletal muscle, a Na⁺/Ca²⁺ exchanger isoform that is a product of a different gene (NCX2) is also expressed (6).

The physiological role of the Na⁺/Ca²⁺ exchanger has been studied most extensively in cardiac muscle. During each action potential, the exchanger rapidly extrudes the Ca²⁺ that has entered the cardiomyocytes via the sarcolemmal L-type Ca²⁺ channels to trigger the release of Ca²⁺ from the SR (7, 8). In addition, the exchanger has been shown to play a much greater role than the sarcolemmal Ca²⁺ pump in the slow extrusion of Ca²⁺ from cardiomyocytes during diastole or under resting conditions (7, 9). On the other hand, the exchanger appears capable of bringing Ca²⁺ into cardiomyocytes during cardiac depolarization, although triggering the release of Ca²⁺ from the SR via the exchanger is much less efficient than via the L-type Ca²⁺ channels (10). Under pathological conditions such as ischemia-associated reperfusion injury, the exchanger is thought to cause Ca²⁺ overloading of cardiomyocytes due to an increase in [Na⁺]_i (11, 12). In other cell types, including nonexcitable cells such as kidney cells, however, the specific contribution of the exchanger to the [Ca²⁺]_i regulation has been difficult to determine, because of the relatively low density of the exchanger in the plasma membranes of these cells and the lack of a specific inhibitor.

A potent and selective inhibitor of the Na⁺/Ca²⁺ exchanger, if available, should be extremely useful to study the reaction mechanism of Na⁺/Ca²⁺ exchange and to clarify its physiological and pathophysiological roles. Moreover, such an inhibitor may serve as a therapeutic agent by virtue of its inotropic, cardioprotective, antiarrhythmic, or antihypertensive effects. Although a variety of natural products, synthetic organic compounds, and inorganic cations have been tested for their ability to inhibit the exchanger (13), few selective inhibitors exist. Amiloride analogues such as 3,4-dichlorobenzamil inhibit Na⁺/Ca²⁺ exchange at micromolar concentrations, but they exert a cytotoxic effect by inhibiting a number of other ion transporters and cell metabolism (14, 15, 16). On the other hand, XIP, a synthetic peptide derived from the amino acid sequence of the cardiac Na⁺/Ca²⁺ exchanger, inhibits Na⁺/Ca²⁺ exchange with high potency (IC₅₀ = 0.11 µM) and relatively high specificity (17) and has thus been used in previous studies (18, 19). However, XIP is highly cationic and interacts with calmodulin. With the latter property, it modulates activities of calmodulin-activated enzymes such as the sarcolemmal Ca²⁺-ATPase (17). Since XIP does not seem to permeate through the cell membrane but acts from the cytoplasmic surface, its use is significantly limited. Therefore, development of a new potent inhibitor that is selective for the Na⁺/Ca²⁺ exchanger in vivo is highly desired.

We report here that a newly synthesized compound, 2-[2-[4-(4-nitrobenzyloxy)phenyl]ethyl]isothiourea methanesulfonate (designated No.7943), is a potent and selective inhibitor of the Na⁺/Ca²⁺ exchanger. This compound (Fig. 1) has been identified by screening a compound library for inhibition of Na⁺_i-dependent Ca²⁺ uptake into isolated cardiac sarcolemmal vesicles. Surprisingly, this compound selectively inhibits the reverse mode of Na⁺/Ca²⁺ exchange in intact cells. We also describe that this compound prevents the excess Ca²⁺ influx evoked by the Ca²⁺ paradox, which has been widely studied as an experimental model for Ca²⁺ overloading in cardiomyocytes.

EXPERIMENTAL PROCEDURES

Primary cultures of neonatal rat cardiomyocytes were prepared by the method described previously (20). Briefly, hearts from 1- or 2-day-old Wistar rats were minced, and cells were dissociated with 0.1% trypsin in buffer A (20 mM Hepes/Tris (pH 7.4), 137 mM NaCl, 5.36 mM KCl, 0.81 mM MgSO₄, 0.44 mM KH₂PO₄, 0.34 mM Na₂HPO₄, and 5.55 mM glucose) containing no added Ca²⁺. After centrifugation, the pellet was resuspended in DMEM (Life Technologies, Inc.) supplemented with 5% heat-inactivated FCS, 1.5 µM vitamin B₁₂, 1 µg/ml insulin, 5 µg/ml transferrin, 50 units/ml penicillin, and 50 µM streptomycin. Dispersed cells were placed in 150-mm dishes (Falcon) for 1 h, and non-attached cells were seeded onto polystyrene dishes or onto sterile glass coverslips. To inhibit growth of fibroblasts, 10 µM cytosine arabinoside was included in the final culture medium for 48 h. Spontaneously beating cells were used after 3-5 days of culture.

Vascular smooth muscle cells were isolated from the thoracic aorta of male Wistar rats (200-300 g) by enzymatic dispersion as described by Chamley et al. (21). The cells were grown for 4 to 5 days in DMEM supplemented with 10% heat-inactivated FCS and antibiotics as above. After reaching confluency, cells were cultured in serum-free medium for an additional 24-48 h to enhance redifferentiation. CCL39 cells (ATCC) were maintained in DMEM supplemented with 7.5% heat-inactivated FCS and antibiotics.

The 3-kilobase SmaI-HindIII fragment of the dog cardiac Na⁺/Ca²⁺ exchanger cDNA (3, 4) was inserted into the mammalian expression vector pKCRH (22). CCL39 cells were transfected with the constructed vector and then screened for high expression of Na⁺/Ca²⁺ exchanger as described previously (23). The selected cells stably expressed about 15-fold higher amounts of the Na⁺/Ca²⁺ exchanger protein as compared with nontransfected cells.

Cells in 24-well dishes were loaded with Na⁺ by incubation at 37 °C for 30 min in 0.5 ml of normal BSS (10 mM Hepes/Tris (pH 7.4), 146 mM NaCl, 4 mM KCl, 2 mM MgCl₂, 1 mM CaCl₂, 10 mM glucose, and 0.1% BSA) containing 1 mM ouabain and 10 µM monensin. ⁴⁵Ca²⁺ uptake was initiated by switching the medium to Na⁺-free BSS, replacing NaCl with equimolar choline chloride, or to normal BSS, both of which contained 0.1-4 mM (370 kBq) ⁴⁵CaCl₂, 1 mM ouabain, and 10 µM verapamil. After a 15- or 30-s incubation, ⁴⁵Ca²⁺ uptake was stopped by washing cells four times with an ice-cold solution containing 10 mM Hepes/Tris (pH 7.4), 120 mM choline chloride, and 10 mM LaCl₃. Cells were solubilized with 0.1 N NaOH, and aliquots were taken for determination of radioactivity and protein. Na⁺_i-dependent ⁴⁵Ca²⁺ uptake was estimated by subtracting ⁴⁵Ca²⁺ uptake in normal BSS from that in Na⁺-free BSS.

⁴⁵Ca²⁺ efflux from cells cultured in a 35-mm dish was assayed as described previously (24). Cells were equilibrated with ⁴⁵Ca²⁺ by incubating them at 37 °C for 4 h in 1 ml of BSS containing 740 kBq of ⁴⁵Ca²⁺. After rinsing cells six times with Ca²⁺- and Na⁺-free BSS for 1 min, ⁴⁵Ca²⁺ efflux was measured for 20 s in Ca²⁺- and Na⁺-free BSS or in Ca²⁺-free BSS; both solutions contained 1 µM thapsigargin to cause a transient increase in [Ca²⁺]_i. Na⁺_o-dependent ⁴⁵Ca²⁺ efflux was estimated by subtracting ⁴⁵Ca²⁺ efflux in Ca²⁺- and Na⁺-free BSS from that in Ca²⁺-free BSS.

[Ca²⁺]_i was monitored using fura-2 as a fluorescent Ca²⁺ indicator. Cells cultured on glass coverslips were loaded with 4 µM fura-2/acetoxymethyl ester for 20 min at 37 °C in buffer A containing 1 mM CaCl₂ and 0.1% BSA (for cardiomyocytes) or in BSS (for smooth muscle cells and transfected CCL39 cells). Loaded cells were then washed twice with the same medium. Glass coverslips were fixed to a mount that was diagonally inserted into a cuvette filled with 2.2 ml of the particular medium. The fluorescence signal was monitored and [Ca²⁺]_i calculated as described previously (25).

Sarcolemmal vesicles were prepared from dog ventricular muscle according to Jones (26). Na⁺_i-dependent Ca²⁺ uptake into vesicles was measured essentially as described previously (27). Briefly, 5 µl of Na⁺-loaded vesicles (1-2 mg/ml) was rapidly diluted into 0.25 ml of uptake medium (20 mM Mops/Tris (pH 7.4), 160 mM KCl, 5-80 µM ⁴⁵CaCl₂ (10 kBq), and 0.5 µM valinomycin) at 37 °C. The reaction was stopped at 1.5 s by adding 4 ml of ice-cold washing medium (160 mM KCl and 1 mM LaCl₃). Vesicles were collected on a glass fiber filter and washed twice with the same medium. Blanks were obtained by measuring ⁴⁵Ca²⁺ uptake in medium containing NaCl instead of KCl. These blanks were subtracted from all data points to correct for Na⁺-independent ⁴⁵Ca²⁺ uptake.

Na⁺_o-dependent Ca²⁺ efflux was quantitated by measuring Na⁺_o-induced ⁴⁵Ca²⁺ loss from vesicles that had been preloaded with ⁴⁵Ca²⁺ by Na⁺_i-dependent Ca²⁺ uptake (28). Briefly, Na⁺-loaded vesicles (5 µl) were diluted with 0.5 ml of the uptake medium containing 10 µM ⁴⁵CaCl₂ for 2 min at 37 °C. ⁴⁵Ca²⁺ efflux was then initiated by addition of 0.5 ml of efflux medium (20 mM Mops/Tris (pH 7.4), 120-160 mM KCl, 0.2 mM EGTA, and 0-40 mM NaCl). The reaction was stopped 10 s later by addition of the washing medium. Blanks were obtained by measuring ⁴⁵Ca²⁺ loss in the efflux medium containing no NaCl.

L-type Ca²⁺ channel activity was assayed by measuring DHP-sensitive ⁴⁵Ca²⁺ uptake into cultured smooth muscle cells as described previously (25). Briefly, cultured smooth muscle cells in 24-well dishes were preincubated at 37 °C for 30 min in 0.5 ml of normal BSS. To initiate ⁴⁵Ca²⁺ uptake, cells were rinsed with BSS containing 1 mM (370 kBq) ⁴⁵CaCl₂ in the presence or absence of 1 µM (+)-PN200-110. After 2 min, ⁴⁵Ca²⁺ uptake was stopped, and radioactivity and protein were determined. DHP-sensitive ⁴⁵Ca²⁺ uptake was estimated by subtracting ⁴⁵Ca²⁺ uptake in the presence of (+)-PN200-110 from that in the absence of (+)-PN200-110.

Na⁺/H⁺ exchange activity was assayed by measuring 5-(N-ethyl-N-isopropyl)amiloride-sensitive ²²Na⁺ uptake as described previously (29). Cultured cardiomyocytes in 24-well dishes were preincubated with BSS containing 30 mM NH₄Cl for 30 min at 37 °C and subsequently washed twice with Na⁺-free BSS for 40 s. ²²Na⁺ uptake was then initiated by adding the Na⁺-free BSS containing 1 mM (37 kBq) ²²NaCl, 1 mM ouabain, and either 0 or 0.1 mM 5-(N-ethyl-N-isopropyl)amiloride. After 40 s, cells were washed four times with an ice-cold phosphate-buffered saline to stop ²²Na⁺ uptake. Passive ²²Na⁺ uptake into cultured cardiomyocytes was measured for 30 min at 37 °C in BSS containing ²²NaCl (740 kBq/ml) and 1 mM ouabain.

Na⁺,K⁺-ATPase activity was measured by incubating cardiac sarcolemmal vesicles (100 µg) for 20 min at 37 °C in a 1-ml reaction medium containing 20 mM Hepes/Tris (pH 7.4), 100 mM NaCl, 10 mM KCl, 5 mM MgCl₂, 3 mM Na₂ATP, and 1 mM EGTA. The reaction was stopped by addition of 10% trichloroacetic acid, and P_i liberated was determined (30). The difference between activities in the presence and absence of 0.2 mM ouabain was taken as Na⁺,K⁺-ATPase activity.

Ca²⁺-ATPase activity was measured using cardiac sarcolemmal vesicles or SR vesicles that were prepared from dog ventricular muscle (31). The ATPase reaction was performed at 37 °C with 50 µg of sarcolemmal vesicles for 20 min or with 10 µg of SR vesicles for 1 min in 0.5 ml of standard medium (20 mM Hepes/Tris (pH 7.2), 100 mM KCl, 5 mM MgCl₂, 0.1 mM CaCl₂ (or 1 mM EGTA), 1 mM ATP, 2 mM phosphoenolpyruvate, 0.2 mg/ml pyruvate kinase, and 2 µM A23187), to which 5 µg/ml calmodulin, 1 µM thapsigargin, and 1 mM ouabain were added further when ATP hydrolysis by sarcolemmal vesicles was measured. After the reaction was terminated by adding 3 N HCl (0.1 ml) containing 2.5 mM 2,4-dinitrophenyl hydrazine, pyruvate produced was determined as described previously (32). The difference between activities in the presence and absence of CaCl₂ was taken as Ca²⁺-ATPase activity.

Cardiac action potential was measured according to the standard method (33). Briefly, papillary muscle bundles were isolated from guinea pig right ventricle, which were mounted in a chambered organ bath and superfused with Tyrode's solution (137 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl₂, 1 mM MgCl₂, 0.5 mM NaH₂PO₄, 11.9 mM NaHCO₃, and 5.5 mM glucose, gassed by 95% O₂/5% CO₂) at 10 ml/min at 36 °C. Muscle preparations were stimulated at a constant rate of 2 Hz through bipolar electrodes with square-wave pulses of 0.5 ms and an intensity 2 times above threshold. Transmembrane electrical activity was recorded with conventional glass microelectrodes with tip resistances of 5-15 M. Transmembrane potentials were measured by a high input impedance preamplifier (MEZ-7200, Nihon Kohden), displayed on dual beam oscilloscope (VC-11, Nihon Kohden), and stored on videotape recorder. The maximum rate of rise of the action potential (_max) was obtained by an electronic differentiator with linear differentiation.

Protein was measured by the modified Lowry method (34) with BSA as a standard.

Data are expressed as the means ± S.E. Differences for multiple comparisons were analyzed by one-way analysis of variance followed by the Dunnett's test. Values of p < 0.05 were considered statistically significant.

No.7943 (2-[2-[4-(4-nitrobenzyloxy)phenyl]ethyl]isothiourea methanesulfonate), 3,4-dichlorobenzamil, and 5-(N-ethyl-N-isopropyl)amiloride were synthesized by the New Drug Research Laboratories, Kanebo Ltd. XIP (RRLLFYKYVYKRYRAGKQRG) was synthesized by the Peptide Institute. ⁴⁵CaCl₂ and ²²NaCl were purchased from DuPont NEN. Ouabain, monensin, valinomycin, A23187, verapamil, nicardipine, and BSA (fatty acid-free) were from Sigma. Thapsigargin, cytosine arabinoside, and calmodulin were from Wako Pure Chemical Co. Fura-2/acetoxymethyl ester was obtained from Dojindo Laboratories. (+)-PN200-110 was from Sandoz Ltd. All other chemicals were of analytical grade.

RESULTS

No.7943 dose-dependently inhibited Na⁺_i-dependent ⁴⁵Ca²⁺ uptake into rat cardiomyocytes, rat aortic smooth muscle cells, and cardiac NCX1-transfected CCL39 cells (Fig. 2A). The IC₅₀ values for individual cell types were 2.4 ± 0.3, 2.0 ± 0.1, and 1.6 ± 0.2 µM (n = 3), respectively, the complete inhibition occurring at 30 µM of this agent. Thus, the inhibitory potency of No.7943 was very similar among these cell types. Under identical conditions, 10 µM 3,4-dichlorobenzamil inhibited Na⁺_i-dependent ⁴⁵Ca²⁺ uptake into rat cardiomyocytes by 30 ± 5% (n = 3). At 30 µM, however, 3,4-dichlorobenzamil exhibited cytotoxicity, causing cell rounding and detachment from dishes, whereas the same concentration of No.7943 did not exert such cytotoxicity. In contrast, XIP did not affect the Na⁺_i-dependent ⁴⁵Ca²⁺ uptake into cardiomyocytes at concentrations up to 100 µM. Of note, the inhibitory potency of No.7943 was identical whether ⁴⁵Ca²⁺ uptake was measured for 15 s in cardiac NCX1-transfected CCL39 cells after preincubation with this agent for 5 min (Fig. 2A) or without such preincubation (IC₅₀ = 2.1 ± 0.4 µM (n = 3)). No.7943 at up to 30 µM did not affect ⁴⁵Ca²⁺ uptake into nontransfected CCL39 cells (data not shown), which is consistent with the lack of detectable Na⁺_i-dependent ⁴⁵Ca²⁺ uptake in these cells (23).

We examined the effect of No.7943 on the rate of Na⁺_i-dependent ⁴⁵Ca²⁺ uptake into NCX1-transfected cells measured as a function of Ca²⁺_o concentration (Fig. 2B). The observed Ca²⁺_o concentration dependences obeyed Michaelis-Menten kinetics. No.7943 at 1 and 3 µM decreased V_max to 7.2 ± 0.2 and 4.5 ± 0.3 nmol/mg/15 s (n = 3), respectively, compared with the control value of 8.8 ± 0.1 nmol/mg/15 s (n = 3). No.7943 did not affect the K_m (0.14 ± 0.01 mM (n = 3) for control and 0.15 ± 0.01 mM (n = 3) for cells treated with 1 or 3 µM No.7943). We also examined the effect of No.7943 on Na⁺_o-dependent inhibition of Na⁺_i-dependent ⁴⁵Ca²⁺ uptake into NCX1-transfected cells (Fig. 2C). Na⁺_o inhibited Na⁺_i-dependent ⁴⁵Ca²⁺ uptake in a concentration-dependent manner. The IC₅₀ for Na⁺_o was similar in control cells (69 ± 3.1 mM, n = 3) and in cells treated with 1 µM (68 ± 3.4 mM, n = 3) or 3 µM (67 ± 2.2 mM, n = 3) No.7943. Thus, the extent of inhibition by No.7943 was similar at different Na⁺_o concentrations. These results indicate that No.7943 does not compete with Ca²⁺_o or Na⁺_o.

Fig. 3 shows the inhibitory effect of No.7943 on Na⁺_i-dependent increase in [Ca²⁺]_i measured in aortic smooth muscle cells in the presence of 10 µM verapamil. A transient increase of [Ca²⁺]_i was induced in these cells by switching the medium from a Na⁺-containing to a Na⁺-free solution for 2 min. The IC₅₀, as estimated from the plot of the initial rate of rise in [Ca²⁺]_i versus the No.7943 concentration, was 1.6 ± 0.2 µM (n = 3). This is very close to the IC₅₀ values obtained from the ⁴⁵Ca²⁺ uptake measurements (see above). Interestingly, the inhibition by this agent completely disappeared after washing cells with fresh medium for 1 min (Fig. 3), indicating that the effect is fully reversible. The transient nature of [Ca²⁺]_i rise observed here might have arisen, at least in part, from a time-dependent decrease in Na⁺/Ca²⁺ exchange activity, which was caused by the exchange-induced reduction in [Na⁺]_i and development of deeper negative membrane potential, as well as by rapid removal of cytoplasmic Ca²⁺ by the sarcolemmal and SR Ca²⁺ pumps. In cardiomyocytes, a similar IC₅₀ value (1.2 ± 0.1 µM, n = 3) was also obtained for the inhibition by No.7943 of the initial rate of Na⁺_i-dependent increase in [Ca²⁺]_i. In this experiment, cardiomyocytes were preincubated at 37 °C for 20 min with buffer A containing 137 mM NaCl, 1 mM ouabain, 10 µM verapamil, and 0.1 mM Ca²⁺_o, and then an increase in [Ca²⁺]_i was evoked by the addition of 1 mM Ca²⁺_o to the medium. A steady increase in [Ca²⁺]_i was observed at least for 2 min under these conditions (data not shown).

We examined the effect of No.7943 on ⁴⁵Ca²⁺ efflux from ⁴⁵Ca²⁺-labeled NCX1-transfected cells in a Ca²⁺- and Na⁺-free medium or in a Ca²⁺-free medium containing 146 mM Na⁺ (Fig. 4). The cells were treated with thapsigargin for 20 s to raise [Ca²⁺]_i. In control cells, the rate of Na⁺_o-independent ⁴⁵Ca²⁺ efflux was 0.22 ± 0.03 nmol/mg/20 s (n = 4), whereas the rate of Na⁺_o-dependent ⁴⁵Ca²⁺ efflux was 0.70 ± 0.13 nmol/mg/20 s (n = 4), as estimated by subtracting ⁴⁵Ca²⁺ efflux in the absence of Na⁺_o from that in the presence of Na⁺_o. No.7943 at 10 and 30 µM did not affect Na⁺_o-independent ⁴⁵Ca²⁺ efflux but decreased the rate of Na⁺_o-dependent ⁴⁵Ca²⁺ efflux by 7 and 38%, respectively. This weak inhibition could not be increased by preincubation of cells with No.7943 for up to 30 min (data not shown). Thus only high doses of No.7943 inhibited Na⁺_o-dependent ⁴⁵Ca²⁺ efflux.

We studied the effect of No.7943 on the time course of Na⁺_o-induced [Ca²⁺]_i decline in smooth muscle cells under conditions where the sarcolemmal and SR Ca²⁺ pumps were inhibited by thapsigargin (1 µM) and a Ca²⁺- and Na⁺-free, high pH (pH 8.8) medium containing 20 mM MgCl₂ (24, 35). Under these conditions, [Ca²⁺]_i increased to a relatively high level (700-800 nM), and we observed spontaneous decline of [Ca²⁺]_i (30 ± 3 nM/10 s, n = 4) (Fig. 5). Addition of 50 mM Na⁺_o accelerated the [Ca²⁺]_i decline with a resultant initial rate of 157 ± 7 nM/10 s (n = 4), whereas addition of 50 mM choline chloride had no effect. No.7943 at 10 and 30 µM decreased the Na⁺_o-dependent portion of [Ca²⁺]_i decline by 12 ± 5 and 37 ± 2% (n = 4), respectively, although the same concentrations of this agent did not affect the background [Ca²⁺]_i decline (Fig. 5).

The inhibitory effect of No.7943 on Na⁺/Ca²⁺ exchange was studied using cardiac sarcolemmal vesicles. As in cells, No.7943 completely and dose-dependently inhibited the initial rate of Na⁺_i-dependent Ca²⁺ uptake into Na⁺-loaded sarcolemmal vesicles (IC₅₀ = 5.4 ± 0.3 µM (n = 3)) (Fig. 6A). Eadie-Hofstee plots of the rate of Na⁺_i-dependent Ca²⁺ uptake versus extravesicular [Ca²⁺]_i (Fig. 6B) revealed that 5 µM No.7943 decreased the V_max to 11 ± 0.2 nmol/mg/1.5 s (n = 3) from the control value of 22 ± 0.6 nmol/mg/1.5 s (n = 3), whereas it did not affect the K_m for Ca²⁺ (35 ± 4.3 µM for control and 34 ± 2.8 µM for the presence of the agent). The results indicate that inhibition is noncompetitive with respect to Ca²⁺. In contrast, inhibition by XIP was incomplete (about 70%) with an IC₅₀ value of 1.0 ± 0.1 µM (n = 3) (Fig. 6B). 3,4-Dichlorobenzamil, on the other hand, completely inhibited Na⁺_i-dependent Ca²⁺ uptake with an IC₅₀ of 19 ± 0.7 µM (n = 3). Absence of the effect on the K_m for extravesicular Ca²⁺ was also seen with 1 µM XIP (Fig. 6B) or 20 µM 3,4-dichlorobenzamil (data not shown).

Fig. 6C shows that No.7943 inhibits Na⁺_o-dependent ⁴⁵Ca²⁺ efflux from cardiac sarcolemmal vesicles that had been preloaded with ⁴⁵Ca²⁺ by Na⁺_i-dependent Ca²⁺ uptake. The IC₅₀ values for the inhibition by this agent were 13 ± 4.0, 12 ± 2.3, and 11 ± 0.6 µM (n = 3) in the presence of 10, 20, and 40 mM [Na⁺]_o, respectively. Thus, the inhibitory effect of No.7943 on Na⁺_o-dependent Ca²⁺ efflux from sarcolemmal vesicles was not affected by a change in [Na⁺]_o.

We studied the effects of No.7943 on other ion transport systems. As shown in Table I, 10 µM No.7943, which inhibited Na⁺_i-dependent ⁴⁵Ca²⁺ uptake by about 90% (Fig. 2A), did not significantly influence Na⁺/H⁺ exchange, DHP-sensitive ⁴⁵Ca²⁺ uptake, passive ²²Na⁺ uptake, sarcolemmal and SR Ca²⁺-ATPases, and Na⁺,K⁺-ATPase. On the other hand, 30 µM No.7943 inhibited only the DHP-sensitive ⁴⁵Ca²⁺ uptake by 35 ± 2% (n = 3, p < 0.05). We also tested the effect of No.7943 on the action potential parameters of guinea pig papillary muscle. Treatment of muscle preparations with up to 10 µM No.7943 for 30 min did not significantly affect the resting membrane potential (88 ± 1.6 mV in control versus 87 ± 3.3 mV (n = 6) with No.7943), action potential amplitude (107 ± 2.2 mV in control versus 106 ± 2.6 mV (n = 6) with No.7943), the maximum rate of rise of action potential (_max) (174 ± 8.5 V/s in control versus 148 ± 3.8 V/s (n = 6) with No.7943), or the action potential duration at 90% repolarization (144 ± 8.2 ms in control versus 156 ± 9.1 ms (n = 6) with No.7943). However, 30 µM No.7943 decreased the _max by 28 ± 4% (n = 6, p < 0.05). Since _max reflects activity of cardiac voltage-gated Na⁺ channels (33), No.7943 at 30 µM appears to inhibit the Na⁺ channels. All these results support the notion that the inhibition by low concentrations (0.3-10 µM) of No.7943 is selective.

Table I. Effect of No. 7943 on ion transporters Activities were measured as described under ``Experimental Procedures.'' For the cell experiments, cells were preincubated with No.7943 for 5 min before the activity measurement. Results are shown as means ± S.E. of three or four determinations. No.7943 Control 10 µM 30 µM Na+/H+ exchange (nmol/mg/min) 6.1  ± 0.3 6.0  ± 0.2 5.9  ± 0.1 DHP-sensitive 45Ca2+ uptake (nmol/mg/2 min) 2.9  ± 0.3 2.5  ± 0.3 1.9  ± 0.1a Passive 22Na+ uptake (nmol/mg/30 min) 310  ± 8.2 313  ± 18 293  ± 4.9 Sarcolemmal Ca2+-ATPase (nmol/mg/20 min) 191  ± 14 203  ± 4.6 171  ± 14 SR Ca2+-ATPase (nmol/mg/min) 828  ± 8.2 871  ± 16 771  ± 28 Na+,K+ATPase (µmol/mg/h) 68  ± 4.1 67  ± 2.0 64  ± 2.5 a  p < 0.05 compared with control.

In untreated fura-2-loaded neonatal cardiomyocytes, small transient increases in [Ca²⁺]_i were induced by spontaneous beating (Fig. 7A). Diastolic [Ca²⁺]_i and the rate of spontaneous beating in these cells were 128 ± 11 nM and 15 ± 5 beats/min (n = 3), respectively. No.7943 at up to 10 µM did not significantly change either diastolic [Ca²⁺]_i or the rate of spontaneous beating. However, 30 µM No.7943 increased diastolic [Ca²⁺]_i by 31 ± 4% (n = 4) and caused cessation of spontaneous beating (Fig. 7A). Under identical conditions, application of 10 µM verapamil or 1 µM nicardipine induced cessation of spontaneous beating without affecting diastolic [Ca²⁺]_i (data not shown). Of note, after the treatment with verapamil, No.7943 (30 µM) failed to induce an increase in diastolic [Ca²⁺]_i, suggesting that the [Ca²⁺]_i increase may be due to both the continued Ca²⁺ influx via verapamil-sensitive Ca²⁺ channels and inhibition of Ca²⁺ extrusion via Na⁺/Ca²⁺ exchanger.

Ca²⁺_o repletion after a period of Ca²⁺_o depletion is known to cause Ca²⁺ overloading of cardiomyocytes and finally cell death, a process called the Ca²⁺ paradox (11). Ca²⁺ overloading in the Ca²⁺ paradox is considered to be due to Ca²⁺ influx via the reverse mode of Na⁺/Ca²⁺ exchange (11). We tested the effect of No.7943 on this experimental model. When cardiomyocytes were treated with a Ca²⁺- and Mg²⁺-free medium for 10 min and then placed in a normal medium containing 1 mM Ca²⁺, [Ca²⁺]_i increased rapidly and markedly in the absence of No.7943, reaching a level of >3 µM (Fig. 7B). In contrast, this increase in [Ca²⁺]_i was remarkably suppressed by 72 ± 15 and 92 ± 7% (n = 3) in the presence of 3 and 10 µM No.7943, respectively (Fig. 7B). Verapamil (10 µM), however, did not prevent this Ca²⁺ overloading (data not shown). Furthermore, during Ca²⁺_o repletion in the presence of 10 µM No.7943, cardiomyocytes showed no evidence of structural change such as the development of hypercontracture, which was characteristically observed in cells not treated with No.7943 under these conditions (data not shown).

DISCUSSION

In this work, we show that No.7943, a new derivative of isothiourea, potently inhibits the Na⁺_i-dependent Ca²⁺ influx (⁴⁵Ca²⁺ uptake and [Ca²⁺]_i increase) into rat cardiomyocytes, rat aortic smooth muscle cells, and cardiac NCX1-transfected cells (IC₅₀ = 1.2-2.4 µM) (Figs. 2 and 3). In contrast, the same agent produced only a weak inhibitory effect on the Na⁺_o-dependent Ca²⁺ efflux (⁴⁵Ca²⁺ efflux and [Ca²⁺]_i decrease) even at a high concentration (Figs. 4 and 5). These effects of No.7943 are most likely to reflect a direct action on the Na⁺/Ca²⁺ exchanger, because of the following. (i) In cardiomyocytes and smooth muscle cells, both Na⁺_i-dependent increase and Na⁺_o-dependent decrease in [Ca²⁺]_i are due to the activity of the Na⁺/Ca²⁺ exchanger (1). The IC₅₀ values for No.7943 obtained here by [Ca²⁺]_i measurements (1.2-1.6 µM) were almost identical to those from the ⁴⁵Ca²⁺ flux measurements (2.0-2.4 µM) (Figs. 2A and 3, and see also ``Results''). (ii) Na<sup>+</sup><sub>i</sub>- and Na<sup>+</sup><sub>o</sub>-induced changes in both <sup>45</sup>Ca<sup>2+</sup> fluxes and [Ca<sup>2+</sup>]<sub>i</sub> were observed in NCX1-transfected but not in nontransfected CCL39 cells<sup>2</sup> (23). No.7943 inhibited the Na<sup>+</sup><sub>i</sub>-dependent <sup>45</sup>Ca<sup>2+</sup> uptake into NCX1-transfected CCL39 cells (Fig. 2A) but not <sup>45</sup>Ca<sup>2+</sup> uptake into nontransfected CCL39 cells (see ``Results''). (iii) No.7943 potently inhibited both Na⁺_i-dependent Ca²⁺ uptake into and Na⁺_o-dependent Ca²⁺ efflux from sarcolemmal vesicles (IC₅₀; 5.4 and 11 to 13 µM, respectively) (Fig. 6, A and C). In intact cells, therefore, No.7943 exerts a much greater inhibitory effect on the Na⁺/Ca²⁺ exchanger operating in the reverse mode as compared with the exchanger operating in the forward mode. Furthermore, there is no difference in the inhibitory potency of No.7943 in different exchanger isoforms from cardiac and smooth muscle cells.

No.7943 is an amphiphilic molecule with an isothiourea group whose pK_a is about 10. Thus this agent is protonated and cationic in most conditions (see Fig. 1). This positive charge on the isothiourea moiety seems to be essential for inhibitory activity, as its deletion renders this agent much less active.³ No.7943 at up to 100 µM is soluble in aqueous buffers. Interestingly, inhibition by No.7943 was easily abolished by washing cells with fresh medium for 1 min (Fig. 3), indicating that the effect is completely reversible and that removal of the agent is relatively rapid. Furthermore, the inhibitory potency of No.7943 was almost identical with or without prior preincubation (see ``Results''). We therefore conclude that the agent primarily acts from the extracellular side of the plasma membrane in intact cells under the conditions used in this study. It should be noted, however, that No.7943 potently inhibited Na⁺_i-dependent Ca²⁺ uptake into cardiac sarcolemmal vesicles, when the latter was exposed to the agent for only 1.5 s (Fig. 6A). The agent thus appears able to inhibit exchange activity also from the cytoplasmic side of the membrane, because a majority of sarcolemmal vesicles used in this study seem to have had inside-out orientation as inferred from the extent of inhibition by XIP. XIP at 10 µM, which presumably is membrane-impermeable (17), inhibited Na⁺_i-dependent Ca²⁺ uptake into sarcolemmal vesicles by 70% (Fig. 6A).

We examined the kinetics of inhibition by No.7943 of the reverse (Ca²⁺ influx) mode of Na⁺/Ca²⁺ exchange with respect to either Ca²⁺_o or Na⁺_o concentration using intact cells (Fig. 2, B and C). The exchanger inhibition was noncompetitive with Ca²⁺_o, and inhibition of Na⁺_i-dependent ⁴⁵Ca²⁺ uptake by Na⁺_o was not influenced by 1 or 3 µM No.7943. Similarly, No.7943 inhibited Na⁺_i-dependent ⁴⁵Ca²⁺ uptake into sarcolemmal vesicles by reducing V_max without affecting the K_m for extravesicular Ca²⁺ (Fig. 6B). It also did not influence [Na⁺]_o dependence of Na⁺_o-induced ⁴⁵Ca²⁺ efflux from vesicles that had been preloaded with ⁴⁵Ca²⁺ by Na⁺_i-dependent Ca²⁺ uptake (Fig. 6C). These vesicular exchange reactions (Na⁺_i-dependent ⁴⁵Ca²⁺ uptake and Na⁺_o-dependent ⁴⁵Ca²⁺ efflux) are mostly due to inside-out vesicles (17) and presumably equivalent to the forward and reverse modes of the exchange in intact cells. From all these results, we conclude that No.7943 does not affect the interaction of transport sites on the exchanger with Ca²⁺ or Na⁺ on either side of the plasma membrane and thus probably acts at site(s) distinct from these sites.

It is striking that in intact cells the potency of No.7943 as a blocker of the reverse mode of Na⁺/Ca²⁺ exchange is 15-25-fold greater compared with that for the forward mode. In contrast, a minimum difference was observed for the effect of No.7943 on the corresponding reactions in sarcolemmal vesicles (Fig. 6, A and C, and see above). Thus, the mode-specific inhibition by No.7943 was observed only in intact cells. Such a difference in the inhibitory pattern may be consistent with our view that the agent acts on the exchanger primarily from the extracellular side in intact cells, whereas it acts mainly from the cytoplasmic side in sarcolemmal vesicles. At present, however, we have no information about the mechanism by which this agent causes such a mode-specific inhibition of Na⁺/Ca²⁺ exchange in intact cells.

No.7943 at 0.3-10 µM, while blocking the Na⁺/Ca²⁺ exchanger-mediated Ca²⁺ influx into cells, did not significantly affect activities of other ion transporters such as Na⁺/H⁺ exchanger, DHP-sensitive Ca²⁺ channels, sarcolemmal and SR Ca²⁺ATPases, and Na⁺,K⁺-ATPase, as well as passive Na⁺ permeability (Table I and Fig. 4). In addition, the same concentration range of No.7943 did not significantly alter the action potential parameters such as resting membrane potential, action potential amplitude, the maximum rate of rise of action potential (_max), and action potential duration at 90% repolarization (see ``Results''). However, No.7943 at a high concentration (30 µM) reduced the activities of voltage-dependent Na⁺ channels (measured as _max) and DHP-sensitive Ca²⁺ channels, as well as the forward mode of the Na⁺/Ca²⁺ exchange. In cultured cardiomyocytes, No.7943 at up to 10 µM affected neither diastolic [Ca²⁺]_i nor spontaneous beating (Fig. 7A), the latter being abolished by the Ca²⁺ channel antagonist verapamil. In contrast, 30 µM No.7943 significantly increased the resting [Ca²⁺]_i (Fig. 7A). Thus, inhibition of Ca²⁺ influx via the Na⁺/Ca²⁺ exchanger by low concentrations of No.7943 has virtually no effect on Ca²⁺ mobilization and spontaneous beating in cultured cardiomyocytes. On the other hand, inhibition of Ca²⁺ extrusion via the Na⁺/Ca²⁺ exchanger by a high concentration of this agent causes an increase in resting [Ca²⁺]_i, probably due to the continued influx of Ca²⁺ via verapamil-sensitive Ca²⁺ channels. All these results indicate that No.7943 at relatively low concentrations is a selective inhibitor of the Na⁺/Ca²⁺ exchanger that only minimally affects cell ion metabolism. In this sense, No.7943 clearly is much superior to 3,4-dichlorobenzamil whose specificity is low (see the Introduction).

We explored the therapeutic potential of No.7943 by using the Ca²⁺ paradox model (Fig. 7B). The Ca²⁺ paradox has been studied as an experimental model for Ca²⁺ overloading in cardiomyocytes during the ischemia-associated reperfusion. The Na⁺/Ca²⁺ exchanger operating in the Ca²⁺ influx mode has been implicated in this mechanism. The same mode of the Na⁺/Ca²⁺ exchange is also considered to be responsible for Ca²⁺ overloading during reoxygenation after cardiac hypoxia (11, 12). We found that low concentrations of No.7943 were very effective in preventing Ca²⁺ influx into cardiomyocytes and the resultant structural change under Ca²⁺ paradox conditions (Fig. 7B and see ``Results''). Importantly, this agent at the same low concentrations (up to 10 µM) also effectively blocks mechanical dysfunction of isolated perfused rat hearts that is caused by the ischemia/reperfusion or by hypoxia/reoxygenation insult, whereas it had no effect on mechanical function of normal rat hearts⁴ (36). Thus, No.7943 could have a therapeutic potential as a selective blocker of excessive Ca²⁺ influx mediated via the Na⁺/Ca²⁺ exchanger under pathological conditions, which include cardiac ischemia/reperfusion, hypoxia/reoxygenation, and possibly some forms of essential hypertension.