|Home | About | Journals | Submit | Contact Us | Français|
Reconstitution of KATP channel activity from coexpression of members of the pore-forming inward rectifier gene family (Kir6.1, KCNJ8, and Kir6.2 KCNJ11) with sulfonylurea receptors (SUR1, ABCC8, and SUR2, ABCC9) of the ABCC protein sub-family, has led to the elucidation of many details of channel gating and pore properties, as well as the essential roles of Kir6.2 and SUR2 subunits in generating cardiac ventricular KATP. However, despite this extensive body of knowledge, there remain significant holes in our understanding of the physiological role of the cardiac KATP channel, and surprising new findings keep emerging. Recent findings from genetically modified animals include the apparent insensitivity of cardiac sarcolemmal channels to nucleotide levels, and unenvisioned complexities of the subunit makeup of the cardiac channels. This topical review focuses on these new findings and considers their implications.
From earlier indications of a K conductance activated by metabolic inhibition, the tale of the cardiac KATP channel really began in 1983 with Akinori Noma's first description of “specific K+ channels which are depressed by intracellular ATP (ATPi) at levels greater than 1 mM”. Because this KATP channel is gated directly by intracellular ATP and ADP, it is therefore a strong candidate for coupling the metabolic state of the cell with its electrical activity and hence contractility. Reconstitution of KATP channel activity by coexpression of members of the pore-forming inward rectifier gene family (Kir6.1, KCNJ8, and Kir6.2 KCNJ11) with sulfonylurea receptors (SUR1, ABCC8, and SUR2, ABCC9) of the ABCC protein sub-family, has led to the elucidation of many details of channel gating and pore properties, as well as essential roles of Kir6.2 and SUR2 subunits in generating cardiac ventricular KATP and the detrimental consequences of knocking them out on the whole organism. However, despite this extensive body of knowledge, there remain significant holes in our understanding of the physiological role of the cardiac KATP channel, and surprising new findings keep emerging. Our purpose in this topical review is to focus on these new findings and to consider their implications.
Several reviews have covered in-depth the molecular basis of cardiac KATP channel activity and the accumulated understanding of nucleotide regulation of channel activity[2, 5, 6]. KATP channels are typically half-maximally inhibited by ~10-50 μM ATP, which acts by binding directly to the regulatory Kir6.2 subunit, with or without Mg2+. However, they are also activated by MgATP and MgADP, which are hydrolysed (MgATP) or maintain an activated post-hydrolytic state (MgADP) that counters ATP inhibition, and leads to the prediction that higher [ATP] (in the range of 0.1 – 1 μM) would be necessary to cause half-maximal inhibition under physiological conditions.
The discovery that application of the phospholipid PIP2 to the cytoplasmic side of inside-out patches causes a decrease in sensitivity to ATP[7, 8] demonstrates that nucleotide sensitivity is not fixed, and could be modulated by other cellular factors. This realization means that the underlying question of when KATP channels become active will not simply be a question of what is the cellular phosphorylation potential (i.e. the [ATP]:[ADP] ratio), but the integration of this with the signals that regulate channel responsivity. Phosphorylation of recombinant KATP channels by protein kinase A or C[9-11] modulates nucleotide sensitivity, as do fatty acyl CoA esters, and minute-to-minute modulation of lipid content in the cells may also regulate the channel in vivo. Interestingly, an early study of ATP inhibition of cardiac KATP channels revealed K1/2 values ranging from 9-580 μM in 102 individual excised patches, illustrating the very marked patch-to-patch variability that is present. This variability is likely to reflect at least in part the complex regulation of nucleotide sensitivity by ambient factors, including the level of PIP2 and other activatory lipids[7, 15]. However, it is also possible that the physical make-up of the individual channels varies, and that, as discussed below, there could be varying numbers of SUR subunits associated with each channel[16, 17].
In experiments where KATP was activated in isolated cardiac myocytes by anoxia, the duration of channel activity was short-lived declining in parallel with a fall in the levels of PIP and PIP2, suggesting that the levels of these two phospholipids act in concert with the intracellular nucleotides to control channel function. Such data illustrate the lability of nucleotide inhibition in the intact cell, but even the relevance of nucleotide sensitivity to physiological activation of the channels might be questioned in light of recent findings regarding the in vivo consequences of channel mutations. Many mutations in the pore-forming Kir6.2 subunit have now been identified as causal in human neonatal diabetes mellitus (NDM), a very severe form of diabetes that typically occurs within the first days or weeks of life[19, 20]. All of the identified Kir6.2 mutations result in a reduced channel sensitivity to ATP inhibition, in recombinant channels, leading to channel activation at elevated [glucose], maintained hyperpolarization of pancreatic islet β-cells, and electrical inexcitability, with consequent inhibition of insulin secretion. The same pore-forming Kir6.2 subunit is present in the heart and pancreas, and hence cardiac KATP channels should also be ATP-insensitive, yet there are no reports of any cardiac abnormalities in NDM patients. Moreover, in transgenic mice that express an ATP-insensitive Kir6.2 subunit (Kir6.2[ΔN30,K185Q]) in the heart, sarcolemmal KATP channels are extremely insensitive to ATP-dependent inhibition (K1/2 = 1.4 μM c.f. 25 μM in WT), yet still remain essentially closed in intact cells[21, 22]. As KATP channels have a higher density than other sarcolemmal K+ channels, opening of as few as 1% of KATP channels is expected to shorten cardiac action potential by about 50%, and multiple studies[23-27] predict that this dramatic reduction of ATP sensitivity should ensure that channels are active enough to significantly shorten the action potential in normal conditions, yet the action potential duration is unaffected. Clearly something other than nucleotide sensitivity is at play.
In order to probe the significance of channel composition in KATP function in cardiac myocytes, not only the ATP-insensitive Kir6.2 subunit (Kir6.2[ΔN30,K185Q], but also SUR2A or SUR1 subunits have been overexpressed under cardiac alpha-MHC control. In each case, cardiac function is only minimally affected; in the Kir6.2[ΔN30,K185Q] transgenics there is a very small increase in background KATP, although this is balanced by a ‘pre-stimulated’ Ca current[21, 22]. Conceivably this reflects an intrinsic compensatory mechanism; i.e. shortening of the action potential being compensated by enhanced Ca current, although underlying signaling processes are unknown. Overexpression of either SUR1 or SUR2A is without effect on the ECG, other than a consistent P-R prolongation in SUR1-overexpressing hearts. Surprisingly, overexpression of any single one of these subunits also significantly suppresses sarcolemmal KATP channel density[22, 28]. One possible explanation is that overexpressed SUR1, SUR2A or Kir6.2 subunits interact with endogenous subunits, and thereby disrupt the stoichiometry of the channel and then may not reach the plasma membrane[16, 30]. A simple resolution to this would seem to be that co-overexpression of both subunits (i.e. by crossing the two animals) should restore appropriate stoichiometry of expression, leading to a high density of ATP-insensitive channels in the sarcolemma. The dramatic result, however, is an even greater suppression of total channel density, and a whole constellation of arrhythmias, leading to sudden death. Not only does this surprising result tell us that altered KATP channels can have profound – but unexplained - effects on electrical activity, but that there is something very specific required for ‘correct’ expression of the two subunits. A clue may be found in the fact that endogenous Kir6 and SUR genes are immediately adjacent. Although perhaps heretical to suggest, it may be that common elements of gene regulation ensure that both are normally transcribed with temporal or spatial coordination, leading to correct assembly at a very early stage of synthesis, and that this coordination is absent when transgenes are exogenously expressed under alpha-MHC promoter control.
Another question is exactly which subunits are expressed in different regions of the heart. The pancreatic β-cell KATP channel is formed as a complex of four Kir6.2 subunits each associated with one SUR1 subunit[29, 32]. Together with studies of gene knockout mice (which show that Kir6.2 and SUR2 genes are essential for normal ventricular KATP currents, whereas KATP currents are unaffected in ventricular myocytes from SUR1 or Kir6.1 knockout animals), early studies of recombinant channel pharmacology[33, 34] led to the widely accepted notion that the sarcolemmal KATP channel is a heteromultimer of Kir6.2 and SUR2A, presumably with the same octameric arrangement. However, several studies have demonstrated that both Kir6 subunits (Kir6.1 and Kir6.2) and both SUR subunits (SUR1 and SUR2A) are expressed in the heart[35-37]. Dominant negative coexpression strategies demonstrate that Kir6.1 and Kir6.2 may assemble into functional channel complexes, and in some studies, dominant negative Kir6.1 subunits suppress sarcolemmal KATP currents. The idea that SUR2 is essential for ventricular KATP is also muddied by the finding that some KATP channels are still present in SUR2−/− myocytes. In this case, it is noteworthy that Kir6.2 channels have been shown to be present at the surface membrane in recombinant cells, in the complete absence of expressed SUR subunits[16, 17, 41], albeit at lower levels than are found with SUR subunits expressed. These channels are less sensitive to ATP inhibition than channels associated with SUR subunits, but lack Mg-nucleotide activation It remains a possibility that an octameric arrangement is not absolutely obligatory, and that variations in channel structure at this level may influence channel activation. However, antisense oligonucleotides specific for either SUR1 or SUR2A/B suppress KATP current in neonatal rat ventricular myocytes, which suggests that SUR1 might also participate in forming the channel, either alone or in conjunction with SUR2A, and recombinant channel studies demonstrate that within a single channel, more than one Kir6.x or SURx subunit can clearly co-exist[38, 43-46].
Definitive proof that SUR1 is a significant component of sarcolemmal KATP channels comes from recent study of SUR1 knockout (SUR1−/−) animals. A key finding is that while KATP currents appear normal in SUR1−/− ventricular myocytes, these currents are abolished in SUR1−/− atrial myocytes. Differential pharmacological properties of wild type atrial and ventricular KATP channels are consistent with atrial channels forming as a heteromultimer of SUR1 and Kir6.2, whereas ventricular KATP channel is a heteromultimer of SUR2A and Kir6.2. It is not yet clear how generalizable this finding is to other species; diazoxide activation of KATP in myocytes of other rodent species[37, 48] and in larger animals suggests that SUR1 and SUR2 may be involved in both ventricular and atrial channels. However, it does raise the important issue of differential pathophysiological roles and regulation in different regions of the heart that may need to be considered when choosing sulfonylurea therapy, as well as for understanding the role of diazoxide in preconditioning.
Under normal metabolic conditions, sarcolemmal KATP channels are predominantly closed, and they do not significantly contribute to cell excitability. However, these channels will open when exposed to a severe metabolic stress such as anoxia, metabolic inhibition or ischemia. By shortening the action potential, calcium entry is reduced, and consequently myocyte contraction fails. By reducing Ca2+ entry, the energy stores that would otherwise be used up in the contracting cell would be preserved and Ca2+ overload will be minimized, protecting the cell. Such a preservation ‘strategy’ is of course self-limiting, since if too many myocytes stop contracting, the heart stops pumping and the animal will die, but it seems a reasonable, if unproven, idea that temporary protection of a small number of cells, or region of the heart, against the damage of Ca-overload during ischemia, is likely to be operable in the heart.
In support of the idea that activation of sarcolemmal KATP is protective during ischemia, it has been demonstrated that KATP channel openers enhance the preservation of ATP and KATP channel openers produce anti-ischemic effects by shortening the action potential. Additional support comes from evidence that glibenclamide, a selective KATP channel inhibitor, abolishes the anti-ischemic effects by inhibiting action potential shortening. In Kir6.2−/− mice, sarcolemmal KATP channels are abolished, and KATP channels are reduced in density in rat cardiac myocytes transfected with a dominant-negative fragment of SUR2A. In both of these studies, ischemic cardioprotection is also abolished. In parallel, moderate overexpression of SUR2A reportedly increased sarcolemmal KATP channels and protected hearts against metabolic stress, including hypoxia and ischemia/reperfusion. The general idea that KATP activation provides cardioprotection against ischemic stress is further supported by recent evidence that Kir6.2−/− mice exhibit impaired response to systolic overload following chronic transverse aortic constriction.
However, two further recent studies seem to contradict a cardioprotective role. In these studies, from independent groups, both SUR2- (SUR2−/−) and SUR1-knockout (SUR1−/−) mice were found to be more tolerant of global ischemia-reperfusion than control mice, with reduced infarct sizes[58, 59]. Since the SUR2−/− mice have a great reduction of ventricular sarcolemmal KATP channels, the enhanced cardioprotection is opposite the expected phenotype (i.e. impaired protection). Cardioprotection in SUR2−/− mice might conceivably be due to the concomitant loss of the SUR2B component of vascular KATP channels, but similar cardioprotection in SUR1−/− mice could not be explained by such a mechanism. The SUR2−/− animals were generated by disruption of the first nucleotide-binding domain, which might allow translation of the preceding TM0 and TM1 or following TM2 domains of the protein. Pu et al. designed antibodies to epitopes in these regions and detected novel SUR2 short forms in the sizes of 28 and 68 kDa in both WT and SUR2−/− hearts. These findings raise yet more questions regarding the ultimate complexity of channel makeup.
Ischemic preconditioning refers to a phenomenon whereby brief ischemic periods improve the recovery of contractile function and reduce the size of the infarct that results from a subsequent prolonged ischemia. Preconditioning can be mimicked by adenosine or acetylcholine, as well as by KATP channel openers, such as diazoxide, and glibenclamide inhibits ischemic as well as adenosine- and acetylcholine-induced preconditioning, implicating KATP channels in the phenomenon. However, several lines of evidence have been used to argue that channels in the sarcolemma are not responsible: (1) preconditioning effects of KATP channel openers may not require shortening of the action potential[62, 63]; (2) the channel opener diazoxide mimics ischemic preconditioning, even though diazoxide is relatively ineffective at opening ventricular sarcolemmal KATP channels[47, 65]. On the other hand, diazoxide is reportedly effective at stimulating a so-called ‘mitoKATP’ channel that was identified in the mitochondria. Based on this, and the finding that (3) 5-hydroxydecanoic acid effectively abolishes ischemic preconditioning, and is reportedly specific for ‘mitoKATP’, without acting on sarcolemmal channels, the current dogma that preconditioning results from activation of ‘mitoKATP’ has arisen.
However, it is now clear that at least one central assumption is incorrect: diazoxide can open sarcolemmal KATP channels and mouse atrial sarcolemmal KATP channels are highly sensitive to diazoxide. Although diazoxide might have an action on mitochondrial succinate dehydrogenase or F1F0 ATPase function (which in turn could alter sarcolemmal function through nucleotide levels), the above findings require at the least that the use of diazoxide in preconditioning studies be interpreted with caution. Data from hearts with transgenically manipulated sarcolemmal KATP subunits also points to a role of sarcolemmal KATP in preconditioning: even though the effects of diazoxide on flavoprotein fluorescence are preserved, ischemic- and diazoxide-induced preconditioning are both abolished in hearts from Kir6.2−/− animals[54, 69] as well as from animals with transgenic overexpression of ATP-insensitive Kir6.2 subunits, and in rat cardiac myocytes transfected with a dominant-negative fragment of SUR2A. There is also little evidence that any of these sarcolemmal subunits are present in the mitochondria. More important, the existence of conventional KATP channels in mitochondria has been questioned, as has the specificity of commercial antibodies used to probe for such components. At this point it is probably fair to say that this muddied picture leaves open the possibility that sarcolemmal KATP channels are the relevant ’KATP players’ in preconditioning.
Four years ago, we summarized a topical review of sarcolemmal KATP channels with the statement that we remain ‘largely in the dark regarding the true physiological determinants, and relevance of sarcolemmal KATP activity’. In many ways, subsequent developments have only served to make us realize just how large a darkness we are in – ATP-sensitivity may not really be the key physiological regulator of channel activity, we now realize that the subunit make-up of sarcolemmal channels may be more complex and labile than we thought, and the existence of mitoKATP and the roles of sarcolemmal versus mitochondrial KATP channels in preconditioning are really not clear. At the same time, however, new studies of the whole animal consequences of manipulation of cardiac KATP subunits have continued to reveal dramatic roles in arrhythmia generation, and in the animal responsivity to cardiac and systemic stress. The questing tale of cardiac KATP channel function continues!
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.