INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

INTERVENTIONS THAT ELEVATE intracellular calcium cause normal repolarization of cardiac action potentials to be followed by oscillations in membrane potential. A calcium-activated transient inward current (I_TI) underlies these delayed afterdepolarizations (DAD) and may initiate triggered activity during ventricular tachycardia and ischemia-induced arrhythmias (9, 17, 18, 26, 27).

Kass et al. (19) demonstrated a reversal potential for digitalis-induced I_TI in Purkinje fibers and suggested that some combination of electrogenic sodium/calcium exchange current (I_NaCa) and calcium-activated cation current could underlie I_TI. However, reversal of I_TI was not observed in similarly treated muscle preparations, seemingly ruling out a contribution from ion channels while favoring I_NaCa (1, 3, 9). The finding that I_TI was present after inhibition of I_NaCa and discovery of calcium-activated channels equally permeable to sodium and potassium ions and a calcium-activated chloride conductance [I_Cl(Ca)] in cardiac myocytes rekindled debate that I_TI might be composed of more than one ionic component (4-6, 8, 12, 38). Unfortunately, I_Cl(Ca) is often recorded in sodium-free solutions needed to block I_NaCa, making it impossible to draw conclusions regarding the relative amplitudes of I_NaCa, I_Cl(Ca), and nonselective cation conductances within Purkinje fibers as well as ventricular and atrial myocytes.

Acetylstrophanthidin- and BAY K 8644-induced DAD were elicited in a subset of cells within the midmyocardium with unique electrical characteristics and referred to as M cells (29). I_Cl(Ca) in canine M cells contributes to an oscillating inward current in calcium-overloaded cells in the absence of I_NaCa (38). The present study was undertaken to determine the importance of I_Cl(Ca), I_NaCa, and nonselective cation conductances to the generation of I_TI and triggered activity in calcium-overloaded M cells. Cell shortening, ionic currents, and action potentials were recorded after application of isoproterenol and ouabain to load myocytes with calcium. We find no evidence for calcium-activated nonselective cation channels in M cells. Although I_NaCa is larger than I_Cl(Ca) at the resting potential, either conductance on its own is sufficient to cause DAD and triggered activity. Some of this work has been reported as an abstract (41).

	METHODS

Top Abstract Introduction Methods Results Discussion References

Adult male mongrel dogs were given 200 U/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 (38). Cells from the midmyocardial region of the left ventricle were used 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. Standard patch-clamp technique was used to record whole cell currents or action potentials. The composition of the standard external solution was (in mM) 2 CaCl₂, 4 KCl, 1 MgCl₂, 10 glucose, 140 NaCl, and 10 HEPES, and pH was adjusted to 7.4 with NaOH. When required, external sodium was reduced by equimolar substitution with either lithium or N-methyl-D-glucamine (NMG), and external chloride was reduced with equimolar substitution of methanesulfonic acid. Standard pipette solution contained (in mM) 10 KCl, 10 NaCl, 135 potassium aspartate, 10 HEPES, 1 MgCl₂, 5 MgATP, and pH was adjusted to 7.1 with KOH. When required, internal chloride was altered by equimolar substitution of aspartate for chloride. Potassium-free solutions were made by omitting KCl from the external solution and by substituting cesium for potassium in the pipette solution. When required, internal sodium was reduced by equimolar substitution of cesium. Most experiments in sodium-free solutions were accomplished without the addition of EGTA to the pipette solution. Those instances when EGTA was required are indicated in the text.

Endogenous calcium buffers are dialyzed from cells using standard patch electrodes. The amphotericin B perforated-patch technique was used in a few cells to verify recordings of action potentials and DAD obtained by standard patch technique. Composition of the external solution was (in mM) 2 CaCl₂, 4 KCl, 1 MgCl₂, 10 glucose, 140 NaCl, and 10 HEPES, and pH was adjusted to 7.4 with NaOH. Pipette solution contained (in mM) 0.00026 amphotericin B, 135 potassium aspartate, 10 NaCl, 10 KCl, 10 HEPES, 1 MgCl₂, and 0.01 CaCl₂, and pH was adjusted to 7.1 with KOH. When required, internal chloride was altered by equimolar substitution of aspartate for chloride. We have assumed that potassium, sodium, and chloride equilibrate with pipette solution within minutes of incorporation of amphotericin B in the membrane, at rates similar to those suggested for the nystatin perforated-patch technique (16a).

Amphotericin B (Sigma Chemical, St. Louis, MO) was made in dimethyl sulfoxide (60 mg/ml) and diluted 250-fold into pipette solution to a final concentration of 240 µg/ml. Fresh dilutions into pipette solution were made every 1.5 h. CdCl₂, ouabain, and isoproterenol were made as concentrated stocks in water and diluted 1,000-fold into external solution. Isoproterenol stock was kept under nitrogen. Final concentrations were 300 µM CdCl₂, 1 µM isoproterenol, and 5 µM ouabain. SITS (Sigma Chemical) was added directly to the external solution just before each experiment. Amphotericin B, SITS, and ouabain were used in a darkened room.

With the exception of the experiments shown in Fig. 1, all voltage-clamp protocols were preceded by a train of six 200-ms pulses to 10 mV delivered at a rate of 2 Hz to maintain calcium loading of the sarcoplasmic reticulum. Calcium-activated currents were measured at either 80 or 50 mV after a 5-ms pulse to 50 mV to inactivate sodium current (I_Na) and a 3-ms pulse to 10 mV to activate calcium current (I_Ca) and release of calcium from the sarcoplasmic reticulum (SR). Protocols were repeated at 15-s intervals. Calcium overload was induced by adding 1 µM isoproterenol and 5 µM ouabain to external solution. The current-clamp mode of the amplifier was used to record action potentials. Triggered activity and DAD were evoked after a train of 15 stimulated beats at a rate of 2 Hz in the presence of 1 µM isoproterenol and 5 µM ouabain. Before application of isoproterenol and ouabain, each cell was stimulated at a rate of 2 Hz to check if calcium handling was compromised. Action potential shape was normal, and DAD were never present before applying isoproterenol and ouabain.

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Rapid substitution of external sodium using a sewer pipe. Cell was bathed in potassium-free external solution containing 140 mM sodium, and a sewer pipe placed within 100 µm of cell was filled with external solution modified to replace sodium with N-methyl-D-glucamine (NMG). Sodium current (INa) was recorded in normal sodium (top left, 0 ms), and during subsequent depolarizations, effects of turning on sewer pipe before activation of INa were evaluated. Intervals indicated in each panel refer to period sewer pipe was turned on before activation of INa.

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 single-barrel sewer pipe with an inside diameter of 0.28 mm was used to exchange the solution immediately surrounding voltage- or current-clamped myocytes. This sewer pipe was placed 100 µm from the cell, and flow was controlled by a pinch valve and computer interface (model BPS-4; Adams and List Associates, Westbury, NY). An Axopatch 200A amplifier (Axon Instruments, Foster City, CA) was operated in voltage- or current-clamp mode to record whole cell currents and action potentials at 37°C. Cell capacitance was 185 ± 7 pF (42 cells). Pipette tip resistances were 1.0-3.5 M, and seal resistances were 8-12 G. Electronic compensation of series resistance averaged 77 ± 2% (42 cells), and the series resistance remaining after this compensation averaged 1.57 ± 0.15 M. After series resistance compensation, capacitive current decayed with a single time constant of 290 ± 20 µs (42 cells).

Tip potentials of standard patch electrodes were measured using established techniques and ranged from 7 to 13 mV, depending on chloride content of the pipette solution. To determine tip potentials for perforated-patch electrodes, cells were depolarized in high-potassium (70 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 like-treated cells. Tip potentials of perforated-patch electrodes ranged from 16 to 19 mV, depending on chloride content of the pipette solution. Voltages reported in the text were corrected for these potentials. The seal between cell membrane and patch pipette was initially formed in Tyrode solution containing 1 mM CaCl₂. A 3 M KCl-agar bridge was used between the Ag/AgCl ground electrode and external solution to avoid development of a ground potential when switching to experimental solutions. Protocols required using a sewer pipe to replace sodium surrounding voltage-clamped cells. To evaluate the effects of this substitution on ground potential, the current-voltage relation of the calcium current was determined in the presence of 140 mM sodium or 140 mM NMG applied via sewer pipe. Threshold and peak voltage of the current-voltage relation were unaffected, suggesting that ground potential was not altered by sodium substitution (5 cells).

Whole cell currents and transmembrane potentials were filtered with a four-pole low-pass Bessel filter at 5 kHz, digitized between 0.5 and 2 kHz (Digidata 1200, Axon Instruments), and stored on a computer. Significant differences between means were determined by a paired or unpaired Student's t-test. Unloaded cell shortening 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 0-40 µm. Clampex 6 acquisition software (Axon Instruments) was used to concurrently record cell shortening and either transmembrane potential or ionic current.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

The capacity to rapidly substitute external sodium to inhibit inward I_NaCa is central to our investigation. We used a sewer pipe to limit the interval spent in sodium-free solutions and thus reduce alterations of intracellular sodium, pH, and calcium that might be expected after blocking Na/H transport, I_Na, and inward I_NaCa. In this study, cells were exposed to sodium-free solutions for periods ranging from 1.2 to 2.8 s. We have assumed that the primary effect of eliminating external sodium is to abolish inward I_NaCa, although changes in intracellular pH or calcium might not be eliminated even with these brief exposures. To evaluate the time needed to completely substitute sodium surrounding a single myocyte, I_Na was recorded in normal sodium and after substitution with an ion too large to penetrate sodium channels.

Figure 1 shows that a complete exchange of NMG for external sodium and a concomitant reduction in I_Na was achieved within 200 ms of activating a sewer pipe to deliver NMG-containing solution. Patch pipettes were filled with potassium-free solution containing 5 mM EGTA. Cells were bathed in potassium-free solution containing normal sodium, and the sewer pipe was filled with an external solution in which sodium had been completely replaced with NMG. Because inward I_Na is effectively reduced by replacement of external sodium with NMG, we compared I_Na in normal sodium with those currents recorded after successively longer exposures to NMG-containing solution from the sewer pipe. The cell was held at 80 mV and repetitively pulsed to 40 mV to activate sodium channels at a rate of 0.2 Hz. During the first pulse, the sewer pipe was not turned on, and the current shown in Fig. 1, top left, was recorded. With subsequent pulses, the sewer pipe was activated for progressively longer periods before the step to 40 mV and always turned off at the end of the 40 mV step. Intervals indicated in each panel of Fig. 1 refer to the period the sewer pipe was activated before the step to 40 mV. I_Na was significantly reduced when the sewer pipe was turned on just 100 ms before activation of I_Na, and when turned on 200 ms before activation of I_Na flow from the sewer pipe completely displaced the sodium-containing bath solution and abolished I_Na. Reduction of I_Na did not result from a sewer pipe-induced shift in the ground potential (see METHODS), but from complete elimination of sodium surrounding the cell. In 20 cells, the time required to completely abolish I_Na was 310 ± 4.2 ms.

Figure 2 demonstrates the method used to trigger calcium-activated conductances and verifies that deactivation of I_Ca is too rapid to interfere with our measurements of calcium-activated fluxes, which are made starting 5 ms after stepping back to 80 mV. Currents and cell shortening were measured in standard external and pipette solutions, and the sewer pipe was filled with external solution containing 300 µM CdCl₂. A cell was held at 80 mV before evoking a 5-ms pulse to 50 mV to inactivate I_Na, followed by a 3-ms pulse to 10 mV to expedite release of calcium from the SR, and currents were measured after returning to 80 mV. In Fig. 2A, the voltage template, ionic currents, and cell shortening are shown in the top, middle, and bottom panels, respectively. Lowercase letters identify corresponding current traces and contractions. Commencing with a step to 80 mV, a slowly decaying inward current accompanied by a contraction was recorded (a). CdCl₂ abolished this current and contraction (b), and within 15 s of turning off the flow of cadmium-containing solution from the sewer pipe, both current and contraction had nearly recovered their original amplitudes (c). Similar results were obtained in six additional cells.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Inward calcium-activated currents and time course for deactivation of calcium current (ICa). In A, voltage protocol along with currents and cell shortening are shown in top, middle, and bottom panels. Lowercase letters indicate currents and accompanying contractions: controls (a), after applying 300 µM CdCl2 (b), and 15 s of returning to control solution (c). In B, voltage protocol and associated currents during long and short steps to 10 mV are shown in top and middle panels, respectively. Subtraction of currents immediately after a 300-ms pulse from those currents immediately after a 3-ms pulse will reveal time course of ICa deactivation. Shown in bottom panel are subtracted currents commencing with step to 80 mV.

Although we were tempted to identify all of the cadmium-sensitive current measured at 80 mV as activated by a rise in cytosolic calcium, deactivation of I_Ca might contribute a short-lived element to this cadmium-sensitive inward current. To determine the time course of tail currents attributable to I_Ca, currents were recorded in sodium- and potassium-free solutions. I_Cl(Ca) was inhibited by reducing external chloride to 6 mM and internal chloride to 2 mM. Pipettes contained 250 µM EGTA to compensate for an increase in cytosolic calcium due to inhibition of I_NaCa. Cells continued to contract under these conditions. A cell was held at 50 mV. Currents were measured at 80 mV immediately after either a 300-ms pulse to 10 mV, or a 3-ms pulse to 10 mV. In Fig. 2B, the voltage template and superimposed currents during long and short steps to 10 mV are shown in the top and middle panels, respectively. I_Ca completely inactivated within 75 ms of the beginning of the long pulse but does not have this same opportunity to inactivate during a 3-ms pulse. For this reason, I_Ca tail currents can only be present after a 3-ms pulse, and subtraction of currents immediately after a 300-ms pulse from those currents immediately after a 3-ms pulse will reveal the time course of I_Ca deactivation. Shown in Fig. 2B, bottom panel, are the subtracted currents commencing with the step back to 80 mV. The amplifier was overloaded for a period of 2 ms during which time we can say nothing about deactivation. After this initial period, a rapidly decaying tail was observed, which normally would be partially hidden by decay of capacitive current. Although the inward current might be a capacitance artifact, in six cells subtractions always yielded a rapidly decaying inward deflection. We conclude that inward currents shown in Fig. 2A largely represent conductances activated by a rise in intracellular calcium. To allow for complete decay of capacitive current and I_Ca deactivation, we have been conservative and quantified calcium-activated currents as beginning 5 ms after stepping back to 80 mV. These conductances underlie DAD in calcium-overloaded cells, and their identification is the focus of this study.

Calcium-activated conductances were investigated after first inhibiting I_Cl(Ca) by reducing external chloride to 6 mM and internal chloride to 2 mM. We asked whether both inward I_NaCa and nonselective cation conductance could be evoked by a rise in intracellular calcium, using the protocol first illustrated in Fig. 2A. Currents recorded in normal external sodium were compared with those remaining after complete substitution of external sodium and inhibition of inward I_NaCa (20). Figure 3 illustrates that the calcium-activated nonselective cation conductance is absent from the canine midmyocardium. In Fig. 3A, the external solution was potassium free and contained 140 mM sodium. A sewer pipe was filled with a modified external solution in which all sodium was replaced by NMG. Patch pipettes were filled with an internal solution that was potassium free, contained 10 mM sodium, and contained no EGTA. Shown in Fig. 3A, middle panel, are superimposed currents recorded in normal sodium (trace marked Na⁺), and after replacement of sodium with NMG (trace marked NMG⁺). NMG-containing solution was applied 600 ms before a step to 50 mV for a total period of 1.5 s. Replacement of sodium caused an outward shift in currents at all potentials, abolished I_Na during the brief pulse to 50 mV, and eliminated calcium-activated current at 80 mV. Figure 3A, bottom panel, shows that contraction was increased after application of NMG, so it is unlikely that elimination of calcium-activated currents was precipitated by loss of the underlying calcium transient. If canine myocytes possessed calcium-activated nonselective cation channels, we expected that sodium substitution with a large cation like NMG and inhibition of I_NaCa should have resulted in a transient outward current at 80 mV carried by cesium. We propose that the steady outward current at 80 mV results from reverse-mode I_NaCa after replacement of all external sodium, consistent with a larger contraction after sodium substitution.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Nonselective cation current is absent from canine midmyocardial myocytes. Calcium-activated currents were investigated after first inhibiting calcium-activated chloride conductance [ICl(Ca)] by reducing external and internal chloride. Currents in normal external sodium were compared with those remaining after inhibition of inward sodium/calcium exchange current (INaCa). In A, middle panel shows currents in normal sodium (trace marked Na+) and after replacement of sodium with NMG (trace marked NMG+). Bottom panel shows that contraction was increased after application of NMG. In B, pipette solution was sodium free to inhibit reverse-mode INaCa. Isoproterenol (1 µM) and ouabain (5 µM) were added to external solutions to increase intracellular calcium. Solutions were potassium free, and as before, ICl(Ca) was inhibited by reduction of chloride. Middle panel, currents in normal external sodium (Na+) and after replacement of all sodium with lithium (Li+); bottom panel, superimposed recordings of cell shortening in normal sodium and after replacement with lithium.

DAD and triggered activity are typically recorded under conditions in which cells become overloaded with calcium. To examine whether a greater calcium load could evoke a nonselective cation conductance, protocols were repeated in the presence of both 1 µM isoproterenol and 5 µM ouabain to duplicate conditions later used to evoke DAD. Solutions were potassium free, and as before, I_Cl(Ca) was inhibited by drastic reduction of chloride. A sewer pipe was filled with a modified external solution in which all sodium was replaced by lithium. Pipette solution was sodium free to inhibit reverse-mode I_NaCa. Shown in Fig. 3B, middle panel, are superimposed currents in normal external sodium and after replacement of all sodium with lithium. Lithium inhibited inward I_NaCa and abolished all calcium-activated current without causing an outward shift in currents or affecting cell shortening. If canine myocytes possessed calcium-activated nonselective cation channels, we expected that a I_TI carried by lithium would remain after blocking I_NaCa. We conclude that under conditions in which I_Cl(Ca) has been blocked, I_NaCa will be the only remaining calcium-activated conductance in these cells. The outward shift in currents that occurred when internal solution contained sodium most likely results from reverse-mode I_NaCa, since it was absent when internal sodium was replaced by cesium.

In succeeding experiments, we have used chloride-containing solutions to determine the relative amplitudes of I_Cl(Ca) and I_NaCa and their contributions to DAD and triggered activity. Figure 4 shows the amplitudes of I_Cl(Ca) and I_NaCa for controls (A) and after isoproterenol and ouabain were applied to increase intracellular calcium loading (B). Solutions were potassium free and contained normal external and internal sodium. External chloride was 146 mM, internal chloride was 22 mM, and the calculated chloride reversal potential (E_Cl) was 49 mV. In Fig. 4A, the top panel shows currents in normal sodium and after replacement with lithium to block inward I_NaCa. Lithium caused an outward shift in current at 80 mV and substantially reduced the calcium-activated current even though the cell shortened to a greater degree. The lithium-insensitive current is I_Cl(Ca), since it was absent when chloride was drastically reduced (Fig. 3) and could be blocked by 2 mM SITS. The sewer pipe was turned off, and currents recovered completely before 1 µM isoproterenol and 5 µM ouabain were applied and the protocols repeated. Figure 4B shows that these agents increased contraction and the calcium-activated current in normal sodium, but this increase over control levels was largely lithium sensitive. I_Cl(Ca) did not increase in the presence of these agents, even though contraction increased significantly.

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Relative amplitudes of ICl(Ca) and INaCa. In A, top panel shows currents in normal sodium (Na+) and after replacement with lithium to block inward INaCa (Li+). Lithium was applied 600 ms before stepping to 50 mV for a total period of 1.5 s. Contraction is shown in bottom panel. In B, sewer pipe was turned off and currents recovered completely, before 1 µM isoproterenol and 5 µM ouabain were applied and protocols repeated in same cell. Currents and contractions are shown in top and bottom panels, respectively.

After we adjusted for the lithium-induced shift in baseline, the lithium-sensitive traces (I_NaCa) and those resistant to lithium [I_Cl(Ca)] were integrated starting 5 ms from the beginning of the step to 80 mV, and ending when these currents decayed to a steady baseline. Traces from experiments in which the lithium-induced outward shift was unstable were not analyzed (2 cells). In 25 control cells, the net flux carried by I_NaCa was 132 ± 19 × 10³ pC/pF and was similar to that carried by I_Cl(Ca), 134 ± 16 × 10³ pC/pF. In 26 cells treated with isoproterenol and ouabain, I_NaCa was 161 ± 15 × 10³ pC/pF and I_Cl(Ca) was 124 ± 23 × 10³ pC/pF. I_NaCa and I_Cl(Ca) were not statistically different from each other for either the control or the calcium-overloaded groups (P > 0.05).

I_NaCa and I_Cl(Ca) were measured with 10 mM sodium in the pipette to duplicate conditions in which action potentials and triggered activity were recorded. As indicated by a larger contraction, this ultimately results in accumulation of intracellular calcium when external sodium is replaced by lithium and could lead to an overestimate of I_Cl(Ca), as well as confound the interpretation of difference currents. To abolish reverse-mode sodium/calcium exchange and limit accumulation of calcium, experiments were repeated without sodium in the pipette. After isoproterenol and ouabain treatment, inward I_NaCa was 168 ± 30 × 10³ pC/pF, and I_Cl(Ca) was 114 ± 24 × 10³ pC/pF, a difference which reached the level of significance (11 cells; P < 0.05). We conclude activation of reverse-mode sodium/calcium exchange does exaggerate the amplitude of I_Cl(Ca) and that I_NaCa is nearly 60% of the total calcium-activated current at 80 mV after treatment with isoproterenol.

We have previously reported that application of isoproterenol increased outward I_Cl(Ca) in rabbit ventricular cells (39). We have not until now investigated the effect of this agent on inward I_Cl(Ca), although certainly our expectation was that isoproterenol would also increase inward chloride current. To confirm the ineffectiveness of isoproterenol to increase inward chloride current (Fig. 4), I_Cl(Ca) was recorded during an action potential clamp in sodium- and potassium-free solutions, with E_Cl equal to 49 mV. Pipette solution contained 250 µM EGTA. To ensure uniform loading of the SR, six 200-ms pulses to 10 mV delivered at a rate of one pulse every 2 s preceded the action potential clamp; 0.5 mM SITS was applied to reduce I_Cl(Ca), and 30 s later, this protocol was repeated. Currents recovered in control solutions, 1 µM isoproterenol was added to the external solution, and protocols were repeated. In five cells, the SITS-sensitive current reversed over the course of the action potential clamp, similar to what we have reported previously (42). What was surprising was that isoproterenol increased peak outward SITS-sensitive current 30 ± 4% but did not increase peak inward SITS-sensitive current in any of these five cells.

Several authors have hypothesized the presence of a subsarcolemmal space in which communication with the bulk cytosol is restricted (23, 25, 42). For our determination of I_Cl(Ca), we have assumed that measurements made with sodium-free pipette solution abolished reverse-mode I_NaCa. We revisited this assumption regarding sodium-free pipette solutions, choosing conditions that would favor the buildup of sodium in the subsarcolemmal space and determined the upper bounds of reverse-mode I_NaCa that might remain under these conditions.

Whole cell currents were recorded using standard patch-clamp technique. External solution was potassium free, with greatly reduced chloride (extracellular chloride concentration = 6 mM), but contained normal sodium, 1 µM isoproterenol, 5 µM ouabain, and 2 mM CaCl₂. Pipette solutions were potassium free and included no EGTA. The sewer pipe was filled with modified external solution in which sodium was replaced with either lithium (4 cells) or NMG (6 cells). Cells were held at 80 mV, and reverse-mode I_NaCa was elicited by turning on the sewer pipe and replacing external sodium. To favor loading of the subsarcolemmal space with sodium, a train of 15 20-ms pulses to 40 mV at a rate of 2 Hz was delivered 1 s before turning on the sewer pipe. Figure 5 shows superimposed traces from two cells of similar size. The larger I_NaCa was recorded when the pipette contained 10 mM sodium, and the smaller I_NaCa was recorded when pipette solution was sodium free. Amplitudes are reported as the difference between the current recorded 1.2 s after switching to lithium and the baseline current in normal sodium. For five cells in which pipettes contained sodium, the outward current evoked in response to substitution of external sodium was 1.25 ± 0.02 pA/pF at 80 mV and 0.056 ± 0.0012 pA/pF when pipettes contained no sodium (5 cells). When pipettes contained sodium, substitution of external sodium resulted in spontaneous contractions after ~1.5 s. Conversely, when cells were dialyzed by sodium-free pipette solution, spontaneous contractions were still not evident 30 s after activating the sewer pipe.

View larger version (10K): [in this window] [in a new window]   Fig. 5.   Reverse-mode INaCa is significantly reduced when using sodium-free pipette solution. External solution was potassium free, with greatly reduced chloride (extracellular chloride concentration = 6 mM) but contained normal sodium, 1 µM isoproterenol, 5 µM ouabain, and 2 mM CaCl2. Pipette solutions were potassium free and included no EGTA. Sewer pipe was filled with modified external solution in which sodium was replaced with lithium. Cells were held at 80 mV, and reverse-mode INaCa was elicited by turning on sewer pipe and replacing external sodium, indicated by bar labeled Na+ free. To favor loading of subsarcolemmal space with sodium, a train of 15 20-ms pulses to 40 mV at a rate of 2 Hz was delivered 1 s before turning on sewer pipe. Larger lithium-induced outward current was recorded when pipette contained 10 mM sodium; cell capacitance was 178 pF. Smaller lithium-induced current was recorded when pipette solution was sodium free; cell capacitance was 201 pF.

If I_Cl(Ca) and I_NaCa both contribute to currents in calcium-overloaded cells, we should be able to separate them based on their reversal potentials. In particular, when E_Cl is 49 mV, activation of cyclic release of calcium from the SR and a step to 50 mV should result in an oscillating outward I_Cl(Ca) superimposed on an oscillating inward I_NaCa. Substitution of sodium with lithium should block all inward I_NaCa, leaving only an oscillating outward I_Cl(Ca). Figure 6 shows that a lithium-sensitive inward current and a chloride-sensitive outward current both contribute to oscillating currents at 50 mV. Calcium overload was induced by the addition of 1 µM isoproterenol and 5 µM ouabain to all external solutions. Oscillating currents were evoked by a 3-ms step to 10 mV and measured after stepping to 50 mV. In Fig. 6A, currents were recorded in potassium-free solution, with E_Cl equal to 49 mV. During the step to 50 mV, currents in normal sodium are outward and show evidence of overlapping inward and outward currents. Inhibition of I_NaCa shifts the current in an outward direction and abolished inward oscillations, leaving much larger transient outward currents. Similar results were obtained in eight cells. In Fig. 6B, currents were recorded in potassium-free solution, and E_Cl was adjusted to 50 mV in a second cell. During the step to 50 mV, currents in normal sodium are outward. Inhibition of I_NaCa shifts the current in an outward direction and left small transient outward currents. In six cells in which E_Cl was 50 mV, oscillating outward transients in lithium were <100 pA. This compares with oscillating outward currents >1 nA when E_Cl was 49 mV. Taken together, these results suggest that I_NaCa and I_Cl(Ca) contribute to oscillating currents in calcium-overload cells.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   INaCa and ICl(Ca) both contribute to oscillating currents in isoproterenol- and ouabain-treated cells. In A, currents were recorded in potassium-free solution, with chloride reversal potential (ECl) equal to 50 mV. Shown are currents recorded at 50 mV in normal sodium (Na+) and after substitution with lithium (Li+). In B, currents were recorded in potassium-free solution, but chloride was adjusted such that ECl was equal to step potential. Again, currents were recorded at 50 mV in normal sodium (Na+) and after substitution with lithium (Li+). In both cells, lithium was applied 800 ms before step to 50 mV for a total period of 2.5 s.

The most convincing demonstration that these two currents contribute to triggered activity in midmyocardial cells is to choose conditions that eliminate one or the other current and test whether triggered activity can still be evoked. In the absence of EGTA in the pipette solution, cells will not survive elimination of I_NaCa except for brief periods. When pipettes contained 10 mM sodium, lithium applied for 15 s caused hypercontraction of cells, from which there was no recovery. Instead, we have recorded action potentials in calcium-overloaded cells in the absence of inward I_Cl(Ca) to see if I_NaCa is sufficient to support triggered activity. The perforated-patch technique was used to record action potentials in standard external solution to which 1 µM isoproterenol and 5 µM ouabain were added. Standard internal solution was modified to reduce internal chloride to 2 mM, with E_Cl equal to 113 mV. Figure 7 shows that I_NaCa is sufficient to cause DAD in calcium-overloaded cells. In Fig. 7A, DAD were evoked after a train of 15 stimulated beats at a 500-ms basic cycle length (BCL). The last stimulated beat is shown at the left, which was followed by a series of DAD. I_Cl(Ca) is an outward current at a potential of 85 mV and might oppose the depolarizing influence of I_NaCa. In Fig. 7B, 0.5 mM SITS was applied, and the protocol was repeated in the same cell. Again, the last stimulated beat is at the left, which was then followed by two triggered beats and a series of DAD. When E_Cl was 113 mV, isoproterenol and ouabain caused DAD, but not triggered beats in 11 cells. SITS was applied to three of these cells, and in all three cells, SITS caused triggered beats where before there were only DAD. Although this result is consistent with SITS blocking an outward I_Cl(Ca), the increase in triggered activity might also arise if SITS partially blocks inwardly rectifying potassium current (I_K1), leading to an increased membrane resistance and an enhanced ability of I_NaCa to elicit action potentials.

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Isoproterenol- and ouabain-induced triggered activity in absence of inward ICl(Ca). Action potentials were recorded using perforated-patch technique. Standard internal solution was modified to reduce internal chloride to 2 mM, with ECl equal to 113 mV. In A, last stimulated beat is shown at left. In B, last stimulated beat is shown at left. ICl(Ca) was reduced by application of 0.5 mM SITS, and protocol was repeated in same cell.

To differentiate between these two alternatives, we used the perforated-patch technique to record inwardly rectifying potassium current and test the effects of 0.5 mM SITS on this current. Whole cell currents were recorded in normal external and pipette solutions containing 1 µM isoproterenol and 5 µM ouabain. Cells were held at 80 mV and pulsed to 90 mV for 300 ms at a rate of 2 Hz. SITS was applied, and the protocol was repeated. Current density was 0.61 ± 0.07 pA/pF (n = 10) at the holding potential of 80 mV and 4.83 ± 0.12 pA/pF at the end of a 300 ms step to 90 mV and was unaffected by SITS.

Taken together, these results are consistent with our characterization of calcium-activated currents in midmyocardial cells and suggest that I_NaCa has a major role in generating DAD in calcium-overloaded cells. Because our characterization of currents also suggests a role for I_Cl(Ca), we next focused on the effects of transiently abolishing inward I_NaCa while recording triggered activity with normal chloride in the pipette.

Figure 8 shows the effects of inhibiting I_NaCa on DAD and triggered activity. Standard patch-clamp technique was used to record action potentials and cell shortening in standard external and internal solutions, with E_Cl equal to 50 mV. Triggered activity and DAD were evoked after a train of 15 stimulated beats at a 500-ms BCL in the presence of 1 µM isoproterenol and 5 µM ouabain. In Fig. 8A, the last stimulated beat is shown at the left, followed by three triggered beats. The second triggered beat was elicited from a slowly rising DAD, and each beat was associated with a contraction (Fig. 8A, bottom trace). In Fig. 8B, the protocol was repeated in the same cell, and lithium was applied 15 ms after the upstroke of the last stimulated beat for a period of 1.5 s. The results in this cell were typical of eight other cells in which the number of triggered beats was reduced after inhibiting I_NaCa, but DAD remained even at the end of the lithium pulse, and contractions continued to be associated with DAD and triggered beats. Lithium had no effect on resting potential, unless applied for longer than 3 s, at which point cells begin to depolarize. Because complete exchange of NMG for sodium occurred within 310 ms (Fig. 1), we considered it sufficient to apply lithium for no longer than 2.8 s in this study.

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Effects of inhibiting INaCa on delayed afterdepolarizations (DAD) and triggered activity when solutions contained normal internal and external chloride, with ECl equal to 50 mV. In A, last stimulated beat is shown at left, followed by 3 triggered beats. Each beat was associated with a contraction (bottom trace). In B, lithium was applied, and protocol was repeated in same cell. Last stimulated beat is shown at left.

We repeated these experiments using the perforated-patch technique to ensure that dialysis of some intracellular component did not influence results obtained by standard patch-clamp technique. Action potentials were recorded in standard external and internal solutions, with E_Cl equal to 50 mV. Triggered activity and DAD were evoked after a train of 15 stimulated beats at a 500-ms BCL in the presence of 1 µM isoproterenol and 5 µM ouabain. This protocol was repeated, and lithium was applied 15 ms after the upstroke of the last stimulated beat for a period of 1.5 s. As was the case with the standard patch-clamp technique, although the number of triggered beats after a train was reduced, triggered activity and DAD continued to be observed after substitution of external sodium with lithium in seven midmyocardial cells.

Any effect of lithium to reduce the number of triggered beats did not arise directly from a lithium-induced shift of the activation threshold for sodium channels to more positive potentials. Sodium currents were measured under conditions that permitted adequate voltage control. Standard patch-clamp technique was used to record currents at 22°C from a holding potential of 110 mV. Solutions were potassium free, pipettes contained 10 mM EGTA, and both internal and external sodium were reduced to 5 mM. The current-voltage relation for sodium current was measured in 5-mV increments over a voltage range between 80 and 40 mV. The protocol was repeated after replacing external sodium with 5 mM lithium. In nine cells, lithium had no effect on the 60 mV threshold of the sodium current.

Hille (16) reports the ionic permeability ratios of lithium over sodium to range between 0.93 and 1.1, depending on cell type and species. Experiments were performed to investigate whether exchanging lithium for sodium might reduce sodium channel conductance, and thus decrease the ability of these channels to trigger an action potential. Standard patch-clamp technique was used to record action potentials in normal external solution. Pipettes contained normal ions and 10 mM EGTA to reduce calcium-activated conductances. A train of action potentials was elicited at a BCL of 500 ms, and the change was made from external sodium to lithium. In eight cells, action potentials could be elicited for an average of 25.3 ± 0.8 s after switching to lithium. During investigations to determine the mechanisms underlying DAD, our substitutions never exceeded 2.8 s. We conclude that a lithium-induced reduction in sodium-channel conductance should not directly contribute to a decrease in triggered activity.

We have previously reported that oscillating inward currents recorded in sodium-free solutions in calcium-overloaded cells were blocked by SITS (38). Figure 9 shows the effects of external sodium replacement on the I_TI that underlie DAD. Currents and cell shortening were obtained in standard external and internal solutions containing 1 µM isoproterenol and 5 µM ouabain, with E_Cl equal to 50 mV. Calcium-activated currents were recorded at 80 mV after a 500-ms pulse to 0 mV. In Fig. 9A, I_TI are shown accompanied by contractions. In Fig. 9B, lithium was applied 250 ms after the beginning of the step to 0 mV for a period of 2.8 s. Lithium abolished inward I_NaCa, but I_TI due to I_Cl(Ca) persisted and continued to be associated with contractions.

View larger version (18K): [in this window] [in a new window]   Fig. 9.   Effects of external sodium replacement on transient inward currents that underlie DAD when solutions contained normal internal and external chloride, with ECl equal to 50 mV. In A, transient inward currents are shown accompanied by contractions. In B, protocol was repeated, but lithium was applied 250 ms after beginning of step to 0 mV for a period of 2.8 s.

When solutions contained normal potassium, replacement of external sodium did not result in an outward shift in currents, as had been previously observed in potassium-free solutions. This lack of outward shift in potassium-containing solutions is consistent with the failure of lithium to modify resting potential in calcium-overloaded cells (Figs. 8 and 10). We examined whether the outward shift in potassium-free solutions due to reverse-mode I_NaCa (Fig. 3) might be offset by a lithium-induced decrease in net outward current contributed by I_K1. I_K1 was recorded in sodium-free internal and external solutions in the presence of 300 µM CdCl₂ to block all calcium transients. While holding at 85 mV, near the typical resting potential of these cells, lithium was applied, and an inward current continued to develop over the course of application. We did not further investigate the degree that this inward current offsets reverse-mode I_NaCa, although we did observe that cells depolarized when lithium was applied for longer than 3 s in complete Tyrode solution.

View larger version (16K): [in this window] [in a new window]   Fig. 10.   Effects of inhibiting INaCa on DAD and triggered activity when intracellular chloride was elevated. Standard patch-clamp technique was used to record action potentials and cell shortening. Internal solution contained 147 mM chloride, with ECl equal to 1 mV. In A, last stimulated beat is shown at left, followed by 2 triggered beats. Each beat was associated with a contraction (bottom trace). In B, protocol was repeated in same cell, and lithium was applied 15 ms after upstroke of last stimulated beat for a period of 2.8 s. Last stimulated beat was followed by a triggered beat and 2 DAD, each associated with a contraction.

We next examined if increasing the driving force for inward I_Cl(Ca) at resting potentials might permit sustained triggered beats in the presence of lithium. Although E_Cl is often portrayed as being 50 mV, evidence also exists for reversal at a more depolarized potential (34, 36, 42). Figure 10 shows the effects of inhibiting I_NaCa on DAD and triggered activity. Standard patch-clamp technique was used to record action potentials and cell shortening in standard external solution containing 1 µM isoproterenol and 5 µM ouabain. Internal solution contained 147 mM chloride, with E_Cl equal to 1 mV. In Fig. 10A, the last stimulated beat is shown at the left, followed by two triggered beats. The second triggered beat was elicited from a slowly rising DAD, and each beat was associated with a contraction (Fig. 10A, bottom trace). In Fig. 10B, the protocol was repeated in the same cell, and lithium was applied 15 ms after the upstroke of the last stimulated beat for a period of 2.8 s. The last stimulated beat was followed by a triggered beat and two DAD, each associated with a contraction. Increasing the driving force for inward I_Cl(Ca) did not noticeably increase triggered activity in the presence of lithium. In seven cells, multiple triggered beats were recorded in normal sodium, but replacement of sodium converted this activity to a single triggered beat followed by DAD.

Figure 11 shows the effects of external sodium replacement on I_TI when intracellular chloride was adjusted to 147 mM. Currents and cell shortening were obtained in standard external solution containing 1 µM isoproterenol and 5 µM ouabain, with E_Cl equal to 1 mV. In Fig. 11A, I_TI are shown accompanied by contractions. In Fig. 11B, lithium was applied 250 ms after the beginning of the step to 0 mV for a period of 2.8 s. Lithium abolished inward I_NaCa, but I_TI due to I_Cl(Ca) persisted and continued to be associated with contractions. Similar results were obtained in six cells.

View larger version (17K): [in this window] [in a new window]   Fig. 11.   Effects of external sodium replacement on transient inward currents that underlie DAD when solutions contained elevated internal and normal external chloride, with ECl equal to 1 mV. In A, transient inward currents are shown accompanied by contractions. In B, protocol was repeated, but lithium was applied 250 ms after beginning of step to 0 mV for a period of 2.8 s.

In this experiment, and that shown in Fig. 9, lithium does not appear to inhibit oscillating inward current, as one might predict for an inhibitor of inward I_NaCa. This is even more confusing given the similar degree of cell shortening after applying lithium. We observed that this was a feature of taking a 500-ms step to a depolarized potential. Under these conditions, substitution of external sodium and activation of reverse-mode sodium/calcium exchange caused undue loading of the cell with calcium, making comparisons between currents qualitative rather than quantitative. This increased contractility was rarely reflected in the cell shortening record because not all movement was along the longitudinal axis of the cell. One aspect of this increased calcium loading that is obvious in Fig. 11B is a lithium-induced increase in frequency of oscillations.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

We are the first to report the relative amplitudes of I_NaCa and inward I_Cl(Ca) under conditions that provoke DAD and triggered activity in canine M cells. Our decision to characterize I_NaCa as the sodium-sensitive conductance was made based on a need to preserve I_Cl(Ca). Whereas others have used millimolar concentrations of nickel to characterize I_NaCa, such applications could have posed problems under our experimental conditions. Because of its inhibitory effects on I_Ca, nickel might indirectly reduce I_Cl(Ca) (37). Nickel-sensitive current would then represent components of both I_NaCa and I_Cl(Ca) and result in overestimation of I_NaCa. Sodium substitution effectively blocks I_NaCa, without reducing I_Cl(Ca) (40). Unfortunately, defining I_NaCa as the sodium-sensitive current is also not benign, since substitution-induced increases in contractility can complicate interpretation of the difference current and lead to a larger I_Cl(Ca). These concerns were addressed, first by limiting the time spent in sodium-free solution and second by comparing the results obtained with and without sodium in the pipette. The difference between I_NaCa and I_Cl(Ca) was greater after reducing reverse-mode exchange, suggesting that determinations of I_Cl(Ca) made with sodium in the pipette were exaggerated. When controlling for changes in contractility, I_Cl(Ca) was found to be ~70% of the amplitude of I_NaCa. One further intricacy must be considered. Contractility is reduced when sodium is removed from the pipette. We propose that the relative amplitudes of I_Cl(Ca) and I_NaCa are best determined in sodium-free pipette solution, whereas the absolute amplitude of I_Cl(Ca) might very well be closer to that found in sodium-containing pipette solution.

For our determination of I_Cl(Ca), it is critical to know whether sodium is eliminated from the subsarcolemmal space while dialyzing a cell with sodium-free pipette solution. Although concentrations of ions in the subsarcolemmal space must certainly be different from concentrations in the bulk cytosol, this restricted space does communicate with the cytosol. The question becomes, How long can sodium that enters through sodium channels be maintained in the subsarcolemmal space in the face of an infinite sink with sodium-free solution? Kimura et al. (20) and Hilgemann (15) found that reverse-mode exchange was abolished by eliminating internal sodium. For cells dialyzed with sodium-free pipette solution, we found that reverse-mode I_NaCa was nearly abolished 1 s after the end of a train of pulses designed to increase intracellular sodium. I_Cl(Ca) was 124 ± 23 × 10³ pC/pF when pipettes contained sodium and 114 ± 24 × 10³ pC/pF in the absence of internal sodium. The substantial reduction in reverse-mode I_NaCa due to elimination of sodium in the pipette equates to an 8% reduction in chloride flux. As shown by Fig. 3, we were unable to detect a lithium-induced increase in cell shortening when pipettes contained no sodium. If we modify our initial assumption to say that sodium-free pipette solutions do not completely abolish reverse-mode I_NaCa, any lithium-induced increase in intracellular calcium must be quite small, and our "overestimate" of I_Cl(Ca) will be much less than 8%.

I_NaCa was the largest contributor to calcium-activated current at the resting potential in isoproterenol-treated cells. Consonant with these voltage-clamp findings, I_NaCa was large enough to cause DAD and triggered activity in the absence of other calcium-activated conductances. Our results are consistent with those reported by Kass et al. (19), who determined the sodium dependence of I_TI in cardiac Purkinje fibers. A correlation between I_NaCa and I_TI or DAD has also been demonstrated for human atrial cells, canine ventricular and atrial cells, and the sinoatrial node of the rabbit (2, 3, 32, 33, 35). I_TI in guinea pig ventricular cells was abolished when external sodium was reduced to 25% of normal, although mechanical oscillations continued, suggesting that I_NaCa is the major component of I_TI in this species (9). In canine coronary sinus, transient replacement of sodium with either lithium or sucrose caused a 30% decrease in I_TI amplitude, leaving open the possibility of further contributions from additional calcium-activated conductances (33). Conversely, a dissociation between forward-mode sodium/calcium exchange and I_TI was demonstrated for rabbit ventricular cells and rabbit Purkinje fibers (10, 12-14, 28).

Han and Ferrier (13) determined that isoproterenol-induced I_TI was dependent on calcium entry via reverse-mode sodium/calcium exchange. They suggested that isoproterenol directly activates reverse-mode sodium/calcium exchange, although a recent study by Main et al. (23) has found no evidence for -adrenergic stimulation of the exchanger. Calcium entry via reverse-mode sodium/calcium exchange was observed in the present study, but this loading of the cell with calcium only occurred after external sodium was reduced and the reversal of the exchanger was significantly altered. In effect, reverse-mode sodium/calcium exchange was an artifact of our experimental procedure. Moreover, calcium entry via the exchanger was abolished by elimination of internal sodium, without dramatically reducing the ability of isoproterenol to evoke I_TI.

I_Cl(Ca) represented ~40% of the total calcium-activated current at the resting potential in isoproterenol- and ouabain-treated cells. It is probable that the relative amplitudes of I_Cl(Ca) and I_NaCa will vary based on the method used to induce calcium overload and DAD. Whereas ouabain will elevate internal sodium and reduce the contribution of I_NaCa to the total calcium-activated conductance, isoproterenol appears to increase the relative contribution of I_NaCa. Other studies have used high extracellular calcium or BAY K 8644 to induce DAD, and the relative importance of the two calcium-activated conductances might be different from what we report. Our choice of isoproterenol and ouabain was made because we did not want to dramatically alter extracellular divalent ions, and isoproterenol alone did not reproducibly produce multiple triggered beats after a train of stimulated action potentials.

Positive to E_Cl, oscillating currents consisted of both inward and outward components, indicating that more than one conductance contributed to I_TI. We attributed the oscillating outward component to I_Cl(Ca), since it was reduced significantly when E_Cl was made equal to the step potential. In the absence of I_NaCa, we have previously shown that calcium overload-induced oscillating inward current is blocked by SITS and absent at E_Cl (38). Reversal of I_TI has been demonstrated for rabbit, sheep, and calf Purkinje fibers and rabbit ventricular myocytes (12-14, 19, 21).

As was the case for I_NaCa, I_Cl(Ca) was sufficiently large to cause DAD in the absence of other calcium-activated conductances. We were unable to record multiple triggered beats after substituting lithium for external sodium, even though a substantial oscillating inward I_Cl(Ca) was present. Furthermore, increasing the driving force for inward chloride current also failed to elicit multiple triggered action potentials. These results suggest that the threshold for sodium channels has been shifted in the presence of lithium. Although we examined the effects of lithium on action potentials and sodium current threshold, these experiments were always performed using pipette solutions containing EGTA to inhibit calcium-activated conductances. Although these protocols weigh the direct effects of lithium, they abolish any subsequent outcome of a lithium-induced rise of intracellular calcium. A rise in intracellular calcium will cause a shift in the voltage dependence of sodium channel gating, such that fewer channels are available at the resting potential (7, 11). Lithium might act in this indirect manner to change membrane surface charge and contribute to a shift in sodium channel threshold. We suggest that this shift contributes to a lack of repeated triggered beats under conditions in which I_Cl(Ca) is known to be large.

When recording action potentials, our method used to inhibit I_NaCa exaggerates I_Cl(Ca) due to an increase in contractility. However, given similar amplitudes of I_Cl(Ca) and I_NaCa when protocol-induced changes in contractility were abolished, we propose that I_Cl(Ca) contributes to the normal formation of DAD. Trafford et al. (31) demonstrated that in the ferret ventricle the time courses of I_NaCa and I_Cl(Ca) are different. We concur with these findings, since oscillating currents positive to E_Cl appear to have a biphasic waveform (see Fig. 6), and I_Cl(Ca) cannot simply be scaled to equal I_NaCa. Trafford et al. (31) suggested that in the ferret, these differences in kinetics result from subsarcolemmal compartmentalization of calcium. Similar compartmentalization might occur in the dog, although we cannot rule out differences in calcium affinity between the exchanger and chloride channel contributing to differences in time course of the two currents.

Our results regarding I_Cl(Ca) are consistent with an earlier demonstration of a SITS- and chloride-sensitive oscillating inward current in calcium-overloaded M cells and conform to similar findings in calcium-overloaded rabbit Purkinje fibers and ventricular cells (12-14, 21, 38). Of the very early investigations of the ionic basis of I_TI, some did not examine the contributions of a calcium-activated chloride conductance (1, 3). However, Kass et al. (19) found that substitution of external chloride did not appreciably affect strophanthidin-induced I_TI reversal and proposed that a nonselective leak conductance or electrogenic sodium/calcium exchange was the ionic basis of I_TI. In an interesting investigation of early afterdepolarizations and DAD in canine myocytes, Volders et al. (35) demonstrated that isoproterenol-induced DAD were abolished by nickel, a nonspecific blocker of I_NaCa. Although they hypothesized that I_NaCa alone caused DAD in these cells, the effects of nickel on I_Cl(Ca) were not examined. It is important to establish in such a study that the time course of the block of I_NaCa by nickel is the same time course as the block of DAD by nickel. Such a comparison would begin to establish whether nickel secondarily blocks additional conductances.

We have considered that the ability of isoproterenol to increase outward but not inward I_Cl(Ca) suggests that this is not a single conductance. Different ionic channels might underlie these outward and inward currents. However, we have shown that this inward current is calcium activated, blocked by removal of chloride, and sensitive to SITS. From these experiments, we conclude that all of the current remaining after inhibition of I_NaCa is I_Cl(Ca). We can only speculate on the failure of isoproterenol to increase inward I_Cl(Ca). Our results can be understood if two populations of chloride channels with different calcium sensitivities, and perhaps different rectification characteristics, are present in the canine midmyocardium. Papp et al. (25) have demonstrated two components of I_Cl(Ca) in rabbit cardiac Purkinje cells, suggesting that these components could arise from spatial and temporal inhomogeneities of calcium transients, or different chloride channel populations with a different calcium sensitivity. If a second population of chloride channels exists, our results indicate a higher calcium sensitivity for those channels carrying inward current, such that saturation of calcium binding sites obviates any potential isoproterenol-induced increase in current.

DAD and triggered activity were elicited in M cells, without activation of a nonselective cation conductance. With I_Cl(Ca) blocked, substitution of external sodium and concomitant inhibition of I_NaCa abolished all calcium-activated current, even though a substantial gradient persisted for flux through nonselective channels. Our inability to detect a leak conductance was not due to channel rectification, since we looked for both inward and outward currents carried by monovalent cations. Moreover, this conductance was absent despite application of isoproterenol and ouabain to elevate intracellular calcium, and with a clear demonstration that contractility was maintained. The lack of a nonselective cation conductance has been reported in rabbit atrial and ventricular myocytes, rabbit Purkinje cells, and canine ventricular myocytes (21, 25, 30, 38-40).

Our conclusions are in disagreement with those presented for the ionic basis of I_TI in rabbit Purkinje fibers. Han and Ferrier (13) propose that reverse-mode sodium/calcium exchange contributes to calcium overload, but that I_NaCa does not act as a charge carrier for I_TI. We are puzzled by this suggestion, since a rise in intracellular calcium should immediately result in activation of forward-mode sodium/calcium exchange and an oscillating inward current. They conclude that an inwardly rectifying nonselective cation conductance plays the most prominent role in generation of inward I_TI with some additional contribution from I_Cl(Ca) and that outward I_TI is primarily carried by I_Cl(Ca) (12, 14). If this were the case, their substitution of external sodium with impermeant NMG or sucrose should have resulted in oscillating outward currents at 55 mV, rather than the inward currents they ascribe to the nonselective cation conductance (14; Figs. 5 and 6). Voltage-clamp studies have characterized the nonselective cation conductance as linear, so it is expected that this conductance should contribute to both inward and outward I_TI if it is present in Purkinje fibers (6, 8). Consonant with our results in the dog, investigations of calcium-activated conductances in rabbit Purkinje cells have failed to find activation of nonselective cation channels, despite elicitation of a robust I_Cl(Ca) (25, 30).

A model of triggered activity under a maintained calcium overload similar to our experimental conditions details the relationship between the nonselective cation conductance and I_NaCa in the generation of DAD (22). In this model, the nonselective cation conductance is much more important than I_NaCa in generating triggered activity, in large part because of elevated internal sodium and concomitant reduction of I_NaCa. Internal sodium was also presumably elevated in our experiments due to blockade of the sodium pump. However, I_NaCa was still larger than I_Cl(Ca) probably because of the smaller electrochemical gradient for chloride.

Evidence from animal and human studies indicates reentry is the primary mechanism underlying arrhythmogenesis but that focal mechanisms might initiate ventricular arrhythmias under certain pathophysiological conditions. Three-dimensional cardiac mapping performed in dogs with ischemic cardiomyopathy induced by intracoronary embolizations revealed that monomorphic ventricular tachycardia was due to focal activation of subendocardial sites (26). Similarly, d-sotalol-induced polymorphic tachyarrhythmias with characteristics of torsade de pointes were initiated by spontaneous premature beats originating in the subendocardium (27). Such focal activation might arise as early afterdepolarizations or DAD originating in M cells or Purkinje fibers. We have determined the underlying mechanisms for DAD in M cells. An important next step to understanding the role these two conductances play during reperfusion- and drug-induced arrhythmias will be to determine if I_NaCa and I_Cl(Ca) are affected differently by intracellular pH, calcium loading of the SR, and the metabolic components of ischemia.