Roles of ATP-Sensitive K+ Channels as Metabolic Sensors: Studies of Kir6.x Null Mice

Kir6.2 and SUR1 both are expressed in all endocrine cells in pancreatic islets (15,16). Because Kir6.2 subunits form the K⁺ ion-selective pore of the channel (3), mice lacking K_ATP channel function can be generated by genetic disruption of Kir6.2 (17). Both whole-cell and single-cell recordings show K_ATP channel activity to be completely absent in the pancreatic β-cells of Kir6.2 null mice, demonstrating that the Kir6.2 subunits are essential for β-cell K_ATP channel function. Although Kir6.2 null mice show transient hypoglycemia as neonates, blood glucose levels of Kir6.2 null mice as adults are not significantly different from those of wild-type mice in the nonfasting state. The insulin response to glucose in vivo is impaired in Kir6.2 null mice, as evaluated by an intraperitoneal glucose tolerance test. The resting membrane potential of Kir6.2 null β-cells is significantly higher than in wild-type β-cells. In the presence of low glucose (2.8 mmol/l), repetitive bursts of action potential were frequently found in Kir6.2 null β-cells. In contrast to control β-cells, high glucose did not alter the membrane potential of Kir6.2 null β-cells. The basal levels of [Ca²⁺]_i in single β-cells were significantly elevated in Kir6.2 null β-cells. However, neither high glucose nor tolbutamide elicited any change in [Ca²⁺]_i in Kir6.2 null β-cells. Thus, the rise in normal pancreatic β-cells of [Ca²⁺]_i elicited by both glucose and tolbutamide requires closure of the K_ATP channels, and the K_ATP channels in pancreatic β-cells are crucial in determining both membrane potential and Ca²⁺ influx. In contrast, acetylcholine or high K⁺ stimulation increased [Ca²⁺]_i in Kir6.2 null β-cells to levels comparable to those in control β-cells, suggesting that intracellular calcium mobilization from inositol 1,4,5-triphosphate-sensitive Ca²⁺ stores and Ca²⁺ influx through VDCCs are independent of K_ATP channel activity. The insulin secretory responses to glucose and tolbutamide were determined using both batch incubation and perifusion of pancreatic islets of adult mice (17). Neither glucose nor tolbutamide elicited significant insulin secretion in Kir6.2 null mice, demonstrating that the K_ATP channels in β-cells are required in both glucose- and sulfonylurea-induced insulin secretion.

Kir6.2 null mice show markedly delayed recovery from insulin-induced systemic hypoglycemia, suggesting impaired secretion of counterregulatory hormones such as glucagon and catecholamines. While epinephrine secretion in response to insulin-induced hypoglycemia in Kir6.2 null mice is similar to that in wild-type in vivo, glucagon secretion is markedly reduced (18). Glucagon secretion from isolated pancreatic islets in response to change from high to low glucose concentration is similar in Kir6.2 null and wild-type mice. In addition, glucagon response to carbachol, a synthetic choline ester, is somewhat enhanced in Kir6.2 null mice compared with wild-type mice. These findings indicate that the primary defect in glucagon secretion in Kir6.2 null mice is upstream of the α-cells. Neuroglycopenia is known to stimulate glucagon secretion through activation of autonomic neurons in the brain. 2-Deoxyglucose (2DG) induces neuroglycopenia in the hypothalamus, thereby stimulating glucagon secretion (19). In contrast to wild-type mice, there is almost no glucagon secretion in response to intracerebroventricularly injected 2DG in Kir6.2 null mice. In the ventromedial hypothalamus (VMH), a subset of neurons known as glucose-responsive neurons (GRNs) increase their firing rate in response to elevated extracellular glucose (20). Wild-type VMH neurons show a twofold increase in spontaneous discharge rate in response to high glucose, while Kir6.2 null VMH neurons display no change in spontaneous discharge rate. Kir6.2 null VMH neurons show higher discharge rate at low glucose than wild-type neurons. In addition, dialysis with an ATP-free solution does not activate K⁺ currents in Kir6.2 null VMH neurons. These findings show that Kir6.2-containing K_ATP channels are required for glucose responsiveness of GRNs. Single-cell RT-PCR analysis demonstrates that VMH K_ATP channels comprise Kir6.2 and SUR1, the same composition as the β-cell K_ATP channel. Accordingly, Kir6.2/SUR1 K_ATP channels in both pancreatic β-cell and hypothalamus participate in the control of blood glucose level (18). As the blood glucose level rises, inhibition of the K_ATP channels in pancreatic β-cells induces insulin secretion, thereby 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 (21), thereby raising the glucose level. In addition to their involvement in glucagon secretion, hypothalamic K_ATP channels also are involved in acute food intake (18).

K_ATP channels also are present in skeletal muscles (22). The pharmacological and electrophysiological properties of the channel have been examined in native skeletal muscles isolated from several vertebrates (23–25). While the unitary conductance in these muscles is similar, the electrophysiological properties of the K_ATP channels recorded in native skeletal muscles vary to some extent in terms of sensitivity to ATP, glibenclamide, and potassium channel openers (PCOs) (25,26). These differences could be due in part to differences in experimental conditions, species, and splice variations of SUR2 (27). Although the electrophysiological properties of the channels reconstituted by Kir6.2 and SUR2A (3) are similar to those of K_ATP channels recorded in native skeletal muscles, the molecular composition of skeletal muscle K_ATP channels in native tissues has not been determined. Some of the sulfonylureas have been shown to improve glycemic control in patients with type 2 diabetes by acting on extrapancreatic tissues to reduce insulin resistance (28). Although there is no direct evidence of the involvement of skeletal muscle K_ATP channels in extrapancreatic effects of sulfonylureas, these findings suggest that the K_ATP channel in skeletal muscle is involved in glucose uptake. Interestingly, the glucose-lowering effect of insulin is enhanced in Kir6.2 null mice (18), indicating that insulin sensitivity is increased. Basal and insulin-stimulated glucose uptake in gastrocnemius of Kir6.2 null mice were both significantly increased in vivo and in vitro compared with wild type, although the effects of genetic disruption of the K_ATP channel vary in the different types of muscle (29). Because there is no K_ATP channel activity in the skeletal muscles of Kir6.2 null mice, these findings demonstrate that the K_ATP channel is involved in glucose uptake. The involvement of these K_ATP channels also has been demonstrated in SUR2 null mice (30). SUR2 null mice show lower fasting and nonfasting glucose levels, improved glucose tolerance during glucose tolerance test, and more rapid and severe hypoglycemia after administration of insulin. Enhanced glucose utilization was confirmed by in vivo hyperinsulinemic-euglycemic clamp studies in which SUR2 null mice required a greater glucose infusion rate to maintain the blood glucose level. Such enhancement of insulin action seems to be characteristic of the skeletal muscles, as in vitro insulin-stimulated glucose uptake was 1.5-fold greater in SUR2 null muscle than in wild-type muscle. Together, these findings clearly show that the K_ATP channels in skeletal muscle are involved in glucose uptake.

To further examine the involvement of K_ATP channels in glucose uptake, we generated double null mice lacking both Kir6.2 and insulin receptor substrate (IRS)-1 (31). IRS-1 null mice show severe insulin resistance in skeletal muscles but exhibit normal glucose tolerance (32,33), apparently because β-cell hyperplasia and the resultant hyperinsulinemia can compensate for the increased peripheral insulin demand (34). In contrast, Kir6.2 null mice show severely impaired glucose-induced insulin secretion but also normoglycemia (18). As mentioned, Kir6.2 null mice exhibit an enhanced glucose-lowering response by exogenous insulin, probably due to the increased insulin-stimulated glucose uptake in skeletal muscles (29) that contributes to the preservation of normoglycemia. Because insulin-stimulated glucose uptake is mediated by IRS-1/phosphatidylinositol 3-kinase (PI3K) signaling, the major pathway of insulin signal transduction, disruption of the IRS-1 gene should impair glucose uptake in Kir6.2 null mice, and the double null mice were expected to develop overt diabetes. However, Kir6.2/IRS-1 double null mice had nonfasting blood glucose levels similar to those of wild-type mice but had lower fasting blood glucose levels like Kir6.2 null mice (Table 1). The glucose-lowering effect by exogenous insulin is similarly enhanced in double null mice compared with Kir6.2 null mice. Moreover, glucose uptake in skeletal muscles is significantly increased in double null mice similarly to Kir6.2 null mice. In terms of blood glucose levels and glucose uptake in skeletal muscles, the phenotype of Kir6.2 null mice is unchanged by absence of the IRS-1 gene (Table 1), demonstrating that K_ATP channel-associated change in glucose uptake is independent of the IRS-1-mediated signaling pathway.

IRS-1-mediated glucose uptake requires activation of PI3K followed by activation of downstream signals. Accordingly, disruption of the IRS-1 gene causes reduction of insulin-stimulated activation of PI3K (31–33). However, lack of the Kir6.2 gene does not increase PI3K activity in either the presence or absence of the IRS-1 gene, indicating that the K_ATP channel-associated increase in glucose uptake is independent of PI3K activity. Furthermore, phosphorylation of Akt/protein kinase B (PKB), a signaling molecule downstream of PI3K, is not affected by lack of the Kir6.2 gene, indicating that the Akt/PKB signal is not involved in K_ATP channel-associated increase in glucose uptake. Glucose uptake in skeletal muscles also is stimulated by an insulin-independent, cAMP-activated protein kinase (AMPK)-dependent mechanism (35,36). Upon phosphorylation, AMPK stimulates glucose uptake that is additive to insulin action, suggesting that if the AMPK signal is enhanced, glucose uptake is increased. However, this is not the case in Kir6.2 null mice, as the phosphorylated form of AMPK was not increased but was decreased in mice lacking the Kir6.2 gene. Although the pathways involved in K_ATP channel-associated glucose uptake remain to be clarified, these findings establish that they are independent of AMPK and the IRS-1/PI3K signal.

Kir6.2/IRS-1 double null mice provide other interesting avenues to be investigated. While disruption of the Kir6.2 gene results in loss of glucose-induced insulin secretion, lack of the IRS-1 gene causes insulin hypersecretion. Kir6.2/IRS-1 double null mice have some but poor insulin response after glucose challenge compared with Kir6.2 null mice (31), indicating that disruption of the IRS-1 gene partially restores glucose-induced insulin secretion even in the absence of the pancreatic β-cell K_ATP channels. Moreover, in the nonfasting state, Kir6.2 null mice exhibit serum insulin levels similar to those in wild-type mice, while Kir6.2/IRS-1 double null mice exhibit marked hyperinsulinemia (31). These findings show that insulin secretion stimulated by secretagogues other than glucose is preserved or even increased in the absence of the pancreatic β-cell K_ATP channels when IRS-1 is lacking. Recently, Doliba et al. (37) reported that insulin secretion from SUR1 null islets is hypersensitive to the muscarine receptor signal and the cAMP signal. The insulin response to mixed-meal ingestion is also observed in Kir6.2 null mice (38). While Kir6.2/IRS-1 double null mice show nonfasting hyperinsulinemia, they exhibit somewhat reduced islet insulin content compared with wild-type mice (31), like IRS-1 null mice (34). In IRS-1 null mice, the extensive β-cell hyperplasia might explain the increased serum insulin concentration despite the reduced islet insulin content (34). Beta-Cell hyperplasia and the resultant hyperinsulinemia are thought to be a compensatory response to insulin resistance. Indeed, IRS-1 null mice show severe insulin resistance (32,33). However, although there is no evidence of insulin resistance in Kir6.2/IRS-1 double null mice, they exhibit hyperinsulinemia (31), most likely due to β-cell hyperplasia. The molecular basis of hyperinsulinemia in Kir6.2/IRS-1 double null mice in the absence of insulin resistance remains to be clarified.

Analyses of the pancreatic islets, hypothalamus, and skeletal muscles of Kir6.2 null mice reveal that the Kir6.2-containing K_ATP channels in these tissues act in concert in the maintenance of glucose homeostasis (Fig. 1).

K_ATP channels also play a role in acute metabolic stress. In the cardiovascular system, K_ATP channels are found in both cardiomyocytes and vascular smooth muscles. Brief and intermittent ischemia paradoxically protects myocardium against prolonged ischemic insult, producing a marked reduction in infarct size, a phenomenon known as ischemic preconditioning (IPC) (39) in which K_ATP channels are thought to participate. Infarct size in wild-type mice with IPC is less than without IPC, while there is no significant difference in infarct size with or without IPC in Kir6.2 null mice (40), demonstrating the involvement of Kir6.2-containing K_ATP channels in IPC. A reconstitution study has suggested that the K_ATP channels in vascular smooth muscle are composed of Kir6.1 and SUR2B (41). Kir6.1 null mice are prone to sudden, premature death. The mice exhibit spontaneous elevation of ST segments followed by atrioventricular blocks of various degrees, indicating that death is due to myocardial ischemia (42). This phenotype closely resembles Prinzmetal angina (vasospastic angina) in humans, a form of unstable angina associated with hypercontractility of coronary arteries (43). In accord with this finding, it has been found that disruption of SUR2 also produces a phenotype similar to Prinzmetal angina (44). These findings indicate that Kir6.1/SUR2B K_ATP channels play a critical role in the regulation of vascular tonus of coronary arteries. Because sulfonylureas are widely used in the treatment of type 2 diabetes, it is important to ascertain the effects of the various sulfonylureas on the Kir6.1/SUR2B channels.

In the brain generalized seizure can be evoked by metabolic stress such as hypoxia, and the SNr acts as a central gating system in seizure propagation (45). Kir6.2 null mice are especially susceptible to generalized seizure after brief hypoxia (46), probably because the K_ATP channel is required to regulate the threshold of seizure.

K_ATP channels belong to the Kir channel family. In distinction from the other inwardly rectifying K⁺ channels, in which the pore-forming subunits themselves function as channels, the pore-forming subunit Kir6.x and the regulatory subunit SUR are both required for K_ATP channel function. Genetic manipulation of the K_ATP channel subunits has clarified many of the physiological and pathophysiological functions of K_ATP channels (Table 2). Studies of Kir6.x null mice have revealed that K_ATP channels play crucial roles as metabolic sensors in many tissues in acute metabolic changes.

The studies in our laboratory were supported by Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture (Japan).

The authors thank several groups for their important contributions to these Kir6.x knockout mice studies, especially the Haruaki Nakaya, Jochen Roeper, Jean-Marc Renaud, Nobuya Inagaki, and Andres Terzic laboratories.