Volatile Anesthetics Mimic Cardiac Preconditioning by Priming the Activation of Mitochondrial K^ATPChannels via  Multiple Signaling Pathways

This study was performed in accordance with the guidelines of the Animal Care and Use Committee of the University of Zurich.

Ca²⁺-tolerant adult rat ventricular myocytes were isolated from hearts of male Sprague-Dawley rats (300 g) as previously described. 11Thirty minutes prior to decapitation, animals were heparinized (500 units administered intraperitoneally). To avoid any putative effects on K^ATPchannel activity, 12,13no anesthetics were used. The hearts were quickly removed into a 4°C HEPES-buffered solution (117 mm NaCl, 5.7 mm KCl, 1.5 mm NaH²PO⁴, 4.4 mm NaHCO³, 20 mm HEPES, 10 mm glucose, 10 mm creatine, at pH 7.4) containing 1.8 mm Ca²⁺and then perfused on a Langendorff apparatus for 5 min at 37°C gassed with oxygen. The perfusion solution was switched to a nominally Ca²⁺-free solution containing 0.1 mm EGTA for 7 min and then to a nominally Ca²⁺-free solution containing 0.15% collagenase A (Roche, Mannheim, Germany). After 30 min of digestion, the enzymatic solution was washed out, and the tissue from the left ventricle was cut into small pieces and gently swirled in the HEPES-buffered solution containing 0.125 mm Ca²⁺. Dissociated cells were filtered through a 200-μm mesh and allowed to settle for 20 min. The cells were then resuspended in the buffered solution containing 5 mg/ml BSA and exposed to a graded series of increasing Ca²⁺concentrations up to 1.8 mm. Each step was followed by a gentle centrifugation with less than 20 g  for 2 min to separate the ventricular myocytes from nonmyocytes. The isolated myocytes were resuspended in serum-free defined culture medium consisting of DMEM with 2 mg/ml BSA, 2 mm l-carnitine, 5 mm creatine, and 5 mm taurine. Notably, to avoid any putative effect on K^ATPchannels, no antibiotics were added to culture medium. Cells were incubated for 3 h before experiments to allow reestablishment of normal electrolyte gradients. Purity of cardiomyocyte cultures was assessed by determining the percentage of myosin positive–staining cells using immunofluorescence with a myosin heavy chain specific antibody, MF-20, as previously described. Ninety-nine percent of cells stained positive. 11 

The fluorescence of FAD-linked enzymes, called flavoproteins, was used to monitor mitochondrial redox state, which is a direct indicator of mitoK^ATPchannel activity. 14,15Isolated myocytes were cultured at a density of 100–150 cells/mm²on 20 mm round glass coverslips precoated with laminin (1 μg/cm²; Sigma, St. Louis, MO) placed in 35-mm plastic culture dishes. After 2 h, the dishes were washed with phosphate-buffered saline to remove cells that were not attached. Experiments were performed over the next 12 h. The treatment protocols and the concentrations of the reagents used were established previously. 16,17Myocytes on glass coverslips were placed in a customized airtight perfusion chamber with a volume of 0.5 ml, covered with a 25-mm glass coverslip, which was tightly sealed with vacuum grease (Fisher, Pittsburgh, PA) and perfused at room temperature (25°C) with a buffer solution at a flow rate of 0.5 ml/min containing (pH 7.4) 140 mm NaCl, 5 mm KCl, 1 mm MgCl², 1.8 mm CaCl², and 10 mm HEPES. Myocytes were exposed in series to the following solutions: plain buffer solution (baseline), buffer with 100 μm diazoxide (Biomol, Plymouth Meeting, PA) followed by a washout with plain buffer, buffer saturated with sevoflurane or isoflurane (Abbott AG, Baar, Switzerland) in the presence or absence of 2 μm chelerythrine (Biomol) for 10 min followed by a washout with plain buffer, and finally buffer with 100 μm 2,4-dinitrophenol (Sigma, St. Louis, MO). Buffer solution was equilibrated with 2.8% (vol/vol) sevoflurane or 1.2% (vol/vol) isoflurane using a Sevotec 5 vaporizer (Datex-Ohmeda, Tewksbury, MA) or an Isotec 3 vaporizer (Datex-Ohmeda), respectively, with an air bubbler. Delivered vapor concentrations of the volatile anesthetics were continuously controlled by the infrared gas analyzer Capnomac Ultima (Datex-Ohmeda). Applied concentrations of sevoflurane and isoflurane were also measured in the buffer solution of the perfusion chamber using a gas chromatograph (Perkin-Elmer, Norwalk, CT): sevoflurane 2.8% (vol/vol) (2 minimum alveolar concentration [MAC] in rats at 25°C) 0.92 ± 0.01 mm, isoflurane 1.2% (vol/vol) (2 MAC in rats at 25°C) 0.63 ± 0.01 mm. Chelerythrine was dissolved in buffer at a concentration of 2 μm. Diazoxide was dissolved at 100 μm in buffer solution containing dimethyl sulfoxide (DMSO) 0.1%. Because DMSO was used as a solvent for diazoxide, a series of experiments was studied in which DMSO alone was administered to myocytes. Importantly, DMSO alone at the concentration used in the experiments had no effect on autofluorescence of myocytes (data not shown). Also, separate experiments showed that chelerythrine (2 μm) alone or followed by diazoxide did not affect autofluorescence of flavoproteins (see Results section). For each experimental group, myocytes of 8 different rat hearts were used (n = 8), and all experiments were performed in duplicate. Data from one heart were averaged. An upright microscope (Axioplan2; Zeiss, Jena, Germany) equipped with a xenon arc lamp and the appropriate filter set (excitation at 480 nm, emission at 530 nm) was used to monitor the autofluorescence of myocytes. Images were captured using a cooled CCD camera (ORCA-100, 12 bit digital output; Hamamatsu Photonics, Herrsching, Germany) controlled by an image acquisition software (Openlab; Improvision, Lexington, MA). Fluorescence intensity was recorded every 15 s by exposing myocytes for 125 ms using a computer-controlled high-speed shutter (Openlab; Improvision). Calibration of flavoprotein fluorescence was achieved by setting fluorescence obtained after 2,4-dinitrophenol exposure to 100%. 2,4-Dinitrophenol uncouples oxidative phosphorylation and subsequently leads to fully oxidized flavoproteins. All measurements of flavoprotein oxidation were then expressed as percentage of 2,4-dinitrophenol–induced maximal autofluorescence. In each experiment, the fluorescence of 5–10 myocytes was monitored with a 20× objective lens (LD Achroplane, NA = 0.4; Zeiss). Processing of the flavoprotein signals was achieved using a binary mask, which was separately drawn for each myocyte using Openlab software (Improvision). This allowed exclusion of artifacts from other myocytes and background, and the time course of flavoprotein fluorescence of multiple individual myocytes could be selectively and simultaneously traced. A pseudocolor palette was finally used to visualize the relative intensity of mitochondrial flavoprotein oxidation states. Myocytes with an increased initial autofluorescence greater than 30% of the peak 2,4-dinitrophenol–induced autofluorescence were excluded from analysis and considered as damaged cells.

Myocytes were suspended in the incubation buffer containing 119 mm NaCl, 25 mm NaHCO³, 1.2 mm KH²PO⁴, 4.8 mm KCl, 1.2 mm MgSO⁴, 1.8 mm CaCl², 10 mm HEPES, 11 mm glucose, 24.9 mm creatine, and 58.5 mm taurine and supplemented with 1% basal medium Eagle amino acids (GIBCO, Paisley, Scotland, UK) and 1% minimum essential medium nonessential amino acids (Sigma) at pH 7.4. Volatile anesthetics at 0.5, 1.0, 1.5, and 2.0 MAC were administered to myocytes for 15 min before simulated ischemia by gently bubbling a 10-ml tube in a standard incubator at 37°C, at the bottom of which 1.5 ml of a myocyte suspension was placed. Control samples were only gassed with air. Applied concentrations were continuously monitored and in addition measured by gas chromatography: sevoflurane: 0.5 MAC 0.22 ± 0.02 mm, 1.0 MAC (2.5%[vol/vol]) 0.50 ± 0.03 mm, 1.5 MAC 0.71 ± 0.01 mm, 2.0 MAC 0.96 ± 0.01 mm; isoflurane: 0.5 MAC 0.16 ± 0.02 mm, 1 MAC (1.4%[vol/vol]) 0.34 ± 0.02 mm, 1.5 MAC 0.50 ± 0.02 mm, 2 MAC 0.72 ± 0.01 mm. In separate experiments, for each of the drugs listed below, dose–response curves with respect to myocyte death were established in the presence and absence of volatile anesthetics to determine the optimum concentration, which guaranteed the maximum inhibitory or stimulatory effect on the specific signal transduction component under investigation. Accordingly, depending on the treatment group, myocytes were exposed to 100 μm 8-sulfophenyl theophylline, 100 μm 8-cyclopentyl-1,3-dipropylxanthine (DPCPX), 100 μm adenosine, 0.1 μm prazosin, 5 μm propranolol, 2 μm chelerythrine, 100 μm 5-hydroxydecanoate (all drugs purchased from Sigma), 50 μm HMR-1098 (a gift from Aventis AG, Frankfurt am Main, Germany), 100 μm 2-(4-carboxyphenyl)-4,4′,5,5′-tetramethylimidazole-1-oxyl-3-oxide (PTIO), 1 μg/ml pertussis toxin (PTX), 50 μm N  ^G-nitro-l-arginine methyl ester (l-NAME), 50 μm l-N6-(1-iminoethyl)lysine (l-NIL), 1 mm 8-bromo-cGMP, and 100 μm S -nitroso-N -acetyl-dl-penicillamine (SNAP) (all drugs purchased from Biomol). Drugs were administered 20 min prior to simulated ischemia and 5 min prior to administration of volatile anesthetics, respectively, except for PTX, which was added to myocytes 2 h before ischemia. Diazoxide, adenosine, SPT, SNAP, and PTIO were dissolved in DMSO containing buffer solution (final concentration of DMSO 0.1%). Because DMSO was used as a solvent for these drugs, a series of experiments was performed in which DMSO alone was administered to myocytes. DMSO alone, at the concentration used in the experiments, had no effect on survival of myocytes. In separate experiments diazoxide was administered at 1 μm, 10 μm, and 100 μm in the presence and absence of 1 MAC sevoflurane or isoflurane 15 min before 120 min of ischemia. After the various treatment modalities, 1 ml of the myocyte suspension was pipetted from the bottom of the 10-ml tube into a 1.5-ml microcentrifuge tube and centrifuged for 20 s at 15 g . The supernatant was discarded, and 0.25 ml of mineral oil (Sigma) was layered onto the myocyte pellet to inhibit gaseous diffusion of oxygen, as previously described. 17,18Following 60 and 120 min of incubation in a standard incubator at 37°C, 15 μl of myocytes was sampled through the mineral oil layer and mixed with 150 μl of a hypotonic trypan blue staining solution containing 11.9 mm NaHCO³, 0.4 mm KH²PO⁴, 2.7 mm KCl, 0.8 mm MgSO⁴, 1.8 mm CaCl², 0.5% glutaraldehyde, and 0.5% trypan blue. 19,20In the various protocols, for each experimental group, cells of 8 different rat hearts (n = 8) were used, and experiments were performed in duplicate. Myocyte viability was assessed by counting the number of myocytes staining clearly after 3 min of exposure to the hypoosmolar trypan blue staining solution. Myocytes, which stained dark blue, were considered irreversibly damaged. Five randomly chosen fields (1 mm²) were counted at 10 × 10 magnification with a phase-contrast microscope in duplicate and expressed as the percentage of total viable myocytes before ischemia. The percentage of myocytes viable at the beginning of the experiments was 90 ± 3% (n = 30). The small percentage of nonviable myocytes was due to the enzymatic isolation procedure. Importantly, the number of viable cells remained unchanged in untreated myocytes over the 1- or 2-h period of the experiments, and administration of volatile anesthetics did not affect myocyte viability over this period. Also, in the preparation used, the myocytes excluded trypan blue and retained their rod-shaped configurations over a 6-h period following isolation (89 ± 4%, n = 30).

Data are expressed as mean ± SD. Analysis of variance with post hoc  Scheffé test for multiple comparisons was performed to determine statistical significance of multiple treatments. P < 0.05 was considered to be significant (StatView Version 4.5; Abacus Concepts, Berkeley, CA).

To evaluate the potency of sevoflurane and isoflurane to protect myocytes from undergoing irreversible ischemic damage, isolated myocytes were exposed to increasing concentrations of sevoflurane and isoflurane (0.5 MAC, 1 MAC, 1.5 MAC, 2.0 MAC) 15 min before 60 or 120 min of ischemia. Cellular injury was determined by counting myocytes, which were not capable of excluding trypan blue staining under hypotonic conditions. Both volatile anesthetics dose-dependently reduced the percentage of trypan blue–positive myocytes, with isoflurane slightly more effective than sevoflurane at equipotent MAC values (fig. 1). After 120 min of ischemia, pretreatment of myocytes with 2 MAC sevoflurane and isoflurane decreased the percentage of trypan blue–positive myocytes markedly, from 67 ± 5% to 30 ± 3% and 26 ± 3%, respectively (P < 0.0001, n = 8). Importantly, no substantial further protection was achieved by sevoflurane and isoflurane with MAC values greater than 1.5.

To test whether isoflurane- and sevoflurane-induced protection in myocytes is mediated by the sarcK^ATPchannel or the mitoK^ATPchannel, myocytes were exposed to 50 μm HMR-1098, a selective sarcK^ATPchannel blocker, or 100 μm 5-hydroxydecanoate, a selective mitoK^ATPchannel blocker, 5 min before exposure to volatile anesthetics (1 MAC sevoflurane or isoflurane for 15 min). Subsequent ischemia was maintained for 60 or 120 min. HMR-1098 did not diminish protection elicited by volatile anesthetics (after 120 min of ischemia, sevoflurane: 45 ± 4%vs . sevoflurane + HMR-1098: 44 ± 3%; isoflurane: 39 ± 3%vs . isoflurane + HMR-1098: 41 ± 4%; not significant, n = 8). In contrast, 5-hydroxydecanoate completely abolished this protection, as assessed by hypoosmolar trypan blue staining (after 120 min of ischemia, sevoflurane: 45 ± 4%vs . sevoflurane + 5-hydroxydecanoate: 63 ± 4%; isoflurane: 39 ± 3%vs . isoflurane + 5-hydroxydecanoate: 61 ± 3%;P < 0.0001, n = 8, sevoflurane + 5-hydroxydecanoate and isoflurane + 5-hydroxydecanoate vs . control (ischemia alone); not significant, n = 8) (fig. 2). These experiments clearly indicate at a cellular level that mitoK^ATPchannels play a pivotal role in ischemic protection elicited by volatile anesthetics.

Diazoxide is known as a highly selective potent opener of the mitoK^ATPchannels (2,000-fold more potent at the mitochondrial channel than at the surface channel). To test whether volatile anesthetics further potentiate the opening effect of diazoxide on mitoK^ATPchannels, myocytes where exposed to increasing concentrations of diazoxide (1 μm, 10 μm, 100 μm) in the presence of sevoflurane or isoflurane (1 MAC for 15 min) before 60 or 120 min of ischemia. Volatile anesthetics markedly potentiated the protective effect of diazoxide and significantly reduced the percentage of trypan blue–positive myocytes (100 μm diazoxide: 22 ± 3%vs . 100 μm diazoxide + 1 MAC sevoflurane: 11 ± 2% and 100 μm diazoxide + 1 MAC isoflurane: 10 ± 2%, P < 0.0001, n = 8), with isoflurane slightly more potent than sevoflurane at the diazoxide concentrations of 1 μm and 10 μm (fig. 3). The results of these experiments suggest that volatile anesthetics enhance opening of mitoK^ATPchannels.

The redox state of flavoproteins directly reflects mitoK^ATPchannel activity. To test whether volatile anesthetics directly open mitoK^ATPchannels, the effect of sevoflurane and isoflurane on flavoprotein oxidation was visualized using fluorescence microscopy. Notably, exposure of sevoflurane and isoflurane as long as 30 min did not alter flavoprotein oxidation as compared with baseline values. Conversely, diazoxide administration to myocytes pretreated with sevoflurane (2.8%[vol/vol]) and isoflurane (1.2%[vol/vol]) clearly accelerated as well as markedly enhanced diazoxide-induced increases in flavoprotein-mediated autofluorescence, which indicates that volatile anesthetics prime activation of mitoK^ATPbut do not open the channel directly (fig. 4). Importantly, coadministration of chelerythrine (2 μm), a specific protein kinase C (PKC) inhibitor, to sevoflurane abolished the priming effect of sevoflurane on diazoxide. Similar results were obtained for isoflurane.

Chelerythrine was previously reported to enhance fluorescence unspecifically. 17However, at the low concentration of 2 μm used in the present work, chelerythrine did not affect baseline fluorescence of the flavoproteins (flavoprotein oxidation at emission wavelength of 530 nm of chelerythrine alone was 18 ± 3% as compared to control [16 ± 2%] in the absence of chelerythrine). We also tested whether chelerythrine alone would modify the subsequent diazoxide-induced fluorescence peak. Pretreatment with chelerythrine did not alter the diazoxide-induced fluorescence peak (41 ± 3%) compared to diazoxide without chelerythrine pretreatment (40 ± 4%).

Conversely, l-NAME, l-NIL, and PTIO could only be used in the experiments using the model of simulated ischemia on isolated myocytes. The latter three substances were found to significantly suppress baseline flavoprotein fluorescence: flavoprotein oxidation was 4 ± 2% for 100 μm PTIO, 7 ± 2% for 50 μm l-NAME, and 7 ± 3% for 50 μm l-NIL compared to baseline (16 ± 2%).

A number of modulators known to be involved in ischemic preconditioning were used to test their effects on volatile anesthetic–induced cellular protection. At the same time, this allowed us to define the main signaling pathways, which are known to contribute to the preconditioned state. Isolated myocytes were exposed to 2 MAC sevoflurane since the dose–response with regard to cellular protection leveled off between 1.5 and 2 MAC (fig. 1). The concentrations of the modulators used in these experiments were established in separate experiments and represent the concentrations required to obtain the maximum inhibitory or stimulatory effects on the signaling components under investigation. The percentage of trypan blue–positive myocytes as a measure for cell viability was evaluated after 120 min of ischemia. Concentration dependence of individual modulators and controls are summarized in table 1. The data concerning the specific effects of individual modulators at maximum inhibitory or stimulatory effects from table 1is visualized for comparison in figure 5.

The unselective adenosine receptor blocker SPT and the selective adenosine 1 receptor blocker DPCPX significantly attenuated the sevoflurane-induced protection but did not completely block its protective effect (fig. 5). Furthermore, the Gⁱ-inhibiting PTX significantly decreased protection by sevoflurane, and adenosine at 100 μm elicited significant protection, which was, however, less pronounced than administration of 2 MAC sevoflurane (2 MAC sevoflurane: 30 ± 3%vs . adenosine 42 ± 4%, P < 0.001, n = 8). Notably, SPT, DPCPX, and PTX alone did not affect the percentage of trypan blue–positive myocytes exposed to 120 min of ischemia. Taken together, these results indicate that mechanisms other than pure stimulation of adenosine receptors contribute to the protection elicited by volatile anesthetics in rat ventricular myocytes.

To determine the role of nitric oxide (NO) in mediating sevoflurane-induced ischemic protection, myocytes were treated with the NO-scavenger PTIO (100 μm) and the NO synthase (NOS) inhibitors l-NAME (50 μm) and l-NIL (50 μm) (relatively specific for inducible NOS). PTIO and both NOS inhibitors significantly diminished but did not abolish the protective effect of sevoflurane. PTIO, l-NAME, and l-NIL alone did not affect the percentage of trypan blue–positive myocytes exposed to 120 min of ischemia. In accordance with this observation, SNAP at 100 μm, a NO donor, and 8-bromo-cGMP at 1 mm, a membrane-permeable analog of cGMP, which is an important mediator of many NO effects, prevented myocytes from undergoing irreversible ischemic damage. Notably, prazosin at 0.1 μm and propranolol at 5 μm did not affect sevoflurane-induced protection. The PKC inhibitor chelerythrine at 2 μm alone and the coadministration of 100 μm SPT and 100 μm PTIO completely abolished the cytoprotection by sevoflurane. Chelerythrine, prazosin, and propranolol alone did not affect the percentage of trypan blue–positive myocytes exposed to 120 min of ischemia.

The results of these experiments suggest that, in rat ventricular myocytes, multiple signaling pathways may contribute to sevoflurane-induced protection against ischemia, which presumably culminate in activation of PKC and thereby prime mitoK^ATPchannels. Figure 6depicts a simplified scheme of the signaling circuits hypothetically involved in the priming of the mitoK^ATPchannel in response to volatile anesthetics.

The principal new findings of this investigation are as follows. Volatile anesthetics do not directly open mitoK^ATPchannels but prime mitoK^ATPchannel activity. First, volatile anesthetics did not affect basal flavoprotein fluorescence in myocytes but potentiated the oxidative effect of the highly specific mitoK^ATPchannel opener diazoxide. Their coadministration to diazoxide also abbreviated the latency to peak mitoK^ATPchannel activity. Second, diazoxide-induced myocyte protection was potentiated by volatile anesthetics in a cellular model of simulated ischemia. Opening of the mitoK^ATPchannels, but not sarcK^ATPchannels, elicited this protection. The specific PKC inhibitor chelerythrine completely abrogated both the enhanced flavoprotein oxidation and the cellular protection induced by volatile anesthetics. This suggests that PKC represents a key signaling component upstream of the mitoK^ATPchannel and that these two events are closely associated. Lastly, our experiments indicate that the cardioprotective potency of volatile anesthetics parallels MAC values, but that different volatile anesthetics may confer varying degrees of protection at equipotent MAC values.

Recent studies focused on the elucidation of the molecular mechanisms that are involved in anesthetic-induced preconditioning. Since simultaneous administration of ischemic preconditioning and volatile anesthetics does not induce additional protection over that provided by each intervention alone, cardioprotection by volatile anesthetics is thought to be mediated by the same end effector as classic preconditioning. 21Using blockers of the K^ATPchannel, many studies identified this channel in a multitude of animal models and experimental conditions as a key element in mediating the anesthetic-induced preconditioned state. 2,3,9,22,23Most recently, effective preconditioning of human atrial trabeculae by isoflurane, but not halothane, was reported, and this effect was clearly abolished by glibenclamide, a nonspecific K^ATPchannel blocker. 24Furthermore, desflurane reduced myocardial infarction size through mitoK^ATPchannels as well as sarcK^ATPchannels, 9and sevoflurane was found to protect stunned myocardium, 25which was blocked by the specific mitoK^ATPchannel blocker 5-hydroxydecanoate. Although recent data suggest that the mitoK^ATPchannel is even more important in mediating preconditioning-like effects, so far, no data were available with respect to interactions between volatile anesthetics and mitoK^ATPchannel activity. Only one study investigated the effects of isoflurane on sarcK^ATPchannel activity in rabbit ventricular myocytes and revealed that isoflurane inhibits channel activity without a change in single channel conductance. 10However, isoflurane decreased the ATP sensitivity of the channel, leading to increased probability of channel opening for a given concentration of ATP. The results of our study are in line with these observations in sarcK^ATPchannels in so much as no direct opening of the mitoK^ATPchannel is recorded in response to volatile anesthetics.

The use of autofluorescence emerged as a new tool with great impact in the endeavor to understand cellular mechanisms. 26In mammalian cells, autofluorescence is caused largely by the reduced pyridine nucleotides [NAD(P)/H] and by the oxidized flavoproteins (FAD/H²). These endogenous fluorophores transfer electrons to oxygen in the inner mitochondrial membrane, ultimately leading to formation of H²O and synthesis of ATP. The excitation of flavoproteins (excitation maximum at 480 nm) is maximal under full oxidation and minimal under full reduction, whereas the opposite is true for NAD(P)H (excitation maxima at 290, 336, 351 nm). The ratio of the concentration of oxidized and reduced electron carriers or of their fluorescence, respectively, therefore gives a measure of the cellular metabolic state. On the other hand, selective monitoring of flavoprotein-induced autofluorescence directly reflects mitoK^ATPchannel activity, as previously shown. 27This noninvasive technique has been successfully used by Marbán et al . to uncover important molecular mechanisms underlying the phenomenon of preconditioning. 14,16,17A previous study investigated the effect of halothane (0.27 mm) on pyridine nucleotide [NAD(P)H]-induced autofluorescence in nonbeating rat ventricular myocytes. 28Due to the opposite behavior of flavoprotein and pyridine nucleotides with respect to the redox state-dependent generation of autofluorescence, the results of this study could give some indications on how flavoproteins and thereby mitoK^ATPchannel activity could be affected by volatile anesthetics. Interestingly, no change in NAD(P)H-mediated autofluorescence was recorded in response to halothane in the absence of electrical stimulation, which indirectly implies that halothane, at the indicated concentration, did not affect basal flavoprotein-induced fluorescence or the mitoK^ATPchannel activity, respectively. This is in clear accordance with the results of our study.

How does activation of K^ATPchannels mediate cardioprotection? A cardioplegic-like effect with action potential shortening, decreased energy consumption, and reduced cytosolic Ca²⁺overload has been proposed as the protective mechanism caused by increased sarcK^ATPchannel activity. 6Conversely, dissipation of the inner mitochondrial membrane potential in response to opening of the mitoK^ATPchannels blunts Ca²⁺overload of mitochondria 29and leads to restoration of the mitochondrial intermembrane space, 30which reestablishes functional coupling between adenine nucleotide translocase and creatine kinase as well as energy processes from mitochondria to ATP-utilizing cytosolic sites. 31Interestingly, considerable cross-talk was documented between sarcK^ATPand mitoK^ATPchannels, whereby increased ATP consumption through uncoupled mitochondria leads, in turn, to activation of sarcK^ATPchannels. 32The results of the current study emphasize the importance of the protective effects of mitoK^ATPchannel in mediating volatile anesthetic–induced protection, but the role of the sarcK^ATPchannel should not be totally dismissed.

Recently, Sato et al . proposed a 3-state model, including resting, primed, and open state of the mitoK^ATPchannel. 17This was motivated by the observation that adenosine did not affect basal mitoK^ATPchannel activity but significantly enhanced opening by diazoxide, a direct mitoK^ATPchannel opener. The primed receptor state allows easy and rapid opening at the initiation of ischemia, and most probably represents a specifically phosphorylated state of the receptor. In our study, isoflurane and sevoflurane similarly potentiated diazoxide-mediated effects, which were blocked by chelerythrine, a specific PKC inhibitor. The important role of adenosine receptors and PKC in the cardioprotection elicited by volatile anesthetics was repeatedly reported in previous studies. 24,33,34Of note, volatile anesthetics may even directly activate PKC, although their reported actions on PKC are contradictory. The so-far hypothetical model of a primed receptor state could explain the recently observed decrease in time threshold by sevoflurane for ischemic preconditioning. 35In this context, an antiprimed state of the mitoK^ATPchannel may also underlie the observed phenomenon of “antipreconditioning,” where the inability of the channel to open at initiation of ischemia may lead to the reported potentiated ischemic damage. 36Taken together, priming of the mitoK^ATPchannel may reflect a general characteristic of this channel.

Using a cellular model of simulated ischemia, we evaluated the effects of specific blockers on various putative signaling pathways involved in the preconditioning-like state induced by volatile anesthetics. This in vitro  model mimics tissue ischemia by restriction of oxygen and extracellular fluid as well as accumulation of metabolites. 18,19Subsequent superposition of an artificially high hypotonic stress upon cardiomyocytes allows sensitive detection of even latent ultrastructural ischemic lesions in sarcolemmal membrane integrity by trypan blue permeability exclusion. 20The results of these experiments allowed us to delineate the signaling pathways involved in the preconditioned state in response to volatile anesthetics (fig. 6). Specifically, the involvement of G-protein–linked signaling, as previously shown in a canine model, 37could be confirmed in our experiments. Two observations, however, need further consideration. First, although initially it was felt that adenosine receptors were not involved in preconditioning in rat, more recent evidence reveals that adenosine antagonists can at least blunt the preconditioning-induced protection. This is consistent with the results of our study, where two adenosine receptor antagonists, SPT and DPCPX, diminished the protective effects of volatile anesthetics. Also, a recent study in rat ventricular myocytes demonstrated that activation of adenosine 1 receptors reduces reactive oxygen species and significantly attenuates myocardial stunning. 38Second, the presented results also suggest NO–cGMP as elements in volatile anesthetic–induced protection (fig. 6). Administration of the NO scavenger PTIO or the NOS inhibitors l-NAME and l-NIL clearly inhibited protection afforded by volatile anesthetics. Notably, NO and its metabolite peroxinitrite are known to activate PKC and the K^ATPchannels. 16NO–cGMP signaling and basal NOS activity were previously reported to play an important role in pacing-associated preconditioning in the isolated rat heart. 39,40Moreover, a recent study in chicken myocytes demonstrated the abrogation of preconditioning protection using 5-hydroxydecanoate or l-NAME. 41It may well be that volatile anesthetics differentially modulate the activity of the various isoenzymes of NOS (nNOS, eNOS, iNOS), which are ubiquitous but heterogeneously distributed in myocytes. 42Although the role of NO in late preconditioning (second window of protection) is well established, its role in early preconditioning, specifically in anesthetic-induced preconditioning, needs further investigations.

Apoptosis, the programmed cell death, plays a key role in myocardial infarction and the various forms of cardiomyopathies. Volatile anesthetics were recently reported to inhibit catecholamine-induced apoptosis in rat ventricular myocytes by modulation of cellular Ca²⁺homeostasis and inhibition of the apoptosis initiator caspase-9, which is closely related to mitochondrial integrity. 11A recent study now links opening of the mitoK^ATPchannel to significant antiapoptotic effects in myocytes, 43thereby raising the interesting possibility that cardioprotection by volatile anesthetics during ischemia may be caused by their priming effect on mitoK^ATPchannels. Taken together, the preconditioning effects of volatile anesthetics are sensitive to adenosine receptor– and NO-coupled signaling.

The results of the present study should be interpreted with caution. Specifically, we recognize that mechanistic information on preconditioning in rat myocytes may not be transferable to other species, in particular to humans. Also, isolated myocyte models have limitations with respect to the choice of external solutions, substrate selection, and unphysiologically low workload. In addition, the effects of only a limited number of putative signaling pathways involved in anesthetic-induced preconditioning were assessed in the present study. Importantly, the use of basal anesthesia, in particular the use of barbiturates 12and ketamine, 13has a great potential to affect experimental results and was therefore carefully avoided in our experiments. Even α-chloralose, which is appreciated for its neglectable effects on experimental results, can potentially affect K^ATPchannel activity by its main metabolite trichloroethanol, since ethanol is known to induce preconditioning. 44Finally, although the question of whether sarcK^ATPor mitoK^ATPchannels would be the more important contributors to anesthetic-induced preconditioning in vivo  could not be clarified by our experimental model, the results of these studies pinpoint the mitoK^ATPchannel as a potential therapeutic target for cardioprotection.

In summary, volatile anesthetics prime mitoK^ATPchannels through multiple PKC-coupled signaling pathways in a model of isolated rat ventricular myocytes. Since anesthetic-induced preconditioning not only affects the heart but also may protect a variety of other tissues, 45appropriate clinical studies are now needed to ascertain the utility and efficacy of this promising therapeutic strategy in perioperative medicine, specifically in patients at high risk for perioperative ischemic injury.