Activation of Adenosine Triphosphate-regulated Potassium Channels 

Potassium (K+) channels are membrane-spanning proteins that are selectively permeable to K+ and represent the largest and most diverse group of any ion channel family that has been identified (Table 1). The K^ATPchannel has a single-channel conductance of [nearly =] 25 pico-siemens at physiologic ion concentrations [9]and is one member of the inwardly rectifying K+ channel super-family (K^ir). The K^irchannel superfamily is one of at least six different superfamilies of K+ channels that have been defined based on their distinct amino acid sequence and structure and derived from molecular biologic isolation and cloning. K^irchannels are formed by tetrameric combinations of individual subunits, each of which is thought to have two transmembrane domains with intervening cytoplasmic segments, in between which lies the ion pore forming region. [10–12]Recent findings suggest that the K^ATPchannel is formed by fusion of four specific K^irsubunits, denoted as K^ir6.2, and four sulfonylurea receptors (SURs) to form the active K^ATPchannel. [11,13–15]At least two different types of SURs have been identified, and evidence suggests that SUR^2acombines with K^ir6.2 to form the channel in the heart. SUR¹, with K^ir6.2, forms the channel in beta cells, whereas SUR^2bplus K^ir6.2 forms the channel in smooth muscle. [11,16]The presence of different SUR types appears to confer differential ATP sensitivity and pharmacologic properties on K sub ATP channels in different tissues. [11,16] 

K^ATPchannels were originally discovered in cardiac myocytes by Noma [3]and have also been identified in vascular smooth muscle, [17–22]skeletal muscle, [23]pancreatic beta cells, [24,25]neurons, [4,5]renal epithelium, [26]and tracheal [27]and urinary tract smooth muscle. [28]K^ATPchannels have received considerable attention recently because they appear to regulate cellular responses to hypoxia and ischemia. Ion flow through K^ATPchannels occurs passively down an electrochemical gradient. Opening of the K^ATPchannel hyperpolarizes the cell or stabilizes the resting membrane potential. [29]Unlike other types of K+ channels, the open state of K^ATPchannels is modulated by intracellular nucleotide concentrations. [30]Thus, the K^ATPchannel may link cellular metabolism and membrane excitability. [31] 

Regulation of K^ATPchannels is complex and incompletely understood. Experimental evidence has led to a proposed model of K^ATPchannel function. [32,33]This model consists of a gated K^ATPchannel, an ATP regulatory site, a phosphorylation subunit, and nucleotide diphosphate binding sites (Figure 2). K^ATPchannel activity is also modulated by inhibitory G protein subunits, including G sub alpha o, G^alphai-1, and G^alphai-2 and G^alphai-2 by the SUR. [11,12,34–36]The defining attribute of K^ATPchannels is the direct inhibition of channel permeability to K+ by the intracellular ATP concentration (ATPⁱ). [32,37,38]As intracellular energy stores decrease in cardiac and smooth muscle cells during hypoxia or ischemia, the K^ATPchannel opens and produces a cascade of events, including membrane hyperpolarization, decreases in Ca sup 2+ influx, and subsequent relaxation of vascular smooth muscle or shortening of the cardiac action potential duration. [39–41]Because the initial level of ATP (micromolar concentrations) that completely inhibits K^ATPchannels in excised membrane patches from the heart is much lower than that occurring during ischemia in vivo (millimolar concentrations), it has been proposed that ATPⁱ, which regulates the channel, is functionally compartmentalized. [42,43]In the heart, glycolytically generated ATP is much more effective than oxidative phosphorylation in inhibiting K^ATPchannels. [43]In addition, depletion of subsarcolemmal ATP enhances K^ATPchannel activation in ventricular myocytes, [42]and intracellular metabolites generated during ischemia enhance activation of K^ATPchannels. For example, increases in adenosine concentration enhance K^ATPchannel activation even at moderately high ATPⁱvia activation of type receptors. [44]Although higher concentrations of ATP inhibit K^ATPchannels, a low concentration of ATP ([nearly =] 1–5 micro Meter) appears to be required to maintain the channel in its functional, phosphorylated state. K^ATPchannel activity depends on the presence of magnesium ATP. This intracellular nucleotide is presumed to modulate channel action through a kinase-dependent channel phosphorylation. [32,45]Nucleotide diphosphates, including adenosine diphosphate, play important roles in the determination of K^ATPchannel activity by antagonizing ATPⁱ-induced inhibition of channel opening. Nucleotide diphosphates may also directly stimulate the opening of K^ATPchannels by binding to a different subunit that is not responsible for ATPⁱinhibition but requires occupation of a second phosphorylation site. [32,46,47]Ligand-receptor-G protein interactions also regulate K^ATPchannels. Adenosine and acetylcholine act through a membrane-delimited pathway to activate K^ATPchannels via stimulation of inhibitory G (Gⁱ) proteins. [34,48,49]Experimental evidence suggests that G proteins activate these channels by antagonizing the inhibitory effect of ATPⁱon channel gating. [34,49]Protein kinase C has also been demonstrated to affect K^ATPchannel activity by changing the stoichiometry of ATP binding. [50]Thus, protein kinase C activation causes stimulation of K^ATPchannels, increasing the ATPⁱconcentration required for inhibition. [32] 

The activity of the K^ATPchannel responds to a diverse group of pharmacologic agents that act as either agonists or antagonists (Table 2). K^ATPchannel agonists, including levcromakalim, cromakalim, aprikalim, bimakalim, pinacidil, and nicorandil, have been advocated for use in the management of hypertension, myocardial ischemia, and congestive heart failure. [51,52]The exact mechanism of K^ATPchannel activation by these drugs has yet to be clearly defined; however, experimental evidence suggests that K^ATPchannel agonists block the ATPⁱbinding site, enhance the actions of nucleotide diphosphates to activate K^ATPchannels, or directly stimulate channel opening independent of ATPⁱregulation. [30,32,33]Sulfonylureas have been shown to act as antagonists of K^ATPchannels and have been used in a wide variety of studies as specific K^ATPchannel blockers. [2,24,25,36,53]These drugs, including tolbutamide, glyburide (glibenclamide), and glipizide, have been used for decades to treat patients with type II diabetes mellitus. K^ATPchannels set the resting membrane potential in pancreatic beta cells. The channels are normally open; they close after glucose metabolism and generation of ATP. [54,55]K^ATPchannel closure causes cell membrane depolarization and enhanced Ca²+ influx through voltage-regulated Ca²+ channels. These actions initiate insulin release from beta cells by exocytosis. [24,25,32,54,55]Inhibition of K^ATPchannels by sulfonylureas also initiates this sequence of events and increases insulin release. Recent evidence demonstrates the important link between molecular characterization of K^ATPchannels and their function in vivo. Persistent hyperinsulinemic hypoglycemia of infancy is characterized by the loss of K^ATPchannel function in pancreatic beta cells, [56]and specific mutations of the SUR may account for the observed dysregulation of K^ATPchannels and persistent insulin secretion observed in this disorder. [57] 

Since the discovery of K^ATPchannels, [3]numerous investigations have supported and extended Noma's original hypothesis that these channels are endogenous mediators of myocardial protection. Ischemic preconditioning was initially described in dogs by Murry et al. in 1986. [58]These investigators demonstrated that four 5-min periods of coronary artery occlusion interspersed with reperfusion, before a prolonged coronary artery occlusion (40 min) and reperfusion (72 h), markedly decreased (70–80%) the size of the resulting myocardial infarct. [58]Subsequently, a single 5-min period of coronary artery occlusion and reperfusion was shown to be a sufficient preconditioning stimulus to reduce myocardial infarct size, and the cardioprotective effect of this single brief occlusion was equivalent to that of multiple (6–12) preconditioning episodes. [59]Preconditioning with a brief period of ischemia has consistently been found to decrease myocyte death in every species and experimental model in which it has been evaluated. [59–64] 

Intense investigation in recent years has attempted to clarify the mechanism(s) responsible for ischemic preconditioning. Ischemic preconditioning has been mimicked by the administration of adenosine [60,65,66]and acetylcholine, [41,67–69]and by pharmacologic stimulation of protein kinase C or other protein kinase C-coupled receptors (e.g., alpha-adrenergic, angiotensin II, bradykinin B²). [70–74]Several of these endogenous mediators are known to activate guanine nucleotide regulatory proteins and are coupled to K^ATPchannels. [48,73,75]The selective K^ATPagonists aprikalim and bimakalim also mimic the effects of ischemic preconditioning [76–80]to reduce myocardial infarct size markedly (Figure 3). In contrast, glyburide and the structurally unrelated K^ATPantagonist sodium 5-hydroxydecanoate (5-HD)[81]abolished the cardioprotective effects of ischemic preconditioning in dogs (Figure 4)[76,81]and pigs. [77,82]Neither glyburide nor 5-HD, however, affected myocardial infarct size in the absence of a preconditioning stimulus.

The relative importance of the K^ATPchannel in ischemic preconditioning in rabbit and rat models of myocardial injury is more controversial. Glyburide did not prevent reductions in myocardial infarct size due to ischemic preconditioning and actually increased infarct size when administered alone in pentobarbital-anesthetized rabbits. [83]Glyburide [84]and 5-HD [85]did not attenuate the protective effects of ischemic preconditioning on recovery of contractile function after global ischemia in isolated rat hearts. In contrast, glyburide blocked the protective effects of ischemic preconditioning in rats [69]and rabbits anesthetized with ketamine/xylazine. [86,87]These data suggest that glyburide may not block ischemic preconditioning in some experimental preparations because pentobarbital was used as the primary anesthetic agent. [83]Barbiturates in high concentrations block K^ATPchannels. [88]Other experiments, however, have shown that glyburide completely inhibits preconditioning in dogs [76]and rats [69]anesthetized with pentobarbital. Glyburide inhibits ischemic preconditioning in a time-dependent fashion in the rat. [89]Glyburide abolished preconditioning when the drug was administered at 30 but not 5 min [83]before initiation of brief periods of coronary artery occlusion and reperfusion. [89]In addition, a high glyburide dose was required to block the protective effects of preconditioning on sarcolemmal enzyme activity (measured as a correlate of myocardial ischemic injury) in pentobarbital- compared with ketamine/xylazine-anesthetized rabbits. [90]Therefore, differences in the time interval between the administration of glyburide and the preconditioning stimulus or dose of glyburide used may explain the inability of this K^ATPchannel antagonist to inhibit ischemic preconditioning in some rat and rabbit models.

Reduction of myocardial infarct size is the classic end point used to evaluate the efficacy of ischemic preconditioning. Other secondary end points, such as enhanced contractile function after ischemia and reperfusion, also have been evaluated, however. Although these different end points may represent a continuum of myocardial injury, mechanisms affording cardioprotection during infarction versus stunning may or may not be similar. The interpretation of studies using glyburide as a pharmacologic antagonist of K^ATPchannels is also complicated by findings in guinea pig myocytes that concentrations of glyburide in excess of 1 micro Meter may nonspecifically block other K+ channels. [91]In contrast, 10 micro Meter glyburide specifically blocked K^ATPchannel currents elicited in cat atrial myocytes while having no effect on baseline acetylcholine-induced K+ currents. [92]Other K+ channel currents, such as the transient outward current and delayed rectifier K+ currents, do not participate in the protection afforded by ischemic preconditioning. [93,94]Administration of dofetilide, a blocker of the fast component of the delayed rectifier, and 4-aminopyridine, a blocker of transient outward current, did not abolish the cardioprotective effects of ischemic preconditioning or cromakalim. [93,94]Although glyburide may block K^irchannels nonspecifically when administered at high concentrations, most evidence suggests that K^ATPchannels, specifically, play a role in mediating ischemic preconditioning. Further, although the SUR in pancreatic beta cells has a higher affinity for glyburide than does the receptor in cardiac myocytes, [11]extracellular acidification, as occurs during ischemia, markedly enhances sensitivity of K^ATPchannels to inhibition by even low concentrations of glyburide. [95]Interpretation of studies using some K^ATPchannel agonists may also be complicated by the presence of other non-K^ATPchannel-mediated pharmacologic effects. For example, nicorandil possesses nitrate-like properties in addition to actions as a K^ATPchannel agonist; however, the effect of nicorandil in reducing infarct size in dogs is independent of its nitrate-like properties. [96] 

Ischemic preconditioning has been suggested to occur in patients with coronary artery disease. [64,97–101]For obvious reasons, classic ischemic preconditioning to reduce myocardial infarct size cannot be demonstrated experimentally in patients. The presence of prodromal angina, however, confers important beneficial effects to reduce subsequent myocardial infarct size, the incidence of congestive heart failure, and mortality. [97–99]These clinical observations have been attributed to ischemic preconditioning. Ischemic preconditioning produced by brief periods of aortic cross clamping before a prolonged episode of cross clamp-induced ventricular fibrillation improves myocardial high-energy phosphate content during coronary artery bypass graft surgery. [102]The severity of myocardial ischemia evaluated by ST segment changes during percutaneous transluminal coronary angioplasty in response to a second balloon inflation was less than that occurring during the first balloon inflation, [100,101]suggesting that the myocardium was preconditioned by the first brief occlusion of the affected vessel. Reduction of ST segment deviation and anginal symptoms during percutaneous transluminal coronary angioplasty were abolished by pretreatment with a single oral dose of glyburide, [64]providing the first evidence to suggest that ischemic preconditioning is linked to K sub ATP channel activation in humans. These findings, although compelling, must be interpreted with caution because most clinical studies did not control for differences in collateral blood flow, used shorter ischemic periods than those used in animal studies, and evaluated end points other than myocardial infarct size. Nonetheless, ischemic preconditioning has subsequently been simulated by administration of cromakalim and inhibited by 1 micro Meter glyburide in isolated human atrial myocytes. [63]In addition, atrial myocytes isolated from patients with long-term exposure to sulfonylurea hypoglycemic agents were not protected by preconditioning, in contrast to the protection afforded by preconditioning in myocardium isolated from control or insulin-treated patients. [103]These findings strengthen the contention that ischemic preconditioning represents an important endogenous means to reduce myocardial injury in humans and that this action is mediated by the opening of K^ATPchannels. Recent clinical experience with nicorandil also indicates that administration of a K^ATPchannel agonist produces salutary effects in patients with coronary artery disease. Nicorandil reduced the extent of myocardial ischemia during angioplasty, an action attributed to K^ATPchannel activation, because nicorandil caused no changes in systemic or coronary hemodynamics. [104]Nicorandil was effective in the treatment of chronic stable angina [52,105]and may be effective in the treatment of unstable angina unresponsive to oral nitrates or calcium channel blockers. [106]Thus, early clinical experience with nicorandil for the treatment of coronary artery disease indicates that this agent is both efficacious and safe. [107] 

K^ATPchannel activation has been proposed to be the end effector of the cardioprotective signal transduction cascade that occurs during and after ischemic preconditioning. Adenosine, an endogenous purine, mimics ischemic preconditioning via activation of A¹receptors [60,65,108]and is coupled to K^ATPchannels through G sub i proteins in ventricular myocytes. [48]The important link between A¹receptor activation and K^ATPchannel stimulation has also been established in vivo. [65,66,108–110]The actions of the selective A¹agonist, R-PIA, [108,110]or adenosine [110,111]to decrease myocardial infarct size and simulate ischemic preconditioning were antagonized by pretreatment with glyburide or 5-HD in dogs, [66,108]rabbits, [111]and pigs. [110]Allosteric enhancement of adenosine binding to the A¹receptor by PD 81,723 lowered the threshold for ischemic preconditioning, [112]and this effect was blocked by glyburide. [112]Recent evidence demonstrates that activation of type 3 adenosine receptors (A³) also mimics ischemic preconditioning [113]; however, whether A³receptors are coupled to K^ATPchannels is unknown. Acetylcholine mimics ischemic preconditioning by activating type 2 muscarinic receptors (M sub 2) in dogs [67,68]and rats, [69]and these actions were attenuated by glyburide [68,69]or 5-HD. [67]K^ATPchannels also mediated the beneficial effects of simulated preconditioning with hypoxia to decrease ventricular myocyte death, [114]providing further evidence in vitro that K^ATPchannels are critical for cardioprotection during periods of reduced oxygen supply.

Although the involvement of protein kinase C in ischemic preconditioning is controversial, [115]K^ATPchannel activation has been linked to increases in protein kinase C activity in vitro. [50,63,116,117]ATP-dependent K+ current (I^K,ATP) was measured in response to phorbol esters using patch clamp techniques in isolated myocytes. Protein kinase C activation elicited I^K,ATP [116]and shortened the latency to elicit I^K,ATP. [117]Increases in I^K,ATP depended on protein kinase C-induced reductions in channel sensitivity to blockade by ATPⁱ[116]and were synergistically enhanced by concomitant administration of adenosine. [117]A recent study confirmed these results and suggested that protein kinase C-catalyzed phosphorylation of K^ATPchannels enhances their activation despite the presence of millimolar concentrations of ATPⁱthat would normally inhibit K^ATPchannel activity. [50]In an isolated human atrial myocyte model of ischemic preconditioning, [63]increased protein kinase C activity simulated the protective effects of preconditioning to enhance recovery of contractile function, and these effects were blocked by glyburide. [63] 

G proteins play a central role in ischemic preconditioning. [48,75,118]Pertussis toxin, the putative blocker of Gⁱ/G^oproteins, attenuated the protective effects of ischemic preconditioning in rabbits [118]and rats. [73]The infarct-reducing actions of the M²receptor agonist, carbachol, [118]were also blocked by pertussis toxin. Gⁱproteins have previously been shown to couple A sub 1 and M²receptors to K^ATPchannels. [34,35,48]Using patch clamp techniques, Gⁱprotein activation increased K^ATPchannel opening, which was further increased by acetylcholine and adenosine. [34,35,49]Activated Gⁱproteins, adenosine, and acetylcholine [92]also increased I^K,ATP in the presence but not in the absence of ATP. [34]These results indicate that Gⁱproteins activate K^ATPchannels by antagonizing ATP-dependent gating of the channel. [34,35]Thus, experimental findings demonstrate that ischemic preconditioning triggers a cascade of events that protect the myocardium by decreasing the sensitivity of K^ATPchannels to ATPⁱand promoting K^ATPchannel opening.

The mechanism of K^ATPchannel-mediated cardioprotection is incompletely understood. Unlike pancreatic beta cells, myocardial K sub ATP channels are normally closed and do not contribute to myocyte repolarization. As tissue ATP content and pH decrease, however, and as adenosine diphosphate, lactate, and adenosine concentrations increase during ischemia, K^ATPchannels open [38,119]and cause a rapid decrease in action potential duration. [41,80,120]This decrease in action potential duration may protect ischemic myocardium by inhibiting Ca²+ influx via the voltage-regulated Ca²+ channel and by maintaining operation of the Na sup +-Ca²+ exchanger in the forward mode. [120]These effects may reduce intracellular Ca²+ accumulation, preserve energy stores, [121]and delay ischemic injury. It has been suggested that because the density of K^ATPchannels in myocardium is high and because intracellular ATP may be nonuniform, only small decreases in ATPⁱmay be required to cause substantial changes in action potential duration. [9,122]In contrast, Yao and Gross [80]and Grover et al. [93,123]have shown a dissociation between action potential shortening and the cardioprotective effect of bimakalim, a K^ATPopener, or ischemic preconditioning in the canine heart. These findings suggest that other unknown cellular mechanisms may be responsible for the cardioprotection observed. A potential target may be the recently discovered mitochondrial K^ATPchannel. [124,125]The precise mechanism through which mitochondrial K^ATPchannels mediate cardioprotection is unknown. Opening of mitochondrial K^ATPchannels causes transient K+ uptake and matrix swelling, effects that modulate metabolic processes and cellular signaling. [124–126]Nonuniform increases in intracellular Ca²+ concentration, known as Ca²+ waves, associated with myocardial ischemia also may contribute to cellular injury and recently have been shown to be prevented by aprikalim. [127]This beneficial effect of K^ATPchannel activation in cardiomyocytes was abolished by K^ATPchannel antagonism. [127] 

An acute memory phase is associated with preconditioning. The heart remains resistant to infarction if the preconditioning stimulus precedes the prolonged period of ischemia by 30 min to 2 h. [128,129]Recent evidence suggests that, although A¹receptor stimulation may trigger ischemic preconditioning, occupation of A¹receptors alone is insufficient to sustain this acute memory phase of preconditioning. [94]Glyburide, but not the A¹receptor antagonist 8-cyclopentyl-1,3,dipropylxanthine (DPCPX), abolished ischemic preconditioning when administered 1 h after brief ischemic stimuli. These results indicated that K^ATPchannel activation was required for maintenance of the acute memory phase of ischemic preconditioning. [94]It is noteworthy that administration of either adenosine or bimakalim alone 60 min before prolonged coronary artery occlusion was not cardioprotective. Simultaneous administration of these drugs, however, produced a synergistic decrease in myocardial infarct size after a 60-min drug-free interval, extending the time window of memory that was otherwise limited with either drug alone. These results demonstrated that pharmacologic activation of K^ATPchannels and A sub 1 receptors closely simulates ischemic preconditioning and are probably both required for the associated acute memory phase.

Ischemic preconditioning is also characterized by a late memory phase or second window of but considerably less is known about the mechanism or precise temporal relation of the second window of preconditioning. Kuzuya et al. [130]demonstrated that dogs preconditioned with multiple brief coronary artery occlusions and reperfusions sustained smaller myocardial infarcts 24 h, but not 3 or 12 h, after the preconditioning stimuli. Reductions in myocardial infarct size [131]and resistance to postischemic contractile dysfunction [132]during late preconditioning were also associated with expression of heat shock proteins. The role of K^ATPchannel activation or changes in K^ATPchannel expression in the mechanism of this second window of cardioprotection have been incompletely evaluated. The nontoxic endotoxin derivative, monophosphoryl lipid A, produces a second window of cardioprotection, reducing myocardial infarct size when administered 24 h before a 60-min coronary artery occlusion and reperfusion. The cardioprotective effect of monophosphoryl lipid A depended on K^ATPchannel activation and was abolished by glyburide or 5-HD, results that provided the first evidence linking K^ATPchannels with late preconditioning. [133] 

Activation of K^ATPchannels attenuates reversible and irreversible myocardial injury. Brief coronary artery occlusion (< 20 min) followed by reperfusion produces prolonged contractile dysfunction in the absence of tissue necrosis. This condition is referred to as myocardial stunning. [134,135]K^ATPchannel activation has been convincingly demonstrated to reduce stunning and enhance recovery of postischemic, reperfused myocardium during experiments using single (15 min)[136–141]or multiple periods of coronary artery occlusion and reperfusion. [41]Intravenous administration of several different K sub ATP agonists markedly enhanced functional recovery of stunned myocardium, actions that were blocked by glyburide [41,137,139–141](Figure 5). Adenosine also plays a key cardioprotective role during myocardial stunning. Adenosine reduced myocardial contractile dysfunction after a 15-min coronary artery occlusion and reperfusion [142–144]or after multiple brief coronary artery occlusions and reperfusions. [109]Functional recovery of stunned myocardium improved after administration of exogenous adenosine, by increases in endogenous adenosine concentration attained through blockade of adenosine deaminase [143,144]and after inhibition of nucleotide transport. [143]It is noteworthy that the enhanced functional recovery of postischemic, reperfused myocardium produced by a selective A¹receptor agonist was blocked by glyburide, [109]confirming an important interaction between A¹receptors and K^ATPchannels. Therefore, substantial experimental evidence indicates that K sub ATP channel activation represents the end effector in a cardioprotective signal transduction pathway modulated by A¹receptors, Gⁱproteins, and protein kinase C.

K^ATPchannel activation clearly limits myocardial necrosis after ischemia; however, modulation of K^ATPchannels by pharmacologic agonists and antagonists also causes cardiac electrophysiologic effects. K^ATPagonists have been shown to have proarrhythmic [145,146]and antiarrhythmic activity. [147,148]Similarly, glyburide has both proarrhythmic [140,149,150]and antiarrhythmic [145,146,149]effects. Action potential shortening in response to K^ATPagonists may reduce Ca²+ influx and prevent triggered activity, [147,148]or conversely, these agents may exacerbate reentry. Glyburide prevents action potential shortening and may thus suppress reentrant arrhythmias but may exacerbate triggered activity. [149,151]In addition to the direct effects of K^ATPchannel modulation on cardiac electrophysiology, changes in myocardial infarct size also indirectly influence the occurrence of arrhythmias after myocardial ischemia and reperfusion. Thus, pharmacologic modulation of K^ATPchannel activity may promote either beneficial or detrimental effects on arrhythmogenesis, mechanisms that require further investigation in patients with cardiovascular disease.

Short-term cerebral ischemia initiates a series of events that ultimately produces neuronal injury and death. An increase in the extracellular concentration of K+ secondary to ischemia causes neuronal depolarization [152]and contributes to intracellular Ca su 2+ overload. [153,154]Depolarization of ischemic neurons stimulates excessive glutamate release and decreases reuptake of this excitatory neurotransmitter. [155]This glutamate cascade contributes to and sustains neuronal injury by stimulation of N-methyl-D-aspartate receptors and the associated N-methyl-D-aspartate channel, effects that increase Ca²+ influx and cell death. [154–156] 

New therapeutic strategies designed to reduce neuronal injury during cerebral ischemia are the subject of intense investigation. The role of K^ATPchannels as potential modulators of neuronal injury during ischemia has only been examined recently. Several techniques, including in situ hybridization using specific radiolabeled antisense oligonucleotides for K^ir6.2 and SUR¹, [157]autoradiographic identification of sulfonylurea binding sites, [4]and identification of K^ATPchannel currents in hippocampal neurons, [158]have demonstrated the presence of K^ATPchannels throughout the brain. [4,158]K^ATPchannels are located in pre- and postsynaptic neurons, [5,159,160]and are particularly concentrated in mossy fibers that are associated with glutamate release. [5]It has been hypothesized that activation of K^ATPchannels during cerebral ischemia protects by hyperpolarizing presynaptic neurons and reducing Ca²+ influx and glutamate release. [159–163]Hyperpolarization of postsynaptic neurons may also occur as a result of K^ATPchannel activation, rendering these neurons less receptive to depolarizing stimuli. Neuronal hyperpolarization may enhance voltage-dependent binding of magnesium to N-methyl-D-aspartate receptor-coupled channels and reduce the risk of intracellular Ca²+ overload known to be associated with neuronal death. [155,156,163,164] 

Evidence for K^ATPchannel-mediated neuroprotection has recently been obtained in rats during transient fore-brain ischemia. [7,165]This model is produced by a 20-min occlusion of all four major extracranial arteries, which causes neuronal death in 78% of CA1 neurons 7 days after the ischemic insult [165]and induces expression of the immediate early genes c-fos and c-jun. Intraventricular administration of the K^ATPchannel agonists cromakalim, nicorandil, or pinacidil before the ischemic event markedly reduced neuronal death and abolished ischemia-induced changes in gene expression. The beneficial effects of the K^ATPchannel agonists were blocked by glipizide. [165]These findings were confirmed and extended in an investigation that examined the role of K^ATPchannel and A¹receptor activation in cerebral ischemic preconditioning. [7]Heurteaux et al. [7]demonstrated that a 3-min ischemic episode protected against neuronal death during a subsequent 6-min ischemic period when the interval between the first and second periods was 3 days but not 1 h. The beneficial effects of this “delayed” cerebral ischemic preconditioning were mimicked by levcromakalim and blocked by glyburide. [7]These results are similar to those observed in ischemic myocardium, with the notable exception of the time course, and suggest that an important link between A¹receptors and K^ATPchannels results in protection of neural tissue against ischemic injury. Protection of CA1 pyramidal cells against death after ischemia was also demonstrated after administration of the A¹receptor agonist cyclopentyladenosine, and this protection was abolished by glyburide. [7]Thus, preliminary experimental evidence suggests that activation of K^ATPchannels plays an important endogenous role in reducing neurologic injury during ischemia and can be modified by pharmacologic agents.

Skeletal muscle, although more resistant to ischemia than cardiac muscle, may infarct when subjected to a prolonged ischemic insult. The efficacy of ischemic preconditioning in reducing myocyte injury was examined in pig skeletal muscle after 4 h of global ischemia and 48 h of reperfusion. [166]Ischemic preconditioning reduced the extent of infarction and preserved high-energy phosphate content of postischemic, reperfused skeletal muscle. [166]Recently, myocyte/endothelial cell adhesive interactions have been shown to play an important role in mediating tissue injury after ischemia and reperfusion, and this process may be modulated by adenosine and K^ATPchannels. [78,136,167,168]Ischemic preconditioning markedly attenuated the increase in leukocyte adhesion and emigration in postischemic, reperfused murine cremaster muscle, an action that was mimicked by continuous superfusion of adenosine during ischemia and reperfusion. [167]Importantly, the K^ATPagonist pinacidil attenuated ischemia-induced leukocyte adhesion and emigration consistent with previous evidence suggesting that leukocyte function was modulated by K^ATPchannel activation. [78,136]Adhesive interactions between circulating neutrophils and the endothelium may contribute to tissue injury after ischemia and reperfusion by reducing capillary perfusion (the “no reflow” phenomenon). [168]Ischemic preconditioning improved microvascular patency after 4 h of ischemia and 30 min of reperfusion in canine gracilus muscle, and a role for K^ATPchannels in mediating this protective effect was demonstrated by findings that glyburide abolished improvements in microvascular patency associated with ischemic preconditioning. [168]Further, pinacidil attenuated capillary no reflow, and this action was also reversed by glyburide. [168]Thus, K^ATPchannel opening produces beneficial effects in skeletal muscle (and possibly other tissues) during ischemia and reperfusion, which may be mediated, at least in part, by attenuating leukocyte/endothelial cell adhesive interactions. Whether these effects are specifically mediated by leukocyte K^ATPchannels or changes in reperfusion blood flow or via activation of skeletal muscle K^ATPchannels with subsequent reduced ischemic intensity is unknown.

Considerable investigation has been directed toward determining the role of K^ATPchannels in the regulation of blood flow in the coronary and cerebral circulations. The presence of K^ATPchannels in vascular smooth muscle was first identified during experiments in which cromakalim and aprikalim produced vasodilation concomitant with an increase in K+ efflux from and hyperpolarization of smooth muscle cells. These actions were inhibited by glyburide. [20]Activation of K^ATPchannels in coronary vascular smooth muscle cells during hypoxia or ischemia results in a marked vasodilator response. [19]Autoregulatory responses to changes in coronary perfusion pressure have been linked to K^ATPchannel activation in experiments in which the left anterior descending coronary artery was cannulated and perfusion pressure reduced in a stepwise fashion. [169]Glyburide abolished autoregulatory vasodilation in this preparation and blocked microvascular responses to graded coronary artery stenosis and complete occlusion. [169,170]Further, reactive hyperemic responses after coronary artery occlusion and reperfusion are attenuated by glyburide. [171–173]During coronary artery occlusion, collateral vessels < 100 micro meter in diameter dilate progressively. Glyburide abolished this adaptive response to ischemia. [174]These important results identified a critical role for K^ATPchannels in the regulation of coronary blood flow in response to ischemia. Resting coronary blood flow is partially modulated by K^ATPchannels, because glyburide reduced resting coronary blood flow in a dose-dependent fashion and did so to a greater extent in subepicardium compared with subendocardium. [175–177]G proteins also play a role in activating K^ATPchannels in both large and small coronary microvessels during autoregulation and ischemia. [178,179]Thus, K sub ATP channels are intimately involved in regulation of coronary blood flow at rest and during conditions of inadequate oxygen delivery in vivo.

Pharmacologic activation of K^ATPchannels in the cerebral circulation relaxes the middle cerebral, [180]basilar, [181]and cerebellar arteries [182]and pial arteries and arterioles, [183,184]in vitro and in vivo. Similar to findings in the coronary vasculature, cerebral arterioles dilate in response to hypoxia, and this dilation has been shown to be inhibited by K^ATPchannel blockade. [184,185]Thus, activation of K^ATPchannels in the cerebral vasculature serves as a homeostatic response to conditions of inadequate tissue oxygen delivery. Recently, K^ATPchannels have been identified not only in vascular smooth muscle but also in brain microvascular and aortic endothelial cells. [186]It has been proposed that activation of endothelial K^ATPchannels may modulate vascular tone via subsequent nitric oxide release. [186]Further investigation is required, however, to determine the relative importance of endothelial versus vascular smooth muscle K^ATPchannels in the regulation of regional blood flow.

Activation of K^ATPchannels by volatile anesthetics was first demonstrated by studies indicating that halothane [187]and isoflurane [188]produce coronary vasodilation via a glyburide-sensitive mechanism. Larach and Schuler [187]demonstrated that halothane-induced increases in coronary blood flow were attenuated by glyburide in isolated rat hearts. Cason et al. [188]showed that intracoronary administration of glyburide abolished increases in coronary blood flow produced by isoflurane but not sodium nitroprusside in anesthetized swine, indicating the specificity of isoflurane for K sub ATP channels. Recently, it has been demonstrated in canine hearts in situ that glyburide reversibly attenuated but did not totally abolish increases in coronary blood flow during intracoronary administration of halothane, isoflurane, or enflurane. [189]Glyburide reduced coronary vasodilation during K^ATPactivation with cromakalim and in response to adenosine but did not alter vascular reactivity during administration of the endothelium-independent vasodilator, sodium nitroprusside, or the endothelium-dependent vasodilator, acetylcholine. [189]The selective A¹receptor antagonist, DPCPX, also attenuated isoflurane-induced decreases in coronary vascular resistance, results that suggest that stimulation of A¹receptors contributes indirectly to anesthetic-induced coronary vasodilation, possibly via coupling to K^ATPchannels. [190] 

The mechanisms responsible for the cardioprotective effects of isoflurane in postischemic, reperfused myocardium have been evaluated recently. Isoflurane markedly improved functional recovery of stunned myocardium in chronically [191,192]or acutely [190]instrumented dogs after single [191,192]or multiple [190]brief periods of coronary artery occlusion and reperfusion. Low doses of glyburide (50 micro gram/kg) completely abolished the cardioprotective effects of isoflurane and did not affect the time course of recovery of contractile function when administered alone [190](Figure 6). These results are consistent with the hypothesis that K^ATPchannels are activated by isoflurane and partially mediate the cardioprotective effects of this volatile anesthetic in stunned myocardium. Preliminary findings using patch clamp techniques demonstrate that isoflurane stimulates outward K+ current through K^ATPchannels in isolated ventricular myocytes. [193]Isoflurane may also diminish the sensitivity of K^ATPchannels to inhibition by ATP, thus increasing the open state probability of the channel. [194] 

The role of A¹receptor stimulation in the enhancement of recovery of contractile function of stunned myocardium by isoflurane was recently examined. [195]DPCPX reduced but did not totally abolish the beneficial effects of isoflurane after multiple brief episodes of ischemia and reperfusion (Figure 6). Isoflurane eliminated increases in interstitial adenosine, however, measured with a myocardial microdialysis technique during brief periods of coronary artery occlusion and reperfusion. [195]Attenuation of ATP breakdown and the subsequent reduction of interstitial adenosine by isoflurane were findings similar to those observed after ischemic preconditioning [196]or administration of K^ATPchannel agonists. [78]These collective findings suggest that isoflurane either directly activates A sub 1 receptors or indirectly enhances sensitivity of A¹receptors to reduced amounts of endogenously released adenosine. The results support the hypothesis that K^ATPchannel activation represents the end effector of the cardioprotective signal transduction pathway activated during isoflurane anesthesia.

Experimental evidence suggests that isoflurane causes a cardioprotective effect in vivo by activating A¹receptors coupled to K^ATPchannels in a fashion similar to that which occurs during ischemic preconditioning. Isoflurane has previously been shown to reduce the extent of myocardial infarction [197]; however, the mechanism(s) responsible for this beneficial effect was unclear. The actions of halothane, enflurane, and isoflurane in reducing myocardial infarct size in in situ and isolated rabbit hearts were evaluated recently. [198]These volatile anesthetic agents caused cardioprotective effects and reduced myocardial infarct size to an extent similar to that with ischemic preconditioning. Protective effects of halothane were abolished by pretreatment with either an A¹receptor antagonist or the specific protein kinase C antagonist, chelerythrine, [198]findings which demonstrate that halothane activates similar mechanisms compared with ischemic preconditioning. In an in vivo canine model, isoflurane also markedly decreased the extent of myocardial infarction compared with pentobarbital-anesthetized dogs (Figure 7). [199]This protective effect was equivalent to that produced by ischemic preconditioning, occurred despite discontinuation of isoflurane 30 min before prolonged coronary artery occlusion, could not be explained by beneficial alterations in hemodynamics, and was abolished by glyburide. [199]These data indicate that protection afforded by isoflurane in experimental myocardial infarction was characterized by a short-term memory phase similar to that of ischemic preconditioning and is consistent with previous findings that pharmacologic stimulation of both A¹receptors and K^ATPchannels mimics the conditions present during ischemic preconditioning. [94]The findings also support the contention that activation of K^ATPchannels by isoflurane is the end effector of a volatile anesthetic-induced cardioprotective signal transduction pathway.

The antinociceptive actions of micro- and delta-opioid receptor agonists have recently been linked to K^ATPchannel activation. [200–205]Ocana et al. [201]first demonstrated that the analgesic effects of morphine are blocked by glyburide. Subsequently, the antinociceptive effects of several other micro-receptor opioid agonists were shown to be mediated by K^ATPchannels. [202,203]The K^ATPagonists cromakalim and lemakalim produce direct antinociceptive effects [200,202]and also synergistically enhance the analgesic efficacy of morphine. [202] 

Activation of micro- and delta-receptors by endogenous opioids cause pial artery dilation during cerebral hypoxia, [185,206]an action that was blocked by glyburide. [185]Similarly, stimulation of opioid receptors coupled to K^ATPchannels plays a role in myocardial ischemic preconditioning. [204]Morphine mimicked the effects of ischemic preconditioning in decreasing myocardial infarct size, and this action was inhibited by either glyburide or naloxone. [205]These initial exciting results suggest that opioid-induced modulation of K^ATPchannels may cause important protective effects during tissue hypoxia and ischemia similar to those produced by volatile anesthetic agents.

The K^ATPchannel is a member of the K^irchannel super-family that is regulated by intracellular nucleotide concentration and may couple intracellular metabolism with membrane excitability. K sub ATP channels are critically posed to mediate the regulation of cellular responses to pathophysiologic states such as hypoxia and ischemia. Recent investigation has established the importance of K^ATPchannel activation in ischemic preconditioning of myocardium and neural tissue, during skeletal muscle ischemia, and in the regulation of vascular smooth muscle tone. K^ATPchannels are closed by sulfonylurea hypoglycemic agents, actions that may inhibit endogenous cellular protective mechanisms that depend on K^ATPchannel activation. Thus, an urgent reevaluation of the risks and benefits of sulfonylurea use in patients with type II diabetes has recently been advocated. [8]In contrast to the detrimental effects of K^ATPchannel blockade, the beneficial effects of K^ATPchannel agonists, including anesthetics, in providing organ protection intraoperatively during cardiac, vascular, and neurosurgical procedures are suggested by recent experimental findings. Further investigation is required to determine the relevance of ischemic preconditioning and the mechanisms responsible for this phenomenon in humans and the role of anesthetic agents in improving morbidity and mortality in patients at risk for ischemic or hypoxic injury.

The authors thank Todd Schmeling and David Schwabe for technical assistance and Angela Barnes for assistance in preparation of this manuscript.