Dynamic Ca2+-Induced Inward Rectification of K+ Current During the Ventricular Action Potential
Abstract—Inward rectification, an important determinant of cell excitability, can result from channel blockade by intracellular cations, including Ca2+. However, mostly on the basis of indirect arguments, Ca2+-mediated rectification of inward rectifier K+ current (IK1) is claimed to play no role in the mammalian heart. The present study investigates Ca2+-mediated IK1 rectification during the mammalian ventricular action potential. Guinea pig ventricular myocytes were patch-clamped in the whole-cell configuration. The action potential waveform was recorded and then applied to reproduce normal excitation under voltage-clamp conditions. Subtraction currents obtained during blockade of K+ currents by either 1 mmol/L Ba2+ (IBa) or K+-free solution (I0K) were used to estimate IK1. Similar time courses were observed for IBa and I0K; both currents were strongly reduced during depolarization (inward rectification). Blockade of L-type Ca2+ current by dihydropyridines (DHPs) increased systolic IBa and I0K by 50.7% and 254.5%, respectively. β-Adrenergic stimulation, when tested on I0K, had an opposite effect; ie, it reduced this current by 66.5%. Ryanodine, an inhibitor of sarcoplasmic Ca2+ release, increased systolic IBa by 47.7%, with effects similar to those of DHPs. Intracellular Ca2+ buffering (BAPTA-AM) increased systolic IBa by 87.7% and blunted the effect of DHPs. Thus, IK1 may be significantly reduced by physiological Ca2+ transients determined by both Ca2+ influx and release. Although Ca2+-induced effects may represent only a small fraction of total IK1 rectification, they are large enough to affect excitability and repolarization. They may also contribute to facilitation of early afterdepolarizations by conditions increasing Ca2+ influx.
Several known K+ channels have a lower resistance to inward than to outward current flow, a property referred to as “inward rectification.” Inward rectification, particularly pronounced in IK1, is of pivotal importance in determining membrane excitability in cardiac myocytes.1 IK1 inward rectification is currently interpreted as a voltage-dependent blockade of the channel by intracellular cations.2 3 4 5 Small ions, such as Mg2+ and Ca2+, block the channel with fast kinetics; polyamines, such as spermine and spermidine, account for a slower component of rectification.4 6 Experiments performed in Mg2+-free conditions (excised inside-out patches) show that micromolar Ca2+ can block IK1 from the intracellular side of the membrane at depolarized potentials.3 5 Moreover, in an elegant study on neonatal avian ventricular myocytes, Mazzanti and DeFelice7 reported that outward current through single IK1 channels could be observed during the action potential if Ca2+ was omitted from the extracellular (pipette) solution. Nonetheless, a physiological relevance of Ca2+-induced rectification of IK1 in the mammalian heart is either disputed or overlooked,3 8 mainly on the basis of the tenet that intracellular Mg2+ and polyamine concentrations are well above those required to saturate IK1 rectification in excised-patch experiments.
The aim of the present study was to reconsider the role for Ca2+-induced IK1 rectification in mammalian ventricular myocytes by adopting experimental conditions preserving cell integrity and simulating, as closely as possible, physiological electrical activity. The evidence obtained suggests that the transient rise in subsarcolemmal Ca2+, occurring during the plateau phase of the action potential as a consequence of both influx and release from the sarcoplasmic reticulum, may significantly contribute to IK1 rectification.
Materials and Methods
Guinea pig ventricular myocytes were isolated by using a coronary perfusion method similar to the one described by Levi and Alloatti.9 In brief, guinea pigs weighing 200 to 300 g were anesthetized by exposure to a cotton wad soaked in tribromoethanol solution (200 mg of tribromoethanol in 10 mL of ether), killed through cervical dislocation, and exsanguinated. Hearts were quickly removed, and the ascending aorta was connected to the outlet of a Langendorff column, perfused with Tyrode’s solution (37°C) containing (mmol/L) NaCl 154, KCl 4, CaCl2 2, MgCl2 1, HEPES-NaOH 5, and d-glucose 5.5, adjusted to pH 7.35, and equilibrated with 100% O2. Perfusion with Tyrode’s solution was maintained until vigorous mechanical activity resumed and blood was completely removed. The heart was then perfused, until arrest occurred, with a nominally Ca2+-free solution containing (mmol/L) NaCl 33.5, KCl 10, d-glucose 22, sucrose 132, KH2PO4 1, MgSO4 5, HEPES KOH 10, and taurine 50, adjusted to pH 7.3, followed by the same solution to which 140 U/mL collagenase (Sigma type V), 2.5 U/mL trypsin (Sigma type III), and 1 mg/mL bovine serum albumin were added. When the effluent became slightly viscous, the heart was moved to Ca2+-free solution (0.1 mmol/L EGTA added), the atria were dissected, and the ventricles were chopped to 1-mm fragments. The fragments were exposed to gentle mechanical agitation in Ca2+-free solution with the addition of 0.3 U/mL protease (Sigma type XIV) and 1 mg/mL bovine serum albumin. Ten-milliliter samples of the suspension were collected every 5 minutes, filtered through a nylon mesh, and centrifuged at 750 rpm for 3 minutes. The pellets were resuspended in Tyrode’s solution and stored in this solution at 4°C until use.
Rod-shaped Ca2+-tolerant myocytes were usually obtained after 10 to 15 minutes of mechanical agitation. Only quiescent myocytes with clear-cut striations were used for measurements; to limit the error generated by large currents through uncompensated series resistance, smaller cells were generally preferred.
Experimental Solutions and Recording Apparatus
The myocyte suspension was placed in a 30-mm polylysine-coated Petri dish with a plastic ring (to reduce total volume to ≈1 mL) and mounted on the stage of an inverted microscope. The dish was perfused at 2 mL/min with standard external Tyrode’s solution containing (mmol/L) NaCl 154, KCl 4, CaCl2 2, MgCl2 1, HEPES-NaOH 5, and d-glucose 5.5, adjusted to pH 7.35. In a set of experiments testing the effect of low extracellular Ca2+, the concentration of this ion was reduced from 2 to 0.5 mmol/L; to compensate for the resulting decrease in surface charge shielding, extracellular Mg2+ was increased to 4 mmol/L. The cell under study was held within 300 μm from the tip (1 mm) of a thermostated multiline pipette, which was connected to solution reservoirs through electrically driven valves. This system allowed us to expose the cell to different solutions, with changes completed in ≈1 second. The solution temperature was monitored at the pipette tip with a fast-response digital thermometer (BAT-12, Physitemp) and kept at 35±0.1°C. Nifedipine (Sigma) and nisoldipine (Bayer) stock solutions were prepared by dissolving the substances in ethanol; the same ethanol concentration was present in all experimental solutions and did not exceed 0.05%. Nisoldipine was a generous gift from Bayer Italia.
Membrane potential and current were measured in the whole-cell configuration, under ruptured- and perforated-patch conditions (Axopatch 200-A, Axon Instruments), by borosilicate glass pipettes with tip resistances between 3 and 5 MΩ. The pipette solution contained (mmol/L) potassium aspartate 110, KCl 23, CaCl2 0.4 (calculated free Ca2+, 10−7 mol/L), MgCl2 3 (calculated free Mg2+, ≈0.1 mmol/L), HEPES KOH 5, EGTA KOH 1, GTP sodium salt 0.4, ATP sodium salt 5, and creatine phosphate 5, pH 7.3. In perforated-patch experiments, 130 μmol/L amphotericin B was added, whereas ATP, GTP, and creatine phosphate were omitted from the pipette solution. The composition of the pipette solution was also changed in selected experiments as specified in the appropriate section of “Results.” Series resistances (<5 MΩ in ruptured-patch and <10 MΩ in perforated-patch experiments) and membrane capacitance were measured in every cell but were not compensated. An average junction potential of ≈5 mV, measured on moving the electrode tip from Tyrode’s solution to “intracellular” (potassium aspartate) solution, was also left uncompensated.
The command potentials for the V-clamp amplifier were supplied through a 12-bit D/A converter (Labmaster TL-40) by an IBM-compatible PC (Pentium, 75 MHz) driven by custom-made software. Potential and current signals were filtered at 2 kHz and were tape-recorded, together with trigger signals, with an adapted VCR system. Recorded signals were later acquired through a 12-bit A/D (sampling rate, 5 kHz) for analysis.
Membrane currents were studied by the AP-clamp technique,10 11 12 as illustrated in Figure 1⇓. Transmembrane potential was recorded (I-clamp configuration) during steady-state stimulation at a cycle length of 1 second in a single cell during superfusion with Tyrode’s solution. The membrane potential waveform, corresponding to a single stimulation cycle, was digitized (5 kHz, 12-bit resolution) and stored in computer memory. After switching to V-clamp mode, the acquired waveform was used as the command signal to drive membrane potential in the same cell, thus allowing measurement of total membrane current during the cycle. Sustained activity was simulated by repeatedly applying the same waveform at a cycle length of 1 second. After a short stabilization period under AP-clamp conditions, total membrane current recorded during superfusion with Tyrode’s solution (control) settled to a value close to zero, except for a short glitch corresponding to the stimulation artifact. The current recorded during subsequent exposure to a channel blocker (compensation current) provided a mirror image of the contribution of the blocked component to action potential.
Recordings were considered acceptable only if (1) the current recorded in control conditions was negligible, indicating that reproduction of the spontaneous activity was satisfactory, and (2) washout of the blocker, performed after each exposure, was followed by reversal of the compensation current.
Identification of K+ currents required means to inhibit their conductance selectively. To compensate for the unavailability of blockers absolutely specific for IK1, we used two different interventions that share a strong inhibitory effect on K+ conductances (particularly IK1): blockade by 1 mmol/L Ba2+ 13 and removal of K+o.12 14 The rationale was that if the observed effect was identically reproduced with both the approaches, it could be ascribed to a change in K+ currents with a reasonable confidence. Either Ba2+ blockade or removal of K+o has been previously used to identify IK1 in AP-clamp experiments11,12; however, in previous studies, conductances other than IK1 (INa, ICa, Ito, and INaK) were preventively blocked with the appropriate agents. Such an approach was not adopted in the present case, because ICa was an essential player and other contaminations, with the exception of INaCa, were substantially irrelevant to the interpretation of data (see “Discussion”). Interference by INaCa was evaluated in a specific set of experiments (see “Results”).
Throughout text and figures, data from AP-clamp experiments are presented in terms of “blocker-sensitive current” (IBa and I0K). The latter was obtained by subtraction of compensation current traces from control ones. Current traces from three to five cycles at steady state in each condition were averaged before subtraction. Because of the complexity introduced by the subtraction procedure, the meaning of changes in “blocker-sensitive” currents is often not intuitive. It may help to consider that according to the procedure used in the present study, any current that is reduced by the blocker will contribute to the blocker-sensitive current with its original sign. For instance, if nifedipine selectively blocked ICaL, the blocker-sensitive current Inif (Figure 1⇑, bottom panel) would be identical to true ICaL. The opposite would be true for currents either directly or indirectly enhanced by the “blocker.”
Currents measured during the activation cycle were divided, according to the cycle phase, into (1) diastolic current, measured during the diastolic interval; (2) systolic current, measured from the upstroke (excluding the artifact) to 50% of repolarization; and (3) peak current during fast repolarization. Whereas the latter represented a single point value, diastolic and systolic currents were measured as “average current”; this was done by integrating the current signal over the appropriate interval and dividing the result by the integration interval. Current density was estimated by dividing the current value by membrane capacitance. In all figures, membrane potential and current traces are aligned.
Means were compared by Student t test for paired or unpaired observations as appropriate. ANOVA for paired measurements, in some cases with a grouping factor (two-way model), was performed whenever more than two means were compared. A probability level of P<0.05 was used to define significance throughout the study. In the text and figures, values are presented as mean±SEM.
I-V Relations of IBa and I0K
Quasi–steady-state I-V relations of IBa and I0K were studied, under V-clamp conditions, by applying voltage ramps from −105 to +55 mV with a steepness of 0.03 V/s. The ramp protocol was applied in control conditions and during exposure to either 1 mmol/L Ba2+ (Figure 2a⇓) or to K+-free solution (Figure 2b⇓). IBa and I0K were obtained by subtraction of the traces recorded in control and test solutions. In some of the experiments, potential contamination by ICaL was prevented by performing the measurements in the presence of 0.2 μmol/L nisoldipine. Since the results obtained were superimposable, data obtained in the presence and absence of nisoldipine were pooled. As shown in Figure 2⇓, the I-V relations of both IBa and I0K displayed strong inward rectification up to ≈−10 mV. Positive to this potential, while IBa began to rise again, I0K underwent a further decay that, in a few cases, led to a small net inward current. Such mirror-like behavior of IBa and I0K at positive potentials can be attributed to divergent effects of the two experimental interventions on IK. IK is reduced by Ba2+,1 thus appearing at potentials positive to IK threshold as outward IBa. Vice versa, as a consequence of the balance between incomplete depression of conductance of its slowly activating component (IKs) and the increased chemical gradient, IK may be unchanged or even slightly increased by removal of extracellular K+.14 Thus, particularly under conditions in which IKs is larger (eg, during longer depolarizations), IK might contribute to I0K with a reversed sign, ie, as a small inward component. It is uncertain to which extent this may apply to I0K recorded under AP-clamp conditions, because the proportion of IK contributed by IKs during the action potential cannot be determined a priori. Nonetheless, a paradoxical increase in I0K, but not in IBa, is expected from selective inhibition of any amount of IKs activated during the action potential.
The reversal potential was slightly more positive for IBa (−83.1±0.89 mV, n=9) than for I0K (−85.9±0.82 mV, n=9 [P<0.05 versus IBa]). When taking into account the offset introduced by the uncompensated junction potential (≈5 mV), the reversal potential was for both currents in reasonable agreement with the expected K+ equilibrium potential (−94 mV). The maximal conductance, estimated by fitting the linear portion of the I-V relation, was similar between IBa (0.65±0.13 nS/pF) and I0K (0.61±0.13 nS/pF [P=NS versus IBa]). However, the maximal outward current, occurring in both cases between −70 and −60 mV, was smaller for IBa (2.41±0.23 pA/pF) than for I0K (3.36±0.32 pA/pF [P<0.05 versus IBa]). Overall, these results might be interpreted as Ba2+ causing complete blockade of inward IK1 but only submaximal inhibition of outward IK1 (see “Discussion”); accordingly, IK1 during the action potential might be underestimated by IBa.
IBa and I0K During the Ventricular Action Potential
IBa and I0K recorded during the ventricular action potential are compared in Figure 3⇓; although with minor differences, both currents had a time course similar to the one predicted by numerical simulations of IK1.15 The tracings in Figure 3⇓ were selected from cells with relatively long action potentials, in which the contribution of IKs to IK would be enhanced. Whereas IBa progressively increased during the plateau phase, the opposite was true for I0K; this might suggest that IK contamination may affect IBa and I0K in opposite directions also under the conditions of AP-clamp measurements. As can be appreciated from the remaining figures, differences between I0K and IBa were less obvious at shorter action potential durations, ie, in the majority of cases.
Although the time courses of IBa and I0K were substantially similar, significant differences emerge from the quantitative analysis of currents during specific phases of the action potential. Average systolic and diastolic current densities and peak current density during fast repolarization are compared between IBa and I0K in the Table⇓. As predicted by the data presented in the previous section, I0K was almost 2-fold larger than IBa over most of the excitation cycle, with the exception of the plateau phase (systolic current). This may be in apparent contrast with the larger IBa present at positive membrane potentials during steady-state depolarization (see Figure 2⇑). However, it should be considered that activation of IKs, the current likely to account for the difference between IBa and I0K at positive potentials (see above), would be largely incomplete during short-lasting action potentials. The average cell capacitance was 58.7±2.52 pF.
Effect of ICa Inhibition on IBa and I0K During the Ventricular Action Potential
To minimize the possibility of effects resulting from ancillary drug properties,16 two DHPs with widely different potencies were used in assessing the effect of Ca2+ channel blockade in separate sets of experiments. Nifedipine and nisoldipine were applied at concentrations of 5 and 0.2 μmol/L, respectively, both sufficient to achieve almost complete blockade of ICaL,17 as also verified in the setting of the present study (data not shown). The effect of inhibiting Ca2+ influx was also tested by lowering extracellular Ca2+ concentration from 2 to 0.5 mmol/L (low Ca2+ solution) and replacing it with Mg2+ (see “Materials and Methods”).
Ca2+ channel blockade by either nisoldipine (Figure 4⇓) or nifedipine (Figure 5⇓) resulted in a significant increase of IBa. As also apparent from the dynamic I-V relations shown in panel c of each figure, such an effect was restricted to positive potentials, ie, those occurring during the action potential plateau and early repolarization phases; diastolic current and the peak current during fast repolarization were unmodified. Average data from the individual experimental groups are as follows: nisoldipine (n=8, Figure 4d⇓) increased systolic IBa by 58.1±21.9% (0.689±0.066 pA/pF versus 0.502±0.082 pA/pF [P<0.05]); diastolic IBa (0.865±0.177 pA/pF versus 0.850±0.138 pA/pF [P=NS]) and peak IBa (2.756±0.212 pA/pF versus 2.768±0.213 pA/pF [P=NS]) remained unmodified. In the presence of nifedipine (n=11, Figure 5d⇓) systolic IBa was increased by 50.7±14.6% (0.767±0.091 pA/pF versus 0.583±0.099 pA/pF [P<0.05]); diastolic IBa (0.583±0.106 pA/pF versus 0.559±0.093 pA/pF [P=NS]) and peak IBa (2.846±0.311 pA/pF versus 2.885±0.304 pA/pF [P=NS]) were unchanged by nifedipine.
Whereas nifedipine-induced changes in IBa were fully and quickly reversible, the effect of nisoldipine reversed more slowly. Since the effects of the two DHPs were similar and complete washout could be more easily obtained with nifedipine, this drug was chosen to test the effect of Ca2+ channel blockade on I0K (Figure 6⇓).
In the presence of nifedipine, systolic I0K (n=6, Figure 6d⇑) was reversibly increased by 254.5±66.6% (1.217±0.202 pA/pF versus 0.413±0.088 pA/pF [P<0.05]); diastolic I0K (1.922±0.463 pA/pF versus 2.071±0.584 pA/pF [P=NS]) and peak I0K (5.566±0.562 pA/pF versus 5.404±0.617 pA/pF [P=NS]) were not affected by nifedipine. Thus, albeit of larger magnitude, the effect of nifedipine on I0K was qualitatively comparable to that exerted on IBa.
The changes in I0K induced by a reduction in extracellular Ca2+ (Figure 7⇓) were more complex than those of Ca2+ channel blockade. An increase of systolic I0K was consistently observed; however, this was associated with a small reduction of peak I0K and to relatively large, although inconsistent, variations of diastolic I0K. In spite of such complexities, on average, the effect of low extracellular Ca2+ was consistent with that of Ca2+ channel blockade. In the presence of low Ca2+ (n=7, Figure 7d⇓), systolic I0K increased by 108.8±43.1% (0.454±0.072 pA/pF versus 0.255±0.049 pA/pF [P<0.05]). Diastolic I0K, although obviously reduced in some cells, underwent opposite changes in others; thus, on average, this change did not achieve significance (−36.5±11.8%, 0.682±0.189 pA/pF versus 1.009±0.143 pA/pF [P=NS]). Peak I0K during fast repolarization was reduced by 8.1±1.9% during low Ca2+ superfusion (2.806±0.201 pA/pF versus 3.075±0.244 pA/pF [P<0.05]). The effects of low Ca2+ superfusion were readily reversed on washout.
It can be anticipated that besides reducing K+ channel conductance, removal of K+o will also inhibit the Na+-K+ pump,18 thus resulting in a reduction of INaK and in a progressive dissipation of the Na+ transmembrane gradient. A decrease in the Na+ gradient, in turn, will reduce the inward component of the current (INaCa) generated by the Na+-Ca2+ exchanger (ie, it will move the reversal potential of INaCa in the negative direction). As a consequence, I0K might contain both INaK and INaCa; whereas the former would appear as an outward component, contamination by INaCa would shift I0K in the inward direction at all potentials. Although DHP-induced enhancement of systolic I0K could not be accounted for by an action on INaK (see “Discussion”), INaCa contamination might be seriously confusing. Indeed, a reduction in intracellular Ca2+ might cause INaCa to become less inward, thus potentially contributing to the extra outward I0K appearing on Ca2+ channel blockade. To test this hypothesis, we studied the effect of nifedipine on I0K in the presence of Na+-K+ pump inhibition by 20 μmol/L ouabain (added to all solutions). Immediately on ouabain superfusion, the compensation current became more inward over the whole cycle, a mirror image of the expected INaK time course15 ; more sustained ouabain superfusion caused the current to become progressively outward until a new, almost stable state was achieved. Under such a condition, further changes in current were minimal over the time required for I0K measurements (full reversal of the effects of zero-K+ superfusion was obtained in each condition; see “Materials and Methods”). In the presence of ouabain (n=10, Figure 8⇓), both systolic (0.18±0.04 pA/pF) and diastolic I0K (0.36±0.22 pA/pF) were reduced with respect to control (P<0.05). Nifedipine (5 μmol/L) increased systolic I0K to 0.53±0.071 pA/pF (467.8±157.7% [P<0.05]), a change larger that the one induced in the absence of ouabain (P<0.05, Figure 8d⇓). Whereas, in the presence of ouabain, nifedipine also increased peak I0K during fast repolarization by 18.1±5.3% (P<0.05), diastolic I0K was not affected by the drug. Because of the instability generated by ouabain superfusion and the relatively long time required for washout, reversibility of the effects of nifedipine on I0K could be tested in only 5 of 10 cells. In these cells, 68.9±12.9% of the nifedipine-induced increase of systolic I0K (377.7±114.9% [P<0.05]) was reversed on washout (example in Figure 8⇓). These experiments show that enhancement of systolic I0K by Ca2+ channel blockade persisted in the presence of ouabain, thus arguing against the possibility that it originated from a change in either INaK or INaCa.
Effect of β-Adrenergic Stimulation on I0K
If inhibition of Ca2+ influx enhanced IK1 during the action potential plateau phase, the opposite effect should be expected from an increase in Ca2+ influx. To test for this hypothesis I0K was measured during β-adrenergic stimulation by 0.1 μmol/L isoproterenol, an intervention known to increase ICaL19 (Figure 9⇓). Isoproterenol reversibly reduced systolic I0K from 0.42±0.07 to 0.16±0.06 pA/pF (n=8 [P<0.05]); diastolic current and peak current during fast repolarization were unchanged by the agonist.
IBa and Inhibition of Sarcoplasmic Reticulum Ca2+ Release
Rather high subsarcolemmal Ca2+ concentrations, on the order of 1 μmol/L, are required to induce IK1 rectification in cell-excised membrane patches,3 5 thus suggesting the involvement of Ca2+ released by the sarcoplasmic reticulum. To test whether reticular Ca2+ release might contribute to IK1 rectification, IBa was measured before and after sustained (≈5-minute) superfusion with 1 μmol/L ryanodine, an agent that causes depletion of intracellular Ca2+ stores.
Figure 10⇓ shows that ryanodine superfusion was followed by an increase in systolic IBa qualitatively and quantitatively comparable to the one induced by Ca2+ channel blockade. Average results from five myocytes (Figure 10d⇓) indicate that ryanodine treatment increased systolic IBa by 47.7±6.5% (0.51±0.1 pA/pF versus 0.35±0.076 pA/pF [P<0.05]) without altering either diastolic IBa (0.73±0.13 pA/pF versus 0.72±0.11 pA/pF [P=NS]) or the current peak during fast repolarization (1.88±0.33 pA/pF versus 1.99±0.32 pA/pF [P=NS]). The effect of ryanodine superfusion was only partially reversible over long recovery periods.
Effect of Intracellular Ca2+ Buffering on IBa and on Its DHP-Induced Changes
A further proof of the contribution of subsarcolemmal Ca2+ to IK1 rectification would be the observation of an increased systolic IK1 following strong intracellular Ca2+ buffering. We initially performed such a test by removing Ca2+ from the pipette solution and replacing 1 mmol/L EGTA with 10 mmol/L BAPTA; although there was a trend to an increase in systolic IBa (0.63±0.04 pA/pF [n=7] versus 0.49±0.06 pA/pF [n=25] [P=NS]) the change was not statistically significant. However, such a negative finding might result from inadequate diffusion of BAPTA to subsarcolemmal spaces. Thus, in a further set of experiments, besides removing Ca2+ and including 10 mmol/L BAPTA in the pipette, cells were preincubated (>60 minutes) in a solution containing 10 μmol/L BAPTA-AM, a membrane-permeable analogue of BAPTA. In BAPTA-AM–treated cells (Figure 11⇓), diastolic potential was depolarized (−63.7±1.1 versus −73.2±1.1 mV [P<0.05]), plateau potential was unchanged, and action potential duration, although apparently longer in some cells, was not, on average, significantly affected (229.4±25.3 versus 192.0±12.4 milliseconds [P=NS]). Compared with untreated cells, BAPTA-AM–treated cells had a larger diastolic IBa (1.04±0.09 versus 0.71±0.07 pA/pF [P<0.05]). The increase in diastolic IBa was similar to that measured from steady-state I-V relations of IBa (Figure 2c⇑) for a change in membrane potential from −73.2 to −63.7 mV (1.75±0.14 versus 2.42±0.23 pA/pF, n=9); thus, changes in currents other than those reflected by IBa should account for the depolarization induced by BAPTA-AM (see above).
Consistent with an inhibition of IK1 by subsarcolemmal Ca2+ transients, systolic IBa was increased in BAPTA-AM–treated cells (0.92±0.10 pA/pF, n=10 [P<0.05]) (Figure 11d⇑).
In the presence of BAPTA-AM, nifedipine failed to increase average systolic IBa (+18.45±7%, n=9 [P=NS]) (Figure 11⇑). Although nifedipine still induced an outward shift of IBa, this was limited to the early portion of the plateau phase, when IBa was initially smaller; in some cells, as the one shown in Figure 11⇑, late-systolic IBa was even reduced below control values by nifedipine (see “Discussion”). Similar to the effect of low extracellular Ca2+ on I0K, nifedipine reduced diastolic IBa and peak IBa during fast repolarization; however, because of the inhomogeneity of effects among cells, significance was achieved only for the change in diastolic current (0.82±0.08 versus 1.00±0.09 pA/pF, n=9 [P<0.05]).
Dependence of IBa Modulation on Intracellular Mg2+ Activity
This set of experiments was designed to verify whether the effect of Ca2+ channel blockade on IBa might be affected by the intracellular levels of Mg2+, the other divalent cation known to contribute to physiological inward rectification of IK1. To this end, IBa and its modulation by nifedipine were evaluated in the presence of pipette Mg2+ activities of 0.1 mmol/L (n=9) and 1 mmol/L (n=9), respectively. Average results from these experiments are shown in Figure 12⇓. Whereas in control conditions IBa was reduced by higher intracellular Mg2+, the nifedipine-induced increase in systolic IBa was enhanced by the same intervention (Figure 12a⇓). Increasing intracellular Mg2+ from 0.1 to 1 mmol/L reduced systolic IBa from 0.563±0.088 to 0.347±0.022 pA/pF (P<0.05). Nifedipine increased IBa by 0.209±0.026 pA/pF (54.7±16.9% [P<0.05]) in 0.1 mmol/L Mg2+ and by 0.313±0.037 pA/pF (93.6±11.8% [P<0.05]) in 1 mmol/L Mg2+. Both absolute and percent nifedipine–induced changes of systolic IBa were significantly larger in the presence of higher intracellular Mg2+ activity. As shown in panels b and c of Figure 12⇓, diastolic and peak IBa also tended to be reduced in the presence of higher intracellular Mg2+; however, such a change did not achieve statistical significance (see “Discussion”).
In order to verify whether inhibition of Ca2+ influx might also affect IK1 under conditions preserving an intact intracellular environment, the effects of nifedipine on IBa were tested by the perforated-patch technique in 5 cells. Under such conditions, systolic IBa was increased from 0.69±0.22 to 1.01±0.23 pA/pF, a change (+71.17±20.5%) similar to the one observed, under ruptured-patch conditions, with 1 mmol/L intracellular Mg2+.
The results of the present study show that inhibition of either Ca2+ influx or Ca2+ release from the sarcoplasmic reticulum may substantially increase both IBa and I0K during the plateau phase of the guinea pig ventricular action potential. Opposite changes were observed during β-adrenergic stimulation, probably as a result of enhanced Ca2+ influx.
Although IK1 is the main constituent of IBa and I0K, other channels may contribute, obviously to different extents, to each of the blocker-sensitive currents. Thus, the interpretation of the present findings depends on whether they could be accounted for by changes in conductances other than IK1.
Do Changes in IBa and I0K Reflect Changes in IK1?
Ba2+-Sensitive Current, IBa
At the concentration used, Ba2+ may inhibit other K+ currents, such as IK and Ito13 20 and, possibly, ICaL; thus, all these currents might contaminate IBa. If IBa included ICaL, block by DHPs might result in an outward shift of IBa, possibly similar to the one observed. However, the effect of DHPs on systolic IBa was reproduced by ryanodine and intracellular Ca2+ buffering (interventions that should not affect ICaL channels directly) and was qualitatively very similar to the one exerted on I0K, for which a contamination by ICaL is very unlikely. On the other hand, both Ito and IK should be reduced by DHPs as a consequence of either direct channel blockade21 22 or reduced intracellular Ca2+.23 24 Thus, the effects expected from DHP-induced changes in these currents are opposite those observed in IBa.
Zero-K+–Sensitive Current, I0K
Removal of K+o may decrease the conductance of all K+ channels, inhibit the Na+-K+ pump, and, secondary to the latter effect, change the Na+ gradient driving the Na+-Ca2+ exchanger (see “Results”). Thus, I0K putative contaminants include Ito, IK, INaK, and, indirectly, INaCa. For what concerns Ito contamination, the same reasoning exposed for IBa applies to I0K. However, decreased Ca2+ influx should selectively depress IKs, thus possibly inducing a paradoxical increase in I0K independent of changes in IK1 (see below).
Either direct or indirect inhibition of INaK by DHPs would reduce systolic I0K, an effect opposite the one observed. Conversely, as already discussed in “Results,” INaCa contamination might well account for the observed changes. However, the persistence of systolic I0K enhancement by nifedipine in the presence of a large concentration of ouabain, a condition in which I0K contamination by both INaK and INaCa should be largely removed, argues against this interpretation. In the presence of ouabain, Ca2+ extrusion through the sarcolemma should be inhibited,25 thus resulting in the persistence of elevated subsarcolemmal Ca2+. This might account for the extension of nifedipine-induced enhancement of I0K to the fast repolarization phase (see Figure 8⇑).
In light of these considerations and of the previously described effects of intracellular Ca2+ on the current,3 5 7 the most likely explanation of the effects of all the interventions tested on IBa and I0K is a removal of Ca2+-induced rectification of IK1.
Differences Between IBa and I0K and in Their Response to DHPs
During diastole and fast repolarization, I0K was significantly larger than IBa; moreover, although the two systolic currents were similar in control conditions, DHP-induced enhancement was larger for I0K than for IBa. A possible interpretation of these findings is a more complete inhibition of outward IK1 by K+o removal than by Ba2+, whose blocking effect on outward IK1 appeared to be incomplete. A dependence of blocking potency on the direction of current is a common property of multi-ion pores and has been previously described for blockade of IK1 by monovalent and divalent cations.1 2 The different sensitivity of systolic IBa and I0K to DHPs might reflect selective inhibition of contaminating IKs by reduced intracellular Ca2+,23 an effect that might enhance drug-induced increase in I0K only (see above). Thus, the effect of subsarcolemmal Ca2+ on IK1 might be slightly overestimated by the effect of DHP on I0K and underestimated by those on IBa.
Effects of DHPs in the Presence of Intracellular Ca2+ Buffering
The concentrations of intracellular Ca2+ reported to inhibit IK1 in excised membrane patches vary considerably (0.1 to 10 μmol/L at plateau potentials, Mg2+-free conditions, 20°C) and may be higher than those measured in the cytoplasm at the peak of physiological Ca2+ transients (1 μmol/L).3 5 26 However, during electrical activity, Ca2+ concentrations in the restricted subsarcolemmal space may exceed those in the bulk cytoplasm. This would be consistent with the evidence suggesting that Ca2+-induced rectification may actually be a very localized phenomenon that would require IK1 channels to be near Ca2+ influx and release sites. In chick myocytes, outward elementary IK1, measured in the cell-attached configuration during a stimulated action potential, was affected by Ca2+ concentration in the cell-attached pipette only; the persistence of Ca2+ influx through the rest of the cell membrane appeared as irrelevant.7 In the present study, diffusion of a Ca2+ chelator (BAPTA-AM) through the whole membrane surface was required to increase systolic IBa and to blunt its nifedipine-induced changes. Moreover, even in the presence of the chelator, nifedipine still caused an early increase in IBa, thus suggesting that at the time of peak Ca2+ influx, buffering might have still been incomplete. At this concern, it should also be considered that the effects of Ca2+ buffering might have been partly offset by increased Ca2+ influx, due to lesser inactivation of ICaL27 and reverse operation of the Na+-Ca2+ exchanger.28 29 Later on during the course of action potential, IBa was even reduced by nifedipine in some cells; this reduction was perhaps due to drug actions unmasked by the removal of Ca2+-mediated effects. On the other hand, in the presence of extensive intracellular Ca2+ buffering, similar to what observed during exposure to low Ca2+ (see Figure 7⇑), the effects of nifedipine on IBa were not limited to the plateau phase. Thus, under conditions in which the supply of Ca2+ from other sources (intracellular stores, Na+-Ca2+ exchanger) might also be reduced, the effects of ICaL blockade on IBa might be more complex than usual, and their interpretation might be more difficult.
Dependence of the Effects of DHPs on Free Intracellular Mg2+
An increase in intracellular Mg2+ from 0.1 to 1 mmol/L was associated with a reduction of systolic IBa. The decrease in IBa did not reach significance during diastole or fast repolarization (Figure 12b⇑ and 12c⇑). However, it should be stressed that the blocking effect of Mg2+ may be easier to detect at depolarized potentials3; moreover, smaller effects at diastolic potentials might be overlooked (because of the lower power of the statistics) when internal comparisons are impossible. This finding suggests that Mg2+-induced IK1 rectification might still be incomplete at 0.1 mmol/L, a concentration two orders of magnitude above the Kd measured in cell-excised patches at plateau potentials (≈3 μmol/L).3 This suggests that Mg2+ affinity for IK1 channels may be affected by experimental conditions, such as cell integrity and temperature.
Inhibition of Ca2+ influx had a larger effect on IBa at 1 than at 0.1 mmol/L intracellular Mg2+. This would be incompatible with the competition of Ca2+ and Mg2+ ions for a common binding site on the channel. On the other hand, previous data show that Ca2+-induced, but not Mg2+-induced, IK1 rectification might be abolished by disruption of the cytoskeleton,30 an observation also consistent with the two cations acting on different sites. As an alternative, the rectifying effect of Ca2+ observed in the present study may not represent a simple blockade, such as the one described in Mg2+-free conditions,3 5 but may be a novel, possibly more complex, effect.
Functional Implications of Ca2+-Induced Rectification of IK1
According to the results of the present study, removal of Ca2+-induced rectification may increase systolic IK1 from 2- to 4-fold, corresponding to increases of 0.2 to 0.8 pA/pF in average systolic current. As shown by the effect of isoproterenol, opposite changes in IK1 can result from enhancement of Ca2+ influx. In ventricular myocytes with a capacitance of ≈60 pF, membrane input resistances in the order of 30 to 50 MΩ can be measured during the action potential plateau (authors’ unpublished data, 1997). Thus, shifts of several millivolts in plateau potentials might conceivably result from changes in the extent of Ca2+-induced IK1 rectification. This may be relevant to the role of Ca2+ influx/release in the modulation of action potential duration. For instance, changes in IK1 might contribute to the marked action potential shortening induced by Ca2+ channel blockers or to action potential prolongation often resulting from exposure to catecholamines.31
Ca2+-induced IK1 rectification might reduce the amount of inward current necessary to the induction of early afterdepolarizations, thus contributing to the facilitation of these phenomena by catecholamines. Increased subsarcolemmal Ca2+ may also lead to enhancement of IK23; thus, in normal conditions, the effects of Ca2+ on IK1 may be offset by those on IK. However, either drug-induced or genetic impairment of IK conductance32 33 34 might unmask the arrhythmogenic potential of the Ca2+-induced decrease in IK1.
Even if Ca2+-induced rectification may be important functionally, it represents only a small fraction of total IK1 rectification. Indeed, on the basis of the conductance measured from the linear portion of the I0K I-V relation, unrectified IK1 should amount to >70 pA/pF at plateau potentials compared with <1 pA/pF measured in the presence of Ca2+ channel blockade. Thus, the effect of Ca2+ observed in the present study may account for <2% of total IK1 rectification, with the remaining part being attributable to Mg2+ and polyamines. Nonetheless, whereas Mg2+ and polyamine-induced rectification is present to the same extent during the whole cardiac cycle, Ca2+-induced rectification may represent a dynamic phenomenon exquisitely sensitive to modulation. The small proportion of total IK1 rectification accounted for by Ca2+ influx may also explain why this phenomenon could not be observed in single-channel studies on guinea pig ventricular myocytes.3
Selected Abbreviations and Acronyms
|AP-clamp||=||action potential clamp|
|I 0K||=||current sensitive to K+o removal|
|I Ba||=||Ba2+-sensitive current|
|I Ca||=||Ca2+ current|
|I CaL||=||L-type Ca2+ current|
|I K||=||delayed rectifier K+ current|
|I K1||=||inward rectifier K+ current|
|I Ks||=||slowly activating K+ current|
|I Na||=||Na+ current|
|I NaCa||=||Na+-Ca2+ exchanger current|
|I NaK||=||Na+-K+ pump current|
|I nif||=||nifedipine-sensitive current|
|I to||=||transient outward current|
The present study was partly supported by Telethon grant No. 396 and by Ministero Università Ricerca Scientifica e Tecnologica funding. We are grateful to Dario DiFrancesco and Michele Mazzanti for reading the manuscript and providing constructive criticism and to Gaspare Mostacciuolo for expert technical assistance.
- Received November 25, 1997.
- Accepted March 4, 1998.
- © 1998 American Heart Association, Inc.
Hille B. Potassium channels and chloride channels. In: Hille B, ed. Ionic Channels of Excitable Membranes. 2nd ed. Sunderland, Mass: Sinauer Associates Inc; 1992:115–139.
Levi RC, Alloatti G. Histamine modulates calcium current in guinea-pig ventricular myocytes. J Pharmacol Exp Ther. 1988;246:377–383.
Luo C-H, Rudy Y. A dynamic model of the cardiac ventricular action potential, I: simulations of ionic currents and concentration changes. Circ Res. 1994;74:1071–1096.
Perkins KL, Wong RKS. Intracellular QX-314 blocks the hyperpolarization-activated inward current Iq in hippocampal CA1 pyramidal cells. J Neurophysiol. 1995;73:911–915.
Daleau P, Turgeon J. Triamterene inhibits the delayed rectifier potassium current (IK) in guinea pig ventricular myocytes. Circ Res. 1994;74:1114–1120.
Josephson IR, Sanchez-Chapula J, Brown AM. Early outward current in rat single ventricular cells. Circ Res. 1984;54:157–162.
Zhang X, Anderson JW, Fedida D. Characterization of nifedipine block of the human heart delayed rectifier, hKv1.5. J Pharmacol Exp Ther. 1997;281:1247–1256.
Hume JR. Comparative interactions of organic Ca++ channel antagonists with myocardial Ca++ and K+ channels. J Pharmacol Exp Ther. 1985;234:134–140.
Nitta J, Furukawa T, Marumo F, Sawanobori T, Hiraoka M. Subcellular mechanism for Ca2+-dependent enhancement of delayed rectifier K+ current in isolated membrane patches of guinea pig ventricular myocytes. Circ Res. 1994;74:96–104.
Bers DM. Excitation-Contraction Coupling and Cardiac Contractile Force. Boston, Mass: Kluwer Academic Publishers; 1993:95–96.
Kohmoto O, Levi AJ, Bridge JHB. Relation between reverse sodium-calcium exchange and sarcoplasmic reticulum calcium release in guinea pig ventricular cells. Circ Res. 1994;74:550–554.
Mazzanti M, Assandri R, Ferroni A, DiFrancesco D. Cytoskeletal control of rectification and expression of four substates in cardiac inward rectifier K+ channels. FASEB J. 1996;10:357–361.
Sanguinetti MC, Jurkiewicz NK, Scott A, Siegl PKS. Isoproterenol antagonizes prolongation of refractory period by the class III antiarrhythmic agent E-4031 in guinea pig myocytes: mechanism of action. Circ Res. 1991;68:77–84.