| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
The School of Biological Sciences, University of Manchester, Manchester; and Institute of Child Health, Great Ormond Street Hospital, London, United Kingdom
ABSTRACT I. HISTORICAL RELEVANCE AND TERMINOLOGY II. INTRODUCTION III. GLUCOSE HOMEOSTASIS AND FETAL NUTRITION IV. EARLY COMPLICATIONS OF HYPERINSULINISM IN INFANCY V. IONIC AND METABOLIC DETERMINANTS OF INSULIN RELEASE A. Depolarization-Response Coupling and Glucose-Stimulated Insulin Secretion B. Glucose Metabolism, Anaplerosis, and Mitochondria C. KATP Channels 1. Structure and function 2. Pharmacology of KATP channels 3. Trafficking of KATP channels to the cell surface D. PDX1, HNF-3{beta}, and GSIS VI. ATP-SENSITIVE POTASSIUM CHANNEL CHANNELOPATHIES AND HYPERINSULINISM A. Clinical Features and Diagnosis of HI-KATP B. Genetic and Histopathophysiological Diversity of HI-KATP: Fo-HI Versus Di-HI C. KATP Channel Gene Defects and HI-KATP D. Depolarization-Response Coupling in HI-KATP {beta}-Cells E. Amplification Pathways of Regulated Insulin Release From HI-KATP {beta}-Cells F. Clinical Implications of an Advanced Understanding of the HI {beta}-Cell VII. METABOLOPATHIES OF ATP-SENSITIVE POTASSIUM CHANNELS AND HYPERINSULINISM A. Glutamate Dehydrogenase and Hyperinsulinism B. Glucokinase and Hyperinsulinism C. Short-Chain L-3-Hydroxyacyl-CoA Dehydrogenase and Hyperinsulinism D. Exercise-Induced Hyperinsulinism VIII. BECKWITH-WIEDEMANN SYNDROME AND HYPERINSULINISM IX. THE THERAPEUTIC ARMAMENTARIUM FOR HYPERINSULINISM IN INFANCY A. Glucagon B. Somatostatin C. Corticosteroids D. Diazoxide and Diazoxide Analogs E. Nifedipine X. ANIMAL MODELS OF HYPERINSULINISM IN INFANCY XI. PROSPECTS AND CONCLUDING REMARKS
 |
ABSTRACT |
|---|
 Top NextReferences |
|---|
-cell metabolism and anaplerosis, this review describes the relationship between pathogenesis and clinical medicine. Until recently, HI was generally considered an orphan disease, but as parallel defects in ion channels, enzymes, and metabolic pathways also give rise to diabetes and impaired insulin release, the HI paradigm has wider implications for more common disorders of the endocrine pancreas and the molecular physiology of ion transport. |
I. HISTORICAL RELEVANCE AND TERMINOLOGY |
|---|
|
TopPreviousNextReferences |
|---|
|
For the severe, early-onset disease, two broad subtypes of the disorder are considered in this review: focal adenomatous hyperplasia associated with Ch.11p15 gene defects and diffuse -cell abnormalities which (so far) are associated with Ch.11p15, 7p15-p13, 10q23.3, and 4q22-q26. We have described the pathological occurrences of hyperinsulinism in neonates and infants using the generic term HI. Distinction will be made between focal (Fo) and diffuse (Di) HI and the differing genetic origins of the disease by reference to the proteins they encode: HI-KATP, HI-GK, etc.
|
II. INTRODUCTION |
|---|
|
TopPreviousNextReferences |
|---|
-cell development prevailed for many years, and it was only in the 1970s and 1980s that the concept of "hyperinsulinism" was proposed and gradually accepted.
HI represents a group of clinically, genetically, morphologically, and functionally heterogeneous disorders. HI is potentially a devastating condition and one of the most difficult medical problems to face the pediatric endocrinologist. The term hyperinsulinism can, however, be misleading since it suggests "hypersecretion of insulin" from the -cell, and this is only very rarely recorded in patients. Indeed, because most patients present with moderately elevated serum levels of insulin that are entirely inappropriate for the level of blood glucose, HI is best described as "inappropriate insulin release for the level of glycemia" rather than "hypersecretion" per se. Clinical diagnosis is thus based on evidence of the effects of excess insulin, which include hypoglycemia, inappropriate suppression of lipolysis and ketogenesis, and (more traditionally) positive glycemic responses after the administration of glucagon when hypoglycemic (83, 285).
Over the past 5–6 years, a combination of advances in the genetic determinants of insulin release, a greater understanding of the morphofunctional basis of hyperinsulinism, and the availability of functional data obtained from patient tissues have provided unparalleled insights into the pathogenesis of hyperinsulinism syndromes. As a consequence, the established long-term problems of HI are now being successfully confronted, and even with our current limited options of less-than-satisfactory treatments, the life-long sequelae of children with HI are being improved.
|
III. GLUCOSE HOMEOSTASIS AND FETAL NUTRITION |
|---|
|
TopPreviousNextReferences |
|---|
1% of the neonate's energy reserves. Fat oxidation is quantitatively thought to be less important than amino acid/glucose oxidation during fetal life, and rates of ketone body production are low (116).
The endocrinology of fetal nutrition is dominated by insulin. Insulin does not cross the placenta, and the fetal insulin axis is therefore independent of the mother's. Fetal insulin secretion is influenced by concentrations of both glucose and amino acids, but -cells of the fetal pancreas become progressively more responsive to glucose relatively late in gestation, with -cell mass increasing markedly in the last trimester of pregnancy. This occurs even though mRNA for key components of the glucose-sensing apparatus of the -cells is already present from weeks 7–10 of fetal life, i.e., SUR1, Kir6.2, GLUT2, voltage-gated Ca2+ channels, etc. (M. J. Dunne, H. D. Moore, A. Naterajan, A. M. Gonzalez, L. Ruban, and P. W. Andrews, unpublished data).
As in adults, insulin promotes anabolism in the fetus by stimulating uptake of glucose into muscle and adipose tissue. Thus, in the final trimester of pregnancy, a period of rapid fetal growth occurs, particularly in the deposition of fat in adipose tissue, as the fetus stores energy reserves in preparation for birth. The induction of physiological competence during the third trimester of pregnancy is thought to be a critical period during which substrate provision "induces" or "programs" pancreatic islet development in an irreversible manner. These changes may subsequently influence the metabolic response to glucose in later life and also predispose to certain patterns of adult disease (112, 119, 134).
At birth, the newborn must switch abruptly from a state of net glucose uptake and glycogen synthesis to one of independent glucose production and homeostasis. The maintenance of normoglycemia is dependent on those conditions which determine nutrient status throughout life: the adequacy of glycogen stores, maturation of glycogenolytic and gluconeogenic pathway, and an integrated endocrine response. The endocrine events believed to trigger neonatal glucose production and the mobilization of fat from peripheral stores include increased epinephrine secretion and a rapid fall in the insulin-to-glucagon ratio as would occur during the first few hours of life. This change is accounted for by both a fall in the plasma insulin concentration and a surge in the plasma glucagon concentration that occurs at this time (157).
|
IV. EARLY COMPLICATIONS OF HYPERINSULINISM IN INFANCY |
|---|
|
TopPreviousNextReferences |
|---|
In addition to genetic disorders, early-onset hyperinsulinism is also associated with the children of diabetic mothers. During gestation, glucose is transferred freely across the placenta. Prolonged hyperglycemia in poorly controlled maternal diabetes results in fetal hyperglycemia. Fetal hyperglycemia induces expansion of the fetal pancreatic -cell mass with resultant hyperinsulinemia and macrosomia. Withdrawal of the transplacental supply of glucose after birth leads to an abrupt fall in the concentration of glucose, and in the absence of any underlying medical condition in the newborn, hyperinsulinism will typically resolve within 1–2 days after birth. Aside from diabetes, Rhesus incompatibility, perinatal stresses such as birth asphyxia, maternal toxemia, intrauterine growth retardation, or exogenous drug or insulin administration (e.g., Munchausen syndrome, Munchausen syndrome by proxy, ingestion of oral hypoglycemic agents) and insulin-secreting adenoma are also associated with hyperinsulinism in infants (34, 38, 230), but these conditions will not be featured as part of this review. Rather, we present an overview of the physiology and pathophysiology of hyperinsulinism and address the histological and genetic diversity of this condition in the context of molecular medicine.
|
V. IONIC AND METABOLIC DETERMINANTS OF INSULIN RELEASE |
|---|
|
TopPreviousNextReferences |
|---|
ATP-sensitive potassium (KATP) channels determine the metabolic sensing of glucose in -cells and coordinate stimulus-secretion coupling (see Ref. 233). Their function determines both "first phase" and "second phase" insulin release from the pancreas. KATP channels are open in resting, unstimulated -cells and, along with the Na+-K+-ATPase, establish a resting membrane potential of approximately –65 mV. The mechanism responsible for spontaneous channel openings has been proposed to involve the low intracellular ATP/ADP ratio that exists in resting -cells. In support of this, ADP has been shown to be both a potent agonist of KATP channels and to reverse the inhibitory effects of ATP even when there is a >20-fold excess in the concentration of ATP relative to ADP at the cell membrane (72, 74, 137). Consequently, despite the fact that the intracellular concentration of ATP in -cells is in the millimolar range and KATP channels are inhibited in vitro by <10 µM ATP, a high sensitivity to ADP ensures channel opening in resting cells. More recently, a role for phosphatidylinositol 4,5-bisphosphate (PIP2) in the endogenous activation of KATP channels has been suggested (13, 258, 276, 280). However, because other phosphoinositides and lipids such as acyl-CoA and arachidonate also stimulate KATP channels, the sensitivity and selectivity of these responses in the physiological context of KATP channel activity requires further consideration (13, 81, 107, 156, 162, 250).
Following both the uptake and metabolism of glucose by glucokinase and mitochondrial events, closure of KATP channels arises when the intracellular ATP/ADP ratio increases (see Refs. 69, 94, 118, 233). Estimations of the free concentrations of ATP and ADP in the -cell cytoplasm suggest that while ADP concentrations are in the range of 40 µM, the free ATP concentration is 100-fold higher (94). In islets and insulin-secreting cell lines glucose induces a 30 (94) and 50% (251) decrease in the concentration of ADP, respectively, leading to an increase in the ATP/ADP ratio. Glucose-dependent closure of KATP channels facilitates depolarization of the cell membrane and the subsequent opening of voltage-dependent Ca2+ channels, which in turn leads to an abrupt rise in the cytosolic Ca2+ concentration close to the plasma membrane. Insulin release is then initiated by the process of Ca2+-dependent exocytosis (Fig. 2). These events describe the KATP channel-dependent or "triggering" pathway of glucose-stimulated insulin secretion (GSIS) and are likely to account for the first phase of insulin release, i.e., the secretion of preformed insulin-containing granules located at the plasma membrane (23, 118). KATP channels also determine second-phase insulin secretion from -cells which is brought about by the gradual augmentation and potentiation of Ca2+-triggered insulin release: a process that entails the preparation of previously nonreleasable granules for exocytosis. This pathway, which is termed the "amplification" or "augmentation" pathway, is also referred to as the "KATP channel-independent pathway of GSIS," a description that is somewhat misleading in terms of -cell function, but an accurate description of the conditions necessary to examine those signaling events downstream of KATP channel closure (see Refs. 4, 118).
|
The precise molecular mechanisms by which glucose metabolism augments distal signaling are yet to be fully resolved (see Ref. 180). Several coupling factors have been proposed including increased ATP/ADP and GTP/GDP ratio, cytosolic levels of long-chain acyl-CoA, the pyruvate-malate shuttle, and glutamate export from mitochondria. ATP is likely to have several roles as a local mediator of exocytotic mechanisms, including, for example, regulation of ATPases such as N-ethylmaleimide-sensitive fusion protein (NSF), V-type H+-ATPases and Hrs-2, and by phosphorylation of specific proteins such as SNAP-25 (11, 15, 114, 273). The stimulation of protein kinases A (PKA) and C (PKC), and Ca2+/calmodulin-dependent kinase II (CaMKII), as would occur during exposure to hormones, neuropeptides, and neurotransmitters (some of which are collectively known as the "incretins") or Ca2+ following its influx is also likely to be important (4, 18, 23, 92, 153, 207, 239, 240, 292, 293, 295, 333). The involvement of G protein-coupled receptor (GPCR) events in distal signaling is supported by observations that mice lacking receptors for glucagon-like peptide 1, or gastric inhibitory polypeptide, show impaired insulin secretion and abnormal glucose tolerance without insulin resistance (206, 260). Equally, GTP can affect both small G proteins such as rab3A (223) and the larger heterotrimeric G proteins (332). Finally, although there is evidence to suggest that "amplification" of GSIS can occur independently of a rise in the cytosolic concentration of Ca2+ in both model -cells and human islets (293), depolarization-evoked Ca2+ influx is sufficient by itself to cause diacylglycerol formation, PKC activation, and substrate phosphorylation (207). (For recent reviews: GSIS pathways, see Refs. 4, 23, 118, 294; mechanisms of exocytosis, see Ref. 27.)
The discrete actions of glucose on the triggering and amplification pathways of insulin secretion can be largely explained by the coordination of events that involve different pools of insulin-containing granules. Several dynamic granule pools are recognized in -cells: reserve, docked, readily releasable, and immediately releasable pools. Activation of the triggering pathway results in exocytosis of the immediately releasable pool of granules giving rise to the first phase of GSIS. In mouse -cells, this has been estimated to involve fewer than 50 of the 10,000 or so insulin-containing granules that are positioned for secretion by clustering within the vicinity of voltage-gated Ca2+ channels (10). As an additional level of complexity, in neurons, defined regions of exocytosis have been described that contain cytoskeletal matrix associated active zone proteins (CAZ), which are important for the modularization of exocytosis and the mobilization of secretory vesicles (91). Although analogous zones are less well characterized in -cells, recent reports suggest the presence of CAZ-associated proteins that are responsible for the integration of Ca2+ sensing and PKA-independent cAMP-induced exocytosis of insulin-containing granules (90, 142, 227). A key rate-limiting step following first-phase insulin release is the conversion of readily releasable granules to the state of immediate releasability, or "priming." This stage is known to be acutely dependent on the ATP/ADP ratio, which appears to be required in the acidification process of granules, a prerequisite to priming, and involves the granular membrane proton pump and Cl– uptake (11). Whilst this general model is considered relevant for all types of insulin-secreting cells, there are some important differences between species. In rat and human -cells, for example, augmentation pathways induce a time-dependent increase in the rate of pool priming resulting in a rising second-phase response, whereas in mouse -cells and insulin-secreting cell lines, despite biphasic insulin secretion profiles and secretory granule mobilization, the rate of this conversion does not appear to be significantly changed by glucose (see Refs. 10, 293).
In the model described above, KATP channels are considered as major determinants of glucose-regulated electrical activity in -cells, and this has important downstream consequences for GSIS, the neurohormonal regulation of insulin release and the manipulation of secretion by pharmacological agents. The central role of KATP channels suggests that drugs which act as channel inhibitors such as sulfonylureas (e.g., glibenclamide, tolbutamide, etc.; Ref. 296) and glinides [e.g., repaglinide (51) and nateglinide (123)] function in vivo as antidiabetic compounds by triggering first-phase insulin release and by maintaining KATP channels in a closed state in order for glucose metabolism to act distally. Conversely, KATP channel agonists or "openers," such as diazoxide and BPDZ 154, have hyperglycemia-inducing capabilities since they inhibit secretion by preventing membrane depolarization and hence reducing voltage-gated Ca2+ channel activity (see Ref. 73; see sect. IXD). In terms of pathophysiology there is now compelling evidence that defects in the genes encoding KATP channels will cause both hyperinsulinism ("KATP channelopathy," see sect. VI) and type 2 diabetes mellitus (125, 259) and that defects in other proteins which control glucose metabolism ("KATP channel metabolopathies," see sect. VII) will induce inappropriate KATP channel closure or activity and thereby promote either hyperinsulinism or reduce insulin release, respectively (Table 1).
|
B. Glucose Metabolism, Anaplerosis, and Mitochondria
Coupling GSIS to regulated changes in the cytosolic Ca2+ concentration ([Ca2+]c) is dependent on metabolic events (Fig. 3). Glucose is transported into -cells by facilitative transporter(s) which allow(s) for the subsequent rapid equilibration of glucose across the -cell membrane. In rodent insulin-secreting cells, GLUT2 is the principal transporter, but this is not the case in human -cells. Enriched populations of human -cells have been shown to have a 100-fold lower GLUT2 abundance than rat -cells, while studies using isolated cells from patients with insulinomas report the expression of GLUT1/3 mRNAs but not GLUT2 (64, 264). After uptake, phosphorylation of glucose by glucokinase (GK) is the first enzymatic process in the glycolytic pathway and, since the Michaelis constant (Km) of GK for glucose is 10 mM, it is this constraint that determines the range of glucose concentrations that are physiologically relevant to the stimulation of insulin release (194, 195). Consequently, even small changes in GK activity can be significant and can directly affect the threshold for downstream events in GSIS. This is most clearly demonstrated by the fact that mutations in GK cause either hyperinsulinism or diabetes depending on the nature of the impairment in enzymatic activity (see sect. VII) (Table 1). Under physiological conditions the activity of GK is thought to be regulated by an endogenous factor, a precursor of the propionyl-CoA carboxylase -subunit, which raises both the affinity of GK for glucose and the enzyme's Vmax (275). Glycolytic flux is also limited by the activity of phosphofructokinase. The predominant form of this enzyme in -cells is an oscillatory isoform that is allosterically regulated for bursts in the production of ATP by oscillations in the ATP/ADP ratio (307). It is activated by the AMP/ADP ratio and fructose 2,6-bisphosphate and inhibited by ATP and citrate (Fig. 3). These enzymes are therefore important for determining fluctuations in the cell membrane potential via KATP channels.
|
Glucose increases the production of cytosolic NADH by the action of glyceraldehyde-3-phosphate dehydrogenase, and the reducing equivalents are then transported into the mitochondria by the 
-glycerolphosphate (-GP) and malate/aspartate shuttles for ATP synthesis (176, 177). This transport helps to maintain glycolytic flux in the direction of ATP production by preventing a decrease in the cytosolic NAD+/NADH ratio which is required for the activities of both glyceraldehyde-3-phosphate dehydrogenase and KATP channels (70). Activity of the -GP shuttle is unusually high in -cells, and this is thought necessary to maintain high levels of NAD+ for glycolysis (265). Insulin-secreting cells have low levels of lactate dehydrogenase, and this serves to direct most of the pyruvate produced from glycolysis into the mitochondria (265).
Other signals generated from glucose metabolism in the -cell play a major role in controlling the relative rates of glucose and FFA oxidation (FFAox) and the shift from FFAs to glucose as a fuel source. Thus, at low levels of glucose, mitochondria fulfill the energy needs of the -cell by FFAox, but as glucose levels begin to rise, FFAox is decreased and glucose oxidation then supplies the cellular energy requirements (183). The first step in glucose-induced inhibition of FFAox is an increase in mitochondrial anaplerosis, resulting in citrate formation (24) (Fig. 3). Elevated mitochondrial citrate leads to an increase in the cytosolic concentration of citrate, which is then converted to malonyl-CoA by the actions of citrate lyase and acetyl-CoA carboxylase (ACC). Malonyl-CoA is a potent allosteric inhibitor of carnitine palmitoyltransferase-1 (CPT1), a mitochondrial membrane enzyme that is responsible for the transport of long-chain acyl-CoA (LC-CoA) from the cytosol into the mitochondria. The isoform of CPT1 in -cells is the same as that in the liver (337). This enzyme is located in the outer mitochondrial membrane and has one inhibition site that faces the cytosol, inhibitable by dicarboxylic CoA esters (e.g., malonyl-CoA, L-3-hydroxybutyryl-CoA, see sect. VIIC), and a second site facing the intermembrane space that is inhibited by shortchain monocarboxylic CoA esters (141). Inhibition of CPT1 blocks the entry of LC-CoA into the mitochondrion, elevating the cytosolic concentration of the lipid which then acts as a signaling molecule with diverse actions that are directly related to the release of insulin, including stimulation of the insulin exocytotic machinery (55) and activation of "classic" and "novel" PKCs (182, 335). In addition, LC-CoA will inhibit mitochondrial adenine nucleotide translocase (331) and the enzyme ACC (238) and cause the activation of both sarco/endoplasmic reticulum Ca2+-ATPases (57) and KATP channels (22, 107, 162). While the latter effect will tend to inhibit insulin release by lowering [Ca2+]c and restoring basal cytosolic Ca2+ levels, this mechanisms may be important either for providing feedback inhibition and/or facilitating the development of glucose-induced oscillations in metabolism and Ca2+ signaling. LC-CoA is also a modulator of ceramide- and/or nitric oxide-mediated apoptosis and the binding to nuclear transcriptional factors in insulin-secreting cells (for recent review, see Ref. 111).
In the scheme outlined above, glucose underpins the generation of two key intracellular mitochondrial signals: acetyl-CoA, which is used for ATP synthesis (see below), and citrate, which is used in the production of malonyl-CoA. The proposed role of malonyl-CoA in the inhibition of CPT1 is pivotal to GSIS and although supported by several studies (45, 61, 243) has remained controversial in the literature and much debated. More recently, it has been reported that malonyl-CoA decarboxylase overexpression in insulin-secreting cell lines lowers the levels of malonyl-CoA but has no effect on GSIS (6, 211). Even though these findings suggest that disruption of malonyl-CoA can be dissociation from GSIS, the same protocols have also been reported to cause a 50% decrease in insulin production (241).
ATP production and an elevation of the [Ca2+]c are important signals for GSIS, but they also have influences on -cell metabolism. It has been estimated that >95% of ATP in -cells is produced in the mitochondria which is regulated by substrate availability (-GP and pyruvate) and by [Ca2+]c (79). -GP and pyruvate concentrations are determined, in part, by ADP and the actions of the ATP/ADP ratio on phosphofructokinase (PFK) (160). In insulin-secreting cells, glucose-induced rises in the ATP/ADP ratio and mitochondrial NADH precede rises in [Ca2+]c, suggesting that glucose metabolism initiates mitochondrial respiration prior to, and independent of, any rise in cytosolic Ca2+ (36). Once elevated, cytosolic Ca2+ can affect mitochondrial ATP synthesis by the activation of the -GP shuttle delivering more reducing equivalents to the electron transport chain and also by increasing the mitochondrial matrix free Ca2+ concentration ([Ca2+]m) (37, 176). These events increase ATP production via the activation of pyruvate dehydrogenase and mitochondrial dehydrogenases (113, 193, 253). Significantly, however, as [Ca2+]m continues to rise, this will act to depolarize the inner mitochondrial membrane, inhibiting ATP synthesis and resulting in a fall in the ATP/ADP ratio. This is one mechanism by which glucose metabolism results in the generation of oscillations of ATP and ADP availability which directly influence the activity of KATP channels, the cell membrane potential, and the rhythmical openings and closures of voltage-gated Ca2+ channels.
-Cells express a KATP channel complex formed by subunits belonging to at least two distinct families of proteins (Fig. 4) (for reviews, see Refs. 2, 5, 263). The K+-selective pore is formed by the weak inward rectifier K+ channel Kir6.2. Comprising 390 amino acids, this protein has a predicted membrane topology with two -helical transmembrane domains linked by a highly conserved sequence of amino acids that shares sequence homology with the P region or K+ selectivity domain of voltage-gated K+ channels. The other subunit is an ATP binding cassette protein and receptor with high affinity for sulfonylureas, designated SUR1 (3). Human SUR1 consists of 1,581 amino acids and has 17 predicted transmembrane regions that are organized into three discrete domains, designated transmembrane domains (TMD): TMD0, TMD1, and TMD2, 9 cytoplasmic loops (CL1-CL9), and 2 intracellularly disposed nucleotide-binding folds (NBF) (312). In most recombinant systems Kir6.2 will not form operational K+ channels independently of SUR1; however, when coexpressed, K+ channel currents are generated that closely resemble those of the native -cell KATP channel complex (128, 129, 255). KATP channels present in other tissues are heteromultimeric complexes of different Kir6.x and SURx proteins; e.g., cardiac KATP channels: Kir6.2 + SUR2A, smooth muscle KATP channels: Kir6.2 + SUR2B, and the smooth muscle nucleotide-activated K+ channel Kir6.1 + SUR2B (see Refs. 2, 5, 263).
|
KATP channels are thought to be organized as an octameric complex of four Kir6.2 subunits arranged around a central pore, coupled to four SUR1 subunits (SUR1+Kir6.2)4 (41, 279) (Fig. 4). Each subunit is known to be differentially regulated. Kir6.2 which determines the biophysical properties of the channel complex including K+ selectivity, rectification, and "gating" is inhibited by ATP but activated by acyl-CoA and phosphoinositide lipids. ATP-induced channel closure appears to involve the cytoplasmic domains, but there is also evidence that actions of ATP can be almost completely abolished by individual point mutations at other sites on Kir6.2 (67, 167, 277, 309–311). SUR1 acts as a conductance regulator of Kir6.2; therefore, the sensitivity of channels to ATP, ADP, and guanosine (GTP, GDP) nucleotides involves both subunits (Fig. 4). Nucleotide hydrolysis is possible at both NBF sites of SUR1 (314). NBF1 and NBF2 are thought to form a close association leading to two binding pockets; site 1 is thought to be more selective for ATP than ADP, and this interaction is stabilized by MgATP hydrolysis at site 2. Recent data suggest that NBD1 of SUR1 has minimal (if any) ATPase activity (189, 190) and that although NBD1 plays a small but significant role in KATP channel activation by MgADP, it is not as important as NBD2 that has ATPase activity (191). It also appears that the linker regions between the Walker A and Walker B motifs of the NBDs are not required for nucleotide binding but are involved in transducing nucleotide binding into channel activation. Thus the functional regulation of KATP channels induced by changes in the ATP/ADP ratio involves cooperative interactions of nucleotides at both subunits with the actions of ATP-induced inhibition of Kir6.2 being countered by the activatory effects of ADP at SUR1.
2. Pharmacology of KATP channels
The SUR1 subunit plays a key role in determining the pharmacological regulation of KATP channels. -Cell-selective potassium channel openers (KCOs) such as diazoxide, BPDZ 154, BPDZ 73, NNC 55–0118, and other less selective KCOs (e.g., levcromakalim, pinacidil, nicorandil, etc.), bind to SUR1 and activate KATP channels in a nucleotide-dependent manner (16, 21, 46, 50, 68, 163, 164, 216). The differential sensitivity of the heterogeneous family of KATP channels to these compounds is explained by the location of the binding sites for KCOs on SUR1 and SUR2. With the use of a series of chimeric SUR proteins, it has been shown that two sequences of SUR1 are critical for determining KCO binding and activation (315). These are located within the final series of transmembrane helices at amino acids 1059–1087 (KCO1) and 1218–1320 (KCO2) (SUR2B nomenclature), which correspond to part of the linking sequence between the membrane-spanning helices 13–14 and 16–17 (Fig. 4). These regions flank a putative sulfonylurea binding site located in CL8 and are thought to form a binding pocket for diazoxide on the -cell KATP channel (315). Sulfonylureas also interact with CL3.
3. Trafficking of KATP channels to the cell surface
Each protein subunit of the KATP channel possesses a number of signaling motifs that influence the assembly of the channel protein and its subsequent expression at the cell surface (Fig. 4). The first of these to be described was the "RKR" motif, which is composed of three amino acids: arginine-lysine-arginine, and found within the sequences of both SUR1 and Kir6.2. The RKR motifs act as endoplasmic reticulum (ER) retention signaling sequences; they must be shielded by the process of channel folding and assembly to permit the expression of KATP channels at the cell surface (338). When SUR1 and Kir6.2 fold correctly and coassemble with the correct stoichiometry (4:4) within the ER, the RKR motifs become cloaked by the other subunits, thus permitting the export of the octameric KATP channel complex from the ER. The RKR motif is located close to NBF1 of SUR1 in amino acid positions 648–650 and close to the COOH-terminal region of Kir6.2 at positions 369–371. Mutagenesis of the RKR motif of SUR1 to AAA permits expression of SUR1 at the membrane in the absence of Kir6.2 (338). Similarly, truncation of the COOH-terminal region of Kir6.2, which removes the RKR motif, allows cell surface expression of the subunit in the absence of SUR1 (311, 338). Kir6.2 also possesses a second retrograde signaling sequence, "L355-L356," a dileucine motif which functions in endosomal targeting in other cells. Mutation or truncation of this motif leads to an increase in membrane targeting of Kir6.2 (311, 338).
In addition to the retrograde signaling sequences outlined above, SUR1 possesses a number of anterograde signaling sequences. The COOH-terminal region of SUR1 possesses a dileucine motif (L1566-L1567) and downstream phenylalanine (F1574), which are required for KATP channels to exit the ER (270). Mutation of these amino acids (or COOH-terminal truncations, as observed in some HI-related mutations, see below) leads to reduced surface expression of functional KATP channels, although the downstream receptors for these signaling motifs have not yet been identified. However, a third leucine residue located close to the COOH-terminal region of SUR1 (L1544) has been implicated in the formation of the RKR shield which allows the protein to exit the ER, thus forming another anterograde signal for forward trafficking (303). N-linked glycosylation sites of SUR1 (N10, N1050) may also be important sites for forward trafficking of KATP channels, since in vitro mutations of these residues result in entrapment of the protein within the ER (45).
Concurrent with the mechanisms of insulin release, glucose metabolism is also coupled to increases in the levels of insulin mRNA through increased transcription of the insulin gene. Although the first description of the actions of glucose on the insulin gene was described some 30 years ago, it was not until recently that the mechanisms were elucidated (132). A key determinant of this is the homeodomain transcription factor PDX1, which becomes phosphorylated during glucose metabolism and is translocated from the cytoplasm to the nucleus (186, 199, 222, 232). Additionally, PDX1 is a transcriptional modulator of a number of other genes preferentially expressed in -cells including the glucose transporter GLUT2 (324), SUR1 (7), Kir6.2 (7), glucokinase (327), islet amyloid polypeptide (32), and the proinsulin processing enzymes PC1/3 and PC2 (for detailed reviews, see Ref. 200). PDX1 also determines -cell lineage and development (133, 290); hence, defects in PDX1 not only alter the identity of -cells causing maturity-onset diabetes in the young type 4 and neonatal diabetes mellitus, but also exocrine insufficiency and pancreatic agenesis (133, 200, 290). At this stage the role of PDX1, or indeed other transcription factors, in determining the pathogenesis of hyperinsulinism in humans has not been critically examined. However, it is of note that impaired PDX1 activity is a phenotypic feature of a hyperinsulinism-derived human -cell line NES2Y (178) and that in mouse -cells ablation of HNF-3 (Foxa2), an upstream transactivator of PDX-1, resulted in hyperinsulinemic hypoglycemia (297). As in both transgenic mice and insulin-secreting cell lines HNF-3 defects led to a marked decrease in the mRNA expression levels of the KATP channel subunits SUR1 and Kir6.2 (297, 3326), an examination of gene defects in the transcriptional regulation of KATP channels and/or other elements of the GSIS pathway is likely to produce new candidate genes and causes of HI.
|
VI. ATP-SENSITIVE POTASSIUM CHANNEL CHANNELOPATHIES AND HYPERINSULINISM |
|---|
|
TopPreviousNextReferences |
|---|
A. Clinical Features and Diagnosis of HI-KATP
HI per se cannot be detected in utero, and there are no characteristic visual, auscultatory, or tactile findings to suggest hyperinsulinism. Infants with prenatal hyperinsulinism may have a characteristic appearance of macrosomia which reflects the anabolic effects of prolonged hyperinsulinemia in utero resulting in increased muscle, fat, and liver mass. However, many are born with appropriate or even low birth weights, yet others are preterm children (8). It has also been reported that many patients with hyperinsulinism have a distinct facial appearance that could be the consequence of fetal intoxication by insulin (59). The first clinical manifestations of hyperinsulinism-induced hypoglycemia are mainly experienced shortly after birth. These may include cyanosis, respiratory distress, sweating, hypothermia, irritability, poor feeding, hunger, jitteriness, lethargy, apnea, which can progress to vomiting, seizures, tachycardia, and averted neonatal death. In older children and adults symptoms tend to be typical of those of hypoglycemia including confusion, headaches, dizziness, syncope, and when severe, loss of consciousness.
The definition of a glucose requirement to maintain normoglycemia is a key indicator as well as therapeutic step in HI, and the demonstration of an increased glucose requirement is the sign of underlying hyperinsulinism. For patients with severe early-onset HI, there is now general agreement that the diagnostic criteria are: 1) a glucose requirement of >6–8 mg · kg–1 · min–1, which is needed to maintain blood glucose above 2.6–3 mM; 2) laboratory blood glucose values <2.6 mM; 3) detectable insulin at the point of hypoglycemia with raised C-peptide; 4) inappropriately low blood FFA and ketone body concentrations at the time of hypoglycemia; 5) a glycemic response after the administration of glucagon when hypoglycemic; and 6) the absence of ketonuria (8).
Most infants with hyperinsulinism present during the first postnatal days, with others during the first year. Rarely, older children present de novo with symptoms of hyperinsulinism-induced hypoglycemia. Failure to recognize and to promptly treat hypoglycemia carries a substantial risk of severe brain damage and mental retardation because of a lack of fuels to sustain normal brain metabolism (for review, see Ref. 100). Medical therapy for the disorder involves an increased carbohydrate intake to meet the elevated requirement, and usually one or more drugs that inhibit insulin secretion (see sect. IX). Unfortunately, the responsiveness of children with HI to these agents is inconsistent and variable as a result of the heterogeneous defects in the mechanisms that control insulin release, and patients who do not show an adequate and immediate response require surgery in the form of a pancreatectomy to prevent recurrent neuroglycopenia (1, 9, 19, 46, 71, 136, 139, 140, 171, 224, 225, 228, 291). Surgical interventions can vary from a sub-total to near-total pancreatectomy and, although these procedures may prevent subsequent hypoglycemic episodes, they also lead to pancreatic insufficiency and iatrogenic diabetes mellitus. The extent of a pancreatic resection necessary to achieve a cure for HI has been the subject of several studies. In a recent review of the literature, Shilyansky et al. (272) recorded resolved hypoglycemia in only 54% of 220 patients undergoing a subtotal pancreatectomy (<95% resection), compared with 64% of 83 patients undergoing a 95% pancreatectomy, and 97% of 74 after a 98% pancreatectomy. To achieve euglycemia, most patients therefore undergo a 95% (or more) resection, but this is associated with complications, since a high incidence of diabetes, from 75 to 85%, has been reported (166, 272).
Most cases of HI are sporadic, and familial forms, although rare, are well documented. Sporadic HI has an estimated incidence of 1 in 27,000 live births in Ireland and 1 in 50,000 live births in Finland (103), compared with 1 in 20,000 in Kuwait (247). However, in some isolated communities the disease incidence is much higher, e.g., 1 in 3,200 in the central area of Finland and 1 in 2,500 in the Arabian peninsular (103). The diffuse form of HI (see below) has a male-to-female incidence ratio of 1.2:1 and the focal form 1.8:1.
B. Genetic and Histopathophysiological Diversity of HI-KATP: Fo-HI Versus Di-HI
The SUR1 gene ABCC8 comprises 39 exon boundaries and is clustered with the Kir6.2 gene KCNJ11, a single open reading frame lying immediately 3' of ABCC8, on the short arm of chromosome 11, Ch.11p15 (96, 305). This is a genetic locus linked to both Di-HI and Fo-HI (reviewed by Refs. 100, 103).
Although autosomal dominant mutations have been described (126), Di-HI predominantly arises from the autosomal recessive inheritance of KATP channel gene mutations (103). This condition affects all of the islets of Langerhans of the affected pancreas (Fig. 5), and surgical treatment usually requires the removal of a minimum of 95% of the pancreas (Di-HI can also be treated by long-term conservative therapeutic regimes, see below). Until 1997 Di-HI was widely believed to be the main cause of congenital hyperinsulinism, despite the fact that cases of "focal hyperinsulinism" were first described more than 25 years ago (105, 117, 131, 152, 244, 298). Fo-HI has a non-Mendelian mode of inheritance of KATP channel dysfunction and -cell hyperplasia leading to HI (60–62, 86, 102, 322). In this condition, epigenetic phenomena involving gene silencing lead to a loss of heterozygosity of a region of the maternal chromosome 11p15 and a reduction to homozygosity of paternally derived genes, and this results in a somatic lesion of defective -cells within the pancreas. Focal lesions appear histologically as small regions of islet adenomatosis measuring 2–5 mm in size (Fig. 5) and appear to develop through imbalanced expression of maternally imprinted tumor suppressor genes H19 and p57kip2, and the paternally derived insulin-like growth factor II gene (Fig. 6). Also encoded by chromosome 11p15.5 is Pidd, a p53-induced protein with a death domain (170). Pidd is thought to promote apoptosis by acting as a downstream regulator of the tumor suppressor p53 (170). It is not known whether Pidd is subject to imprinting, but since antisense inhibition of Pidd expression attenuates apoptosis in response to p53 activation and DNA damage (170), loss of Pidd through gene silencing may contribute to the development of focal lesions. Importantly, away from the lesion, the remainder of the pancreas retains a relatively normal histological appearance (see Ref. 267).
|
|
Recent estimates from France, Israel, and the United States now suggest that 40–65% of all patients with HI have the focal form of HI-KATP (62, 102, 284). Because this represents a significant number of patients that are potentially curable through a limited surgical intervention to remove just the focal lesion, much effort is now being placed on attempts to diagnose Fo-HI to avoid the obvious long-term complications of radical pancreatic surgery (49, 173, 192, 201, 299). Suggested diagnostic procedures include preoperative interventional radiography such as intra-arterial calcium stimulation tests, pancreatic venous sampling or transhepatic portal venous insulin sampling, positron emission tomography, laparoscopy, and intraoperative histological examination of biopsies of the resected pancreas (14, 20, 25, 33, 62, 82, 220, 245, 246). A greater understanding of the pathophysiology of -cells in HI has also led to the introduction of a noninvasive procedure to monitor acute insulin-response profiles to glucose and tolbutamide in patients to distinguish focal forms of HI from diffuse disease (108, see sect. VID).
C. KATP Channel Gene Defects and HI-KATP
More than 100 mutations in the ABCC8 and KCNJ11 genes have so far been described (85). For a number of these mutations it has been possible to demonstrate that they result in differing abnormalities of recombinant KATP channels including protein folding defects, assembly and trafficking defects, and alterations in both nucleotide regulation and open-state frequency (Fig. 7). However, since in >50% of Di-HI and 30% of Fo-HI patients, screening has failed to define the genetic basis of disease (85, 86, 213) and in Japan SUR1 mutations account for only 20% of HI cases, our overall understanding of the pathogenesis of HI is far from complete (103, 282, 284). To address this and to define the causal relationship between ABCC8, KCNJ11, KATP channel dysfunction, and inappropriate insulin release, we undertook studies on -cells from more than 110 patients with early-onset hyperinsulinism who required pancreatectomy. In this group 85% of all patients carried functional defects in channels, which confirms that HI-KATP is the principal cause of early-onset, aggressive HI, but also reveals that in nearly 15% of patients with this clinical phenotype, hyperinsulinism is unrelated to loss of channel function per se (1, 19, 46, 71, 139, 140, 171, 224, 228, 291). These patients therefore represent a group likely to yield novel HI disease-causing mutations unrelated to the known KATP channel genes.
|
There are numerous mechanisms that could result in defects in KATP channel function, and in most cases the relationship of molecular mechanisms, cellular defect, and clinical hyperinsulinism has yet to be established. Figure 7 attempts to summarize these with reference to SUR1/Kir6.2 gene transcription, translation, protein formation (synthesis, folding, and assembly), exit from the ER, and the function of SUR1/Kir6.2 at the membrane. Potentially, gene defects in KATP channel subunits could influence any of the steps that ultimately result in protein expression at the plasma membrane. From functional studies of HI-KATP, patients have been classified into two groupings.
In type 1 disease, functional KATP channel currents were ablated in -cells. Accounting for 10% of all Di-HI patients and 55% of Fo-HI patients, this condition is typified by abnormalities in gene expression, protein synthesis, maturation, assembly, or trafficking (Fig. 7, D–H). Type 1 HI-KATP was originally described in two patients from interrelated Saudi Arabian families who carried the autosomal recessive R1437Q(23)X mutation in exon 35 of ABCC8 (71). Probands presented with severe, early-onset, and drug-resistant Di-HI and no functional KATP channels in -cells. In this case, the SUR1 mutation truncates 200 amino acids from the COOH-terminal region of the protein, an area that contains the L1566, L1567.F1574 anterograde signaling sequence and residue L1544 which is part of the cloaking region for the RKR sequence. Because this defect will affect the exit of channel subunits from the ER compartment, the functional loss of KATP channels in these patients appears to arise from a trafficking abnormality, which was confirmed when a parallel mutation was engineered and then coexpressed with wild-type Kir6.2 in COS cells (71). In addition to SUR1 gene defects that cause major protein truncations, trafficking abnormalities have been reported in recombinant cells for the 
F1388 and R1394H disease-causing SUR mutations (30, 229, 302).
More than 60% of Di-HI and 45% of Fo-Hi patients have a type 2 KATP channelopathy. This reflects a diverse group of patients in which KATP channel currents were recordable in -cells but were found either to express defects in function or were present in limited numbers (Fig. 7, B–D). The V1479R SUR1 mutation located in NBF2 is a typical example of a gene defect leading to a regulatory abnormality in KATP channels (217). In recombinant cells this mutation leads to loss of ADP-dependent KATP channel gating and sensitivity to diazoxide. A number of other HI-associated mutations have similar actions (F591L, T1139M, R1215Q, G1382S, E1506K, I446T, R1420C, and R1436Q) and are linked to the clinical phenotype, since the loss of ADP-dependent gating results in the constitutive inhibition of KATP channels by ATP (126, 191, 278, 301).
In the first study of nonconsanguineous familial disease in the United Kingdom cohort of HI patients, we recently described three caucasian neonates from two separate families with early-onset, drug-resistant HI (291). Genotyping revealed a novel defect in intron 16 of ABCC8-2154+3a>g in all probands which was paternally inherited in one family and located on the maternal allele in the other. The intronic mutation is approximate to the NBF1 coding region, and functional studies were used to describe a >95% loss of KATP channel function (Fig. 8), with no responses to ADP and diazoxide. It is not entirely clear how the intronic defect alters the amino acid sequence of SUR1, but one suggestion is that the abnormality is able to introduce a cryptic splice site in the ABCC8 reading frame. Because the defect is proximal to the RKR retention sequence R648-K649-R650, it is tempting to speculate that the folding of the mutated SUR1 fails to effectively cloak the RKR motifs, leading to a cell surface trafficking abnormality in the -cell. In insulin-secreting cells from patients, we were also able to demonstrate that the effects of the SUR1 mutation were selective for KATP channels, as there were no defects in the expression or function of voltage-gated K+ channels, nor the actions of sulfonylureas on Ca2+-dependent exocytosis (291). The V187D SUR1 mutation, located in TMD0 of SUR1, leads to early-onset HI and is one of the two founder mutations associated with 50% of HI in the Finnish population (224). In -cells from patients, >95% of the normal current magnitude was lost, and residual channels were unresponsive to ADP and diazoxide (224). These types of experiments indicate that the outcome of mutations located towards NBF1 and the NH2-terminal sequence of SUR1 can be equally as severe as COOH-terminal truncations and mutations located within the regions of ABCC8 that encode for the NBFs. Interestingly, HI resulting from the V187D mutation is equally as severe in patients who are homozygous or heterozygous for the mutation, yet carriers can also be asymptomatic with normal insulin secretion profiles, tissue sensitivity to insulin, and appropriate insulin release profiles during hypoglycemia (127).
|
