Abstract Transient currents are activated by spontaneous Ca2+ oscillations in rabbit ventricular myocytes. We investigated the ionic basis for these transient currents under conditions in which K+ currents would be expected to be blocked. Holding cells under voltage clamp at positive potentials leads to a rise in intracellular Ca2+ via reversal of the Na+-Ca2+ exchanger and subsequently to the initiation of spontaneous Ca2+ transients, presumably from a Ca2+-overloaded sarcoplasmic reticulum. The current transients associated with these Ca2+ transients reversed at about +10 to +15 mV under conditions of approximately symmetrical Cl−. In the absence of Cl−, this current was inward at all potentials examined over the range from −88 to +72 mV, consistent with a Na+-Ca2+ exchanger current. In the absence of Na+, the repetitive spontaneous Ca2+ transients could be initiated by a brief train of depolarizations to activate the inward Ca2+ current. Under such conditions, the current was found to reverse at −3 mV when the equilibrium potential of Cl− (ECl) was −2 mV, and the reversal potential shifted to −32 mV when internal Cl− was lowered, to make ECl −33 mV. Thus, in the absence of Na+, it appears that the current is exclusively a Ca2+-activated Cl− current. There is no evidence to indicate the presence of a Ca2+-activated cationic conductance. Further, our results demonstrate that the Ca2+-activated Cl− conductance can carry inward current at potentials more negative to ECl in rabbit ventricular myocytes and is therefore likely to contribute to the arrhythmogenic delayed afterdepolarizations that occur in Ca2+-overloaded cells.
Triggered release of Ca2+ from the SR has been demonstrated to underlie normal excitation-contraction coupling in cardiac myocytes. Fluorometric measurements of [Ca2+]i in voltage-clamped ventricular cells indicate that Ca2+ influx through voltage-gated channels regulates this release.1 However, under conditions to promote Ca2+ overload, for instance, inhibition of the Na+-K+ pump by cardiac glycosides2 3 or enhancement of Ca2+ influx by catecholamines,4 spontaneous recycling of Ca2+ between the SR and the cytosol is possible. Spontaneous Ca2+ oscillations in cardiomyocytes at the resting membrane potential are associated with a transient inward current (Iti) that causes delayed afterdepolarizations and may underlie some forms of triggered activity. Two possible mechanisms for the Iti accompanying a rise in cytosolic [Ca2+]i have been advanced: electrogenic Na+-Ca2+ exchanger Ca2+ extrusion and a Ca2+-activated membrane conductance.
The first possibility, that of INa-Ca, was proposed by Kass et al5 in an examination of Iti in strophanthidin-poisoned Purkinje fibers. Because an increase in [Ca2+]i would cause a positive shift in the equilibrium potential of the exchanger,6 an inward current would be expected to result from a Ca2+ oscillation; ie, because of the 3 Na+:1 Ca2+ stoichiometry of the exchanger, a transient Ca2+ elevation would produce either increased inward current (by the exchanger operating in Ca2+ efflux mode) below ENa-Ca or decreased outward current (by the exchanger operating in Ca2+ influx mode) at higher potentials. Consistent with this hypothesis, Lipp and Pott7 failed to detect a reversal of Iti in voltage-clamped cultured guinea pig myocytes over a wide range of membrane potentials.
Other investigators have proposed the existence of Ca2+-activated cation currents in cardiac myocytes, most typically nonselective in nature.5 8 This conductance appears to persist even under ionic conditions in which Na+-Ca2+ exchange would seem unlikely, such as replacement of all external Na+ with isotonic CaCl2.9 More recently, Zygmunt and Gibbons10 have demonstrated ICl(Ca) in rabbit ventricular myocytes. However, as described by these authors, ICl(Ca) is strongly rectifying, making it unlikely to contribute significantly to Iti.
In this study, we describe the characteristics of the oscillatory current (Iosc) that is activated by spontaneous Ca2+ transients in rabbit ventricular myocytes. We examined this current to determine its ion selectivity and voltage dependence. When studied under conditions designed to block K+ currents, the whole-cell current activated by spontaneous Ca2+ transients appeared to consist of at least two separate currents, a nonreversing inward INa-Ca and a Ca2+-activated Cl−-selective current. Under such conditions, we found no evidence to suggest the presence of a Ca2+-activated nonselective current. In contrast to the original characterization of ICl(Ca), we found that substantial inward current was present at potentials negative to ECl. The ICl(Ca) does not appear to be voltage gated. We conclude that both currents will contribute to the arrhythmogenic afterdepolarizations that result from spontaneous SR Ca2+ release.
Materials and Methods
Ventricular myocytes were generously provided by Dr Rajiv Kumar from the laboratory of Dr Ronald W. Joyner in the Department of Pediatrics, Emory University School of Medicine. All procedures performed on the animals were in accordance with institutional guidelines. The cells were isolated from 2- to 3-kg New Zealand White rabbit hearts by a procedure described previously.11 Briefly, heparinized rabbits were anesthetized with pentobarbital (100 to 150 mg IV). The chest was opened and the heart rapidly removed. The aorta was cannulated and the heart then perfused for 6 minutes with an oxygenated normal Tyrode’s solution containing (mmol/L): 149 NaCl, 4 KCl, 1.8 CaCl2, 0.5 MgCl2, 0.33 NaH2PO4, 5 glucose, 5 HEPES, pH 7.4 with NaOH. The heart was then perfused for 6 to 7 minutes with a nominally Ca2+-free Tyrode’s solution containing (mmol/L): 100 NaCl, 10 KCl, 1.2 KH2PO4, 5 MgSO4, 50 taurine, 20 glucose, 10 HEPES, pH 7.2 with KOH. Subsequently, perfusion was switched to the same solution containing 15 mg/200 mL collagenase (Yakult) and 7.5 mg/200 mL pronase E (Sigma) for 6 to 12 minutes at a flow rate of about 15 mL/min. After digestion was complete, the heart was perfused for 5 minutes with KB solution12 consisting of (mmol/L): 140 potassium glutamate, 5 MgCl2, 1 EGTA, 10 glucose, 10 HEPES, pH 7.4 with KOH. The heart was then cut into small pieces and myocytes were released by gentle trituration. Cells were stored in KB solution at 4°C for 1 to 5 hours before use.
Cells were patch clamped in the ruptured-patch whole-cell configuration13 using Corning 7052 borosilicate glass electrodes (2 to 5 MΩ) filled with one of the fura 2–containing pipette solutions (4 through 7) listed in the Table⇓, generally pipette solution 4 unless noted. Voltage clamp was achieved with an AxoPatch-1D amplifier (Axon Instruments). During recordings, the cell chamber contained one of the bath solutions listed in the Table⇓, generally bath solution 1 unless noted. When reversal potentials were measured under conditions in which Cl− was varied, the liquid junction potentials were measured and accounted for by using the method of Neher14 for each of the pipette and bath solution pairs employed. These measurements agreed to within 15% of the junction potentials calculated using the computer program JPCalc.15
The magnitude of the transient current associated with the Ca2+ oscillations was taken as the change in the current from the mean holding current at any given potential. In general, the maximum current observed to be associated with a particular Ca2+ oscillation was used in assessments of the voltage dependence and drug dependence. The reversal potential for the Iosc for each particular cell was estimated by a linear interpolation between the inward and outward current data points that bracketed the reversal point, and the mean reversal potential was the average of these individual estimates.
Fura 2 Fluorescence Measurements
Fura 2 fluorescence was measured with a high–temporal-resolution microfluorimeter consisting of a modified Zeiss IM-35 inverted fluorescence microscope outfitted with a Hamamatsu model 3460-04 PMT. A special filter wheel (fabricated by Omega Optical) controlled the excitation wavelengths. The filter wheel has two large transmissive wedges (150° each) that pass 340 and 380 nm light, respectively. Two small opaque wedges (30° each) reside between the larger transmissive sectors to prevent simultaneous excitation with two different wavelengths. The filter wheel is mounted to the shaft of a motor, which spins the filter in the excitation light path of the microscope. The excitation light from a xenon lamp source was thus rapidly alternated (≈110 Hz) between 340 and 380 nm. The fluorescence emission image of the cell was focused onto the photocathode of a PMT. A dichroic reflecting mirror in the emission path diverted the emission light (<590 nm) to the PMT but allowed longer wavelengths to pass to the microscope oculars. This technique allowed us to monitor the cell and its environment with bright-field optics using red light without significantly compromising the fluorescence detection efficiency. An emission filter centered at 530 nm (bandwidth 40 nm) was placed in front of the PMT to select the emission wavelengths. The field stop diaphragm in the excitation path was narrowed to illuminate only about two thirds of the cell (≈100 to 150 μm). In addition, a mask placed at an intermediate image plane in the emission path limited the field of view that was focused on the PMT to limit the background light signal. Generally, this mask was wider than a cell but about two thirds of a cell length long and positioned so as not to image the fura 2–filled patch pipette. Note that the recorded FI ratio represents the spatially averaged signal from the illuminated region of the cell. Thus, when the signal derives from propagating Ca2+ waves, the characteristics of the recorded transient are determined not only by the rate of Ca2+ release and reuptake but also by the path and propagation speed of the wave front within the cell.
The PMT was connected to a custom-designed photon-counting circuit interfaced to an IBM-compatible computer. Control and data acquisition software was written to synchronize the operation of the photon counter to the revolution of the filter wheel. The program stored fluorescence data for each wavelength once per revolution, sampled and stored membrane potential and membrane current three times per revolution, and sampled and stored other parameters (eg, stimulus markers) once per revolution. The 340- and 380-nm fluorescence counts used to compute the 340/380 FI ratio were corrected for the background fluorescence and cell autofluorescence measured before rupturing the patch.
DNDS was generously provided by Dr Roger Worrell, Emory University. DNDS was made as a 1-mmol/L extracellular solution and applied locally to cells by pressure ejection from a pipette positioned near the cell. Fura 2–free acid was purchased from Molecular Probes. All other reagents were obtained from Sigma Chemical Co. Except where noted, values for various measurements are reported as the mean±SEM.
Positive Polarization Induces Ca2+ and Current Oscillations
We initially observed Ca2+-dependent current oscillations under conditions in which cells were held under voltage clamp at positive membrane potentials for prolonged intervals. Cells were voltage clamped under conditions designed to block K+ currents (see the Table⇑, bath solution 1 and pipette solution 4) and subjected to positive polarization to elevate the [Ca2+]i via reverse Na+-Ca2+ exchange. Under these conditions, holding cells at positive potentials for several seconds resulted in the elevation of the cytosolic [Ca2+]i and the initiation of spontaneous Ca2+ oscillations. Such Ca2+ oscillations have been observed by many other investigators in cardiac myocyte preparations from many species3 16 17 18 and are believed to result from the overloading of the SR with Ca2+. At positive membrane potentials, conditions favor the influx of Ca2+ via the Na+-Ca2+ exchanger. For example, given the concentrations of Na+ and Ca2+ in the pipette and bath, holding the membrane potential at +36 mV would favor Ca2+ influx whenever the [Ca2+]i is <≈10 μmol/L. This elevated [Ca2+]i leads to an enhanced accumulation by the SR and eventually to spontaneous SR Ca2+ release by a mechanism that is not fully understood.18 19 20
An example of such an event is shown in Fig 1A⇓. Note that when the membrane potential was stepped from −64 to +36 mV, there was a relatively slow elevation of the [Ca2+]i, as indicated by the increase in the 340/380 FI ratio, due to the reverse Na+-Ca2+ exchanger influx. This elevated SR-pump substrate concentration led to Ca2+ loading of the SR, followed by the initiation of spontaneous Ca2+ transients. Outward currents were associated with these Ca2+ transients at +36 mV. Returning to negative potentials generally resulted in the suppression of the Ca2+ oscillations, although occasionally one or two transients would occur before cessation. On these occasions, inward currents were observed. Although it is not evident from the data shown in Fig 1⇓, the [Ca2+]i elevation is not synchronous throughout the cell. Rather, as many other investigators have shown,18 20 21 the [Ca2+]i rise appears to be initiated in one discrete location and subsequently propagates throughout the cell at a rate on the order of 100 μm/s.18 Such oscillations have been termed “calcium waves,” and they initiate contractile waves that follow in their wake. Because fluorescence was recorded from a rather large part of the cell on our setup, this relatively slow propagation speed resulted in Ca2+ transient durations that appear prolonged relative to transients that were initiated synchronously.
As is also evident in the example shown in Fig 1A⇑, these Ca2+ oscillations are associated with oscillations, or transients, in the membrane holding current. As described below, the currents could be inward or outward depending on the holding potential. Because of the common mechanism of activation of the inward and outward components under these conditions, we have chosen to refer to such currents as Iosc to reflect this fact. Specifically, we have avoided the more familiar terms Ito(Ca) and Iti that are commonly used to refer to currents elicited under similar circumstances.
These spontaneous Ca2+ oscillations occurred more frequently at positive potentials (where Ca2+ influx via the Na+-Ca2+ exchanger is favored) and were absent or rare at negative potentials. However, by first holding the cell for several seconds at a positive potential to load the SR, we would often observe an oscillation on stepping back to a more negative potential. By this method of using positive polarization to initiate spontaneous Ca2+ oscillations, we were able to examine the voltage dependence of this current. The result from this series of experiments is shown in Fig 1B⇑. The average reversal potential of Iosc was somewhere between +5 and +15 mV, although it was different in each cell. A difficult variable to control in these experiments was the change in[Ca2+]i at different potentials. For example, the [Ca2+]i oscillations that occurred at negative potentials started at lower Ca2+ levels (because the [Ca2+]i would start to fall immediately on stepping to the negative potential; see Fig 1A⇑) and were somewhat smaller in magnitude. Thus, one variable that affects these currents (the [Ca2+]i) was not independent of potential. Furthermore, even at a single potential, there was marked variability in both the Ca2+ and current magnitudes (eg, note the variability in the current magnitude in Fig 1A⇑ on return to +36 mV). We attribute one aspect of this variability to differences in the starting location and propagation path of the Ca2+ wave. Because of this variability, the current magnitudes plotted in Fig 1B⇑ were normalized to the current magnitude observed during the immediately preceding +36-mV holding period. In this way, we attempted to normalize for cell-to-cell variability and to account, as best we could, for time-dependent processes such as the degree of SR loading produced by holding the cell at the positive potential. Because of the variations in the Ca2+-transient signal observed at different potentials and our inability to record the spatial characteristics of the Ca2+ transients, we hesitate to draw firm conclusions from the shape of the I-V relation. Nevertheless, it is clear that Iosc can carry significant charge in either direction depending on the membrane potential.
What is Iosc?
Iosc is undoubtedly a current activated by the elevated [Ca2+]i. A number of Ca2+-activated currents have been described or postulated in cardiac myocytes, including specific K+22 and Cl− currents,10 23 24 a nonselective cation current,8 and INa-Ca.5 6 7 Under the conditions used in our study (eg, Cs+ and TEA on both sides of the membrane), it is unlikely that the current oscillations we observed were due even in part to Ca2+-activated channels selective for K+. From the data to be presented below, we feel that it is likely that Iosc consists of at least two distinct Ca2+-activated currents: a Cl− current and INa-Ca.
The existence of ICl(Ca) has previously been demonstrated in rabbit ventricular myocytes by Zygmunt and Gibbons.10 We examined the involvement of a Cl− conductance in these Ca2+-activated current oscillations with a number of approaches. We first examined the action of the reversible anion channel blocker DNDS and found that it attenuated the outward current observed with Ca2+ oscillations at +36 mV. An example of such an experiment is seen in Fig 2⇓. A 1-mmol/L DNDS solution was pressure ejected from a nearby pipette onto a cell that was exhibiting spontaneous Ca2+ and current oscillations at a holding potential of +36 mV. Shortly after the application of DNDS, there was a reduction in the peak outward current amplitude. Overall, the mean peak current amplitude fell to 45±10% of the pre-DNDS control value and recovered within 2 minutes of terminating the DNDS application to 90±30% of control (six observations in five cells). Further, we also observed that the holding current becomes less outward in the presence of DNDS. This finding suggests that there was a sustained Cl− conductance at +36 mV that contributed to the holding current.
The pressure and aim of the puffer pipette were adjusted to avoid disturbing the position of the cell with respect to the patch pipette. Thus, we are not confident that the DNDS was applied to the whole cell surface, nor can we be certain that the delivered concentration was sufficient to completely suppress Cl− currents. Indeed, because we rarely observed inward currents under these conditions (an expectation based on experiments to be described later), we feel it is likely that such application did not completely block the Cl− currents.
DNDS application also lowered the overall 340/380 FI ratio and reduced the amplitude of the fluorescence ratio transients associated with the current oscillations. However, DNDS is fluorescent at both 340 and 380 nm excitation; therefore, in the presence of DNDS, the fura 2 FI ratio was no longer a reliable indicator of the [Ca2+]i. However, there is good circumstantial evidence to suggest that the Ca2+ transients continued unabated in the presence of DNDS. All cells exposed to DNDS continued to contract in a wavelike manner during DNDS application, indicating that the [Ca2+]i rise was sufficient to initiate this process. Moreover, an examination of the fluorescence resulting from excitation at both wavelengths used to compute the FI ratio (ie, 340 and 380 nm) indicates that the amplitude of the fluorescence change that occurred during the oscillations in all these cells was not significantly altered by DNDS. If these fluorescence transients were due solely to the rise and fall of [Ca2+]i (and not, for example, to an abrupt change in DNDS concentration during the time course of the transients) and if the DNDS did not significantly affect the “resting” [Ca2+]i, then the amplitude of the transients at any one wavelength should remain proportional to the [Ca2+]i change. Although this assessment is problematic during the pressure ejection period owing to movement of the cell with respect to the illumination spot, an analysis of the 380-nm fluorescence during oscillations that followed the termination of DNDS confirms that the Ca2+ transients were not significantly attenuated. The mean Δ380 FI for the three transients (labeled with dot pairs) that occurred immediately after the termination of the DNDS ejection in Fig 2⇑ was 96% of the mean Δ380 FI of the three transients (labeled with single dots) just before DNDS application, whereas the mean peak current amplitude was attenuated 50%. Thus, DNDS application significantly reduced the peak outward current amplitude but did not appear to significantly alter the Ca2+ transient amplitude. This fact strongly suggests the presence of a Cl− conductance that was activated by the elevated [Ca2+]i.
Although this evidence suggests that Iosc involves ICl(Ca), the reversal potential indicated in Fig 1B⇑ (ranging between +5 and +15 mV) is significantly more positive than the −2-mV ECl under these conditions. Further, while ICl(Ca) can account for the outward component of Iosc, ICl(Ca) has been described as a strongly outwardly rectifying current in rabbit ventricular myocytes. Another current that should be altered by a change in the [Ca2+]i is INa-Ca.5 Indeed, such currents have previously been found to correlate to spontaneous [Ca2+]i rises.7 INa-Ca is in the direction of the movement of Na+ (since the exchange is believed to be three sodium ions for one calcium ion25 ), and so an increase in the [Ca2+]i would either attenuate an outward current or enhance an inward current, depending on in which direction the exchanger was operating when the [Ca2+]i changed. Hence, the change in INa-Ca with an abrupt increase in the [Ca2+]i will always be in the inward direction, regardless of the holding potential or the gradients of the two ions. To examine the contribution of INa-Ca to Iosc, we conducted a series of experiments in the absence of extracellular or intracellular Cl− (Table⇑; solutions 2 and 5, respectively). Although we were concerned about the accuracy of the voltage measurements under these conditions (Ag/AgCl electrodes were used for both pipette and bath coupling), the voltage dependence of the L-type Ca2+ current indicated that it was close: peak ICa normally occurs at +10 mV, and in these experiments in the absence of Cl−, we found it was within 10 mV of this value. As shown in Fig 3⇓, when we eliminated Cl− and then elicited Ca2+ oscillations (by intermittent voltage-clamp intervals at positive potentials), an inward Iosc was observed at all potentials (from −88 mV to +72 mV). As was the case in the presence of Na+ and Cl− (Fig 1⇑), the peak current magnitude was variable from transient to transient even in the same cell. However, although the current magnitude demonstrated considerable variability, no reversal potential or strong voltage dependence was evident, a result that would be expected if the current was due to the Na+-Ca2+ exchanger.
Do Other Currents Contribute to Iosc?
In the presence of Cs+ and TEA, it appears that Iosc might consist of just two components, ICl(Ca), as has been described by Zygmunt and Gibbons,10 and INa-Ca. To test this hypothesis, we examined Iosc in the presence of Cl− under conditions that block the Na+-Ca2+ exchanger. We reasoned that if these two currents are solely responsible for Iosc, then eliminating INa-Ca should leave just ICl(Ca). We used NMDG as a Na+ substitute in the extracellular and electrode solutions (Table⇑, solutions 3 and 6) to prevent Na+-Ca2+ exchange. Although this approach prevented us from overloading the SR by reverse Na+-Ca2+ exchange, we found that a short train of repetitive 300-millisecond depolarizations (≈10 at 2 Hz) from a holding potential of −80 to +10 mV to activate the inward ICa was sufficient to overload the SR to the point of inducing Ca2+ oscillations. In fact, we found that the oscillations induced under these conditions occurred with a remarkably regular frequency and with an amplitude that was for the most part independent of voltage, and the subsequent current transients were less variable as a result. Fig 4A⇓ shows a representative series of currents activated by spontaneous Ca2+ transients at a range of holding potentials in the absence of Na+ under conditions of approximately symmetrical Cl− (ECl=−2 mV). Although the magnitude of the Ca2+ transients was fairly invariant with potential, the associated Iosc varied in both magnitude and direction. Fig 4B⇓ shows the I-V relation from this series of experiments. As anticipated, the current was outward at potentials more positive to ECl. Moreover, the current reversed very close to ECl (−3±2 mV, n=3) and appeared to be only somewhat outwardly rectifying. Note that it was clearly inward at negative potentials. This sizable inward component was not expected, as ICl(Ca) described by Zygmunt and Gibbons10 in rabbit ventricular myocytes was quite strongly outwardly rectifying.
We considered two possible explanations for the persistence of an inward current in the absence of Na+: either an additional Ca2+-activated inward current was also present or ICl(Ca) can indeed carry significant inward current in rabbit ventricular myocytes. To further clarify this question, we tried shifting ECl to see how the reversal potential of Iosc changed. ECl was set equal to −33 mV in the absence of Na+ using solutions 3 and 7 from the Table⇑, and Ca2+ oscillations were induced as described above. Fig 5⇓ shows the resulting I-V relation from this series of experiments. Iosc was observed to reverse at −32±4 mV (n=4 cells), very close to the new ECl. In contrast to the situation of approximately symmetrical Cl−, the current under these conditions was more strongly rectifying, but it was again clearly inward at more negative potentials. Comparing the observations in Figs 4B⇑ and 5⇓, for a −31-mV shift in ECl, the reversal potential of Iosc shifted −29 mV. No other ionic species experienced an appreciable change in its equilibrium potential except the Cl− substitute aspartate. These observations are entirely consistent with the hypothesis that Cl− is the charge carrier of this current under these conditions and argue against the significant involvement of other conductances. Further, in a separate series of experiments, under Cl− conditions where Iosc was found to reverse at −49±2 mV (n=4; data not shown), the isoproterenol-induced time-independent ICl(cAMP)26 was found to reverse at −51±4 mV (n=3), indicating that these two currents have a similar Cl− selectivity. Collectively, these data strongly support the hypothesis that, in the absence of Na+ and in the presence of the K+ channel blockers Cs+ and TEA, Iosc is the ICl(Ca). Therefore, it appears that ICl(Ca) can carry inward current in rabbit ventricular myocytes when activated at a potential more negative to ECl.
[Ca2+]i Sensitivity of Iosc
Our fluorescence recordings of the [Ca2+]i changes lack the spatial information required to make precise correlations between the current and the [Ca2+]i, and thus we did not explore this relation in detail. However, several observations are worth noting. We monitored the averaged fura 2 fluorescence in only about two thirds of the cell volume, although the voltage clamp amplifier should measure a current whenever and wherever it is activated. Thus, if a spontaneous Ca2+ wave was initiated at the nonilluminated end of the cell, the current would be detected immediately, but the fluorescence signal would change only after the wave reached the illuminated portion of the cell. Knowing this, our expectation would be that if Iosc was proportional simply to the mean [Ca2+]i indicated by the fura 2 FI ratio, then the duration of the current transient should be either equal to or longer than the Ca2+ transient. But, to the contrary, the current transients (measured as the deflection from the mean holding current) were shorter in duration than Ca2+ transients. This occurrence is evident in several of the records presented here (eg, Figs 1A⇑ and 2⇑) and is true even in the absence of the Na+-Ca2+ exchanger (eg, see Fig 4A⇑), which suggests that either the current inactivates or is sensitive to a higher [Ca2+]i than is indicated by the fura 2 FI ratio. Although our data don’t rule out inactivation as a mechanism for this difference in time course, we note that in addition to suppressing the transient current, DNDS also attenuated the basal holding current recorded at +36 mV, suggesting that a Cl− conductance was present even at times between the Ca2+ transients. Since holding at +36 mV elevated the mean “resting” [Ca2+]i, it is possible that the ICl(Ca) was partially activated by this modest [Ca2+]i elevation.
In any case, the discrepancy between the time course of the current and Ca2+ transients indicates that the magnitude of the current may be more indicative of the ongoing release process, perhaps reflecting the surface area of the Ca2+ wave front (where, as a consequence of the release process, the [Ca2+]i gradient is large) rather than of the fraction of the cell volume that has an elevated [Ca2+]i. In further support of this hypothesis, we note that the current observed is often variable and “spiky” in amplitude (eg, see initial five current transients in Fig 2⇑). Recently, Lipp and Niggli,21 using confocal microscopy to obtain detailed spatial information, noted the sometimes rather complex and tortuous paths that Ca2+ waves can travel within cells. The variability and spikiness of the currents we measure would be qualitatively consistent with the variations in Ca2+ wave front surface observed by these investigators.
In this report, we have examined the ionic basis for the transient currents (Iosc) that are associated with spontaneous Ca2+ oscillations in rabbit ventricular myocytes. Under conditions designed to block K+ currents, Iosc can be either inward or outward, depending on the holding potential, with a reversal potential more positive to ECl. The anion-channel blocker DNDS attenuated the outward component of the current observed at +36 mV. In the absence of Cl−, the current was inward at all potentials, consistent with INa-Ca. In the absence of Na+, the current again was either inward or outward, depending on the potential, with a reversal potential virtually identical to ECl, consistent with a Cl− current. Hence, two currents likely contribute to Iosc: ICl(Ca) and INa-Ca. The duration of Iosc was briefer than the initiating Ca2+ transient, suggesting that these currents are both sensitive to a higher [Ca2+]i than the weighted mean value indicated by the fura 2 FI ratio. Furthermore, the activation and time course of these two currents appear to be similar, suggesting a roughly similar Ca2+ sensitivity. Under ionic conditions that permit both currents to be present, the reversal potential was variable. Such variability would be expected if the current is composed of two separate and opposing currents and if there is cell-to-cell variability in the relative proportion of the two component currents. The reversal potential of Iosc was much less variable when INa-Ca was eliminated.
We have deliberately avoided using the terms “Iti” and “Ito(Ca)” that are commonly used to classify the transient currents observed in cardiac cells. Over the last several years, it has become increasingly evident that several underlying currents are common to both, and thus such classifications based on the experimental parameters used to elicit them appear increasingly artificial. We feel it is reasonable to conclude that Iti consists of an inward INa-Ca along with (depending on species and tissue) ICl(Ca) and/or a Ca2+-activated cationic current. Ito very likely consists of outward manifestations of the Cl− and cationic currents (if present in Iti) and INa-Ca (possibly varying in direction with time), in addition to a 4-aminopyridine–sensitive transient K+ current.
The use of spontaneous Ca2+ oscillations to study Ca2+-dependent currents presents a number of advantages for assessing the voltage and ionic dependence of these currents. The most important as it relates to this investigation is that detection of the Ca2+-activated transient currents was not complicated by capacitance transients or by the simultaneous activation of voltage-dependent transient currents. Thus, the subtraction of current measurements made in the presence of blockers was not required to reveal the specific Ca2+-activated component. This made the task of estimating the reversal potential of the current rather straightforward and less uncertain. On the other hand, activation of the current is not spatially and temporally uniform, since it occurs as a consequence of a relatively slowly propagating wave of Ca2+ release, which in turn often had different characteristics at different potentials. Because our high–time-resolution recording apparatus was not capable of recording the spatial characteristics of the Ca2+ transients, we were unable to assess directly how these characteristics modulated the current. This variability in the underlying Ca2+ transients undoubtedly contributes to the variability in the general appearance of the I-V relation; therefore, assessments of the degree of rectification of the currents are less reliable when they are activated in this way.
Characteristics of ICl(Ca)
In the absence of Na+, Iosc appeared to be exclusively ICl(Ca). The current was moderately outwardly rectifying under conditions of approximately symmetrical Cl− and more strongly rectifying when ECl was −33 mV. The reversal potential under either Cl− gradient condition was virtually identical to ECl. Since no other ion gradient changed in a manner that would be consistent with this observation, we conclude that the current in the absence of Na+ is likely to be almost exclusively ICl(Ca). Our results also demonstrate that ICl(Ca) can be activated at all physiological potentials. Under approximately symmetrical Cl− gradient, the current appeared somewhat outwardly rectifying, but not extremely so. Thus, voltage does not appear to alter the gating characteristics of the underlying channels significantly. Therefore, this current will be present whenever [Ca2+]i is elevated, irrespective of the reason.
Our observations demonstrate that ICl(Ca) is briefer in duration than the associated Ca2+ transient indicated by the average fura 2 FI ratio. We suggest that this is largely because of the lower sensitivity to [Ca2+]i for activating this current, such that it is activated to a significant extent primarily by the very high [Ca2+]is that exist for a brief time at the Ca2+ release sites. Nonetheless, a small current may be activated by elevations in the resting [Ca2+]i that can be achieved by reverse Na+-Ca2+ exchange influx at positive membrane potentials, as indicated by the ability of DNDS to attenuate the outward holding current observed at +36 mV.
The characteristics of this current as observed here differ in one important respect from the initial description by Zygmunt and Gibbons10 : we observed substantial inward current at potentials negative to ECl, whereas those investigators did not. Zygmunt and Gibbons activated the current by step depolarizations that activated ICa and then quantified the SITS-sensitive component of the tail currents evident on repolarization to more negative potentials. They observed a strongly outward-rectifying current; indeed, they observed no current at potentials more negative to +20 mV under conditions in which ECl was as negative as −23 mV, which implied a voltage dependence to the gating of these channels. It is unclear what the explanation is for this difference in our observations. In contrast to our protocol, the subtraction of current records made at different times in the presence of the irreversible anion channel blocker SITS was required to account for other components of the tail current, and one possibility is that perhaps rundown of ICa (and its tail current) might have obscured ICl(Ca) at these potentials at which its magnitude was small. On the basis of the close agreement we observed between ECl and the reversal potential of Iosc under conditions of different internal Cl−, we feel it is unlikely that another Ca2+-activated current is responsible for the inward component.
Since their initial report on rabbit ventricular myocytes, Zygmunt and Gibbons24 and Zygmunt23 have characterized similar Cl− currents in rabbit atria and in canine ventricular cells. In both studies, these investigators did observe inward ICl(Ca) that reversed at potentials close to ECl. In light of our observations, it is reasonable to assume that ICl(Ca) in rabbit ventricular myocytes is the same channel present in rabbit atrial myocytes and that this channel is very similar to that expressed in canine ventricular myocytes.
In the absence of Cl−, the observed current was always inward, even at potentials more positive than +70 mV. With the exception of Ca2+, all other ionic species had an equilibrium potential more negative than this; therefore, the observed current is unlikely to represent the flux of a single ion species. On the other hand, an apparent inward transient current would be expected from the Na+-Ca2+ exchanger, with a transient elevation of the internal [Ca2+]i, although the observed current may not necessarily represent a true inward current.5 7 ENa-Ca is equal to 3 ENa−2 ECa,25 and thus the effect of a [Ca2+]i elevation would be to shift this equilibrium potential in the positive direction. If the [Ca2+]i rose when the membrane potential was held more negative than ENa-Ca, there should be an enhancement of the inward current. However, if the [Ca2+]i rose when the membrane potential was held more positive than ENa-Ca, a decrease in the outward current would be expected (perhaps becoming inward, depending on the potential and the extent of [Ca2+]i rise). In all cases, however, a deflection from the steady state current in the inward direction would be expected. As we showed in Fig 3⇑, elimination of chloride ions did result in a current that was inward at all potentials examined.
Contribution of Other Ca2+-Activated Currents to Iosc
The role of other Ca2+-activated currents cannot be completely excluded by our results, although we feel the evidence makes it unlikely that other such currents were present under the conditions employed. The existence of a Ca2+-activated K+ current has not been convincingly documented in rabbit ventricular myocytes, and because of the use of TEA and Cs+ in our pipette and bathing solutions, they likely would not have contributed to the currents we observed if they did exist. We are confident that under the zero Na+ conditions, the current we observe is overwhelmingly ICl(Ca): the current’s reversal potential shifted with a change in the Cl− gradient, and under both conditions, it reversed very close to the calculated ECl. The cardiac nonselective cationic channel described by Ehara et al8 does conduct Cs+. Therefore, the inability of Cs+ to influence the reversal potential of the transient current in our hands suggests that either such a conductance is absent in rabbit ventricular myocytes or it is blocked by TEA. However, we cannot completely rule out the contribution of other currents under conditions in which Na+ was present. Cannell and Lederer9 proposed the existence of a Ca2+-activated nonselective current that can pass K+ and Ca2+ in sheep Purkinje fibers on the basis of behavior of “delayed afterdepolarization” currents. However, this earlier work did not consider the possibility of the more recently described ICl(Ca), and some of their cationic substitutions resulted in [Cl−] changes as well. In a more recent study, Han and Ferrier27 detected the Cl− dependence of this Ca2+-activated current in rabbit Purkinje fibers but noted that the reversal potential was also altered by changes in the calcium ion gradient, supporting the view that a Ca2+-activated cationic conductance capable of passing Ca2+ was also present. Although there are significant methodological differences between these studies and ours, it seems reasonable to conclude that Purkinje fibers differ from ventricular cells in this respect. Our results showing (1) virtual identity of the current’s reversal potential to ECl under two different Cl− gradients and (2) the similarity of the reversal potential of ICl(Ca) and the sustained ICl(cAMP) are not compatible with the presence of a significant Ca2+-activated cationic conductance in rabbit ventricular cells under the conditions of our study.
Role of Ca2+-Activated Currents
Although Ca2+-activated transient currents undoubtedly influence the characteristics of the normal ventricular action potential, perturbations in Ca2+ cycling make them especially likely to contribute to triggered activity under pathological conditions. Estimates of the internal Cl− activity in rabbit heart range from ≈15 to 20 mmol/L.28 Thus, ECl is likely to be in the range of ≈−40 mV. Under these conditions, one would expect that an outward ICl(Ca) would normally be transiently activated shortly after the initial action potential depolarization. However, because this current is most sensitive to the high [Ca2+]is that exist at moments of SR Ca2+ release, it would probably subside within a few tens of milliseconds, well before the mean [Ca2+]i declined. As a consequence, this current should contribute to the initial repolarization after the action potential spike, and perhaps for a short period it would also contribute to a lowering of the plateau potential. This ICl(Ca) will subside later during the plateau phase of the action potential as the internal Ca2+ release process is terminated and thus will likely have only a minor influence on the latter part of the action potential waveform.
On the basis of thermodynamic calculations of ENa-Ca, it has been postulated that this transporter might operate in reverse mode early in depolarization (before a significant Ca2+ rise),29 30 and experimental evidence,30 31 albeit controversial,32 has led to the suggestion that this influx of Ca2+ could contribute to triggering SR Ca2+ release. In reverse mode, INa-Ca would be outward, but on SR Ca2+ release, the magnitude of this outward current should be attenuated and likely reversed. Thus, as a consequence of the SR Ca2+ release (however triggered), a net inward current will be activated, having the effect of driving the membrane potential more positive. The simultaneous activation of ICl(Ca) would tend to oppose this action of the change in INa-Ca. One possible consequence of this effect would be to enhance the magnitude of ICa. During the action potential, the initial depolarization of the cell membrane is likely sufficient to activate most available Ca2+ channels. By acting to keep the membrane potential less positive, the electrochemical driving force for Ca2+ would be enhanced, resulting in a larger current for a given number of open Ca2+ channels, and possibly a larger SR Ca2+ release.23
Delayed afterdepolarizations are frequently observed after repolarization under conditions favoring SR Ca2+ overload. These afterdepolarizations have been the focus of a number of studies, and results from several laboratories have implicated the Na+-Ca2+ exchanger and a Ca2+-activated cation current as possible contributors to the depolarization. Because the Ca2+-activated Cl− conductance was not thought to carry inward current in rabbit ventricular cells, it seemed unlikely to contribute to these events. However, our results clearly indicate that this current can carry inward current, and thus a role for this current in afterdepolarizations is likely.
The presence of a Ca2+-activated current has special implications for ICa and [Ca2+]i studies that require or rely on quantitative assessment of the magnitude and time course of ICa. As was illustrated by Zygmunt and Gibbons,10 the apparent rate of ICa inactivation will be altered due to activation of ICl(Ca). This alteration will frustrate efforts to quantify the net Ca2+ entry from the ICa time course. Precautions should therefore be taken in such studies on cardiac preparations that exhibit this current, such as setting ECl close to the typical command potentials employed. So far, ICl(Ca) has been confirmed in rabbit atria,24 ventricles,10 and Purkinje fibers,27 34 in dog ventricles,23 in ferret ventricles,35 and in cultured chick myocytes.36 On the other hand, Cl− substitutions had little effect on the reversal potential of Iti in bovine Purkinje fibers.5 In preliminary experiments, we have observed little evidence for a significant ICl(Ca) in rat ventricular myocytes.
Selected Abbreviations and Acronyms
|ECa, ECl, ENa, and ENa-Ca||=||equilibrium potential of Ca2+, Cl−, Na+, and Na+-Ca2+ exchanger, respectively|
|380 FI||=||FI at 380-nm excitation|
|340/380 FI ratio||=||ratio of FIs computed at 340 and 380 nm|
|ICa, INa-Ca, and Iosc||=||Ca2+, Na+-Ca2+ exchanger, and oscillatory currents, respectively|
|ICl(Ca)||=||Ca2+-activated Cl− current|
|ICl(cAMP)||=||cAMP-activated Cl− current|
The authors wish to thank Drs Rajiv Kumar and Ronald Joyner for preparing and providing us with the rabbit ventricular myocytes, Dr Roger Worrell for generously providing the DNDS and for helpful discussions, and Linda Hereford for valuable technical assistance.
- Received May 4, 1995.
- Accepted January 11, 1996.
- © 1996 American Heart Association, Inc.
Priori SG, Corr PB. Mechanisms underlying early and delayed afterdepolarizations induced by catecholamines. Am J Physiol. 1990;258:H1796-H1805.
Zygmunt AC, Gibbons WR. Calcium-activated chloride current in rabbit ventricular myocytes. Circ Res. 1991;68:424-437.
Osaka T, Joyner RW. Developmental changes in calcium currents of rabbit ventricular cells. Circ Res. 1991;68:788-796.
Lakatta EG. Functional implications of spontaneous sarcoplasmic reticulum Ca2+ release in the heart. Cardiovasc Res. 1992;26:193-214.
Berlin JR, Cannell MB, Lederer WJ. Cellular origins of the transient inward current in cardiac myocytes: role of fluctuations and waves of elevated intracellular calcium. Circ Res. 1989;65:115-126.
Takamatsu T, Wier WG. Calcium waves in mammalian heart: quantification of origin, magnitude, waveform, and velocity. FASEB J. 1990;4:1519-1525.
Cheng H, Lederer WJ, Cannell MB. Calcium sparks: elementary events underlying excitation-contraction coupling in heart muscle. Science. 1993;262:740-744.
Zygmunt AC. Intracellular calcium activates a chloride current in canine ventricular myocytes. Am J Physiol. 1994;267:H1984-H1995.
Zygmunt AC, Gibbons WR. Properties of the calcium-activated chloride current in heart. J Gen Physiol. 1992;99:391-414.
Sheu SS, Blaustein MP. Sodium/calcium exchange and control of cell calcium and contractility in cardiac and vascular smooth muscles. In: Fozzard HA, Haber E, Jennings RB, Katz AM, Morgan HE, eds. The Heart and Cardiovascular System: Scientific Foundations. 2nd ed. New York, NY: Raven Press; 1991:903-943.
Harvey RD, Hume JR. Isoproterenol activates a chloride current, not the transient outward current, in rabbit ventricular myocytes. Am J Physiol. 1989;257:C1177-C1181.
Baumgarten CM, Fozzard HA. Cardiac resting and pacemaker potentials. In: Fozzard HA, Haber E, Jennings RB, Katz AM, Morgan HE, eds. The Heart and Cardiovascular System: Scientific Foundations. 2nd ed. New York, NY: Raven Press; 1991:963-1001.
Levi AJ, Spitzer KW, Kohmoto O, Bridge JH. Depolarization-induced Ca entry via Na-Ca exchange triggers SR release in guinea pig cardiac myocytes. Am J Physiol. 1994;266:H1422-H1433.
Leblanc N, Hume JR. Sodium current–induced release of calcium from cardiac sarcoplasmic reticulum. Science. 1990;248:372-376.
Sham JS, Cleemann L, Morad M. Gating of the cardiac Ca2+ release channel: the role of Na+ current and Na+-Ca2+ exchange. Science. 1992;255:850-853.
Liu S, Lieberman M. Electrodiffusive movements of chloride ion in sodium-free solution: a possible calcium activated chloride conductance in cultured heart cells. Cardiovasc Res. 1994;28:1629-1634.