Gene targeting approach to clarification of ion channel function: studies of Kir6.x null mice

Abstract

ATP-sensitive potassium (K_ATP) channels are present in many tissues, including pancreatic β-cells, heart, skeletal muscle, vascular smooth muscle and brain, in which they couple the cell metabolic state to membrane potential. K_ATP channels are hetero-octameric proteins composed of the pore-forming subunits Kir6.x (Kir6.1 or Kir6.2) of the inwardly rectifying K⁺ channel family and the regulatory subunits SURx (SUR1, SUR2A or SUR2B), the receptor of the sulphonylureas widely used in treatment of type 2 diabetes mellitus. Different combinations of Kir6.x and SURx comprise K_ATP channels with distinct electrophysiological and pharmacological properties, but their physiological functions in the various tissues are unclear. Our studies of Kir6.2 null (knockout) and Kir6.1 null mice have shown that K_ATP channels are critical metabolic sensors in protection against acute metabolic stress such as hyperglycaemia, hypoglycaemia, ischaemia and hypoxia.

Introduction

ATP-sensitive K⁺ (K_ATP) channels are found in many tissues, including heart, pancreatic β-cells, brain, skeletal and smooth muscles, and kidney (Ashcroft, 1988). K_ATP channels are inhibited by intracellular ATP and activated by MgADP, and couple the metabolic state of the cell to its membrane potential by sensing changes in the intracellular adenine nucleotide concentration (Ashcroft, 1988). K_ATP channels play especially important roles in the cellular responses of tissues under metabolic stress such as hyperglycaemia, hypoglycaemia, ischaemia and hypoxia (Yokoshiki et al. 1998). The K_ATP channels in pancreatic β-cells regulate glucose-induced insulin secretion. Increase in the ATP concentration due to increased glucose metabolism closes the K_ATP channels, depolarizing the β-cell membrane, opening the voltage-dependent Ca²⁺ channels (VDCCs) and allowing Ca²⁺ influx. The rise in intracellular Ca²⁺ concentration ([Ca²⁺]_i) in the β-cell triggers exocytosis of insulin-containing granules. Sulphonylureas, widely used in treatment of type 2 diabetes mellitus, stimulate insulin release by closing the K_ATP channels directly (Ashcroft, 1988). In heart, K_ATP channels are involved in increased K⁺ efflux and shortened action potential, both associated with induction of arrhythmias (Terzic et al. 1995). Activation of K_ATP channels in the heart during ischaemia is thought to minimize cardiac damage by ‘ischaemic preconditioning’ (Grover et al. 1992). In the vascular system, K_ATP channels regulate the tonus of vascular smooth muscles, playing an important role in blood pressure regulation (Quayle et al. 1997). In coronary artery, K_ATP channels are involved in vasodilatation during ischaemia (Quayle et al. 1997). In brain, activation of K_ATP channels during metabolic stress protects neurones from damage (Heurteaux et al. 1995).

It is known that different combinations of Kir6.1 or Kir6.2 and SUR1 or SUR2 (SUR2A and SUR2B) constitute K_ATP channels with distinct properties in various tissues (Ashcroft & Gribble, 1998; Aguilar-Bryan et al. 1998; Seino, 1999). We have analysed the functional roles of K_ATP channels by disrupting the genes encoding the pore-forming subunits. To date, there are four K_ATP channel null mice (Table 1). We have generated both a Kir6.2 knockout mouse and a Kir6.1 knockout mouse. Studies of these mice show that K_ATP channels are especially important in protection of cells against acute metabolic stress.

Table 1.

Types of K_ATP channel

Subunit composition	Type	References
Kir6.2/SUR1	Pancreatic β-cell	Inagaki et al. 1995b;
		Sakura et al. 1995
Kir6.2/SUR2A	Cardiomyocyte	Inagaki et al. 1996
Kir6.2/SUR2B	Smooth muscle	Isomoto et al. 1996
Kir6.1/SUR2B	Vascular smooth muscle	Yamada et al. 1997

Structure and function of K_ATP channels

K_ATP channels are hetero-octameric proteins composed of two subunits, the pore-forming Kir6.x (Kir6.1 or Kir6.2) subunit and the regulatory SUR (SUR1 or SUR2) subunit, receptors of the sulphonylureas widely used in treatment of type 2 diabetes mellitus (Ashcroft & Gribble, 1998; Aguilar-Bryan et al. 1998; Seino, 1999). Kir6.1 (Inagaki et al. 1995a) and Kir6.2 (Inagaki et al. 1995b; Sakura et al. 1995) are members of the family of inward rectifier K⁺ channels having two transmembrane domains. There is about 71% amino acid identity between Kir6.1 and Kir6.2 (Inagaki et al. 1995b). Although the Gly-Tyr-Gly motif in the H5 region, which is thought to be critical in K⁺ ion selectivity, is highly conserved in inwardly rectifying K⁺ channels, the amino acid sequence in both Kir6.2 and Kir6.1 is Gly-Phe-Gly, suggesting that the motif is unique to Kir6.x members. Sulphonylurea receptors belong to the ATP-binding cassette (ABC) protein superfamily. There are two isoforms of the sulphonylurea receptor, SUR1 and SUR2, which are derived from two different genes. In addition, there are several variants of SUR2A produced by alternative splicing (Isomoto et al. 1996; Chutkow et al. 1996; Chutkow et al. 1999), the major one being SUR2B. SUR2A and SUR2B differ by only 42 amino acids in the C-terminus due to alternative splicing (Isomoto et al. 1996). The C-terminus of SUR2B is similar to that of SUR1. While SUR1 is expressed at high levels in pancreatic islets, SUR2A is expressed predominantly in heart and skeletal muscle (Inagaki et al. 1996; Isomoto et al. 1996), and SUR2B is ubiquitously expressed, as assessed by reverse transcription-polymerase chain reaction (RT-PCR) assay (Isomoto et al. 1996). Heterologous expression of Kir6.x and SUR subunits in various combinations reconstitutes K_ATP channels with electrophysiological properties and nucleotide and pharmacological sensitivities reflecting the composition of the various K_ATP channels in native tissues. Kir6.2 and SUR1 constitute the pancreatic β-cell K_ATP channel (Inagaki et al. 1995b); Kir6.2 and SUR2A constitute the cardiac K_ATP channel (Inagaki et al. 1996); and Kir6.2 and SUR2B constitute non-vascular smooth muscle K_ATP channels (Yamada et al. 1997). Kir6.1 and SUR2B constitute the vascular smooth muscle K_ATP channel (K_NDP), which is somewhat insensitive to ATP, activated by nucleoside diphosphates, and inhibited by glibenclamide (Yamada et al. 1997) (Table 2).

Table 2.

K_ATP channel knockout mice

Knockout mice	References
Kir6.2 knockout	Miki et al. 1998
SUR1 knockout	Seghers et al. 2000;
	Shiota et al. 2002
SUR2 knockout	Chutkow et al. 2001
Kir6.1 knockout	Miki et al. 2002a

K_ATP channel in pancreatic β-cells

The resting membrane potential of Kir6.2 knockout β-cells (at 2.8 mm glucose) is significantly higher than that of wild-type cells, and repetitive bursts of action potential are found frequently even at low glucose (2.8 mm). In contrast to wild-type β-cells, the basal level of [Ca²⁺]_i is elevated significantly and high glucose does not alter membrane potential in Kir6.2 knockout β-cells. However, neither high glucose (16.7 mm) nor tolbutamide (100 μm) elicits change in [Ca²⁺]_i. In contrast, acetylcholine or high K⁺ stimulation increases [Ca²⁺]_i to levels comparable to wild-type, suggesting that intracellular Ca²⁺ mobilization from inositol 1,4,5-trisphosphate (IP₃)-sensitive Ca²⁺ stores and Ca²⁺ influx through voltage-dependent Ca²⁺ channels are independent of K_ATP channel activity. Importantly, neither glucose (16.7 mm) nor the sulphonylurea tolbutamide (100 μm) elicits significant insulin secretion in Kir6.2 knockout mice. SUR1 knockout mice also lack β-cell K_ATP channel activity, and the insulin secretory responses to glucose and sulphonylureas are markedly impaired (Seghers et al. 2000; Shiota et al. 2002). These findings show clearly that pancreatic β-cell K_ATP channel activity is critical in the regulation of insulin secretion by glucose and sulphonylureas.

K_ATP channel in hypothalamus

Recovery from insulin-induced systemic hypoglycaemia is severely impaired in Kir6.2 knockout mice, suggesting that secretion of counter-regulatory hormones such as glucagon and catecholamines is impaired. While adrenaline secretion in response to insulin-induced hypoglycaemia in Kir6.2 knockout mice is similar to wild-type in vivo, glucagon secretion is markedly reduced. Glucagon secretion from isolated pancreatic islets in response to change from high (16.7 mm) to low (1 mm) glucose concentration is similar in Kir6.2 knockout and wild-type mice. In addition, the glucagon response to carbachol (50 μm), a synthetic choline ester, is unimpaired and somewhat enhanced in Kir6.2 knockout mice compared to wild-type mice. These findings indicate that the primary defect in glucagon secretion in Kir6.2 knockout mice is upstream of the α-cells. Neuroglycopaenia is known to stimulate glucagon secretion through activation of autonomic neurones in the brain. 2-Deoxyglucose (2DG) induces neuroglycopenia in the hypothalamus, thereby stimulating glucagon secretion (Borg et al. 1995; Borg et al. 1999). In contrast to wild-type, there is almost no glucagon secretion in response to intracerebroventricularly injected 2DG in Kir6.2 knockout mice. In the ventromedial hypothalamus (VMH), a subset of neurones known as glucose-responsive neurones (GRNs) increase their firing rate in response to elevated extracellular glucose (Oomura et al. 1969). Wild-type VMH neurones show a twofold increase in spontaneous discharge rate in response to high glucose (25 mm), while no Kir6.2 knockout VMH neurones display any change in spontaneous discharge rate. In addition, Kir6.2 knockout VMH neurones have a higher discharge rate at low glucose than wild-type. These findings show that glucose responsiveness in VMH neurones is lost in Kir6.2 knockout mice. In contrast to wild-type, dialysis with an ATP-free solution does not activate K⁺ currents in Kir6.2 knockout VMH neurones. These findings show that Kir6.2-contaning K_ATP channels are essential for glucose responsiveness in GRNs. Single cell RT-PCR analysis demonstrates that VMH K_ATP channels include Kir6.2 and SUR1, the same composition as the β-cell K_ATP channel.

Based on these findings in Kir6.2 knockout mice, we propose a model for the maintenance of glucose homeostasis by K_ATP channels (Miki et al. 2001). As the blood glucose level rises, inhibition of the K_ATP channels in pancreatic β-cells induces insulin secretion, lowering the glucose level. As the blood glucose level falls, activation of the K_ATP channels in the GRNs of the hypothalamus triggers glucagon secretion, most likely through stimulation of autonomic input to the pancreatic α-cells (Taborsky et al. 1998). The mechanism by which the decreased firing rate of GRNs stimulates glucagon secretion remains unknown, however. Thus, the K_ATP channels of the pancreatic β-cells and the hypothalamus share the same molecular composition and act in concert in the maintenance of glucose homeostasis. In addition to stimulating glucagon secretion, 2DG is known to increase food intake in normal mice, presumably by inhibiting activity of GRNs in the hypothalamus (Bergen et al. 1996). The increment in food intake induced by 2DG is significantly less in Kir6.2 knockout mice than in wild-type, suggesting a functional role for hypothalamic K_ATP channels in control of appetite.

K_ATP channel in skeletal muscles

Certain sulphonylureas have been shown to improve glycaemic control in patients with type 2 diabetes mellitus by acting on extra-pancreatic tissues to reduce insulin resistance (Wang et al. 1989), suggesting that the K_ATP channel in skeletal muscles is involved in glucose uptake. Although it has been shown that K_ATP channel activity in skeletal muscle is important in preventing resting tension during fatigue (Gong et al. 2000), there is no evidence of their involvement in the extra-pancreatic effects of sulphonylureas. Interestingly, the glucose-lowering effect of insulin was found to be enhanced in Kir6.2 knockout mice (Miki et al. 1998), indicating that insulin sensitivity is increased. In vitro measurement of glucose uptake in Kir6.2 knockout muscles shows that disruption of the channel increases basal uptake in extensor digitorum longus (EDL) muscle and insulin-stimulated glucose uptake in soleus muscle, demonstrating that the effect of K_ATP channel disruption differs in muscle types (Miki et al. 2002b). K_ATP channel involvement in glucose uptake also is shown in SUR2 knockout mice (Chutkow et al. 2001). In vitro insulin-stimulated glucose transport is 1.5-fold greater in SUR2 knockout muscle than in wild-type. These findings strongly suggest that K_ATP channel activity is involved in glucose uptake in skeletal muscle. Whether or not K_ATP channel activity in skeletal muscles affects insulin signalling remains to be determined.

K_ATP channel in substantia nigra (SNr)

As the SNr acts as a central gating system in the propagation of seizure (Iadarola & Gale, 1982) and generalized seizure can be evoked by metabolic stress such as hypoxia and hypoglycaemia, K_ATP channels might well be involved in seizure propagation during hypoxia. A study of Kir6.2 knockout mice has found different responses to brief (150 s) hypoxia due to oxygen deprivation in Kir6.2 knockout and wild-type mice (Yamada et al. 2001). Wild-type mice all remain sedated during the challenge and revive normally. In contrast, Kir6.2 knockout mice all respond with a myoclonic jerk followed by severe tonic-chronic convulsion and death. Electroencephalogram (EEG) and electromyogram (EMG) reveal a sequence of seizure patterns in Kir6.2 knockout mice during hypoxic (5.4%O₂) challenge that does not occur in wild-type mice, suggesting K_ATP channel participation in the threshold of seizure. Wild-type SNr neurones, when perfused with hypoxic solution, become hyperpolarized, which is reversed by tolbutamide. Diazoxide also produces hyperpolarization. In contrast, Kir6.2 knockout neurones show no such hyperpolarization but are depolarized. In addition, in acute brain slice preparations, activation of postsynaptic K_ATP channels was found to be critical in hypoxia-induced inactivation of SNr neurones. These findings suggest that hyperpolarization of membrane potential due to opening of the K_ATP channels in SNr neurones prevents seizure propagation during hyoxia (Depaulis et al. 1994), but the mechanism is unknown.

K_ATP channel in cardiomyocytes

Brief, intermittent ischaemia paradoxically protects myocardium against prolonged ischaemic insult, producing a marked reduction in infarct size, a phenomenon known as ischaemic preconditioning (IPC) (Cohen et al. 2000). Infarct size in wild-type mice with IPC (three cycles of 3 min of coronary occlusion before long-term occlusion) is significantly less than without IPC. However, there is no significant difference in infarct size between Kir6.2 knockout mice with or without IPC (Suzuki et al. 2002). In in vitro experiments using Langendorff-perfused hearts following global ischaemia–reperfusion, increase of left ventricular end-diastolic pressure (LVEDP) during ischaemia is more marked in Kir6.2 knockout mice than in wild-type, while recovery of contractile function in Kir6.2 knockout mice is less marked. Treatment with HMR1098, a sarcolemmal K_ATP channel blocker, but not with 5-hydroxydecanoate, a putative mitochondrial K_ATP channel blocker, produces deterioration of contractile function in hearts of wild-type mice comparable to that in Kir6.2 knockout mice. These findings show that sarcolemmal K_ATP channels in the heart are involved in protection against ischaemic injury.

K_ATP channel in vascular smooth muscle

Channels reconstituted from Kir6.1 and SUR2B are activated by nucleoside diphosphates such as UDP and inhibited by the sulphonylurea glibenclamide, properties similar to K_ATP channels in native vascular smooth muscle (Quayle et al. 1997), also called the K_NDP channel (Yamada et al. 1997). Kir6.1 null mice are prone to sudden, premature death. All the knockout mice examined showed spontaneous elevation of ST segments followed by atrioventricular (AV) blocks of various degrees, indicating that death is due to myocardial ischaemia (Miki et al. 2002a). In situ hybridization detects Kir6.1 mRNA widely in heart, but most clearly in cardiomyocytes of wild-type mice. In addition, Kir6.1 mRNA is present in vascular smooth muscles of coronary arteries of wild-type mice. By contrast, no Kir6.1 mRNA is detected in cardiomyocytes or coronary arteries of Kir6.1 knockout mice. In both wild-type and Kir6.1 knockout mice, pinacidil induces outward currents in cardiomycytes that are blocked by glibenclamide, indicating that Kir6.1 is not a component of the sarcolemmal K_ATP channel. Intravenous injection of pinacidil decreases the mean arterial pressure significantly in control mice but not in Kir6.1 knockout mice, indicating lack of vasodilatation in response to pinacidil in Kir6.1 knockout mice. The vasodilatation response of aortae to pinacidil in Kir6.1 knockout mice is also reduced in vitro, as assessed by changes in isometric tension of aortic rings. Pinacidil elicits significant K⁺ currents that are blocked by glibenclamide in aortic smooth muscle cells in wild-type mice, but fails to evoke significant K⁺ currents in knockout mice, indicating that Kir6.1 is a component of vascular smooth muscle K_ATP channels. Vasospasms occur in these knockout mice that directly trigger vasoconstriction of vascular smooth muscles. These findings suggest that the Kir6.1 knockout phenotype resembles Prinzmetal (variant or vasospastic) angina in human (Prinzmetal et al. 1959). It has been shown that disruption of SUR2 also produces the phenotype of Prinzmetal angina (Chutkow et al. 2002). Although vasospastic angina is rare in Caucasians, many reports indicate that it is more common in Japanese (Yasue & Kugiyama, 1997; Beltrame et al. 1999), suggesting that Kir6.1 and SUR2 may be associated with Prinzmetal angina only in certain ethnic populations.

Conclusions

K_ATP channels are critical metabolic sensors in acute metabolic stress such as hyperglycaemia, hypoglycaemia, ischaemia and hypoxia (Table 3). Tissue-specific disruption of each of the K_ATP channel subunits (conditional knockout) is required to determine their specific roles in the various tissues. K_ATP channel knockout mice are especially useful in clarifying physiological and pathophysiological roles of K_ATP channels and for development of novel drugs targeting K_ATP channels.

Table 3.

Physiological roles of Kir6.2-containing and Kir6.1-containing K_ATP channels

Kir6.2-containing KATP channels

Regulation of glucose-induced and sulphonylurea-induced

insulin secretion from pancreatic β-cells

Regulation of glucagon secretion by glucose-responsive

neurones in the hypothalamus during hypoglycemia

Prevention of hypoxia-induced generalized seizure

Protection of skeletal muscles against fatigue

Protection of cardiac function against adrenergic stress*

Ischaemic preconditioning in heart

Kir6.1-containing KATP channels

Regulation of vascular tonus and prevention of

coronary artery spasm

^*Zingmanet al.2002.