Mechanisms of Cardioprotection by Volatile Anesthetics

The antiischemic effects of volatile anesthetics were initially proposed more than 20 yr ago. Lowenstein et al.  1demonstrated that halothane reduced ST segment elevation in a canine model of brief coronary artery occlusion. These data were consistent with the hypothesis that exposure to halothane reduced acute ischemic injury. A subsequent study by this research group also demonstrated that halothane reduced myocardial infarct size when administered before prolonged coronary artery occlusion in dogs. 2Lactate production was decreased in the presence as compared with the absence of enflurane during demand-induced ischemia produced by a critical coronary artery stenosis and ventricular pacing when coronary perfusion pressure was maintained. 3These results suggested that myocardial metabolism may be improved by administration of a volatile agent during an ischemic episode independent of alterations in hemodynamics. The relative importance of these early findings was initially overshadowed 4by a series of reports published in the mid-1980s 5–8suggesting that isoflurane may be capable of producing an abnormal redistribution of coronary blood flow away from ischemic toward normal myocardium. 9–11This “coronary steal” phenomenon was attributed to the coronary vasodilating properties of isoflurane that are known to occur primarily in arterioles of less than 100 μm in diameter. 12Isoflurane was thought to be capable of directly producing myocardial ischemia in susceptible patients with “steal-prone” coronary artery anatomy under certain hemodynamic conditions 13in a fashion similar to that of potent coronary vasodilators (e.g. , adenosine, chromonar, dipyridamole).

The implication that isoflurane might produce myocardial ischemia through such a steal mechanism was subsequently dispelled by several investigations conducted in animal models 14–17and humans with coronary artery disease. 18–20For example, isoflurane did not selectively redistribute blood flow away from the collateral-dependent region in a chronically instrumented canine model of multivessel coronary artery disease. 15In contrast, adenosine produced marked coronary steal by preferentially shunting blood flow away from collateral-dependent myocardium in this model. 15Other studies 14,16–20also suggested that isoflurane-induced hypotension may reduce myocardial perfusion, but true coronary steal did not occur when coronary perfusion pressure was maintained. Subsequent investigations with the newer volatile anesthetics sevoflurane 21,22and desflurane 23showed that these drugs also did not reduce or abnormally redistribute coronary collateral blood flow. Therefore, despite initial concerns, volatile anesthetics were subsequently shown to be relatively weak coronary vasodilators that are incapable of causing coronary steal under the vast majority of clinical conditions. 24 

Contrary to the hypothesis that the use of volatile anesthetics may be potentially deleterious in some patients with coronary artery disease, many laboratory and clinical investigations conducted since the resolution of the coronary steal controversy have convincingly shown that volatile anesthetics protect  the heart against ischemia and reperfusion injury. In addition to previously cited studies suggesting that halothane 1,2and enflurane 3exerted antiischemic effects, halothane was also shown to preserve contractile function and ultrastructural integrity during cardioplegic arrest. 25This latter study was of considerable interest because these data indicated that halothane was capable of exerting a cardioprotective effect completely independent of improvements in myocardial oxygen supply-demand balance. In addition, halothane, 26enflurane, 26desflurane, 26–28and sevoflurane 26–31have been shown to reduce myocardial damage when administered during reperfusion after prolonged coronary occlusion or cardioplegic arrest. Another study showed that preservation of high-energy phosphate concentrations was coupled to the protective effects of enflurane. 32Isoflurane and desflurane did not depress but modestly enhanced left ventricular diastolic function during acute coronary occlusion in dogs. 33Halothane, 34–36enflurane, 34,37,38isoflurane, 34,35,37and sevoflurane 39were shown to improve the functional recovery of isolated hearts subjected to global ischemia and reperfusion. Halothane and isoflurane markedly augmented the recovery of regional contractile function of stunned myocardium in vivo . 40,41Halothane 2and isoflurane 42reduced myocardial infarct size in dogs, and this beneficial action was found to persist despite discontinuation of the volatile anesthetic before coronary artery occlusion. 42The myocardium acted as if it had “remembered” the previous exposure to the volatile agent. This phenomenon was termed anesthetic-induced preconditioning  (APC) 43and was characterized by a short-term memory phase similar to that observed during ischemic preconditioning (IPC).

Anesthetic-induced preconditioning has also been described in other animal species, including rats 44and rabbits. 43The efficacy of APC conferred by isoflurane to reduce infarct size has been shown to be dose dependent in rats, 44an animal model with minimal coronary collateral flow. 45High concentrations of isoflurane may also have greater efficacy to protect myocardium during conditions of low coronary collateral blood flow in dog myocardium. 46Similarly, isoflurane and sevoflurane dose-dependently preserved the viability of isolated cardiac myocytes during ischemia. 47Isoflurane has been shown to elicit cardioprotective effects after discontinuation for 15 min or 30 min before coronary artery occlusion in rats and rabbits 43,44or dogs, 42respectively. In contrast, sevoflurane did not exert antiischemic actions after a 30-min washout period. 48Taken together, these data suggest that the memory period associated with APC may differ between volatile anesthetics and among species. Interestingly, recent findings showed that isoflurane reduced myocardial damage when administered 24 h before coronary artery occlusion and reperfusion in rabbit hearts in vivo . 49Pretreatment with isoflurane also preserved endothelial and vascular smooth muscle cell viability 12–48 h after cytokine-induced injury. 50Therefore, volatile anesthetics also produce a late phase (i.e. , a second window) of myocardial protection similar to IPC. In addition, sevoflurane reduced the duration of a brief ischemic episode required to protect against infarction during IPC. 48Sevoflurane also enhanced cardioprotection when administered 24 h after an initial IPC stimulus. 51These important findings showed that administration of a volatile anesthetic combined with a brief ischemic event synergistically protects myocardium against subsequent damage as well.

Additional data about the effects of volatile agents on the coronary circulation also stand in contrast with the conclusions implicated by the coronary steal hypothesis. These results indicated that volatile agents are certainly not deleterious to but may instead exert beneficial actions on coronary collateral perfusion to ischemic myocardium. Volatile anesthetics have been shown to produce coronary vasodilation by activating K^ATPchannels 39,52–55or by favorably affecting intracellular Ca²⁺homeostasis in vascular smooth muscle. 56Halothane attenuated reductions in coronary collateral perfusion associated with acute coronary occlusion and also improved the myocardial oxygen supply-demand relation in collateral-dependent myocardium. 57In addition, halothane reduced cyclical changes in coronary blood flow and prevented the development of platelet thrombi in the presence of a critical coronary artery stenosis. 58Sevoflurane increased collateral blood flow to ischemic myocardium when perfusion pressure was maintained. 21,59Sevoflurane also improved the functional recovery of coronary vascular reactivity and nitric oxide release in isolated hearts after global ischemia. 39Lastly, volatile anesthetics attenuated neutrophil and platelet aggregation 60and also inhibited cytokine-induced cell death 50,61after ischemia-reperfusion injury in vitro . The results of these studies collectively show the protection against ischemia and reperfusion injury may be at least partially based on favorable effects of volatile agents on coronary perfusion.

The precise mechanisms responsible for volatile anesthetic-induced protection against ischemic injury remain unclear despite extensive study. Although it is clear that volatile anesthetics may indirectly improve myocardial oxygen supply-demand relations or enhance coronary collateral perfusion, it is equally clear that these actions are not entirely responsible for the antiischemic effects of these agents. This contention is emphasized by findings showing that volatile anesthetics conferred protection during cardioplegic arrest 25and during reperfusion, 26–30conditions in which myocardial oxygen supply-demand relations play little if any role. Similarly, isoflurane and sevoflurane increased the viability of isolated cardiac myocytes, 47and sevoflurane 62and desflurane 63improved contractility of isolated cardiac muscle exposed to simulated ischemia. These results were initially attributed to reductions in excessive intracellular Ca²⁺concentrations during ischemia and reperfusion 64produced by partial inhibition of Ca²⁺channel activity. 65–68However, this relatively generic Ca²⁺hypothesis did not address the precise mechanisms or provide deeper insight into the intracellular processes by which volatile anesthetics exert protective effects in the intact heart.

The signal transduction pathways involved in APC bear striking similarity to those responsible for IPC. It is hypothesized that volatile anesthetics stimulate a trigger that initiates a cascade of events leading to activation of an end-effector that is responsible for resistance to injury. To date, adenosine type 1 (A¹) receptors, 34,69,70protein kinase C (PKC), 34,71,72inhibitory guanine nucleotide binding (Gⁱ) proteins, 73ROS, 74–76and mitochondrial and sarcolemmal K^ATP(mito K^ATPand sarc K^ATP, respectively) channels 42,77–79have been shown to mediate APC (fig. 1). K^ATPchannels are heteromultimeric complexes containing an inward-rectifying potassium (K^ir) channel and a sulfonylurea receptor (SUR). 80Pharmacologic and recombinant techniques indicate that sarc K^ATPand mito K^ATPchannels 81,82are composed of the K^ir6.2/SUR2A and K^ir6.1/SUR1 isoforms, 83respectively. K^ATPchannel opening was initially implicated as the central end-effector during APC, 84similar to the findings during studies of the mechanisms responsible for IPC. 85,86Isoflurane and sevoflurane preserved myocardial viability in a cellular model of ischemia, and this protective effect was abolished by the selective mito K^ATPchannel antagonist 5-hydroxydecanoate (5-HD) but not the selective sarc K^ATPchannel antagonist HMR-1098. 47Isoflurane, 69sevoflurane, 62and desflurane 63but not halothane 69enhanced the recovery of contractile force of isolated human right atrial trabeculae after hypoxia and reoxygenation. The nonselective K^ATPchannel blocker glyburide (glibenclamide) or 5-HD inhibited this protective effect. HMR-1098 also attenuated the beneficial actions produced by sevoflurane in isolated human atria. 62Glyburide blocked the enhanced recovery of contractile function produced by isoflurane in stunned myocardium in vivo . 41,87Reductions in canine myocardial infarct size produced by isoflurane 42and the ATP-sparing effects of this agent 88have been shown to be blocked by glyburide as well. 5-HD also inhibited preconditioning by isoflurane in rats 44and rabbits. 77Both 5-HD and HMR-1098 abolished the protective effects of desflurane against ischemia and reperfusion injury in dogs, 78supporting a role for both mito K^ATPand sarc K^ATPchannels in APC. In contrast, another study showed that HMR-1098 did not modify desflurane-induced preconditioning in isolated human right atria in vitro . 63Therefore, some controversy continues to exist about the relative contribution of sarc K^ATPand mito K^ATPchannels in APC.

Carefully conducted in vitro  experiments suggest that volatile anesthetics are capable of modifying K^ATPchannel activity. Isoflurane stimulated outward K⁺current through sarc K^ATPchannels in isolated ventricular myocytes during patch clamping. 89,90Volatile anesthetics also reduced sarc K^ATPchannel sensitivity to inhibition by ATP, thereby increasing open state probability. 91In contrast, other patch clamp results suggested that volatile agents alone did not open K^ATPchannels. Isoflurane did not affect sarc K^ATPchannel current in human atrial cells 69and also inhibited sarc K^ATPchannel activity in rabbit ventricular myocytes. 91However, some volatile anesthetics were able to enhance sarc K^ATPchannel current by facilitating channel opening after initial activation. 89,90Isoflurane enhanced sarc K^ATPchannel opening in the presence of the mitochondrial uncoupler 2,4-dinitrophenol, the K^ATPchannel opener pinacidil, and the protein tyrosine kinase (PTK) inhibitor genistein in a whole cell patch clamp model. 89,92Activation of PKC, 90adenosine receptors, 93and phosphatidylinositol kinase 93seemed to be necessary for this process to occur. Isoflurane also directly opened sarc K^ATPchannels during intracellular acidosis, a condition that is known to occur during ischemia. 94These data suggest that volatile anesthetics may not directly interact with sarc K^ATPchannels but instead may affect other signaling elements that modulate sarc K^ATPchannel activity. In contrast with the findings with isoflurane, halothane had no effect on pinacidil-induced increases in sarc K^ATPchannel current and even inhibited K^ATPchannel current that had been maximally activated by 2,4-dinitrophenol. 89The anesthetic specificity for APC remains to be well characterized, although studies such as these do suggest important differences in efficacy may exist among individual agents.

The ability of volatile anesthetics to directly open mito K^ATPchannels has also been examined. Isoflurane and sevoflurane increased mitochondrial flavoprotein oxidation, an index of mito K^ATPchannel activity, in guinea pig cardiac myocytes. 95This process was inhibited by 5-HD. 95Flavoprotein fluorescence may not be entirely specific for mito K^ATPchannel opening, 96but isoflurane has also been shown to directly activate mito K^ATPchannels reconstituted in lipid bilayers. 97In contrast with these intriguing findings, 97Zaugg et al.  47recently demonstrated that although isoflurane or sevoflurane did not directly enhance flavoprotein oxidation in rat ventricular myocytes, these volatile agents did potentiate increases in fluorescence produced by the selective mito K^ATPchannel agonist diazoxide. These results suggested that volatile anesthetics may not directly open but instead act to prime mito K^ATPchannels, thus enhancing their ability to open in response to an agonist. Sarc K^ATPchannels may also be linked to the function of the mitochondrial inner membrane. For example, ROS generated by mitochondria may act to open sarc K^ATPchannels. 982,4-Dinitrophenol-induced activation of sarc K^ATPchannel current was reversible and accompanied by nicotinamide adenine dinucleotide oxidation, suggesting the existence of cross-talk between mito K^ATPand sarc K^ATPchannels. 99Taken as a whole, the preponderance of evidence collected to date implies that volatile anesthetics do not necessarily directly open K^ATPchannels but instead prime the activation of these channels in both sarcolemmal and mitochondrial membranes.

Adenosine triphosphate-sensitive potassium channels in vascular smooth muscle cells have been shown to be essential regulators of coronary vascular tone when ATP production is reduced. 100Volatile anesthetic-induced coronary vasodilation 39,52–55was attenuated by glyburide, indicating an important role for K^ATPchannels in this process. These data suggest that the beneficial actions of volatile agents during myocardial ischemia may be partially attributed to increased oxygen supply mediated via  K^ATPchannel-dependent coronary vasodilation. However, sevoflurane increased coronary collateral blood flow in the presence of glyburide in vivo , indicating that volatile anesthetics enhance collateral perfusion independent of K^ATPchannel activation. 59In fact, sevoflurane-induced increases in collateral perfusion were recently shown to occur as a result of Ca²⁺-regulated potassium and not K^ATPchannel activation. 101Based on these findings and results obtained in isolated cardiac myocytes where blood flow is not a factor, 47it seems highly unlikely that myocardial protection produced by volatile anesthetics is solely related to favorable alterations in coronary vascular tone mediated by K^ATPchannels.

Volatile anesthetics may activate parallel or redundant signaling pathways that involve K^ATPchannel opening to generate a physiologically meaningful cellular response. The sequential activation of several intracellular elements within a given transduction pathway may facilitate signal amplification and interaction between other redundant signaling systems. For example, administration of isoflurane in the presence of the K^ATPchannel openers nicorandil 102or diazoxide 103markedly enhances protection against ischemic injury beyond that observed with either drug alone. Several receptor-mediated events and intracellular signaling elements that converge on the K^ATPchannel have been implicated in APC. Pretreatment with pertussis toxin abolished any reduction in infarct size produced by isoflurane, indicating that Gⁱproteins are linked to the signal transduction pathways that mediate APC. 73In contrast, pertussis toxin did not alter the beneficial effects of direct K^ATPchannel opening produced by nicorandil. These data strongly support the contention that volatile anesthetics modulate K^ATPchannel activity through second messenger signaling.

Halothane-induced protection against infarction was completely abolished by blockade of the adenosine A¹receptor. 34The nonselective adenosine receptor antagonist 8-(p-sulfophenyl)-theophylline abolished isoflurane-induced preconditioning in rabbits. 79The selective A¹receptor antagonist 8-cyclopentyl-1,3-dipropylxanthine partially attenuated the beneficial effects of isoflurane in canine stunned myocardium. 70A role for adenosine receptors in APC was also identified in isolated human right atrial myocardium in vitro . 69Isoflurane eliminated increases in interstitial adenosine during repetitive periods of coronary artery occlusion and reperfusion using a microdialysis technique. 70These findings suggest that ATP preservation and a subsequent reduction of adenosine released into the interstitium occur during isoflurane anesthesia. 70In addition, the data imply that volatile agents may either directly activate A¹receptors or indirectly enhance A¹receptor sensitivity to diminished endogenous adenosine concentrations. 70These results were also similar to those observed during IPC 104or bimakalim-induced pharmacologic preconditioning. 105The preservation of cardiac myocyte viability during ischemia produced by volatile anesthetics was also sensitive to adenosine receptor and Gⁱprotein inhibition in rats. 47 

Stimulation of the δ¹-opioid receptor has been shown to produce a cardioprotective effect that is abolished by selective opioid antagonists 106–110or K^ATPchannel blockers. 107,108,110–113The acute and delayed phases of IPC are also mediated by activation of the δ¹-opioid receptor. 114Recent results indicated that the combined administration of isoflurane and selective δ¹-opioid receptor agonists TAN-67 or BW373U86 potentiated K^ATPchannel opening and enhanced protection against myocardial ischemia and reperfusion injury. 103Combined administration of isoflurane and morphine, a μ receptor agonist with δ¹receptor agonist properties, also reduced the extent of myocardial infarction to a greater degree than either drug alone. 44This beneficial effect was shown to be mediated by mito K^ATPchannels and opioid receptors. Interestingly, the nonselective opioid antagonist naloxone abolished isoflurane-induced preconditioning. 44These intriguing data suggest an important link between volatile anesthetics and the opioid family of G protein-coupled receptors. Another recent study also indicated that halothane competitively inhibited the ligand-binding site of G protein-coupled receptors. 115Adrenergic receptor blockade was shown to abolish desflurane-induced preconditioning in isolated human right atria 63but had no effect on the antiischemic actions of sevoflurane in isolated rat cardiac myocytes. 47Overall, APC seems to be associated with the activation of separate receptor-mediated pathways that are linked to Gⁱproteins.

Translocation and phosphorylation of multiple protein kinases are known to be involved in signal transduction pathways involved in protecting myocardium against cell death after ischemia and reperfusion. 116–118In particular, PKC is an essential component of the signaling pathways associated with preserving cellular viability. 119The diverse PKC isoform family is a large group of serine/threonine protein kinases that are distinguished by variable regulatory domains and cofactors and also display diverse tissue and species distributions. 120Activation of G protein-coupled receptors (e.g. , A¹, 121,122bradykinin, 123,124δ¹opioid 125,126) stimulate PKC during IPC. Volatile anesthetics have also been shown to stimulate PKC translocation and activity, 127possibly by interacting with the regulatory domain of the enzyme. 128Inhibition of PKC attenuated isoflurane-enhanced recovery of contractile function in canine stunned myocardium. 71The antiischemic actions of halothane were abolished by selective PKC antagonism in rabbits. 34The δ and ε isoforms of PKC translocated to mitochondria and sarcolemma, respectively, 10 min after discontinuation of isoflurane in isolated rat hearts. 129In contrast, isoflurane stimulated translocation of PKC-δ and -ε to sarcolemmal and mitochondrial membranes, respectively, in the in vivo  rat heart. 72These discrepancies may be attributed to differences in experimental model or time of tissue sampling. The microtubule depolymerizing drug, colchicine, prevented isoflurane-induced reductions in myocardial infarct size in rabbits, 130suggesting that an intact cytoskeleton is essential for translocation of these protein kinases. 72,129 

Recent findings strongly suggest that volatile anesthetic-induced PKC activation is required to open K^ATPchannels and produce myocardial protection. For example, the nonselective PKC antagonist chelerythrine abolished sevoflurane-induced increases in mito K^ATPchannel activity in rat ventricular myocytes and prevented protection against simulated ischemia. 47Patch clamp experiments showed that isoflurane did not facilitate K^ATPchannel opening in excised membrane patches but enhanced K^ATPchannel current in a whole cell configuration concomitant with PKC stimulation. 90These observations were supported by other studies showing that adenosine and PKC increased K^ATPchannel activity. 103,131–134Specific PKC consensus sites have been identified on K^ATPchannels, indicating a molecular basis for phosphorylation and activation of the channel by the enzyme. 135Mito K^ATPchannel opening also occurred after PKC activation during IPC in isolated rabbit hearts. 136In contrast, recent evidence indicates that 5-HD inhibited PKC translocation, 72suggesting that mito K^ATPchannel opening may be upstream of PKC activation. Therefore, a possible feed-forward system between PKC and K^ATPchannel activation may occur during APC.

Protein kinase C has been shown to stimulate PTK 137,138and mitogen-activated protein kinases (MAPKs), 139and volatile anesthetics may modulate several of these critical intracellular signaling proteins independent of direct receptor activation as well. Ischemic and pharmacologic preconditioning have been shown to be mediated by activation of PKC, 34,71,106,124,140PTK, 116,137,138,141,142and MAPK. 117,118,143–145A recent investigation showed that the PTK inhibitor lavendustin A and the Src-selective inhibitor PP1 abolished isoflurane-induced preconditioning in rats. 72The MAPK family plays an important role in signal transduction from the cell surface and the nucleus and has been strongly implicated in the initiation and progression of cell death (i.e. , apoptosis). 146–148The p38 MAPK subfamily mediated IPC of myocardium in dogs, and activation of p38 MAPK was also associated with phosphorylation and translocation of heat shock protein 27 in vivo . 144In contrast with the findings during IPC, recent data suggest that p38 MAPK may not play a role in APC in isolated rat hearts. 149Nonetheless, volatile anesthetics modulate activity of one or more intracellular kinases to produce APC, and it seems that activation of PKC is critical to cardioprotection.

Large quantities of ROS are released during reperfusion of ischemic myocardium that damage proteins responsible for intracellular homeostasis, depress contractile function, and produce membrane damage. 150–152Halothane, isoflurane, and enflurane have been shown to attenuate the toxic effects of ROS on left ventricular pressure development in isolated hearts. 153Isoflurane decreased hydroxyl radical generation in the ischemic rat heart, 154and halothane had a similar effect in dogs. 155The protective effects of sevoflurane were associated with reduced dityrosine formation, an indirect marker of ROS and reactive nitrogen species. 76These results support the hypothesis that volatile anesthetics reduce the release of deleterious quantities of ROS associated with coronary artery occlusion and reperfusion. Isoflurane also inhibited superoxide anion production by activated neutrophils, an action that occurred independent of K^ATPchannel opening. 156In addition, isoflurane and sevoflurane have been shown to abolish activated neutrophil-induced myocardial dysfunction. 157These effects were associated with reductions in superoxide anion production and neutrophil adherence to coronary vascular endothelium. Therefore, volatile anesthetics also seem to exert beneficial actions by inhibiting neutrophil-induced injury during reperfusion.

In contrast with data implicating a pathologic role of large amounts of ROS, other findings strongly suggest that a variety of preconditioning stimuli, including brief ischemia, direct mito K^ATPchannel openers, opioids, and volatile anesthetics, stimulate a small burst of ROS that initiate downstream signaling events and produce protection from subsequent ischemic injury. 158For example, pretreatment with low concentrations of ROS have been shown to mimic the beneficial actions of IPC. 159,160Free radical scavengers administered before or during brief ischemia markedly attenuated the protective effect of the preconditioning stimulus on infarct size. 161,162These findings indicate that IPC is mediated by small quantities of ROS released during the preconditioning stimulus. The beneficial actions of sevoflurane against ischemic damage were abolished by scavengers of superoxide anion and inhibition of nitric oxide synthase. 76These results suggest that superoxide anion may act to trigger APC and further indicated that nitric oxide may scavenge superoxide anion on reperfusion to reduce injury. ROS scavengers attenuated isoflurane-induced reductions in myocardial infarct size in rabbits 74,75and also inhibited the beneficial effects of direct mito K^ATPchannel activation. 163Isoflurane has been shown to directly increase superoxide anion formation in vivo  independent of an ischemic episode by use of the fluorescent probe dihydroethidium and laser confocal microscopy. 75These data 75indicated for the first time that volatile anesthetics were capable of producing small amounts of ROS that were correlated with a reduction in myocardial infarct size after prolonged ischemia. Taken as a whole, these reports provide compelling evidence that small quantities of ROS also play a critical role in APC.

Reactive oxygen species have been shown to act as regulatory mediators in many signaling processes that protect the cells against oxidative stress. 164ROS-induced activation of PKC 159and MAPK 165–167have been implicated in both ischemic and pharmacologic preconditioning. Hydrogen peroxide activated all three MAPK subtypes in neonatal rat ventricular myocytes, but stimulation of the p38 MAPK family and the consequent phosphorylation of heat shock protein 25/27 seemed to be of critical importance during cardioprotection. 165ROS have also been shown to activate Gαⁱand Gα^oproteins. 168,169Recent findings showed that sevoflurane-induced ROS generation was unaffected by PKC inhibition, 170but ROS scavengers inhibited isoflurane-induced PKC translocation. 72These findings provide indirect evidence linking ROS production by volatile anesthetics to subsequent activation of protein kinases implicated in the signal transduction responsible for APC.

A controversy continues to exist regarding the temporal relation between mito K^ATPchannel opening and ROS production during ischemic or pharmacologic preconditioning. 171Although sarc K^ATPchannel opening was initially assumed to be the end-effector of IPC, 81,85mito K^ATPchannel activation may instead trigger preconditioning by generating ROS. 163,172Mito K^ATPchannel opening produced by selective agonists generates ROS 163,172that seem to be essential for activation of MAPK 166and are also required for beneficial effects on myocardium. 173The mito K^ATPchannel agonist diazoxide caused oxidation of the ROS probe MitoTracker® orange (Molecular Probes, Eugene, OR) and enhanced cell viability after hypoxia and reoxygenation in vitro . 172These actions were attenuated by pretreatment with 5-HD or ROS scavengers. Therefore, the protective effects of mito K^ATPchannel agonists may occur as a consequence of triggering by ROS that subsequently reduces myocyte injury, including the release of large quantities of these reactive intermediates during reperfusion injury. Morphine increased cardiomyocyte viability and the fluorescence intensity of the hydrogen peroxide-sensitive probe, 2′7′-dichlorofluorescin. 111These actions were abolished by 5-HD pretreatment, suggesting that activation of mito K^ATPchannels by opioids results in ROS production. Conversely, other studies 174,175have indicated that ROS modulate mito K^ATPchannel activity 174to provide a beneficial effect, indirectly suggesting that mito K^ATPchannels may also function as an end-effector of preconditioning. For example, superoxide anion generated by xanthine oxidase activated mito K^ATPchannels from bovine ventricular myocardium reconstituted from lipid bilayers. 174Lebuffe et al.  175demonstrated that ROS may serve as a trigger by opening mito K^ATPchannels, which subsequently generates additional ROS and nitric oxide that are both required for preconditioning in isolated chick neonatal myocytes. Therefore, whether mito K^ATPchannel opening serves as a trigger or end-effector of ischemic or pharmacologic preconditioning remains unclear. 176Nevertheless, the apparently complimentary interaction between ROS and mito K^ATPchannels suggest the intriguing possibility that positive feedback loops may exist between these elements that contribute to myocardial protection.

It also remains unclear whether volatile anesthetic-induced mito K^ATPchannel opening precedes or follows ROS generation. Pretreatment with 5-HD or the ROS scavengers N-acetylcysteine or N-2-mercaptopropionyl glycine before administration of isoflurane abolished ROS generation in vivo . 177In contrast, administration of 5-HD after discontinuation of isoflurane but before prolonged ischemia only partially attenuated this effect. These data suggest that mito K^ATPchannel opening acted as a trigger of APC by generating ROS. Conversely, another recent investigation conducted in isolated guinea pig hearts showed that sevoflurane-induced generation of ROS was not inhibited by 5-HD before ischemia. 178Therefore, experimental findings remain equivocal in support of the hypothesis that mito K^ATPchannel opening is the major trigger of APC.

Reactive oxygen species derived from the mitochondrial respiratory chain have been shown to play important roles during IPC or pharmacologic preconditioning. 111,160,179,180The complex III inhibitor myxothiazol blocked hypoxia- 160or acetylcholine-induced 179ROS generation and abolished preconditioning in isolated chick cardiac myocytes. The precise source of volatile anesthetic-induced production of ROS has yet to be finally established, but volatile agents have been previously shown to inhibit electron transport chain complexes I and II of cardiac mitochondria. 181,182Interestingly, sevoflurane-induced complex I inhibition was attenuated by the superoxide dismutase mimetic Mn(III)tetrakis(4-benzoic acid)porphyrin chloride. 182These results suggested that ROS may inhibit mitochondrial respiration through a positive feedback mechanism to amplify the ROS signal for triggering APC. In contrast, a recent investigation showed that the complex III inhibitor myxothiazol, but not the complex I inhibitor diphenyleneiodonium, abolished isoflurane-induced reductions in myocardial infarct size and generation of ROS. 183These preliminary data indicated that mitochondrial electron transport chain complex III may be the source of ROS production induced by isoflurane during APC. Taken together, it is possible that volatile anesthetics may modulate multiple sites of the electron transport chain either directly or indirectly via  a ROS-mediated feedback mechanism. Despite these compelling results, other potential enzymatic sources of ROS (i.e. , nicotinamide adenine dinucleotide oxidase, cyclooxygenase, lipoxygenase, xanthine oxidase, nitric oxide synthase, cytochrome P450 184–188) may play a role in APC and have yet to be excluded from this process.

The precise identities of the specific ROS involved in APC have yet to be defined, and the signaling pathways that may be modulated by these ROS are also largely unknown. ROS activated PKC, restored contractility, and reduced infarct size in rabbit hearts. 159Superoxide anion also opened mito K^ATPchannels. 174Hydrogen peroxide stimulated PTK-dependent activation of phospholipase C in mouse embryonic fibroblasts, rendering these cells resistant to stress. 189Hydrogen peroxide has also been shown to directly activate Gⁱand G^oproteins 168,169and other protein kinases involved in reducing cellular injury. 166,190–192Hydrogen peroxide may also be converted to other more reactive species that modify cysteine residues of specific G proteins, resulting in their selective activation. 169Recent results showed that isoflurane administration produced ethidium fluorescence in rabbit myocardium. 75,177Dihydroethidium is oxidized by intracellular superoxide anion to produce ethidium that subsequently binds to DNA, further amplifying its fluorescence. 193These results strongly suggest that superoxide anion is the particular ROS involved in isoflurane-induced preconditioning. Sevoflurane also generated superoxide anion before ischemia and reperfusion in isolated hearts. 178Alternatively, different ROS may exert opposing actions on mito K^ATPchannel activity. Dismutation of superoxide anion leads to production of secondary ROS, including hydrogen peroxide, hydroxyl radical, and peroxynitrite, 194and these radicals may differentially alter channel activity. For example, superoxide anion and hydrogen peroxide enhanced but peroxynitrite decreased Ca²⁺-regulated potassium channel activity. 195Therefore, it remains possible that volatile anesthetics may also generate ROS other than superoxide that activate mito K^ATPchannels, or these agents may inhibit the formation of intermediates such as peroxynitrite that adversely affect mito K^ATPchannel function. Further research is needed to clarify this issue.

Opening of sarc K^ATPchannels was originally implicated in IPC and pharmacologic preconditioning by shortening the action potential duration, 81,196thereby reducing intracellular Ca²⁺overload during ischemia. 81However, subsequent studies conducted after the discovery of mito K^ATPchannels 82indicated that the antiischemic actions of K^ATPchannel activation occurred independent of action potential duration. 197–199Nevertheless, IPC did not occur in K^ir6.2-deficient mice, suggesting that the presence of the sarc K^ATPchannel was still required for myocardial protection to occur. 200Despite these latter data, the majority of findings accumulated indicate that preservation of mitochondrial bioenergetic function that occurs as a consequence of mito K^ATPchannel opening seems to be of critical importance for protection against ischemia. 201–204Selective pharmacologic openers of mito K^ATPchannels (e.g. , diazoxide) maintain mitochondrial Ca²⁺homeostasis and inhibit Ca²⁺overload within the organelle. 203,204Alteration of the mitochondrial oxidation-reduction balance by mito K^ATPchannel opening may also act to promote cellular protection. 204,205Membrane depolarization, matrix swelling, and uncoupling of ATP synthesis occur as a result of mito K^ATPchannel opening that may mediate cellular viability during IPC. 205Mito K^ATPchannel opening depolarizes the inner mitochondrial membrane and causes a transient swelling of the mitochondrial matrix, 206resulting from a shift in the ionic balance. 207These actions initially reduce ATP production 204but subsequently stimulate a compensatory increase in respiration that optimizes the efficacy of oxidative phosphorylation in part through energy-dependent matrix volume regulation. 208Therefore, the moderate disturbance of mitochondrial homeostasis caused by mito K^ATPchannel opening (fig. 2) may promote tolerance to subsequent ischemic damage by reducing Ca²⁺overload, 203,204preventing activation of necrotic or apoptotic pathways, 209,210or attenuating oxidative stress. 211 

Mitochondrial ATP synthesis has also been shown to be preserved after prolonged as compared with brief ischemia and reperfusion. 212This beneficial effect was abolished by 5-HD, suggesting that activation of mito K^ATPchannels improves mitochondrial energy production. 213Opening of mito K^ATPchannels has been hypothesized to preserve outer mitochondrial membrane permeability to ATP precursors (e.g. , adenosine, adenosine diphosphate) and cytochrome c. The structure of the intermembrane space may also be maintained as a consequence of mito K^ATPchannel activation despite generalized swelling of the mitochondrial matrix. 201Preservation of ATP substrates and mitochondrial structure may facilitate more efficient energy transfer between the mitochondria and the cytosol immediately after ischemia. Sevoflurane was recently shown to preserve ATP synthesis in isolated cardiac mitochondria obtained during early reperfusion after ischemia in vivo , and this beneficial effect was abolished by pretreatment with a ROS scavenger. 214Sevoflurane-induced preconditioning improved mitochondrial bioenergetics through mito K^ATPchannel activation in isolated guinea pig hearts as assessed using flavoprotein fluorescence. 215Therefore, it seems likely that mito K^ATPchannel opening by volatile anesthetics may be associated with preservation of mitochondrial function during reperfusion and, further, that this maintenance of mitochondrial performance contributes to cardioprotection.

Experiments conducted in isolated mitochondria have shown that a triggering quantity of ROS exceeding a critical threshold results in a transition in mitochondrial inner membrane permeability and the subsequent release of a burst of ROS in a process that has been termed ROS-induced ROS release . 216This mitochondrial permeability transition (MPT) has been shown to precede necrotic or apoptotic cell death, 217and glutathione is a primary defense against this event. 216,218These data suggest that volatile anesthetics or other mito K^ATPchannel agonists may prevent MPT in an oxidant-sensitive fashion, but this intriguing hypothesis has yet to be tested. Opening of an MPT pore using the agonist atractyloside during reperfusion was recently shown to abolish IPC and diazoxide-induced preconditioning in isolated rat hearts. 219These findings suggested that inhibition of MPT pore opening may represent a distal effector responsible for preconditioning, whereas mito K^ATPactivation functions as a trigger or a mediator. Most recently, rabbit hearts pretreated with desflurane before ischemia and reperfusion exhibited resistance to MPT pore opening. 220Future investigations are necessary to delineate the role of MPT during APC.

Cytosolic and mitochondrial Ca²⁺overload during prolonged ischemia and reperfusion have been shown to be associated with mitochondrial damage and myocardial cell death. 221–223Ischemic and sevoflurane-induced preconditioning were shown to reduce cytosolic Ca²⁺overload and improve the recovery of contractile function during reperfusion. 64Administration of sevoflurane after ischemia also reduced cytosolic Ca²⁺and myocardial damage. 31IPC and APC attenuated mitochondrial Ca²⁺overload during ischemia in rat and guinea pig hearts, 224,225actions that were abolished by 5-HD. These data suggested that volatile anesthetics protect against ischemia-reperfusion injury, at least in part, by attenuating cytosolic and mitochondrial Ca²⁺overload through a mito K^ATPchannel-dependent mechanism. Volatile anesthetics have also been shown to suppress sarcoplasmic reticulum Ca²⁺release 226,227and depress myofilament Ca²⁺sensitivity. 227Therefore, modulation of the sarcoplasmic reticulum to reduce cellular Ca²⁺overload and alterations of myofilament Ca²⁺sensitivity under conditions of excess Ca²⁺have been implicated in cardioprotection. 228,229The inhibitory actions of the volatile anesthetics on the voltage-dependent Ca²⁺channel are also well known, 65–68and reductions in cytosolic and mitochondrial Ca²⁺overload during ischemia and reperfusion injury may also occur through this mechanism.

Compelling experimental data in multiple animal models regarding the protective effects of volatile anesthetics remain to be translated into therapeutic approaches to reduce morbidity and mortality in patients with ischemic heart disease. However, evidence accumulated to date strongly suggests that APC occurs in human myocardium. Repetitive, brief (60- to 90-s) balloon inflations and deflations performed during percutaneous transluminal coronary angioplasty were associated with progressive reductions in the severity of chest pain and in the extent of ST segment elevation, decreases in myocardial lactate production, and declines in cardiac enzyme and troponin release. 230–233Adenosine 234or the mito K^ATPchannel opener nicorandil 235administered before the first balloon inflation during percutaneous transluminal coronary angioplasty was also shown to reduce the severity of ST changes during subsequent occlusions. Pretreatment with nicorandil before percutaneous transluminal coronary angioplasty also attenuated the release of troponin T, an indicator of myocyte necrosis. 236These findings suggest that both ischemic and pharmacologic preconditioning can be elicited during percutaneous transluminal coronary angioplasty. Consecutive exercise stress tests (separated by 15 min) performed in patients with critical left anterior descending coronary artery stenoses showed that anginal symptoms, ST segment depression, and myocardial oxygen consumption were reduced during the second as compared with the first exercise period for an equivalent amount of work. 237This “warm-up” phenomenon occurred independent of coronary vasodilation 237and also provides evidence of IPC in humans. 238,239 

Ischemic preconditioning has also been shown during cardiac surgery in patients with coronary artery disease. Yellon et al.  239used intermittent aortic cross clamping to produce IPC during coronary artery bypass graft surgery (CABG) and found enhanced preservation of ATP content in preconditioned hearts as compared with those that did not receive preconditioning stimuli. These investigators also showed that less troponin T was released in the presence as compared with the absence of IPC in this model. 240The incidence of ventricular tachyarrhythmias was reduced after cardiopulmonary bypass in patients undergoing CABG when cold blood cardioplegia was used for myocardial preservation. 241–243Regional IPC produced by brief coronary occlusion has also been shown to result in improved hemodynamic recovery and reduced release of cardiac troponin I during off-pump CABG. 244In contrast with the findings of these investigations, other studies have failed to show that IPC exerts beneficial effects during CABG in the presence of cardioplegia and cardiopulmonary bypass. 245–247Therefore, although the myocardial protective effects of IPC have been clearly identified in the experimental laboratory, further large scale clinical trials are needed to definitively demonstrate the beneficial actions of IPC in humans.

Documentation of APC in patients has been complicated by alterations in systemic and coronary hemodynamics; the use of other anesthetics, analgesics, or vasoactive drugs; preexisting disease states; and the acute influence of surgery on cardiovascular homeostasis. Nevertheless, isoflurane, 69desflurane, 63and sevoflurane 62enhanced the recovery of contractile function of human atrial trabeculae in vitro  by stimulation of adenosine receptors and opening of K^ATPchannels. Other studies have previously shown a role for adenosine receptors, MAPK, 248and ROS 172in other forms of preconditioning concomitant with opening of mito K^ATPchannels in human atrial myocytes. Isoflurane increased the tolerance to pacing-induced ischemia in patients with coronary artery disease. 249Isoflurane also decreased postoperative release of troponin I and creatine kinase-MB in patients undergoing CABG. 250Although the aforementioned results 250were not statistically significant, these data suggest that reduction in the extent of myocardial necrosis had occurred. Administration of isoflurane immediately before aortic cross clamping in patients undergoing CABG was shown to decrease the severity of subsequent ST segment changes and preserve cardiac index to a greater extent than that observed in patients who did not receive pretreatment with the volatile anesthetic. 251Administration of enflurane before cardioplegic arrest enhanced recovery of postischemic contractile function assessed using pressure-area relations in CABG patients. 252Sevoflurane 253–255and desflurane 253,255but not the intravenous anesthetic propofol was shown to preserve myocardial function in patients undergoing CABG as well as a reduction in troponin I release. Most recently, preconditioning with sevoflurane reduced a biochemical marker of myocardial dysfunction (i.e. , N-terminal pro-brain natriuretic peptide) in patients undergoing CABG concomitant with translocation of PKC-δ and -ε. 256This compelling evidence strongly suggests that volatile anesthetics exert beneficial effects against ischemic injury in humans.

In contrast with the aforementioned results, no differences in PKC and p38 MAPK activity or peak troponin I release were observed between patients undergoing cardioplegic arrest in the presence or absence of sevoflurane pretreatment. 257The activities of PKC, PTK, and p38 MAPK were increased equally in both groups, suggesting that cardiopulmonary bypass and cardioplegic arrest may produce a preconditioning-like effect that obscured the antiischemic actions of sevoflurane in this setting. However, sevoflurane was recently shown to reduce myocardial injury to a greater degree than propofol in patients undergoing off-pump CABG, 254a clinical setting that does not require cardiopulmonary bypass. Most investigations conducted in humans seem to indicate that volatile anesthetics represent an important therapeutic modality to reduce the sequelae of perioperative myocardial ischemia and infarction. 258,259A large-scale, randomized clinical trial is clearly needed to firmly establish this conclusion. Given the recent data suggesting that cardiopulmonary bypass and cardioplegia may exert a protective effect, 257such a clinical trial may best be conducted in patients undergoing off-pump CABG 254or those with documented coronary artery disease undergoing noncardiac surgery. 4Furthermore, no clinical study has shown that the use of volatile anesthetics in patients with coronary artery disease contributes to reduced cardiac morbidity or perioperative mortality. Additional multicenter trials also need to be conducted to identify the relative impact of APC on clinical outcome.

Experimental evidence collected indicates that volatile anesthetics exert important cardioprotective effects that reduce the consequences of reversible and irreversible ischemia and reperfusion injury. Differences in the efficacy of APC and timing of administration among volatile anesthetics alone and in combination with other cardioprotective drugs remain to be fully distinguished. Another important aspect is to determine the effect of aging on APC. 255,260Characterization of a late phase or second window of APC may be of special clinical significance to protect against ischemic events that frequently occur in the postoperative period. Several endogenous signaling elements seem to mediate APC, and mito K^ATPchannels, PKC, and ROS have emerged as central features in this process. Future investigations are needed to further delineate and identify essential components in these complex signal transduction cascades that mediate the early and late phases of APC. In this regard, microarray technology may prove useful in ascertaining candidate genes that are responsible for APC. 261Use of other fundamental molecular and biochemical tools is needed to determine whether preservation of mitochondrial integrity and metabolic homeostasis ultimately enhances tolerance to myocardial ischemia. Volatile anesthetics also act to elicit signaling that is probably present in many types of cells. Therefore, it is not surprising that these agents may reduce injury to other tissues. 256,262This may also be based on preventing cytokine-induced injury in endothelial and vascular smooth muscle cells. 50,61Finally, recent investigations have also strongly implied that APC occurs in humans and may represent an important therapeutic approach to reduce morbidity and mortality in patients with coronary artery disease. Nevertheless, further investigation is needed to firmly link this emerging body of clinical evidence to the already strongly established experimental data about the protective effects of volatile anesthetics in myocardium.

The authors thank David Schwabe, B.S.E.E., for technical assistance and Mary Lorence-Hanke, A.A. (Department of Anesthesiology, Medical College of Wisconsin, Milwaukee, Wisconsin), for assistance in preparation of the manuscript.