K⁺ channels that are closed by ATP (K_ATP) were discovered by Aki Noma in the heart, and subsequently reported in many other cell types, including pancreatic β-cells. A paradigm for control of insulin secretion is that glucose metabolism elevates cytoplasmic [ATP]/[ADP] in pancreatic β-cells, closing K_ATP channels, and causing depolarization, Ca²⁺ entry, and insulin release. Sulfonylureas, which inhibit K_ATP channel activity, are in major clinical use for enhancing insulin secretion in diabetic patients, and persistent hypersecretion of insulin is seen in patients with mutations that decrease K_ATP activity. In the heart there is evidence for involvement of K_ATP in action potential (AP) shortening and response to ischaemia, but there remains no consensus of when, and where, cardiac K_ATP channels are active.

Susumu Seino, Joe Bryan and colleagues cloned genes encoding K_ATP in the mid-1990s; the channels are formed of pore-forming Kir6.2 subunits associated with regulatory sulfonylurea receptors (SUR1 in the pancreas, SUR2A in the heart). Each gene has been ‘knocked-out’ by recombinant approaches; knock-out of Kir6.2 or SUR1 impairs insulin secretion and abolishes glucose dependence, but neither fully reiterates the maintained hyperinsulinaemia that is observed in human HI patients. Knock-out of Kir6.2 or SUR2 genes abolishes AP shortening in myocardial ischaemia but does not give an indication of when the channels are likely to activate normally.

As an alternative approach to probing K_ATP function, we are examining the consequences of transgenic expression of mutant channels. Decreased ATP sensitivity is predicted to cause decreased insulin secretion and perhaps diabetes in the pancreas, and to cause AP shortening, and perhaps inexcitability in the heart. We generated transgenic mice expressing mutant Kir6.2 subunits, in pancreatic β-cells under insulin gene promoter (RIP1) control in β-cells, and under α-myosin heavy chain (MHC) promoter control in heart. The constructs we used were Kir6.2[DN] and Kir6.2[DN,K185Q], which generate K_ATP channels that are respectively ~10-fold and ~300-fold less sensitive to ATP inhibition when expressed with SUR1 subunits in COSm6 cells.

Under RIP1-control, we obtained no Kir6.2[DN,K185Q] founders, but five founders expressing the Kir6.2[DN] construct. F1 mice all showed the same dramatic phenotype: they developed severe hyperglycaemia, hypoinsulinaemia and ketoacidosis within 2 days, and typically died within 5 days. Islet morphology, insulin localization, and α- and β-cell distributions were normal (< day 3), pointing to reduced insulin secretion as the single causal mechanism. The data indicate that normal K_ATP channel activity is critical for maintenance of euglycaemia, and that even very minimal overactivity may cause diabetes by inhibiting insulin secretion and in accordance, Matthias Schwanstecher and colleagues have recently reported that even a 2-fold reduction of ATP sensitivity resulting from the E23K mutation in human Kir6.2 may be causal in Type II diabetes.

In contrast, viable founder mice expressing both Kir6.2[DN,K185Q]-GFP and Kir6.2[DN]-GFP in the heart were obtained. K_ATP channels from Kir6.2[DN,K185Q]-GFP transgenic myocytes exhibit a spectrum of ATP sensitivities, due to heteromultimerizations of endogenous and transgenic subunits, but are on average very ATP insensitive. In the highest expressing (line 4) myocytes, ATP sensitivity is reduced almost two orders of magnitude from ~50 mM to about 4 mM! Many studies indicate that given the high density of K_ATP conductance in heart cells, only ~1% of maximal channel activity would be necessary to cause 50% shortening of the AP. Severe AP shortening would thus be expected in these transgenic mice, but none is observed, either in isolated cells or intact hearts. The immediate reason is clear enough – despite being very ATP-insensitive, the channels are not open in the intact cell. But why not? At least a part of the answer may lie in the different SUR isoform found in the heart – SUR2A subunits may fail to activate cardiac channels. There are additional confusing consequences of transgene expression. Heart rate is reduced in Kir6.2[DN,K185Q] transgenic mice, but contractility is elevated in isolated hearts. Why is this? A slight AP prolongation during mid-repolarization correlates with enhanced L-type Ca²⁺ current in these myocytes. Is enhanced Ca²⁺ current driving enhanced contractility in the face of reduced heart rate? Or is enhanced contractility, secondary to enhanced Ca²⁺ current, driving down the heart rate?

In summary, both knock-out of K_ATP channels, and transgenic expression of mutant channels, has generated dramatic phenotypic changes in cardiac and β-cells. As will be discussed, the results have confirmed or enhanced some paradigms, but there are some surprising and as yet unexplained phenomena to be considered.