Cytoplasmic ATP inhibits cardiac K_ATP channels with a half-maximal inhibition (K_1/2) of ∼10–50 μM [59–63]. Recognition of the heteromultimeric architecture of the channel and molecular cloning of the cDNA encoding each subunit allowed the identification of amino acid residues that determine ATP inhibition. Pore-forming Kir6.2 subunits alone, without regulatory SUR subunits do not effectively traffic to the plasma membrane [64–67]. However, truncated Kir6.2, lacking the C-terminal RKR endoplasmic reticulum retention signal, can reach the membrane [64]. These truncated Kir6.2 generated channels retain ATP-induced inhibition, implying that the determinants are contained within the Kir6 subunit [64]. The cytoplasmic N- and C-termini of Kir6 form an ATP binding pocket with a total of four binding pockets per channel [68–72] at the interface of adjacent subunits (see Fig. 16.1c). Phosphorylation is not necessary for channel inhibition because nonhydrolyzable analogues of ATP are still effective [60, 73, 74]. In the absence of Mg²⁺, reducing the number of phosphate groups from three to two (ADP) or one (AMP) also blocks channel activity but with reduced affinity [59, 60, 75–77], suggesting an important role of the γ-phosphate for ATP binding and K_ATP channel inhibition.

In the presence of Mg²⁺, both ATP and ADP stimulate channel activity [14, 60, 63, 78–81]. Mg-nucleotide stimulation and responsiveness to diazoxide and sulfonylurea require expression with SUR subunits [15, 29, 51, 64, 82–86], indicating that the structural elements responsible for MgADP stimulation and other K_ATP modulators are on the SUR subunit. Mutations of key SUR subunit residues disrupt the response of K_ATP channels to MgATP and MgADP [87, 88], as well as K_ATP channel openers [82, 89]. The nucleotide binding domains-NBD1 and NBD2, located between the eleventh and twelfth transmembrane regions and the C-terminus of SUR subunit, are essential structures for nucleotide binding and stimulation [82, 89–92] (see Fig. 16.1a, c). Mutations in NBD1 prevent ATP binding at both NBDs and interfere K_ATP channel openers acting on NBD2 [82, 88, 89, 93, 94]. SUR1 and 2A have intrinsic Mg²⁺-dependent ATPase activity which hydrolyzes ATP at the NBDs, in particular NBD2 [95–97], and evokes a “power stroke” which overcomes ATP inhibition of Kir6 [97, 98]. Experimentally, MgADP stimulation is more effective than MgATP [14, 60, 63, 78–81] suggesting that binding of MgADP at the NBDs maintains an activated state of SUR, reducing Kir6.2 sensitivity to ATP inhibition [14, 82, 98, 99]. Binding of MgADP at NBD2 may also promote ATP binding at NBD1 [91, 98, 100]. Although there are currently no crystal structures of SUR subunits or domains, related bacterial NBDs crystallize as head-to-tail dimmers [101], with two ATP binding sites (ABSs) formed at the dimer interface. A recent mutagenesis study indicates that both NBDs form as a heterodimer and mutations at the dimer interface disrupt stimulation by MgADP and diazoxide, suggesting an important role for the NBDs in SUR-dependent regulation [101].

Membrane phosphoinositides potently stimulate K_ATP activity by interaction with the Kir6.2 subunit [102–106]. Application of ­phosphatidylinositol 4,5-bisphosphate (PIP₂) to the cytoplasmic membrane increases channel open probability and decreases ATP sensitivity [102–104, 107], demonstrating that ATP inhibition is a dynamic property that is dependent on membrane composition. PIP₂ and ATP bind competitively to K_ATP channels [105, 108]. However, this apparent competition may be revealing allosteric regulation instead of competitive ligands [72, 109]. Following the activation of K_ATP channels during metabolic inhibition, a secondary decrease in PIP and PIP₂ causes the channel activity to decline [110]. PIP₂ likely binds to a pocket formed by the N- and C- termini of two adjacent Kir6.2 subunits [72, 111], directly interacting with positively charged residues in the slide helix near the lipid-cytoplasm interface to stabilize channel opening [109, 112–114]. The negatively charged head group and phospholipid tail of PIP₂ are critical for channel activation by PIP₂ because the PIP₂ cleavage products-IP3 and diacylglycerol have limited effect [102].

Another class of anionic lipids, long chain fatty acyl-coenzyme A (LC-CoA) esters (e.g. oleoyl-CoA and palmityl-CoA), also potently activate K_ATP channels [115–121]. Long-chain acyl-CoA esters are β-oxidation products of fatty acids and the principal metabolic substrates of the heart which are thought to link K_ATP channel activity to the cellular metabolism of fatty acids. Through the same mechanism of phosphoinositides stimulation on K_ATP channels, LC-CoA esters increase K_ATP channel activity by decreasing ATP sensitivity and reducing channel rundown [117–120]. Acyl-CoA esters bind at Kir6.2 subunit with different binding sites from those for MgADP [116], but similar residues for PIP₂ binding are probably involved [122] and an acyl-CoA ester binding motif reportedly locates in the C-terminus of Kir6.2 subunit [123].

Dynamically, K_ATP channel regulation is predominantly a balance between ATP-induced inhibition and MgADP-induced activation [12, 59, 124]. At saturating concentrations, MgADP shifts the IC ₅₀ for ATP inhibition from ∼30 to ∼300 μM, which is still below the normal bulk cytosolic ATP concentration. K_ATP channels may respond to sub-sarcolemmal rather than bulk cytosol ATP concentration [125, 126]. Glycolysis has been reported to effectively inhibit channel activity, and functional glycolytic enzymes are associated with K_ATP channels [127], which suggests that glycolytic enzymes located in the membrane or adjacent cytoskeleton near the channels may be important in maintaining a high local cytosolic ATP/ADP ratio near K_ATP channels to inhibit channel activity [126]. In addition, phosphotransfer enzymes are recognized in amplifying small changes in bulk cytosolic ATP concentration, thereby regulating channel activity (Fig. 16.2). For example, the adenylate kinase (AK) phosphotransfer system facilitates conversion of ATP to ADP causing channels to open, whereas creatine kinase (CK) and protein kinase (PK)/glycolytic systems ­promote conversion of ADP to ATP and channel closure [61, 125, 128–130]. Competitive regulation between AK, CK, and PK regulates the conversion between ATP and ADP, which results in local changes of the ATP/ADP ratio in the vicinity of K_ATP channels, thereby regulating channel activity. Protein kinase A (PKA) [131, 132] and protein kinase C (PKC) [133] also increase channel activity and decrease ATP sensitivity through direct phosphorylation of K_ATP channels. Interestingly, AK [134], CK [135], lactate ­dehydrogenase (LDH) [135], glyceraldehyde-­3-phosphate dehydrogenase (GAPDH) [136] and PK [137] all physically interact with K_ATP channel subunits and this channel-enzyme complex may play an integral role in regulating the nucleotide concentration in the microenvironment surrounding the channel, and fine-tuning the response of K_ATP to the energetic state of the cell (Fig. 16.2).

The emerging picture of K_ATP regulation in the heart is complex. While ATP is a primary inhibitor, it is now clear that K_ATP activation is the result of dynamic regulation by multiple factors, which may account for the highly variable ATP sensitivity [62], the high variance of the latency to channel activation during anoxia [138], and the metabolic state-dependent inhibition by K_ATP channel blockers [139–141].

Severe metabolic stresses such as anoxia, metabolic inhibition and global ischemia cause marked channel opening, but under normal metabolic conditions, sarcolemmal K_ATP channels are predominantly closed, and therefore not expected to contribute significantly to cell excitability. However, due to the higher density of K_ATP compared to other sarcolemmal K⁺ channels, opening of as few as 1 % of K_ATP channels is expected to shorten cardiac action potential by about 50 % [142–145]. In so doing, K_ATP activation will reduce calcium entry and cause myocyte contraction to fail [146] and the energy stores that would otherwise be depleted in the contracting cell are preserved, providing cardioprotection.

In accordance with this cardioprotection concept, K_ATP channel openers enhance the preservation of ATP [147] and produce anti-ischemic effects by causing action potential shortening [148–150]. In parallel, glibenclamide prevents cardioprotective action potential shortening [151–153]. Renewed evidence for the cardioprotection by sarcolemmal K_ATP channels has been obtained through genetic manipulation. Moderate overexpression of SUR2A reportedly increases sarcolemmal K_ATP density and protects the heart against metabolic stress including hypoxia and ischemia/reperfusion [154]. Consistently, ischemic protection is abolished in rat cardiac myocytes transfected with a dominant-negative fragment of SUR2A which causes reduced sarcolemmal K_ATP expression [155]. Additionally, SUR1 overexpressing mice and Kir6.2^−/− mice exhibit a disruption of K_ATP activity and impairment of the cardiac response to systolic overload following chronic transverse aortic constriction [156, 157].

However, in contrast to the accepted notion that increased K_ATP channel activity will produce cardioprotection, two recent studies have shown that neither SUR2^−/− mice nor SUR1^−/− mice have impaired response to ischemia reperfusion [158, 159]. SUR2^−/− mice may actually be resistant to cardiovascular stress caused by global ischemia, due to reduced infarct size [158], contrary to the expected phenotype of impaired protection. However, these SUR2^−/− mice were generated by disruption of NBD1, which might allow translation of the preceding TMD0 and TMD1 or following TMD2 of SUR2, and novel short form SUR2 proteins have been identified in both WT and SUR2^−/− hearts [40]. The cardioprotection observed in these SUR2^−/− mice may be due somehow to the presence of these short form proteins, but may also be related to abolition of the SUR2B component of vascular K_ATP channels [158]. Similar cardioprotection was also observed in SUR1^−/− mice [159]. In an in vivo model of myocardial ischemia/reperfusion injury, SUR1^−/− mice exhibit increased protection and reduced infarct size as well as preservation of left ventricular function [159], presenting the first potential pathological role for SUR1 in the heart. Again the mechanism is unknown, but may be related to abolition of SUR1, and hence abolition of K_ATP, in the atria, where SUR1 is an essential subunit [38, 48]. While it is not clear how SUR knockout mice produce enhanced cardioprotection, clearly reduction of either SUR1 or SUR2 expression can modulate cardioprotection, and hence these subunits may be useful cardioprotective targets.

Cardioprotective functions of the K_ATP channel have been best characterized during ischemic preconditioning. First recognized in 1986, this type of preconditioning arises from brief ischemic challenges that precede a prolonged insult, improving recovery of contractile function and reducing infarct size that results from more severe metabolic insult [160]. Among many endogenous ligands, such as adenosine and ­acetylcholine, and associated signaling proteins, such as protein kinase C and MAP or JUN kinases, K_ATP channels may be the common end-target in ischemic preconditioning [161]. Preconditioning can be mimicked by adenosine, as well as by synthetic K_ATP channel openers such as pinacidil and cromakalim [150]. Additionally, glibenclamide inhibits ischemic as well as adenosine- and acetylcholine-mediated preconditioning [150, 162]. Ischemic preconditioning is also abolished in hearts from Kir6.2^−/− mice [156], in hearts from Kir6.2 gain-of-function mutation mice which show decreased K_ATP density [163], and in rat cardiac myocytes transfected with a dominant-negative fragment of SUR2A [155], strongly suggesting a role for sarcolemmal K_ATP channel as a preconditioning mediator.

In 1991, K_ATP channels were also reported in the mitochondria [164]. A succession of pharmacological studies that followed argued that mitochondrial rather than sarcolemmal K_ATP channels are primarily involved in preconditioning. First, action potential shortening cannot account for the cardioprotective effects of K_ATP channel openers [165–167]. Second, diazoxide, reportedly without effect on sarcolemmal K_ATP channels, can mimic ischemic preconditioning [168–170]. Third, 5-hydroxydecanoic acid (5-HD), reportedly a specific mitochondrial K_ATP channel blocker, effectively abolishes ischemic preconditioning and the anti-ischemic effects of K_ATP channel openers without inhibiting action potential shortening [171–173].

Several studies show that diazoxide activates sarcolemmal K_ATP channels [174, 175] and 5-HD suppresses APD shortening in ischemic hearts [152, 176], suggesting diazoxide and 5-HD actually target the sarcolemmal K_ATP channel. Moreover, additional K_ATP-independent targets are recognized as being sensitive to diazoxide and 5-HD: diazoxide inhibits succinate oxidation and succinate dehydrogenase activity; 5-HD serves as substrate for acyl-CoA synthetase [177]. Therefore, the use of the above pharmacological tools as probes for mitochondrial K_ATP requires reassessment.

Genetic manipulation has made it possible to circumvent some of the limitations of pharmacology by examining gene function directly. The molecular structure of the mitochondrial K_ATP channel is presently unknown, but Kir6.2, the primary pore-forming subunit of sarcolemmal K_ATP channels, does not appear to form mitochondrial K_ATP channels [178]. One study, using pharmacological tools and transfection systems suggests that current through the Kir6.1/SUR1 complex mimics the properties of mitochondrial K_ATP channels [50]. In this study, Kir6.1/SUR1 channels expressed heterologously in HEK293 cells were activated by diazoxide but not P-1075, and were inhibited by 5-HD but not HMR1098. The pharmacological properties of Kir6.1/SUR1 channels are similar to those reported for mitochondrial K_ATP channels [50], suggesting a Kir6.1/SUR1 composition in mitochondria. Therefore, genetically controlling Kir6.2 may be a direct means of assessing the role of K_ATP channels in the sarcolemma rather than in the mitochondria. Kir6.2^−/− mice lose ischemia preconditioning [156] as well as diazoxide induced preconditioning [179]. Preconditioning is also abolished in cardiac Kir6.2 gain-of-function mice in which cardiac basal K_ATP activity is increased but overall activity is dramatically decreased [163, 180]. These studies are therefore more consistent with sarcolemmal rather than mitochondrial K_ATP channels mediating preconditioning. Ultimately, the existence of conventional K_ATP channels in mitochondria may be questionable: both Kir6.1 and SUR1 subunits may be expressed strongly at the sarcolemma of ventricular myocytes [41]. Even efforts to determine the molecular composition of mitochondrial K_ATP from immunohistological experiments may be unreliable given the still considerable difficulties in isolation of mitochondrial membrane proteins [181].

Lethal ventricular arrhythmia remains a leading cause of sudden death in patients with ischemic heart disease; 80 % result from ventricular fibrillation accompanied by myocardial ischemia and infarction [182]. The development of malignant arrhythmias during ischemia may be primarily due to the opening of K_ATP channels which cause rapid external potassium accumulation and reduction in action potential duration and therefore reduction in the refractory period in the ischemic myocardium [183, 184]. Additionally, dispersion of action potential duration may result from K_ATP activation: epicardial cells may be preferentially affected by K_ATP opening due to an action potential notch caused by the transient outward current (I_to) [185]. Depression of the membrane potential by I_to in conjunction with K_ATP channel opening will tend to prematurely deactivate Ca²⁺ channels, leading to early action potential termination. Selectively shortened action potentials in epicardium by K_ATP opening will then result in a transmural dispersion of repolarization, causing ST-segment elevation in the ECG and giving rise to phase 2 re-entry, which can precipitate ventricular tachycardia and ventricular fibrillation (VF) [185, 186]. In accord with this idea, Di Diego and Antzelevitch were able to show reentry in isolated ventricular myocardium after adding the K_ATP opener pinacidil, electrical homogeneity was restored and arrhythmias were abolished by blocking I_to (4-aminopyridine, 4-AP) or K_ATP (glibencamide) (Fig. 16.3) [186].

While K_ATP channel activation is potentially proarrhythmic and may predispose to lethal re-entrant arrhythmias, the opening of K_ATP channels during ischemia is also expected to stabilize the resting membrane potential and reduce ectopic pacemaker activity, i.e. an antiarrhythmic effect. Indeed, pharmacological studies provide two contrasting effects of K_ATP channels with respect to arrhythmias. Glibenclamide decreases the incidence of sustained tachycardia and ventricular fibrillation during ischemia-reperfusion [187–189], but increases the occurrence of tachycardia [190]. Similarly, some have observed that K_ATP channel openers promote tachycardia and fibrillation [58, 187, 190], while, others see a reduction in the overall incidence of arrhythmias [149, 191]. Grover et al. [149] reported opposite effects of the K_ATP opener cromakalim in different species: inhibition of fibrillation in anesthetized dogs but induction of fibrillation in isolated rat hearts, implying that the arrhythmic effects of K_ATP channels may be a function of the experimental model.

The significance of K_ATP channel composition in arrhythmia induction has been further explored through genetic manipulation. Kir6.2^−/− mice exhibit high risk for induction of triggered activity and ventricular arrhythmia under catecholamine challenge, due to defective action potential shortening and predisposition to early afterdepolarizations [192]. Thus, sarcolemmal K_ATP is required for tolerance against triggered catecholamine-induced ventricular arrhythmia. Evidence for a proarrhythmic effect of sarcolemmal K_ATP channels comes from double transgenic (DTG) mice which overexpress both SUR1 and an ATP-insensitive Kir6.2 in the heart [193]. These SUR1-expressing DTG mice develop several different types of arrhythmias including AV block, atrial and ventricular tachyarrhythmia, indicating a proarrhythmic effect of K_ATP channels comprised of SUR1 and ATP-insensitive Kir6.2. In marked contrast, SUR2A-expressing DTG mice exhibit no change in arrhythmia incidence [193]. In addition, single transgenic mice which overexpress SUR1 but not SUR2A exhibit PR prolongation, suggesting delayed conduction in the AV nodal or diffuse conduction system. These results may reflect the higher sensitivity of SUR1-based K_ATP channels to metabolic activation [49], and indicate that expression of hyperactive K_ATP in the heart is likely to be proarrhythmic.

Also of interest are K_ATP mutations that confer risk for arrhythmia. The SUR2 T1547I mutation was associated with adrenergic atrial fibrillation originating from the vein of Marshall in a patient with paroxysmal atrial fibrillation [194]. More recent studies also report association of the S422L mutation in Kir6.1 with J-wave syndromes, which include Early Repolarization Syndrome and Brugada Syndrome (Fig. 16.4) [6–8]. As discussed above, the existence and distribution of Kir6.1 in the heart are not yet well established, nor can the consequences of mutations in other tissues be excluded, but these findings will certainly cause a re-evaluation of cardiac K_ATP channel function.

A role for sarcolemmal K_ATP channels during arrhythmias may be masked by the potential that mitochondrial K_ATP inhibition influences the incidence of arrhythmia. 5-HD suppressed the incidence of ventricular fibrillation induced by coronary ligation in rats [195], suggesting an antiarrhythmic effect for mitochondrial K_ATP inhibition. On the other hand, the antiarrhythmic effects of both Nicorandil and Minoxidil in rabbits were abolished by 5-HD but not by HMAR 1883, suggesting a proarrhythmic effect of mitochondrial K_ATP inhibition. The role of 5-HD in arrhythmias became more confusing when it was reported that 5-HD was not anti-arrhythmic in dogs [196] or in rats [197]. Conversely, HMR 1883 decreased the incidence of ischemia-induced irreversible ventricular fibrillation and improved survival during acute myocardial infarction in conscious rats as well as during reperfusion after brief myocardial ischemia in anesthetized rats [198], while in the same study, 5-HD did not protect hearts in these arrhythmia models [198], suggesting arrhythmias during ischemia and ischemia/reperfusion are the result of opening sarcolemmal K_ATP channels rather than mitochondrial K_ATP channels.

Cardiac hypertrophy can be an adaptive response but prolonged hypertrophy can progress into heart failure and sudden death, which may result from consequently disturbed electrical activity [199]. Decreased K_ATP channel responsiveness to ATP inhibition has been reported in hypertrophied cardiomyocytes suggesting hyper-activity [200, 201]. However, impaired K_ATP channel activation has also been reported in hypertrophic myocytes and suggested as a potential explanation for reduced tolerance of hypertrophied hearts to ischemia [202]. Interestingly, SUR2^−/− mice exhibit greater ventricular mass and relative heart size compared to control cohorts, evidence of hypertrophy [158]. Kir6.2^−/− mice also show increased mortality and exaggerated hypertrophy in response to pressure overload [157, 203] and to mineralocorticoid/salt challenges [204], suggesting that disrupting K_ATP channel actually predisposes hearts to hypertrophy under stress.

K_ATP channels appear to be less sensitive to ATP inhibition and more sensitive to K_ATP channel openers in failing human hearts, which may indicate that K_ATP remodeling may favor a beneficial effect in failing hearts [58, 205]. Optical mapping of failing hearts reveals increased APD shortening in response to diazoxide (Fig. 16.5). A heart failure model induced by transgenic expression of cytokine tumor necrosis factor alpha (TNFα) shows a disturbed link between the metabolic state of the cell and K_ATP channel activity and increased cardiomyocyte vulnerability to stress [206]. In this case, application of K_ATP channel openers improved stress tolerance [206]. The same group has reported linkage of human heart failure to K_ATP channel mutations. Mutations in SUR2 that cause decreased catalytic activity of NBD1 and disruption of K_ATP channel gating were identified in patients with dilated cardiomyopathy [9, 124]. Also, the Kir6.2 E23K polymorphism, which causes enhanced channel activity and is associated with type 2 diabetes susceptibility [207] also appears to be linked with subclinical maladaptive cardiac remodeling among individuals with increased stress load due to hypertension [208] and is over-represented in patients with congestive heart failure [5]. More recently, an in-frame deletion (E332del) and a missense mutation (V346I) were identified in two nonrelated sudden infant death cases. Both mutations lead to loss-of-function of Kir6.1 channels which may predispose a maladaptive cardiac response to systemic metabolic stressors [209].

Potential protective effects of K_ATP are not limited to an extreme metabolic insult, such as global ischemia. A physiological role for cardiac K_ATP channels has been recognized during adaptation to physiological stress [1–3]. Short-term exercise up-regulates cardiac sarcolemmal K_ATP channel expression and enhances APD shortening which in turn reduces cardiac energy ­consumption in response to heart rate acceleration [1]. In addition to the lack of protection during ischemia-reperfusion, Kir6.2^−/− mice exhibit intolerance to vigorous exercise in ­treadmill test and impaired survival in response to ­isoproterenol treatment compared with age- and ­gender-matched WT controls [3]. The Kir6.2^−/− mice show less action potential shortening and diastolic calcium overload associated with contractile dysfunction which may explain the decreased adaption to stress [3]. Bulk adenine nucleotide content was not changed under the exercise and isoproterenol challenge, suggesting K_ATP channel activation may be caused by metabolic fluctuations in subsarcolemmal compartments. In addition, during chronic stress induced by long-term swimming training, Kir6.2^−/− mice develop calcium-­dependent structural damage and impaired contractile function, with a significant reduction in left ventricular fractional shortening and impaired cardiac output [2]. The Kir6.2^−/− animals lack K_ATP in multiple tissues and further studies are warranted. However, selective disruption of K_ATP channels in heart (loss-of-function mutation of pore-forming subunit Kir6.1 [αMHC-Kir6.1AAA]) abolishes the enhanced APD shortening in response to the workload imposed by heart rate acceleration and causes increased cardiac energy consumption [1], strongly ­suggesting that K_ATP channels may play a more widespread role in cardiac physiology than generally appreciated, and are not limited to roles in pathological states.