Isoflurane Decreases ATP Sensitivity of Guinea Pig Cardiac Sarcolemmal K^ATPChannel at Reduced Intracellular pH

After approval by the Institutional Animal Use and Care Committee of the Medical College of Wisconsin, single ventricular myocytes were isolated from guinea pig hearts by enzymatic dissociation with collagenase Type II, (Gibco/Invitrogen, Life Technologies, Grand Island, NY) and protease Type XIV, (Sigma, St Louis, MO) as reported previously. 23Ventricular myocytes were stored in modified Tyrode solution, and only calcium-tolerant, rod-shaped cells with distinct cross-striations were used for experiments within 8 h after isolation.

The modified Tyrode solution contained NaCl, 132 mm; KCl, 4.8 mm; MgCl², 1.2 mm; CaCl², 1 mm; HEPES, 10 mm; and glucose, 5 mm; at pH adjusted to 7.4 with NaOH.

For single channel recordings in the inside-out patch clamp configuration, the bath solution facing the intracellular side of membrane patches contained KCl, 140 mm; MgCl², 0.5 mm; EGTA, 1 mm; HEPES, 10 mm; and variable 0–1 mm K²-ATP, at pH 7.4 or pH 6.8 adjusted with KOH or HCl. The pipette solution facing the extracellular side of membrane patches contained KCl, 140 mm; MgCl², 0.5 mm; CaCl², 0.5 mm; and HEPES, 10 mm at pH 7.4 adjusted with KOH. All chemicals were purchased from Sigma (St. Louis, MO).

Isoflurane (Baxter Healthcare, Deerfield, IL) was delivered to the recording chamber in the bath solution. Anesthetic solution was prepared by adding a measured aliquot of isoflurane to a known volume of bath solution and dispersing it by sonication. This solution was then transferred into a gas-tight glass syringe reservoir to be delivered to the recording chamber by a gravity-fed perfusion system. Isoflurane was used at a clinically relevant concentration of 0.53 ± 0.04 mm (n = 94) equivalent to 1.06 vol% at 20–23°C, or 0.9 minimum alveolar concentration (MAC) in guinea pigs and humans. This concentration was chosen because at 1.0 MAC, isoflurane has been shown to protect the myocardium in vivo  and in vitro  3,6and to enhance activity of whole cell I^KATPand single K^ATPchannels. 8,22Isoflurane concentrations were determined by the headspace analysis method using a Shimadzu GC 8A flame ionization detection gas chromatograph (Shimadzu, Kyoto, Japan) from aliquots of the bath solution sampled directly from the recording chamber.

Ventricular cells were placed in a RC-16 recording chamber (Warner, Hamden, CT) on the stage of an inverted IMT-2 microscope (Olympus, Tokyo, Japan). Single K^ATPchannel activity was monitored in the inside-out configuration of the patch clamp technique at 20–23°C. Patch pipettes were pulled from borosilicate glass tubing (Garner Glass, Claremont, CA) with a horizontal PC-84 micropipette puller (Sutter Instruments, Novato, CA). The tips were heat-polished with an MF-83 microforge (Narishige, Tokyo, Japan). Pipettes had resistances of 7–12 MΩ when filled with the extracellular solution. After gigaseal formation, the inside-out patches were excised by rapidly pulling the pipette away from the cell. In this configuration, the intracellular side of the membrane patch was directly exposed to the intracellular bath solution. Channel activity was recorded using a List EPC-7 amplifier (ALA Scientific Instruments, Westbury, NY) interfaced to a personal computer through a Digiata 1200B (Axon Instruments, Foster City, CA). Data were acquired using pClamp8 software (Axon Instruments, Foster City, CA). The current signal was filtered at 500 Hz through an 8-pole Bessel low-pass filter and sampled at 1 kHz. Single channel data were analyzed with pClamp8 software (Axon Instruments, Foster City, CA) and Origin6 software (OriginLab, Northampton, MA).

At symmetric 140 mm K⁺concentration, unitary outward current through single K^ATPchannels was monitored at the transmembrane patch potential of +40 mV. The 60-s recordings were made at each experimental step. The K^ATPchannels were identified by the single channel conductance, sensitivity to inhibition by [ATP]i, and blockade by glibenclamide (1 μm). A 50% threshold criterion was used for detecting the open state. Amplitude of single channel current was determined from the all-points amplitude histograms constructed from data segments of 60-s duration. Channel Po was determined from the ratios of the area under the peaks in the all-points amplitude histograms fitted with a Gaussian function. The number (N) of channels in each patch was estimated during brief exposure to ATP-free internal solution (0 ATP) at the end of the experiments. Po was calculated using the equation Po =[1 − (Pc)^1/n] where Pc is the channel closed state probability. For measurements of ATP sensitivity, Po of each patch was normalized to Po determined at 0 ATP to control for variations in Po among patches. The sensitivity to ATP (1, 10, 50, 100, and 1,000 μm) was determined at pHi 7.4 and pHi 6.8 in the absence or presence of isoflurane. The experimental protocols were completed within 10–12 min after patch excision. To minimize channel rundown, we used the Ca²⁺-free and low Mg²⁺intracellular solution and exposed each patch to 0 ATP only at the end of experimental protocols because even a brief exposure to ATP-free solution immediately after patch excision could accelerate rundown. Therefore, the number of channels in patches could have been underestimated in our study. Channel rundown occurred more frequently at pHi 7.4. Decreasing pHi to 6.8 tended to stabilize the channels and slow rundown. Recordings from patches exhibiting a significant rundown were excluded from analysis. For measurement of ATP sensitivity, the relationship between [ATP]i and Po was fitted by Hill equation:

where Po is channel open probability at any test [ATP]i; Po^maxis open probability at 0 [ATP]i; IC⁵⁰is ATP concentration for half-maximal effect; and nH is Hill coefficient.

Data are presented as mean ± SEM. Comparisons were made using paired or unpaired Student t  test. Multiple group means were compared by analysis of variance with a Student–Newman–Keuls test. Differences with a two-tailed P < 0.05 were accepted as significant.

The effects of isoflurane on the outward current through K^ATPchannels were investigated at pHi 7.4 and pHi 6.8 in the inside-out patches from guinea pig ventricular myocytes at a symmetric 140 mm K⁺concentration and the patch potential of +40 mV.

Membrane patches were excised into the intracellular solution containing 0.2 mm ATP. Multiple channel openings that appeared on patch excision decreased within 30–40 s, and thereafter only the activity of spontaneously operative channels was recorded. To assess ATP dependence of isoflurane effects, we first evaluated ATP sensitivity of the channel during control conditions at pHi 7.4. The following protocol was carried out at each tested [ATP]i: control at pHi 7.4, isoflurane at pHi 7.4, washout of isoflurane at pHi 7.4, and 0 ATP at pHi 7.4. Only one concentration of ATP was tested per patch. ATPi inhibited channel activity and decreased Po in a concentration-dependent manner. Figure 1shows summary data for [ATP]i–normalized Po relationship at pHi 7.4 in the control and during application of 0.5 mm isoflurane. Each data point is a mean from four patches. During control conditions, increasing [ATP]i caused a concentration-dependent decrease in Po. Fitting mean data to the Hill equation yielded an IC⁵⁰for ATP inhibition of 8 ± 1.5 μm and a Hill coefficient (nH) of 0.6 ± 0.05. When applied to the internal side of patches at pHi 7.4, isoflurane decreased Po at [ATP]i less than 50 μm but had no marked effect on Po at [ATP]i greater than 50 μm. Figure 2shows the sample traces of single K^ATPchannel activity recorded at pHi 7.4 in the control and during application of 0.5 mm isoflurane at 50 μm [ATP]i. In a patch containing five channels, isoflurane decreased channel activity in a reversible manner. As shown in figure 3, at 100 μm [ATP]i channel activity was much lower and little affected by isoflurane. The unitary amplitudes of 2.2 ± 0.1 pA (control) and 2.1 ± 0.1 pA (isoflurane) and the conductance were not altered by the anesthetic at pHi 7.4.

To test whether anesthetic–K^ATPchannel interaction is modulated by pHi, the effects of isoflurane were examined by decreasing pHi from 7.4 to 6.8. The following protocol was carried out at each tested [ATP]i: control at pHi 7.4, control at pHi 6.8, isoflurane at pHi 6.8, washout of isoflurane at pHi 6.8, and 0 ATP at pHi 6.8. This protocol also included a control baseline at pHi 7.4 because decreasing pHi itself is known to enhance opening of the cardiac K^ATPchannels. 12–14,16Only one concentration of ATP was tested per patch. Figure 4shows recordings at 100 μm [ATP]i from a patch containing three active channels. Infrequent at pHi 7.4, channel opening increased markedly when decreasing pHi to 6.8. Application of 0.5 mm isoflurane at pHi 6.8 enhanced channel activity and further increased Po. Isoflurane effects were reversible during washout. Figure 5shows summary data for [ATP]i–Po relationship obtained during the above condition where each patch was sequentially exposed to the internal solution at pHi 7.4 and pHi 6.8 and to 0.5 mm isoflurane at pHi 6.8. Each data point is a mean from six patches. At pHi 7.4, fitting the [ATP]i–Po relationship to the Hill equation yielded an IC⁵⁰of 8 ± 1.2 μm and nH of 0.6 ± 0.04. Decreasing pHi from 7.4 to 6.8 caused a rightward shift of the curve with IC⁵⁰of 45 ± 5.6 μm and nH of 0.8 ± 0.1. The IC⁵⁰value was approximately fivefold greater than that at pHi 7.4, and both values were different from each other at P < 0.05. When applied at pHi 6.8, isoflurane further decreased ATP sensitivity, and the curve shifted further to the right, yielding an IC⁵⁰of 110 ± 10.0 μm and nH of 1.05 ± 0.12. The IC⁵⁰value was more than twofold greater than that at pHi 6.8 alone, and the values were significantly different at P < 0.05. Neither decreasing pHi to 6.8 nor application of isoflurane at pHi 6.8 altered the amplitude of unitary current, which was 2.2 ± 0.1 pA at pHi 7.4, 2.3 ± 0.1 pA at pHi 6.8, and 2.3 ± 0.1 at pHi 6.8 with isoflurane. Single channel conductance remained in a range of 55–57 pS.

Results from this study provide direct evidence for isoflurane inhibition of ATP sensitivity of cardiac K^ATPchannels at reduced pHi. In the inside-out patches at pHi 6.8, isoflurane potentiates single K^ATPchannels by increasing Po and reduces channel sensitivity to ATP without affecting amplitude of unitary outward current and conductance.

Isoflurane alone did not activate whole cell I^KATPin human atrial 6and guinea pig ventricular myocytes;22however, isoflurane potentiated the whole cell I^KATPpreactivated by pinacidil or 2,4-dinitrophenol. 22In guinea pig ventricular myocytes, isoflurane increased Po of K^ATPchannels in the cell-attached but not in the inside-out membrane patches. 8By contrast, isoflurane decreased the activity of single K^ATPchannels in the inside-out patches from rabbit ventricular myocytes. 7The pHi in all of these studies was kept at 7.4.

In the present study, in the inside-out patches and at pHi 7.4, isoflurane inhibited single K^ATPchannel activity and decreased Po at [ATP]i less than 50 μm in a concentration-independent manner but had no effect on channel activity at [ATP]i greater than 50 μm. This finding confirms the results of Fujimoto et al.  8, who reported lack of isoflurane effects on K^ATPchannel activity in the inside-out patches at 300 μm [ATP]i. We also found that isoflurane does not decrease ATP sensitivity of the channel at pHi 7.4. This observation differs from that of Han et al.  7, who reported a decrease in ATP sensitivity after exposure of inside-out patches to isoflurane at pHi 7.4. Reasons for this discrepancy are not clear, but taking aside species (rabbit vs.  guinea pig) differences, they could be related to differences in the experimental design. First, Han et al.  7examined the effects of isoflurane at the membrane potential of −70 mV and thus investigated the inward current through K^ATPchannels, whereas our studies focused on the unitary outward current at the membrane potential of 0 mV (Fujimoto et al.  8) and +40 mV (present study). This raises the question whether anesthetics may differentially modulate the inward versus  outward conductance of the K^ATPchannel. Second, our measurements were taken during the exposure to isoflurane, whereas Han et al.  7measured ATP sensitivity before and after application of isoflurane, as shown in their figure 1(middle and lower) and figure 3A and B. Third, from their figures 1, 2A, and 3Ait appears that ATP was absent during the patch exposure to isoflurane. Regardless of these differences, both studies have demonstrated that a volatile anesthetic, isoflurane, may directly interact with the sarcolemmal K^ATPchannel.

Our results suggest that pHi may be one of the factors that modulate isoflurane–channel interaction. The K^ATPchannels are sensitive to changes in pHi, and a decrease in pHi is a potent stimulus for their activation by the mechanism that involves a decrease in sensitivity to inhibition by ATP. The optimal effect occurs at pHi 6.8 to 6.5, and further decrease or increase in pHi leads to channel inhibition. 12,15,17–19,21For instance, intracellular acidification to pHi less than 6.5 inhibits cardiac K^ATPchannels by inducing multiple subconductance levels. 15As anticipated, in our study decreasing pHi from 7.4 to 6.8 increased channel opening. This effect was the result of increased Po and reduced ATP sensitivity, as reflected by the rightward shift in the [ATP]i –Po curve with IC⁵⁰increasing from 8 μm to 45 μm. Decreasing pHi to 6.8 did not affect the amplitude of unitary outward current, as also reported by others. 12The IC⁵⁰values for ATP inhibition obtained by us at near physiologic and mild acidotic pHi are in the range of previously reported concentrations. 12,14However, these values are not identical, and our Hill coefficients are lower. This is not surprising because the experimental conditions vary among studies, and it is well known that many factors may alter ATP sensitivity of K^ATPchannels. These include the variations in experimental protocols and ionic conditions, the presence and concentration of monovalent and divalent ions (Ca²⁺, Mg²⁺) and glucose, differences in the range of pHi under study, outward or inward channel conductance under study, and species differences. In addition, pHi sensitivity of K^ATPchannels may be modified by ATP 21and Mg²⁺ions.

Our study demonstrated that mild intracellular acidosis modulates direct interaction of isoflurane with the K^ATPchannel. At pHi 6.8, isoflurane potentiates channel activity by decreasing ATP sensitivity and shifting IC⁵⁰for ATP inhibition from 45 μm to 110 μm. The mechanism by which intracellular acidosis modulates isoflurane–K^ATPchannel interaction is unknown, and we can only be speculative on this point. It has been established that pH sensing is an inherent property of Kir6.2 subunits of the K^ATPchannel. 17Three separate domains in the Kir6.2 protein—the N terminus, C terminus, and M2 domain—are involved in pH regulation, and the proton-sensing amino acid residues responsible for modulation of channel activity were identified in these domains. 18,19Intracellular protons appear to increase the activity of K^ATPchannel by specifically binding to histidine (His-175) on the C-terminus of the Kir6.2 subunit, 18,19and this site is independent of the ATP-binding site, lysine (K185). 17,21Although independent, these sites appear to interact with each other, and allosteric modulation of the cloned K^ATPchannels by ATP and H⁺has been recently demonstrated. 21Whether the allosteric modulation of channel activity by intracellular protons and ATP may play a role in isoflurane potentiation is not known. Because during ionic conditions of our study at pHi 6.8 isoflurane increased channel activity by reducing ATP sensitivity but did not affect unitary current amplitude or conductance, the channel pore is probably not targeted by the anesthetic. However, there is still a possibility of anesthetic interaction with the C terminus of the Kir6.2 subunit harboring the proton and ATP-binding sites. Furthermore, we cannot exclude a possibility of anesthetic interaction with the SUR2A subunit, modulating channel gating. 24,25Previous findings from our laboratory and results of this study suggest that isoflurane may enhance opening of the K^ATPchannel previously activated or modified by the action of intracellular channel regulators, and that pHi may be one of them.

The functional significance of the mitochondrial versus  sarcolemmal K^ATPchannels for cardioprotection remains controversial. Recent evidence suggests a role for mitochondrial K^ATPchannels in the initiation of cardioprotection, 26but sarcolemmal K^ATPchannels have also been indicated in the protection afforded by ischemic preconditioning. 27Activation of the K^ATPchannels is thought to be crucial for anesthetic preconditioning. However, the precise mechanism by which the enhancement of K^ATPchannel activity by volatile anesthetics protects the myocardium is not yet established. Cardiac K^ATPchannels are closed at physiologic [ATP]i. During pathophysiologic conditions, such as ischemia, K^ATPchannels activate during the first few minutes of the ischemic insult, 28long before any significant decrease in [ATP]i. 29This suggested that other intracellular factors must be involved, and modulation of ATP sensitivity by factors targeting specifically the SUR subunit or the Kir6.x subunit has been demonstrated for ADP, 24,25PIP2, 30,31and pHi. 17,21A transient decrease in pHi that occurs during ischemia 32–34and accompanies other metabolic stresses may promote opening of sarcolemmal K^ATPchannels by decreasing sensitivity to ATP, thus allowing isoflurane interaction with the channel, leading to further enhancement of channel activity. Recent studies implicated an important role of the adenylate kinase and creatine kinase-mediated phosphotransfer in communicating mitochondria-generated signals to the sarcolemmal K^ATPchannels. 35,36Whether volatile anesthetics alter the signal transfer to the cell subsarcolemmal compartment and K^ATPchannels is an open question.

Our results support a possible role of volatile anesthetics in ischemia because of early ischemic acidosis. 32However, as demonstrated in the animal models and in humans, volatile anesthetics precondition the myocardium independently of ischemia. 3,6,37–39In clinical settings, the pHi dependence of K^ATPchannel potentiation by isoflurane would likely play a protective role perioperatively when various metabolic stresses may produce a transient decrease in pHi.

In conclusion, this study provides evidence for pHi-dependent modulation of direct isoflurane interaction with the cardiac sarcolemmal K^ATPchannel in guinea pig ventricular myocytes. Although at near physiologic pHi isoflurane has no effect nor inhibits K^ATPchannel, at reduced pHi of 6.8, isoflurane increases channel Po and decreases sensitivity to inhibition by ATP as reflected by the rightward shift of the [ATP]i–Po relationship.