INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

UNEQUAL DISTRIBUTION of potassium and sodium channels contributes to transmural electrical heterogeneity and thus to the development of a variety of cardiac arrhythmias (for a review, see Ref. 1). Transmural and interventricular differences have been described in the transient outward potassium conductance (I_to), the delayed rectifier potassium conductance (I_Ks), and the sustained component of the sodium conductance (I_Na) (6, 9, 16-18). I_to is larger in the midmyocardium and epicardium than it is in the endocardium, contributing to the characteristic notch of canine epicardial and midmyocardial action potentials and the absence of an endocardial notch (18). Similarly, a deeper notch recorded from the right versus left epicardium of dogs correlates with a greater I_to in right epicardial myocytes (6, 29). Cardiac action potentials of human and canine midmyocardial cells exhibit a pronounced prolongation at slow stimulation rates (1, 16, 18, 22, 23), and drugs that block components of the delayed rectifier or slow the decay of I_Na preferentially lengthen midmyocardial action potential duration (APD) (21, 23, 31). Voltage-clamp studies have revealed that a smaller I_Ks and a larger, late I_Na contribute to the longer APD of canine midmyocardial cells (9, 17).

The extent to which the distribution of electrogenic exchangers contribute to transmural heterogeneity is unknown. Sicouri and Antzelevitch (22, 23) found that calcium overload-induced delayed afterdepolarizations arise from a unique population of canine midmyocardial cells, the M cells. An important contributor to early and delayed afterdepolarizations in canine ventricle is the sodium-calcium exchange current (I_NaCa) (28, 34). The present study investigates the density of the sodium-calcium exchanger across the canine left ventricular wall. Some of this work has been reported in abstract form (33).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Adult male mongrel dogs were given 200 IU/kg heparin (sodium salt) and anesthetized with 35 mg/kg intravenous pentobarbital sodium, and their hearts were quickly removed and placed in Tyrode solution. Single myocytes were obtained by enzymatic dissociation from a wedge-shaped section of the ventricular free wall supplied by the left circumflex coronary artery (32). Cells from the epicardial, midmyocardial, and endocardial regions of the left ventricle were utilized in this study. All procedures followed were in accordance with guidelines established by the Institutional Animal Care and Use Committee.

Tyrode solution used in the dissociation contained (in mM) 135 NaCl, 5.4 KCl, 1 MgCl₂, 0 or 0.5 CaCl₂, 10 glucose, 0.33 NaH₂PO₄, and 10 HEPES, and pH was adjusted to 7.4 with NaOH. Both amphotericin B perforated-patch and standard whole cell voltage-clamp techniques were utilized to record I_NaCa. External solution was potassium free and contained 6 mM chloride, and its composition was (in mM) 0.005 ouabain, 2 CaCl₂, 10 glucose, 1 MgCl₂, 140 sodium methanesulfonate, and 10 HEPES, and pH was adjusted to 7.4 with methanesulfonic acid. When sodium-free solution was required, equimolar lithium was substituted for external sodium. Ouabain was included in the external solution because of a concern that the electrogenic sodium pump was not blocked in potassium-free solution (10, 11). Internal solutions were potassium free and contained 4 mM chloride. Pipette solution for perforated-patch experiments contained (in mM) 2.6 × 10⁴ amphotericin B, 150 cesium aspartate, 10 NaOH, 10 HEPES, 1 MgCl₂, and 0.01 CaCl₂, and pH was adjusted to 7.1 with CsOH. Pipette solution for standard whole cell voltage clamp contained (in mM) 150 cesium aspartate, 10 NaOH, 10 HEPES, 1 MgCl₂, 0.1 EGTA, and 5 MgATP, and pH was adjusted to 7.1 with CsOH. Sodium was replaced by equimolar cesium when required.

Even in potassium-free solutions, substitution of external sodium by lithium caused an outward sodium-pump current that might contaminate measurement of reverse-mode I_NaCa (see Fig. 6). In six cells we determined that 5 µM ouabain was sufficient to completely block the sodium pump, allowing an uncontaminated evaluation of I_NaCa. External solution for experiments designed to determine the appropriate concentration of ouabain to block the sodium pump contained (in mM) 2 BaCl₂, 0 or 4 potassium methanesulfonate, 10 glucose, 1 MgCl₂, 140 sodium methanesulfonate, and 10 HEPES, and pH was adjusted to 7.4 with methanesulfonic acid. Pipette solution for sodium pump experiments contained (in mM) 140 potassium aspartate, 15 NaOH, 10 EGTA, 10 HEPES, 1 MgCl₂, and 5 MgATP, and pH was adjusted to 7.1 with KOH.

Amphotericin B (Sigma Chemical) was made in DMSO (60 mg/ml) and diluted (1:250) into pipette solution to a final concentration of 240 µg/ml. Fresh dilutions into pipette solution were made every 2 h. Ouabain was made as concentrated stocks in water and diluted (1:1,000) into external solution. Final concentration was 5 µM ouabain. Amphotericin B and ouabain were used in a darkened room.

Voltage-clamp protocols were preceded by a train of ten 200-ms pulses to 0 mV delivered at a rate of 1 Hz, followed by a rest of 2 s to maintain calcium loading of the sarcoplasmic reticulum (SR). We have assumed that this protocol resulted in a maximal SR load because stimulation beyond the initial six beats did not increase peak contraction. To trigger I_NaCa by means of the normal calcium transient, we held perforated-patch voltage-clamped cells at a potential of 80 mV before evoking a 5-ms pulse to 50 mV to inactivate I_Na and a 3-ms pulse to 0 mV to activate calcium current (I_Ca) and a calcium transient. This two-step protocol was immediately followed by a pulse to 80, 50, 20, or 50 mV to record I_NaCa. To separate I_NaCa from I_Ca and the capacitance spike, we utilized a train of 3-ms pulses to 0 mV to empty the SR of calcium, with the final 3-ms pulse taken 250 ms before the two-step protocol was repeated (5). Currents remaining after inhibition of the calcium transient were subtracted from those elicited by a normal calcium transient. I_NaCa was characterized as the calcium transient-induced inward difference current (see Fig. 1) and quantified as either total charge transported or peak inward current normalized to cell capacitance.

To trigger I_NaCa without sequential activation of I_Ca and the calcium transient, reverse-mode I_NaCa was evoked by rapid substitution of external sodium with equimolar lithium. Reverse-mode I_NaCa was measured as the peak sodium substitute-induced current at 80 mV, after it was first confirmed that this outward current was completely blocked by 5 mM NiCl₂ (see Fig. 6). These experiments required the addition of 100 µM EGTA to the pipette solution and the use of the standard whole cell voltage-clamp technique. Cells vigorously contracted when depolarized to 0 mV from a holding potential of 80 mV under these conditions.

Dissociated cells were placed in a temperature-controlled 0.5-ml chamber (Medical Systems, Greenvale, NY) on the stage of an inverted microscope and superfused at 2 ml/min. A four-barrel quartz micromanifold (ALA Scientific Instruments, Westbury, NY) was used to exchange the solution immediately surrounding voltage-clamped myocytes. This micromanifold was placed 100 µm from the cell, and flow was controlled by a pinch valve and computer interface (model BPS-4; ALA Scientific Instruments). An Axopatch 200A amplifier (Axon Instruments, Foster City, CA) was operated in voltage-clamp mode to record whole cell currents at 37°C. Cell capacitance was 191 ± 7 pF (54 cells). Pipette tip resistance was 1.0-2.3 M, and seal resistance was 5-11 G. Electronic compensation of series resistance averaged 72 ± 2% (54 cells), and the series resistance remaining after this compensation averaged 2.73 ± 0.26 M. After series resistance compensation, capacitive current decayed with a single time constant of 467 ± 35 µs (54 cells).

Tip potential of standard electrodes was measured using established techniques and, because of low internal chloride concentration, averaged 13 mV. To determine tip potential for perforated-patch electrodes, cells were depolarized in 145 mM KCl, calcium-free external solution, and comparisons were made between the potential recorded by perforated-patch electrodes and those recorded by 3 M KCl-filled microelectrodes in similarly treated cells. Tip potential of perforated-patch electrodes averaged 19 mV. Voltages reported in the text were corrected for these potentials. After tip potential compensation, the peak of the current-voltage relation for calcium channels was 0 mV, regardless of whether the perforated-patch or standard whole cell technique was utilized. The seal between the cell membrane and patch pipette was initially formed in Tyrode solution containing 1 mM CaCl₂. A 3 M KCl-agar bridge was made use of between the Ag/AgCl ground electrode and external solution to avoid development of a ground potential when switching to experimental solutions.

Whole cell currents and transmembrane potentials were filtered with a four-pole low-pass Bessel filter at 5 kHz, digitized between 1 and 10 kHz (Digidata 1200A, Axon Instruments), and stored on a computer. Significant differences among means were determined by an unpaired Student's t-test. Unloaded cell shortening was adopted as a relative measure of the calcium released from the SR and was recorded with a video-edge motion detector (model VED 104; Crescent Electronics, Sandy, UT) coupled with a Phillips type FTM800NH/HGI camera operating at a 60-Hz scan rate. The single-ended output of this detector was linear over the range of 0-40 µm. Clampex 8 acquisition software (Axon Instruments) was employed to concurrently record cell shortening and ionic current.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We first sought a benign method to separate I_NaCa from ionic and capacitive currents that does not require the application of nonselective blockers such as nickel or the attendant calcium-loading of cells in sodium-free external solution. In dog myocytes, the calcium-activated chloride conductance [I_Cl(Ca)] and I_NaCa are the only two conductances activated by release of calcium from the SR (34). Figure 1 shows that a series of 3-ms depolarizing pulses eliminated an inward current identical to that abolished by rapid removal of external sodium, an established method of inhibiting I_NaCa (15). Currents were recorded using the perforated-patch voltage-clamp technique. In these and all subsequent experiments, I_Cl(Ca) was inhibited by reducing the extracellular chloride concentration to 6 mM and the intracellular concentration to 4 mM. Pipette solution was modified to eliminate intracellular sodium and reverse-mode I_NaCa. Figure 1, top, shows three superimposed currents following conditioning trains of either 200 or 3 ms. Currents were first recorded in normal external sodium. After release of calcium from the SR was triggered, an immediate return to 80 mV resulted in a transient inward current that decayed within 200 ms (control). Still in normal external sodium, the protocol was repeated after a series of 3-ms pulses emptied the SR of calcium. Both contraction and the transient inward current were abolished by this intervention (prepulse). Reverting to the original conditioning train of 200-ms pulses restored contraction and inward current to control levels. Finally, after a series of 200-ms pulses, a rapid change to sodium-free solution was made before the basic protocol was repeated. Despite a vigorous contraction, the transient inward current was abolished by removal of sodium (sodium free). Subtraction of the prepulse and sodium-free traces from the control trace yields the calcium transient-induced and sodium-sensitive currents, respectively. Figure 1, bottom, displays these two difference currents beginning 5 ms after the step back to 80 mV to allow for complete decay of the capacitive spike and deactivation of I_Ca. Similar results were obtained in five cells.

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Separation of sodium-calcium exchange current (INaCa) from capacitive and ionic currents. Top, inset: cells were voltage clamped using a conditioning train of either 200-ms pulses (protocol 1) or 3-ms pulses (protocol 2). Top: superimposed traces under 3 conditions. A slowly decaying inward current at 80 mV was evoked immediately following a 3-ms step to 0 mV to trigger a calcium transient (control). This current was abolished by reducing SR calcium stores and contraction (prepulse) or by rapid removal of external sodium before activation of normal contraction (sodium free). Bottom: superimposed difference currents beginning 5 ms after stepping back to 80 mV show that calcium transient-induced and sodium-sensitive inward currents are identical.

These results clearly identify the inward current measured at 80 mV as I_NaCa. We presume that inhibition of I_NaCa in normal sodium by the use of a train of brief pulses is secondary to attenuation of the calcium transient. We found in six cells that cell shortening was abolished by this same protocol of ten 3-ms pulses to 0 mV.

The method outlined above was employed to measure I_NaCa in epicardial, midmyocardial, and endocardial myocytes. Figure 2 indicates that I_NaCa is large in epicardial and midmyocardial cells and significantly smaller in endocardial cells. Currents were recorded using the perforated-patch technique and physiological levels of sodium to permit normal contraction (3, 27). In Fig. 2A, currents were recorded in a typical endocardial cell. The top trace shows the voltage template and current after a conditioning train of 200-ms pulses that maintained contraction. The middle trace shows the voltage template and current after a series of 3-ms pulses that inhibited I_NaCa. The bottom trace shows subtracted currents yielding the waveform of I_NaCa. I_NaCa was inward throughout its time course at this voltage. In Fig. 2B, I_NaCa was normalized by cell capacitance and plotted as a function of cell type. I_NaCa in endocardial cells was ~30% smaller than in epicardial or midmyocardial cells (P < 0.001). Net exchanger flux was 0.210 ± 0.009 pC/pF in 17 endocardial cells versus 0.316 ± 0.013 pC/pF in 18 midmyocardial cells and 0.293 ± 0.016 pC/pF in 18 epicardial cells.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   INaCa measured as net inward flux at 80 mV is smaller in endocardial cells than in midmyocardial and epicardial cells. A, top: voltage template and current after a conditioning train of 200-ms pulses. A slowly decaying inward current is clearly evident after stepping back to 80 mV. A, middle: voltage template and current after a series of 3-ms pulses. A, bottom: subtracted currents yield INaCa, commencing 5 ms after stepping back to 80 mV. B: inward ionic flux was normalized by cell capacitance and plotted for endocardial (Endo; n = 17 cells), midmyocardial (Mid; n = 18 cells), and epicardial cells (Epi; n = 18 cells). Data are means ± SE.

The ability to decide whether I_NaCa varies in the three cell types requires extension of these measurements to additional voltages and evidence that difference currents are uncontaminated by other conductances over this extended range of potentials. Because generation of difference currents requires inhibition of the calcium transient, our concern was that reduction of the calcium transient might affect other currents in addition to I_NaCa, thus contaminating the difference current. To validate this technique over a range of voltages, we compared difference currents acquired in the presence of I_NaCa with those obtained after the substitution of external sodium eliminated all I_NaCa. Internal solution was sodium free to exclude any possibility that outward currents originated from reverse-mode I_NaCa. In Fig. 3A, a transient inward difference current was obtained in normal external sodium at 50 mV in a midmyocardial cell. The external solution was then rapidly changed to one in which sodium was replaced by lithium, and the protocol was repeated to obtain a second difference current. I_NaCa was abolished, and the remaining currents were unaffected by conditioning train pulse duration, as was evident from the insignificant difference current. In Fig. 3B, subtracted currents for the same midmyocardial cell are shown at 10 mV. The difference current in normal external sodium exhibited both outward and inward components at this voltage. Inhibiting I_NaCa abolished the inward component and left only a rapidly decaying outward difference current. Difference currents in the presence (control) and absence of inward I_NaCa (sodium free) for an endocardial cell at 20 mV, an epicardial cell at 50 mV, and a midmyocardial cell at 80 mV are shown in Fig. 3, C-E, respectively. At these voltages, difference currents in normal sodium were inward and nearly zero after the substitution of external sodium. These results are representative of traces obtained in 10 myocytes.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Difference currents over a range of voltages in normal external sodium and after change to a sodium-free external solution. A: a transient inward difference current obtained in normal external sodium (bottom trace) at 50 mV in a midmyocardial cell was superimposed on another (top trace) obtained in sodium-free solutions. B: subtracted currents for same midmyocardial cell as in A are shown at 0 mV. Difference current in normal external sodium (bottom trace) exhibited both outward and inward components at this voltage. Switching to sodium-free external solution and repeating protocol abolished inward component and left only a rapidly decaying outward difference current (top trace). Currents in the presence (control) and absence of inward INaCa (Na free) at 20, 50, and 80 mV are shown in C-E, respectively. At these potentials the difference current (top traces) after inhibition of INaCa was insignificant.

Two criteria must be met to prove that inhibition of the calcium transient affects only I_NaCa. First, difference currents in normal external sodium must contain solely an inward component, and second, switching to sodium-free external solution should result in a negligible difference current. These criteria were met for 80, 50, 20, and 50 mV, and we conclude that measurements of ionic flux and peak I_NaCa will be uncontaminated by other conductances at these voltages. Conversely, it would be inappropriate to present data for 10 mV because this difference current contains elements other than I_NaCa.

The current-voltage relation for I_NaCa in the three cell types is shown in Fig. 4. Protocol and solutions were the same as those used for experiments shown in Fig. 2. Because patch pipettes contained 10 mM sodium, there was an opportunity that difference currents would contain both forward- and reverse-mode exchange currents. Figure 4, top, shows endocardial, midmyocardial, and epicardial flux for selected potentials between 80 and 50 mV. I_NaCa at negative voltages was inward during the entire pulse but consisted of both outward and inward elements at 50 mV, as shown by the inset in Fig. 4, top. For the purposes of this summary plot, I_NaCa was depicted as the net inward flux denoted by the shaded area of this representative trace. Shown in Fig. 4, bottom, is the normalized peak inward I_NaCa. Endocardial I_NaCa was ~25% smaller (P < 0.01) compared with either epicardial or midmyocardial I_NaCa at all negative voltages.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Current-voltage relation for INaCa. Currents were recorded using perforated-patch technique with normal internal and external sodium. Top: inward flux normalized by cell capacitance for Endo, Mid, and Epi cells. Top, inset: at 50 mV INaCa was characterized as net inward flux, shown as shaded area. Bottom: normalized peak inward INaCa for Endo, Mid, and Epi cells.

Table 1 summarizes net inward flux in the three cell types. I_NaCa was similar in epicardial and midmyocardial cells (P > 0.05) at all voltages. Conversely, endocardial flux was ~30% smaller at all negative voltages (P < 0.01).

Activation of I_NaCa is dependent on the underlying calcium transient. To test whether a smaller endocardial I_NaCa ensued from less than maximal activation of the calcium transient, we investigated the effects of lengthening the 0-mV triggering pulse on contraction and I_NaCa. Figure 5 shows that increasing the triggering pulse duration did not systematically increase contraction or I_NaCa in the three cell types. Solutions were the same as those used for experiments shown in Figs. 2 and 4. I_NaCa was initially triggered following a standard 3-ms pulse to 0 mV. This pulse duration was then increased in 3-ms increments, and the protocol was repeated. Epicardial I_NaCa and contractions are shown in Fig. 5, A and B, respectively, following triggering pulses of 3-15 ms. In Fig. 5C, peak contractions following 6- to 15-ms pulses were compared with the peak contraction following a 3-ms pulse, and these contractile changes were plotted as a function of triggering pulse duration for the three cell types. Neither I_NaCa nor peak contraction was significantly affected by lengthening the triggering pulse duration. We conclude that a 3-ms triggering pulse is sufficient to maximally activate contraction and I_NaCa.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Effects of triggering pulse duration on cell shortening and INaCa. Whole cell currents were recorded utilizing perforated-patch technique to ensure normal excitation-contraction coupling. Epicardial INaCa and contractions are shown in A and B, respectively, following triggering pulses from 3 to 15 ms taken in 3-ms increments. C: summary plot of peak contraction as a function of triggering pulse duration for 4 Epi, Mid, and Endo cells. Contraction was expressed as percent deviation from peak contraction (%contractile change) after a 3-ms triggering pulse. Lengthening triggering pulse duration did not result in a consistent increase of peak contraction in any of the 3 cell types.

The results shown in Table 1 are consistent with a smaller endocardial exchanger density, but transmural differences in SR load might also contribute to this result. Activation of reverse-mode I_NaCa by removal of external sodium circumvents a need to trigger I_Ca and a normal calcium transient. The standard whole cell voltage-clamp technique was utilized to record reverse-mode I_NaCa. External and internal solutions were potassium free, and cells were held at 80 mV. A train of ten 200-ms pulses to 0 mV triggered vigorous contractions before each protocol. Figure 6 clearly shows that substitution of external sodium with lithium activates both reverse-mode I_NaCa and the electrogenic sodium pump. A 3-s application of sodium-free solution beginning at the time indicated by the arrow produced an outward current (Fig. 6, control) that decayed to baseline in normal sodium. This application of sodium-free solution was repeated after the addition of 5 µM ouabain to inhibit the sodium pump and in the presence of both 5 µM ouabain and 5 mM nickel. Ouabain reduced, but did not completely abolish, the sodium substitute-induced outward current (Fig. 6, ouabain). This ouabain-insensitive outward current was blocked by 5 mM nickel, an inhibitor of I_NaCa (Fig. 6, ouabain + nickel). Nickel blocked the ouabain-insensitive outward currents in five additional cells. Inhibition by nickel did not result from rundown of the current, because repeated administration of sodium-free solution in the absence of ouabain or nickel produced highly reproducible outward currents.

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Substitution of external sodium activated both reverse-mode INaCa and the electrogenic sodium pump. Cells were held at 80 mV. Sodium-free solution was applied at time indicated by arrow and continued for a period of 3 s. Substitution of external sodium produced an outward current (control) that decayed to baseline in normal sodium. Ouabain (5 µM) inhibited the sodium pump and significantly reduced outward current (ouabain) caused by a second application of sodium-free external solution. Outward current during a third application of sodium-free solution was completely blocked by 5 mM nickel (ouabain + nickel).

We confirmed that 5 µM ouabain was sufficient to completely block the sodium pump in canine ventricular cells and that 5 mM nickel had no effect on this current. Whole cell currents were recorded utilizing the standard patch-clamp technique. External solution was potassium free and contained 2 mM BaCl₂ to block the inwardly rectifying potassium conductance (I_K1) (see METHODS). Internal solution contained potassium and 15 mM sodium. Cells were held at 80 mV. Rapid addition of 4 mM potassium to the external solution caused an outward current that decayed to baseline when external potassium was removed. In six cells, 5 µM ouabain completely blocked the potassium-induced current, whereas this current was unaffected by 5 mM nickel.

We conclude that, after proper precautions have been taken to block the sodium pump, substitution of external sodium evokes only I_NaCa. We compared I_NaCa evoked by sodium substitution with that elicited by the triggering of a normal calcium transient. Cells were held at 80 mV. External solution contained normal sodium and 5 µM ouabain. The standard whole cell voltage-clamp technique was used to record peak outward I_NaCa induced by a 3-s pulse of sodium-free external solution. Figure 7 shows the density of I_NaCa normalized by cell capacitance across the left ventricular wall. Peak I_NaCa was 0.74 ± 0.04, 0.57 ± 0.04, and 0.50 ± 0.03 pA/pF, respectively, in 22 midmyocardial, 13 epicardial, and 12 endocardial cells. Epicardial and endocardial I_NaCa were of similar amplitude, and both were significantly smaller than midmyocardial I_NaCa (P < 0.001).

View larger version (8K): [in this window] [in a new window]   Fig. 7.   Peak reverse-mode INaCa in the 3 cell types. Maximal outward current during a 3-s pulse of sodium-free solution was compared with baseline current before sodium substitution. Peak currents shown are normalized by cell capacitance for 13 Epi, 22 Mid, and 12 Endo cells.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We are the first to report unequal distribution of I_NaCa across the canine left ventricular wall. Although comparisons among epicardial, midmyocardial, and endocardial cells were largely independent of whether peak current or net inward charge was used as a measure of I_NaCa, the relative amplitudes of I_NaCa were sensitive to the protocol utilized to evoke the exchanger.

Sequential activation of calcium channels and the calcium transient resulted in epicardial and midmyocardial I_NaCa of similar amplitudes and, by comparison, a significantly smaller endocardial I_NaCa. Efforts were made to fully load the SR with calcium to provide a brief but maximal trigger for calcium release and to avoid rundown of the calcium current. Such efforts, however, do not assure that SR load is uniform across the ventricular wall, and these experiments alone cannot distinguish whether transmural differences in I_NaCa result from a smaller density of sodium-calcium exchangers, nonuniform SR loading, or some factor limiting SR calcium release in the vicinity of endocardial sodium-calcium exchangers.

To address these issues, we triggered reverse-mode I_NaCa to measure exchanger density independent of normal excitation-contraction coupling and any transmural differences in SR loading. This is a common approach to the study of I_NaCa, and an unanticipated finding was that sodium removal activated both I_NaCa and the sodium pump. After the sodium pump was blocked, substitution of sodium resulted in a monophasic increase of outward current to a new steady state and concomitant accumulation of intracellular calcium. Although internal calcium had to increase with removal of external sodium, outward current did not decay because the normal effect of internal calcium to influence the reversal of the exchanger was abolished in the absence of external sodium. Regardless of the concentration of calcium inside the cell, forward-mode exchange was effectively blocked by the lack of external sodium.

Interestingly, internal solution containing 10 mM EGTA completely inhibited sodium substitution-induced I_NaCa, suggesting that some level of internal calcium is required to support reverse-mode exchange in canine myocytes. Modulation of reverse-mode sodium-calcium exchange by calcium has also been reported in squid giant axon and guinea pig ventricular myocytes (7, 12). In the present study discrepancies in internal calcium among the three cell types were minimized by the use of an internal solution containing 100 µM EGTA, a concentration that supported contraction in epicardial, midmyocardial, and endocardial cells. Moreover, activation of I_NaCa was preceded by a train of conditioning pulses to 0 mV, and peak reverse-mode I_NaCa was unaffected by varying the conditioning pulse duration between 3 and 200 ms, suggesting that internal calcium could be altered without unduly influencing the outcome. We conclude that peak reverse-mode I_NaCa is an accurate measure of sodium-calcium exchanger density and that this density is greater in midmyocardial cells.

Accurate measurements of I_NaCa are dependent on selectively separating I_NaCa from capacitive and ionic currents. We wanted to avoid characterizing I_NaCa as the nickel- or sodium-sensitive current because nickel blocks calcium current and sodium removal causes loading of the cell with calcium if intracellular sodium is normal (30, 34). In potassium-free, very low chloride-concentration solutions containing ouabain, it is evident that removing external sodium without changing voltage evokes only reverse-mode sodium-calcium exchange. It is more problematic to successfully isolate forward-mode I_NaCa over a range of voltages following a 3-ms pulse to 0 mV. Dog ventricular myocytes lack a calcium-activated, nonselective cation conductance, and I_NaCa is the only remaining calcium-activated conductance when chloride has been drastically reduced (34). Under our conditions, any method of eliminating the calcium transient should produce a difference current that accurately reflects the time course of I_NaCa at potentials negative to the calcium-channel threshold. We utilized a train of brief pulses to inhibit the calcium transient primarily because complete recovery of I_NaCa was rapid, permitting a larger quantity of data to be gathered from each cell. Figure 1 clearly shows that difference currents obtained in this fashion reflect the time course of I_NaCa, and Fig. 3 shows that this relation holds for voltages negative to the calcium-channel threshold. This method of isolating I_NaCa is not selective at a voltage at which significant calcium current is present, but it also would not be accurate to define I_NaCa as the nickel-sensitive current under this circumstance.

Unloaded cell shortening was adopted as a proportionate measure of calcium released from the SR on the basis of previously published results of simultaneous measurements of calcium using fluorescent indicators and optical measurements of contraction. In rat myocytes the calcium signal and contraction responded in parallel to increased external calcium, and feline myocytes stimulated after a rest produced a positive staircase in both the calcium transient and twitch contraction (8, 25). When assessing the amplitude of SR calcium release, we chose to measure contraction rather than calcium transients because calcium indicator dyes buffer intracellular calcium. In addition, use of the perforated-patch technique requires cell-permeable calcium indicators, leading to the trapping of dye in subcellular organelles and erroneous signals (8, 25). Moreover, a dissociation between the time course of the calcium signal and that of calcium-dependent conductances has been demonstrated in cardiac myocytes, suggesting that the use of a calcium indicator in the present study would not have measured subsarcolemmal calcium seen by sodium-calcium exchangers (20, 24, 26). Epicardial, midmyocardial, and endocardial recordings were made under identical conditions, and lack of an absolute measure of subsarcolemmal calcium should not negate conclusions regarding transmural differences in I_NaCa.

Interventions that inhibit I_NaCa or drastically shift the reversal potential of the sodium-calcium exchanger to a more negative voltage cause early repolarization of atrial and ventricular action potentials (2, 13, 14, 19). Janvier et al. (14) found that I_NaCa and I_Ca contribute equally to supply inward current to maintain the plateau of the ferret ventricular action potential. If I_NaCa serves a similar role in the canine ventricle, transmural differences in I_NaCa must be reflected in changes in action potential shape across the ventricular wall, although our data suggest that we must separately consider contributions of I_NaCa during normal activation and after spontaneous release from the SR in calcium overloaded cells.

A larger I_NaCa at voltages negative to 20 mV in midmyocardial and epicardial cells will counter repolarization and increase APD, and this tendency will be stronger if transmural differences are maintained at more positive plateau potentials of ~0 mV. APD will of course be dependent on a number of ionic conductances. Although I_NaCa is similar in epicardial and midmyocardial cells, a smaller I_Ks and a larger I_Na in midmyocardial cells will promote a longer M cell action potential (9, 17). Similarly, we predict that interventions such as postextrasystolic potentiation that transiently elevate intracellular calcium will preferentially increase epicardial and midmyocardial I_NaCa but that frequency-dependent effects on a host of ionic currents will ultimately determine APD.

In calcium-overloaded cells, spontaneous release of calcium from the SR occurs without a need to trigger calcium channels. I_NaCa contributes to delayed afterdepolarizations and early afterdepolarizations caused by spontaneous release of calcium from the SR in calcium-overloaded canine ventricular cells (28, 34). A larger density of exchangers in midmyocardial cells will favor development of early and late afterdepolarizations under calcium loading conditions such as ischemia and will contribute to the preferential lengthening of midmyocardial action potentials and generation of early afterdepolarizations in response to accelerated stimulation rates (4). Findings of the present study are consistent with those in tissues in which calcium overload-induced delayed afterdepolarizations could only be recorded from M cells (22, 23).