ATP-sensitive potassium channels (KATP) are widely distributed and present in a number of tissues including muscle, pancreatic beta cells and the brain. Their activity is regulated by adenine nucleotides, characteristically being activated by falling ATP and rising ADP levels. Thus, they link cellular metabolism with membrane excitability. Recent studies using genetically modified mice and genomic studies in patients have implicated KATP channels in a number of physiological and pathological processes. In this review, we focus on their role in cellular function and protection particularly in the cardiovascular system.
Two independent laboratories can lay claim to having first described the ATP-sensitive potassium channels (KATP; channel nomenclature follows Alexander et al., 2013). Noma (1983) observed the appearance of an outward K+ current in heart muscle cells when treated with metabolic poisons or hypoxia. This was reversed by ATP injected into the cell. Similar observations were made by another group (Trube and Hescheler, 1984). Such channels were subsequently described in pancreatic beta cells (Ashcroft et al., 1984), skeletal muscle (Spruce et al., 1985), smooth muscle (Standen et al., 1989) and neurones (Ashford et al., 1988). During this period, the basic electrophysiological and pharmacological properties of the channel were elucidated (Ashcroft, 1988; Noma and Takano, 1991). In inside-out patches in ∼140 mM symmetrical K+ concentrations, the single-channel conductance is ohmic with a conductance of 70–80 pS. The lower values sometimes noted in the literature generally have lower and asymmetric K+ concentrations. The channel is highly selective for potassium (PNa/PK∼0.01). Activity is inhibited by the application of ATP with a Ki of 10–500 μM with a Hill coefficient of more than 1 (generally around 2) depending on the tissue and recording configuration. The ATP inhibition is not dependent on ATP hydrolysis: it is not reliant on Mg2+ and ATP can be substituted by non-hydrolysable derivatives. In the absence of magnesium other adenine nucleotides can inhibit channel activity but they are less potent. However, in the presence of Mg2+ and ATP, ADP is stimulatory.
Even at the beginning of the 1990s the channels were known to have a rich pharmacology (see Edwards and Weston, 1993). Sulphonylureas were discovered accidentally when it was noted that the anti-microbial sulphonamides caused hypoglycaemia in animals. It became apparent that stimulation of insulin release from pancreatic beta cells occurred because of inhibition of KATP channels. There is a family of these drugs: the most widely known are the first-generation agents (e.g. tolbutamide, chlorpropamide) and the more potent second-generation agents (e.g. glibenclamide, gliclazide, glipizide). These agents still have a place in the management of type 2 diabetes mellitus. There are also agents that are able to open KATP channels [KATP channel opening drugs (KCOs)]. Intriguingly, not only are some of these agents selective for KATP but they also exhibit a very broad range of chemical structures: for example, diazoxide is a benzothiadiazine, pinacidil a cyanoguanidine and nicorandil a pyridyl nitrate (Mannhold, 2006). Agents known to block other K+ channels, for example, Ba2+ and 4-aminopyridine, are also active on KATP channels.
The inwardly rectifying family of potassium channels (KIR) resisted cloning even after the elucidation of the primary structure of voltage-gated potassium channels. It was not until expression cloning techniques were employed that the first cDNAs were isolated (Ho et al., 1993; Kubo et al., 1993) and a substantial gene family with seven subfamilies was revealed (see Nichols and Lopatin, 1997). The pore-forming α subunits have two transmembrane domains with an intracellular N- and C-terminus. The only significant area of homology with the voltage-gated family was the pore-forming H5 segment responsible for potassium selectivity. Initially, however, homology cloning approaches did not elucidate an obvious candidate for the KATP channel that functioned as expected in a heterologous expression system. A critical missing component was revealed as the sulphonylurea receptor (SUR) (Aguilar Bryan et al., 1995). Co-expression of the KIR6.0 family of inwardly rectifying potassium channels with the SUR reconstituted the KATP channel (Inagaki et al., 1995a,c). It became apparent that there were two isoforms of KIR6.0 (KIR6.1 and KIR6.2) and two variants of SUR (SUR1 and SUR2 with two splice variants SUR2A and SUR2B) (Inagaki et al., 1996; Isomoto et al., 1996; Yamada et al., 1997). SUR is a member of the ATP-binding cassette (ABC) family of protein (Linton and Higgins, 2007). It is most closely related to the multidrug resistant-related proteins and they are now all classified in the ABCC family (Toyoda et al., 2008). Characteristically, SUR has 17 transmembrane segments grouped into three domains comprised of five (TMD0), six (TMD1) and six (TMD2) helices respectively. The N-terminus is extracellular and each of these domains is connected by intracellular linkers and finally an intracellular C-terminus. The TMD1-TMD2 and C-terminus contain nucleotide-binding domains (NBDs) with Walker A and Walker B motifs and linker regions indicative of ATP binding and hydrolysis (Linton and Higgins, 2007). This topology is well supported by experimental data (Conti et al., 2001). At the genomic level, the genes encoding SUR1 and KIR6.2 and SUR2 and KIR6.1 are adjacent to one another on 11p15.1 and 12p12.1, respectively, and this arrangement suggests a coordinated regulation of the SUR and KIR6.0 subunit (Inagaki et al., 1995a,b; Chutkow et al., 1996). The mature KATP channel complex is a hetero-octamer of four KIR6.0 subunits and four SUR subunits (Clement et al., 1997; Shyng and Nichols, 1997). A diagram of KATP channel assembly is shown in Figure 1.
SUR2A\ KIR6.2 underlies the cardiac KATP channel present in ventricular muscle and SUR2B\ KIR6.1 that in smooth muscle. However, there are qualifications to this simplified picture. KIR6.1 is ubiquitously expressed and thus there exists the potential for heteromultimerization with KIR6.2. Indeed, this can be demonstrated with heterologous expression (Cui et al., 2001) and it might occur potentially in the cardiac conduction system (Yoshida et al., 2004; Bao et al., 2011b). The issue of whether different SUR subunits can heteromultimerize is more controversial but practically there are not many occasions where more than one isoform is expressed and the evidence favours homomultimers (Giblin et al., 2002; Tricarico et al., 2006; Cheng et al., 2008). Secondly, the composition of the channels may show subtle but important anatomical variations. For example, KATP in atrial cardiac myocytes is constituted by SUR1\KIR6.2 and this differential tissue distribution may open the door for selective pharmacological manipulation in the heart in atrial fibrillation (Flagg et al., 2008). Finally, most smooth muscle KATP channels have been known to be functionally different for some time. Some of these channels have a lower single-channel conductance (∼35 pS), an absolute dependence on the provision of nucleotide diphosphates for activity (‘KNDP’) and are less sensitive to ATP inhibition (Beech et al., 1993). These properties were reproduced by the co-expression of SUR2B and KIR6.1 subunits in heterologous systems (Yamada et al., 1997; Cui et al., 2002). However, KIR6.2 alone or together with KIR6.1 and SUR2B might participate in some vascular beds and in non-vascular smooth muscle (Teramoto et al., 2009).
The octameric nature of the channel complex has led to interesting structure-function questions. This area has been reviewed extensively (Rodrigo and Standen, 2005; Flagg et al., 2010). The continuing crystallization of K+ channels is likely to enrich and supersede much of this work. A high resolution structure of the KATP channel complex or that of the KIR6.0 pore-forming subunit alone has not yet been reported. A low resolution cryoEM study was possible and showed a compact structure with the four SUR1 subunits interacting with KIR6.0 in the cytosolic and membrane domains (Mikhailov et al., 2005). One interesting feature was a cleft between the SUR1 subunits by which ATP could access its binding site. The crystal structure of KIR2.2, with and without phosphatidyl 4,5-bisphosphate (PIP2), offers insight into potential gating mechanisms that might be revealed by high-resolution KATP channel structures (Hansen et al., 2011). The most recent structure of the KIR3.0 family of channels shows that these channels possess two gates, one as described for KIR2.2 and another in the C-terminus, potentially gated by G-protein βγ subunits. Both gates need to be open to exhibit channel activity (Whorton and MacKinnon, 2011).
Metabolic regulation and mitochondrial KATP channels
One of the defining features of KATP channels is their sensitivity to metabolic changes. The inhibition by ATP is determined by the KIR6.0 subunit and site-directed mutagenesis has identified key residues in the C- and N-terminus of the KIR6.2, in particular R50, C166, I167, T171 and K185 (Tucker et al., 1997; 1998), and this work underpins a detailed structural model (Antcliff et al., 2005). KIR6.1 may also have more substantial ATP sensitivity than is generally appreciated (Babenko and Bryan, 2001), though it may depend on the recording configuration and cellular environment (Cui et al., 2002). It is certainly true that both KIR6.1 and KIR6.2 containing channel complexes are sensitive to metabolic poisoning (Farzaneh and Tinker, 2008).
The issue of how SUR interacts with nucleotides is complex and not fully resolved. Early work showed that the SUR subunit endows the channel complex with sensitivity to activation by MgADP, and this is a function of the NBDs (Gribble et al., 1997; Shyng et al., 1997). NBDs are asymmetric in function with NBD2 binding and hydrolysing MgATP rapidly while NBD1 binds ATP even in the absence of Mg2+ and hydrolyses it more slowly (Ueda et al., 1997; Bienengraeber et al., 2000). Furthermore, experiments using vanadate and beryllium which mimic the post- and pre-hydrolytic states, respectively, support the idea that ATP hydrolysis at NBD2 is needed for channel activation (Zingman et al., 2001). However, in inside-out patches, MgADP potently activates the channel, suggesting the activated state is directly accessible without prior hydrolysis. More recent work supports the idea that hydrolysis at NBD2 may not be necessary for activation (Ortiz et al., 2013). Another feature that has not been resolved is the potential dimerization of the NBDs during this cycle. In other ABC transporters, dimerization is a necessary prerequisite for ATP hydrolysis (Linton and Higgins, 2007). Thus, there are some outstanding questions in this complex mechanism. In the intact cell, studies have revealed the interaction of KATP channels with enzymes involved in cell metabolism. The cardiac channel complex (KIR6.2\SUR2A) is able to directly interact with adenylate kinase, creatinine kinase and lactate dehydrogenase (Carrasco et al., 2001; Crawford et al., 2002a,b). These interactions may make the channel sensitive to small changes in cytoplasmic ATP within the cell and to ATP derived from glycolysis (Weiss and Lamp, 1987).
There have been suggestions that KATP channels are also present in mitochondria (‘mitoKATP’) (Inoue et al., 1991; Paucek et al., 1992). In the first study, channel activity was directly assayed to demonstrate a 10 pS channel inhibited by ATP and glibenclamide. Subsequent work revealed a possibly unique pharmacology in that mitoKATP was inhibited by 5-hydroxydecanoate and activated by diazoxide, properties not shared by the cardiac sarcolemmal channel (Grover and Garlid, 2000). However, these distinguishing features are no longer so clear cut (Li et al., 2010). The most convincing approach would be to directly clone the subunits. Are the known KIR6.0 and\or SUR subunits the molecular equivalent of mitoKATP? A number of studies have proposed that KIR6.1 might be a mitoKATP subunit specifically showing a mitochondrial localization or comparable pharmacology (Suzuki et al., 1997; Liu et al., 2001); however, the commercial antibodies may be detecting other unrelated proteins in mitochondria (Foster et al., 2008). If KIR6.1 does not underlie mitoKATP channels what are the other possibilities? One group isolated a complex of a succinate dehydrogenase, mitochondrial ABC protein 1, phosphate carrier, adenine nucleotide translocator and ATP synthase (Ardehali et al., 2004). However, none of these proteins contain the canonical elements in the H5 segment of the established potassium channels and structures. Furthermore, some of the drugs used seem to have effects or are metabolized in other pathways (Das et al., 2003; Hanley et al., 2005). The most recent work in this area may address some of these issues (Foster et al., 2012). These authors isolated KIR1.2 from the inner mitochondrial membranes of bovine heart and cell imaging confirmed the mitochondrial localization of KIR1.2, albeit after heterologous expression. In pharmacological studies, overexpression of KIR1.2 along with the use of shRNA to silence the protein, implicated this channel subunit as underlying mitoKATP and having a role in cellular protection. These are potentially exciting findings but await confirmation and validation using more in vivo approaches.
Regulation through cell signalling pathways
KATP channels have a tendency to run down in ATP-free solutions and the channel activity can be ‘refreshed’ with low concentrations of MgATP. This dependence was not understood until channel activity was shown to be absolutely dependent on membrane phosphoinositides, in particular PIP2 (Hilgemann and Ball, 1996; Fan and Makielski, 1997; Shyng and Nichols, 1998). The ATP is needed for the synthesis of plasma membrane PIP2 via PI kinases and the PIP2 antagonizes ATP inhibition and leads to channel opening (Shyng and Nichols, 1998; Pratt et al., 2011). Direct evidence for the involvement of membrane PIP2 has been obtained using an elegant binary membrane recruitment system and this works for KATP channel regulation also (Suh et al., 2006; Quinn et al., 2008). There is likely to be a direct interaction with the protein as the channel, KIR2.1 and KIR2.2 in this case, can be purified and reconstituted in liposomes of fixed lipid composition (D'Avanzo et al., 2010). Furthermore, recent crystal structures show PIP2 binding in the homologous KIR2.2 leads to C-terminal domain translocation to interact with the transmembrane domain and this change opens the helix bundle gate which occludes the lower pore.
In smooth muscle, downstream activation of PKA through various receptors coupled to the stimulatory G-protein Gs, such as adenosine A2, β adrenoreceptors, calcitonin gene-related peptide and prostacyclin, leads to vasodilatation. A major contribution to the vascular smooth muscle (VSM) cell hyperpolarization is the opening of KATP channels (Standen et al., 1989; Nelson et al., 1990; Rodrigo and Standen, 2005). Subsequent molecular studies on KIR6.1\SUR2B revealed that this was likely due to the direct phosphorylation of both channel subunits (T633, S1387 and S1465 on SUR2B; S385 on KIR6.1) (Quinn et al., 2004; Shi et al., 2007). Further regulation may occur through dephosphorylation of these residues via the Ca2+-dependent phosphatase calcineurin (Wilson et al., 2000; Orie et al., 2009). The receptor\PKA\KATP axis may be selectively localized within membrane compartments of the cell. Thus, there is evidence that PKA is largely present in a particulate fashion through interaction with A-kinase anchoring proteins (Hayabuchi et al., 2001a). Furthermore, the channel complex may be localized to caveolae and this may be important for signalling (Sampson et al., 2004; 2007; Davies et al., 2010).
Vasoconstrictors, such as angiotensin II and endothelin-1, activate PKC and there is evidence that KATP channel activity can be modulated through such pathways (Kubo et al., 1997; Cole et al., 2000; Hayabuchi et al., 2001b; Thorneloe et al., 2002; Quinn et al., 2003; Sampson et al., 2007). The regulation is Ca2+ independent and mediated by PKCε (Hayabuchi et al., 2001b; Quinn et al., 2003). Direct channel phosphorylation of KIR6.1 is likely responsible with a cluster of serine residues in the distal C-terminus playing a key role (Quinn et al., 2003; Shi et al., 2008). There may also be effects on channel internalization and recycling perhaps via caveolae (Jiao et al., 2008). Furthermore, vasoconstrictors may inhibit PKA as this would act as an additional inhibitory input to KATP channels (Hayabuchi et al., 2001b). One mechanism that has been little explored is the role of PIP2 depletion in addition to PKC activation. However, it should be noted that the modulation is essentially abolished by PKC inhibitors and that KIR6.1 seems to have a relatively high affinity for PIP2, suggesting channel activity may be maintained in the face of substantial depletion (Quinn et al., 2003; Sampson et al., 2007). It should be emphasized that PKC-dependent modulation of VSM KATP channel activity is only likely to account for a part of the action of vasoconstrictors.
A group of endothelial mediators – the so-called gasotransmitters – may also influence VSM KATP channel function (Zhao et al., 2001; Mustafa et al., 2009b). The action of NO is of course well known but the endothelium also generates two other gasotransmiters, CO and H2S. H2S is produced largely by cystathionine γ-lyase and production is regulated by the activation of Ca2+-calmodulin, in a manner analogous to that of NO (Yang et al., 2008). The effector mechanism is thought to be the activation of KATP channels in VSMs through direct modification of channel cysteines, though other mechanisms are possible (Zhao et al., 2001; Cheng et al., 2004; Mustafa et al., 2009a).
Thus, the vascular KATP channel and the cloned equivalent KIR6.1\SUR2B are subject to prominent hormonal regulation through direct subunit phosphorylation. What is known about the regulation of KIR6.2 complexes particularly in cardiac muscle? This issue is particularly pertinent for PKC modulation as this has been implicated as being central in cellular protection and preconditioning in cardiac cells (Yellon and Downey, 2003). In early studies, PKC modulation was thought to activate sarcolemmal cardiac KATP; however, it now appears that there is a biphasic regulation with activation followed by a slower inhibitory response corresponding to channel internalization (Light et al., 1996; Hu et al., 2003). The phenomena are critically dependent on the prevailing conditions of study. We thought for some time that PKC activation did not modulate KIR6.2 (Quinn et al., 2003). However, in intact cells and with higher pipette Ca2+ in whole cell recordings, we did see biphasic modulation. The inhibitory response was due to channel internalization and occurred because of phosphorylation of S372 in KIR6.2 (Aziz et al., 2012). The regulation of sarcolemmal KATP channels by PKA has been little studied. With KIR6.2\SUR1, PKA phosphorylation leads to increased channel activity through residues in SUR1 and KIR6.2 that are homologous to those in SUR2B and KIR6.1 (Gonoi et al., 1999; Lin et al., 2000; Quinn et al., 2004). Indeed, a plausible a priori case could be made for a contribution of sarcolemmal cardiac KATP in the action potential (AP) shortening occurring with the increased heart rate in exercise, in addition to the slowly activating component of the delayed rectifier current. One interesting feature is the subcellular localization of KATP channels that appear to be concentrated at the neck of the T-tubule, suggesting activation could have a significant influence on excitation-contraction coupling (Korchev et al., 2000).
Pathophysiological function of KATP channels in the cardiovascular system
Generally, three types of study contribute to the understanding of the physiological role of KATP channels: ex vivo and in vivo pharmacological studies, functional studies in genetically engineered mice, and human genetics (Figure 2).
Paradoxically, there has been substantial focus on the role of the cardiac sarcolemmal KATP channel in pathobiology (see below), without really addressing the issue of what physiological role the channel might perform in ventricular myocytes. The use of mice with global genetic deletion of KIR6.2 (KIR6.2 null) has suggested some intriguing possibilities. These animals have an attenuated ability to perform high-intensity exercise and are predisposed to catecholamine cardiotoxicity (Zingman et al., 2002). However, it is not yet clear if it is the deletion in cardiac myocytes that is critical. Channel function will also be impaired in skeletal muscle, pancreatic beta cells and in central neurons, and all of these could have an influence on the integrated physiological function. Mice with cardiac overexpression of a dominant negative KIR6.2 subunit had a proarrhythmic phenotype and impaired exercise tolerance (Tong et al., 2006). The study of mice with conditional deletion of KIR6.2. (and SUR2) in cardiac myocytes is likely to be highly informative.
Sarcolemmal KATP channels are essentially closed under normal metabolic conditions and hence are thought not to contribute towards the coupling of membrane excitation and contraction in the basal physiological state. In keeping with this, studies in ventricular myocytes from KIR6.2 null mice show that AP duration and contractile function are normal (Suzuki et al., 2001). However, exposure to severe metabolic insults such as during hypoxia and ischaemia leads to opening of cardiac KATP channels. Metabolic stress leads to substantial shortening of the AP and attenuated contraction (Lederer et al., 1989; Venkatesh et al., 1991), and these effects are absent when KIR6.2 is not present (Li et al., 2000), or in the presence of a KATP channel blocker (Venkatesh et al., 1991). Primarily, opening of KATP channels is likely to be a protective mechanism because the increase in K+ conductance stabilizes the resting membrane potential, shortens the AP and reduces Ca2+ influx, resulting in the conservation of intracellular energy stores and preventing calcium overload. In support of this proposal, application of the KCO pinacidil enhanced ischaemia-induced AP shortening, early contractile failure and preserved ATP levels (McPherson et al., 1993), and in hearts from KIR6.2 null mice, contraction was prolonged and AP duration unaffected during ischaemia (Suzuki et al., 2002). In addition, pinacidil activation of sarcolemmal KATP channels reduced reperfusion-induced Ca2+ overload in cardiac myocytes, and ablation of KATP channels in vivo gives rise to a greater susceptibility to Ca2+ overload and impairs contractile recovery. Moreover, exercise causes significant remodelling of cardiac KATP channels. Specifically, exercise induces an increase in KATP channel expression (∼40%) in mouse ventricles promoting AP shortening in response to an increased heart rate, and these effects are abolished when non-functional KATP channels are transgenically expressed (Zingman et al., 2011). Interestingly, overexpression of SUR2A in cardiac tissue leads to a phenotype protected from ischaemia (Du et al., 2006). An intracellular pool of KATP channels may serve as a reservoir to modulate membrane surface density in stress conditions (Bao et al., 2011a).
There is a substantial body of work proposing that mitoKATP may also have a role in cardioprotection and this has been reviewed in detail elsewhere (Yellon and Downey, 2003). It is worth stating that there are persuasive data favouring the involvement of sarcolemmal cardiac KATP channels in these phenomena, at least in mice. In KIR6.2 null mice, the protective effect of ischaemic preconditioning was abolished and recovery of contractile function was compromised (Suzuki et al., 2002; 2003; Gumina et al., 2003). Moreover, the preconditioning effect of diazoxide was also absent (Suzuki et al., 2003).
The initial opening of KATP channels in response to a metabolic insult is cardioprotective; however, activation of KATP channels also induces early repolarization, thereby shortening the QT interval and reducing the refractory period, hence predisposing to re-entrant arrhythmias. Typical ECG changes observed during ischaemic insults include ST elevation or depression suggesting changes in repolarization. These ischaemia-induced changes in ECG characteristics are ameliorated by the KATP channel blocker glibenclamide and induced by the KCO pinacidil (in the absence of ischaemia) consistent with KATP channel activation underlying these features (Kubota et al., 1993). In support, KIR6.2 null mice are not prone to ST elevation in response to ischaemia (Li et al., 2000). However, ischaemia-induced ST elevation was observed in SUR2 null mice (Chutkow et al., 2002). KIR6.1 null mice also have episodes of cardiac ischaemia accompanied by ST elevation (Miki et al., 2002). This was originally postulated to be due to the absence of VSM KATP channels; however, it is likely this is not strictly the case as when SUR2B was selectively re-introduced to smooth muscle in SUR2 null mice, the phenotype persisted (Kakkar et al., 2006). These data suggest that KIR6.1 in heart may also be involved in early repolarization and ischaemia-induced arrhythmia and is consistent with recent studies showing heterogeneity of KATP subunit composition in different regions of the heart (Flagg et al., 2008; Bao et al., 2011b). Early repolarization patterns in the ECG has historically been considered to be of little consequence and is commonly observed in healthy males and athletes. However, there is now evidence to suggest that the early repolarization pattern (‘J wave syndromes’) may be associated with increased risk of ventricular fibrillation (Antzelevitch, 2012). A clear association of KATP channels with early repolarization syndromes was made when Haissaguerre et al. (2009) found a rare variant of KCNJ8 (KIR6.1) in a patient with idiopathic ventricular fibrillation and prominent early repolarization. In further independent studies, the same missense mutation of a highly conserved serine to leucine (S422L) was discovered in five more patients (Medeiros-Domingo et al., 2010; Barajas-Martinez et al., 2012). Both studies showed an increased current density when KIR6.1-S422L was co-expressed heterologously with SUR2A. Mechanistically, this gain of function in the mutant channel can be explained by the decreased ATP sensitivity of KIR6.1-S422L channels compared with wild-type (WT) channels (Barajas-Martinez et al., 2012).
Most studies of cardiac arrhythmias resulting from myocardial ischaemia have been focused predominantly on the abnormalities of ventricular rhythm and relatively little is known about the role of KATP channels in abnormal atrial rhythm. A recent study has shown that activation of KATP channels by β-adrenoceptor-induced metabolic stress provides a substrate for atrial tachyarrhythmias in mouse isolated heart (Kim et al., 2012). In support, pinacidil shortens atrial AP duration and increases arrhythmia inducibility in human right atrial and right ventricular wall (Fedorov et al., 2011). Furthermore, atrial electrical remodelling and increased arrhythmia inducibility in a murine model of salt-induced hypertension have been shown to be associated with increased KATP current and SUR1 expression (Lader et al., 2011). A recent screening of patients with atrial fibrillation found two patients with the KIR6.1-S422L variant (Delaney et al., 2012).
Heart failure, hypertrophy and cell swelling
Cardiac hypertrophy is triggered by a prolonged increase in cardiac workload. When transverse aortic constriction was applied in KIR6.2 null mice or in mice with cardiac specific overexpression of SUR1 which paradoxically disrupts cardiac sarcolemmal KATP channel function, increased left ventricular hypertrophy was observed (Yamada et al., 2006; Hu et al., 2008). Interestingly, there seems to be an interaction between cardiac KATP channel expression and the activity of the PPAR-γ coactivator, PGC-1α. Decreased channel function leads to decreased activity at the PGC-1α promoter partly via FOXO-1 repression (Hu et al., 2008). Remodelled ventricular cardiomyocytes from rats subjected to coronary occlusion show up-regulation of KIR6.1, especially around the infarct zone (Isidoro et al., 2007). Congestive heart failure or infarction in human hearts leads to an increased AP duration and sensitivity to potassium channel openers in both atria and ventricles (Fedorov et al., 2011). There are a variety of other molecules/enzymes that have been shown to be involved in the progression from compensated hypertrophy to heart failure possibly by their interactions with KATP channels. Angiotensin II and TNF-α expression is positively correlated to that of KIR6.1 in failing rat myocardium or cultured cardiomyocytes and negatively correlated with KIR6.2 (Isidoro et al., 2009). Furthermore, cardiomyocytes treated with these were responsive to diazoxide, indicating increased expression of KIR6.1/SUR2B in these cells as part of the progression to hypertrophy (Isidoro et al., 2007; 2009).
Various mutations have been identified within KATP channel subunits which confer susceptibility to cardiomyopathy, hypertrophy and heart failure. A cohort of patients with dilated cardiomyopathy allowed the identification of a frameshift mutation leading to a premature stop codon at Leu1524 and a missense mutation A1513T in SUR2A. Both mutations are located in the NBD2 and compromise the ability of ATP to be hydrolysed (Bienengraeber et al., 2004). In KIR6.2, a non-synonymous polymorphism leading to the coding change, E23K, was identified in 18% of heart failure patients (Reyes et al., 2009) and is also known to lead to an increased risk of type 2 diabetes (Gloyn et al., 2003). Both heterozygous and homozygous patients have the same resting heart rates and show similar degrees of left ventricular dysfunction and remodelling. When these homozygous patients are exercised, they show a reduced heart rate, oxygen consumption and peak VO2. Other missense mutations in SUR2A within transmembrane domains have also been identified to cause activation of the channel in the rare Cantu syndrome, characterised by cardiac hypertrophy and cardiomegaly (Harakalova et al., 2012). Although the human and murine experimental data are very different in nature, there seems to be inconsistency in that both reduction and increases in KATP channel activity, can result in cardiac hypertrophy.
Excessive changes in cell volume in the heart can result in the alteration of the structural integrity of the cells affecting cellular functions and cell death. These changes can arise as a result of an intracellular accumulation of metabolites that increase cellular osmolality, allow water to enter the cell, increase the cell volume and alter ion channel function. Reduction of cardiomyocyte swelling during myocardial ischaemia may be a potential mechanism of cardioprotection (Shi et al., 2009). KATP has been shown to be regulated during cell volume changes with atrial KATP channels opening in response to cell swelling leading to AP shortening (Saegusa et al., 2005). The absence of KIR6.2 in cardiac myocytes isolated from KIR6.2 null mice prevents cell swelling from occurring whereas in WT mice there is exaggerated cell swelling which can be disrupted by the addition of diazoxide (Prasad et al., 2006). The use of diazoxide in some of the studies appears to be more complicated. Although cell swelling could be diminished by the addition of diazoxide, the addition of HMR1098 and 5-hydroxydecanoate did not reverse the events initially suggesting that diazoxide may be acting via a mechanism separate from the activation of KATP channels (Maffit et al., 2012).
Vascular reactivity and hypertension
The modulation of VSM KATP currents by vasoactive agents suggests the channel may be important for blood pressure control. In VSM cells from KIR6.1 and SUR2 null mice KATP currents were absent, whereas cells from KIR6.2 null mice exhibited normal KATP currents (Suzuki et al., 2001; Chutkow et al., 2002; Miki et al., 2002). As well as providing direct evidence for the molecular composition of the VSM KATP channel, KIR6.1 and SUR2 null mice exhibit hyper-contractility of the coronary vasculature and are prone to early sudden death due to coronary artery spasm (Chutkow et al., 2002; Miki et al., 2002). SUR2 null mice also show focal narrowing of the coronary arteries and have significantly elevated blood pressure (Chutkow et al., 2002). Interestingly, restoration of the vascular KATP channel in SUR2 null mice does not protect against a rise in baseline coronary artery perfusion pressure suggesting a role for KATP channels from tissues other than VSM (Kakkar et al., 2006). Specifically, SUR2B RNA has been detected in endothelium and it is believed that heteromeric KIR6.1/ KIR6.2 in combination with SUR2B could form an endothelial KATP channel (Yoshida et al., 2004). In mice expressing endothelium-specific dominant-negative KIR6.1 subunits, basal coronary perfusion pressure and ET-1 concentrations were substantially elevated suggesting a role for endothelial KATP channels in the regulation of vascular tone (Malester et al., 2007).
Further evidence for the role of VSM KATP channels in vascular tone regulation comes from studies of hypertensive animal models where there is substantial remodelling of KATP channels in vascular beds (Blanco-Rivero et al., 2008; Tajada et al., 2012). KATP channels in hypertensive phenotypes show altered vascular reactivity probably as a result of impaired and fewer KATP channels. VSM cells from hypertensive animals are significantly depolarized and KCO compounds have little effect on membrane potential compared with normotensive animals (Tajada et al., 2012). The KCO iptakalim has been put forward as a promising anti-hypertensive agent for mild to moderate essential hypertension (Sikka et al., 2012). Interestingly, on a mechanistic level iptakalim has also been shown to inhibit ET-1 release and synthesis and increase NO release and NOS activity in aortic endothelial cells (Gao et al., 2009). Additionally, there are data to suggest that KCO compounds such as iptakalim have therapeutic potential as a treatment for pulmonary hypertension (Sikka et al., 2012).
The role of KATP channels in sepsis is complex (Buckley et al., 2006). There is pharmacological evidence that channel activation occurs in septic shock leading to hypotension and this can be reversed by glibenclamide (Matsuda and Hattori, 2007). There is also evidence for increased activity of these channels but that their pharmacology alters such that they becomes unresponsive to sulphonylureas and only direct pore blockers can inhibit activity (O'Brien et al., 2009), and this is consistent with the lack of clinical efficacy (Warrillow et al., 2006; Morelli et al., 2007). However, animals, and also flies, with global genetic deletion of the channel are predisposed to an early and substantial survival disadvantage in sepsis (Kane et al., 2006; Croker et al., 2007). Furthermore, the expression of KIR6.1 is regulated via Toll-like receptors and NF-κB and the increase in expression of the current is postulated to underlie the poor response to vasoconstrictors in septic shock (Shi et al., 2010). The exact mechanism for the survival disadvantage is unclear but inappropriate coronary artery vasoconstriction during increased cardiac demand is one proposal. However, the pathophysiological circulatory changes in severe sepsis are actually profound and widespread. These include hypotension, hyporesponsiveness to vasoconstrictors, microvascular dysfunction, endothelial dysfunction, and increased vascular and capillary permeability (Matsuda and Hattori, 2007). The absence of KIR6.1 in both smooth muscle and endothelium may promote these adaptations.
The physiological role of KATP channels is well defined in the pancreatic beta cell. Recent work has begun to reveal similar pathophysiological importance in the function of cardiac muscle, specialized conduction tissues in the heart and of VSM (Figure 2). These channels have a rich existing pharmacology that could be exploited to develop novel therapeutic agents for the treatment of cardiovascular disease.
This work was supported by the British Heart Foundation, Wellcome Trust, Medical Research Council and the National Institute for Health Research Barts Cardiovascular Biomedical Research.
Conflict of interest
The authors have no conflicts of interest to declare.