III- Regulation of glucagon secretion

It is often assumed that peripheral glucagon concentrations do not accurately reflect the actual α-cell secretion rate, because of hepatic clearance of glucagon. However, the data are controversial. The concentrations of glucagon in plasma are low, with fasting concentrations in healthy individuals around 10 pmol/L. Being an essential part of glucose regulation, glucagon secretion is tightly linked to plasma glucose concentrations, with increasesup to 40 pmol/L when glucose levels are lowered to 2-3 mmol/L and decreases to 1-2 pmol/L when glucose levels are elevated (to around 10-12 mmol/L) (Holst et al, unpublished). Emerging work reveals a complex but sophisticated regulatory mechanism for the modulation of glucagon output from the α-cells, with effects from pancreatic and endocrine hormones including insulin, somatostatin, epinephrine and incretins, Zn2+, nutrients,and the central and autonomic nervous pathways (Kulkarni, Rorsman, Lectures).

 

1. Regulation of glucagon secretion by insulin and incretins

The distribution and arrangement of different islet cell types are important for physiological regulation between the cells since the blood flows from the center of the islets towards the periphery, ie, from β-cells to non-β-cells, suggesting that secreted insulin regulates hormone secretion by other islet cell types. This architecture is typically preserved in rodent islets, while in humans non-β-cells are often observed both at the periphery and also seemingly in clusters within the center of islet (Figure 1). This implies several possibilities; 1) rodent cellular hierarchy in the islets does not apply to human islets, or 2) human islets consist of several clover leaf−like “rosettes”, with each rosette resembling the basic islet architecture observed in rodent islets. Whether the blood flow is similar to that seen in rodents within each rosette in the human “islet” is currently unclear (Kulkarni, Lecture).

 

a. Role of insulin

In vivo and in vitro studies indicate that insulin suppresses glucagon secretion.17-18 Data on the αIRKO mouse10, 19-20 provided the first direct genetic evidence for a significant role for insulin signaling in the regulation of α-cell function in both normo and hypoglycemic states in vivo. Whereas their glucagon response to hypoglycemiain the presence of fasting-induced hypoinsulinemia was impaired, the knockouts exhibited significantly enhanced glucagon response to hyperinsulinemia-induced hypoglycemia. In contrast, glucagon secretion was enhanced in control mice in the first conditions but suppressed in the second. This suggested that the αIRKO mice were unable to sense the variations in ambient insulin and therefore to adapt their glucagon secretion. Interestingly, streptozotocin (STZ)-induced hyperglycemia secondary to hypoinsulinemia induced a paradoxical similar increase in plasma glucagon levels both in control and αIRKO mice. Further, in STZ-treated mice, normalization of hyperglycemiaby phloridzin treatment decreased plasma glucagon to levels comparable to those of normoglycemic untreated mice. These findings suggested that hyperglycemia itself stimulates glucagon secretion by α-cells in vivo in the absence of insulin, and that the intra-islet effect of insulin plays a central role in physiological suppression of glucagon secretion induced by high glucose.21 It was also demonstrated that hypoglycemia itself is able to stimulate glucagon secretion independently of insulin sinceit induced a large increase in glucagon secretion under supraphysiological hyperinsulinemiain both control and αIRKO mice.10 It is worth noting that this phenomenon might, in part, involve other modulators of glucagon secretion including glutamate, neurotransmitters, and the nervous system, while alternatively a direct effectof low glucose on glucagon secretion is also possible. A model for the regulation of glucagon secretion is proposed in Figure 5A by Kulkarni et al: in states of hyperglycemia the greater insulin secretion by β-cells activates insulin signaling in α-cells and represses glucagon secretion. On the other hand, in hypoglycemic states the consequent low levels of insulin would allow the α-cells to sense the reduction in ambient insulin leading to a lack of activation of insulin signaling proteins which in turn leads to the stimulation of glucagon secretion. This would occur in addition to direct stimulation by low glucose itself. A recent clinical study reported that this mechanism is actually feasible in humans22 (Kulkarni, Lecture).

 

b. Role of incretins

Incretins are hormones released from enteroendocrine cells in the gut that act to potentiate glucose clearance in response to the ingestion of food.23 This potentiation, termed the “incretin effect,” is due to the insulinotropic capacity of incretins. In fact, incretins are thought to be responsible for between 50% and 70% of the insulin release from pancreatic β-cells in response to an oral glucose load.24 There are two known mammalian incretin hormones: glucose-dependent insulinotropic peptide (GIP), and GLP-1.
■ Incretin gene
GIP and GLP-1 are each encoded by distinct genes in mammalian genomes.14 The Gip gene solely encodes GIP and its expression is restricted to the intestinal K-cells. As already mentioned, GLP-1 is encoded by the proglucagon (Gcg) gene.13 Mammalian GLP-2 is primarily involved in intestinal proliferation and function, bone breakdown, and neuroprotection. Interestingly, while glucagon activity is strongly conserved across vertebrate species, the role of GLP-1 differs between mammals and fish and the role of GLP-2 has only been defined in mammalian species.14 The function of the GIP hormone has only been described in a limited number of mammalian species with a few other diverse vertebrate species possessing a GIP-like gene.14 Both Gip and Gcg exist as single copy genes in mammalian genomes, and reside within stable genomic neighborhoods with a strong conservation of the flanking gene order.14 The glucagon and GLP-1 sequences are largely invariant among mammals.
■ Incretin functions
Functionally, these hormones are nearly identical in their ability to stimulate insulin secretion, with each being able to compensate in the event of a loss of function of the other, although each incretin has a specific receptor that is unable to be activated by the other incretin peptides.25 GIP is the first incretin produced following an enteral dose of glucose, and is secreted from the K-cells of the duodenum and the jejunum. GLP-1 is produced shortly there after by the L-type enteroendocrine cells of the ileum and colon. In addition to stimulating insulin secretion, they also promoteβ-cell survival by inhibition of apoptosis and stimulation of proliferation. The incretin hormones also have physiological effects on insulin target tissues, as liver, adipose tissue, and skeletal muscles, thereby increasing insulin sensitivity, glucose uptake, and metabolism.24 The primary difference between GIP and GLP-1 is the non-insulinotropic activity of GLP-1, but not GIP. GLP-1 acts on the stomach to inhibit gastric emptying and thus regulate nutrient uptake, as well as on thebrain to increase satiety and reduce nutrient consumption (Irwin, Lecture; Irwinand Prentice, unpublished).
GIP and GLP-1 act through receptors located on the plasma membrane of target tissues.24 The presence of GLP-1 receptors in α-cells is a highly controversial question. Administration of the GLP-1 receptor antagonist exendin 9-39 is associated with elevations of plasma glucagon concentration, suggesting that endogenous GLP-1 inhibits glucagon secretion.26 The intra-islet hypothesis that inhibition of glucagon secretion is secondary to stimulation of β-cells is incompatible with the observation that GLP-1 inhibition of both basal and stimulated glucagon secretion is also observed in patients with no residual β-cells, or after a treatment with a powerful insulin receptor antagonist (Holst, unpublished). Moreover, GLP-1 treatment significantly suppressed plasma glucagon in IRKO mice to the same extent as in controls.10 Further, an oral glucose load suppressed plasma glucagon more than did intraperitoneal glucose, where GLP-1 secretion is not stimulated, but elevates blood glucose to the same extent. These data imply that GLP-1 can suppress glucagon secretion directly and independently of insulin (Kulkarni, Lecture). Accordingly, De Marinis et al reported that expression of GLP-1 receptors on α-cells is less than 0.2% of its expression in β-cells; thus GLP-1 can induce a small elevation in cAMP activating PKAfollowed by selective inhibition of N-type Ca2+ ion channels, leading to the suppression of glucagon exocytosis (Figure 5B27). Recently, in vitro experiments in rats showedthat GLP-1−induced inhibition of glucagon secretion involves somatostatin secreted from neighboring δ-cells28 (Holst, Lecture; Holst et al, unpublished). In contrast to GLP1, GIP receptors are expressed abundantly in α-cells and their activation stimulates electrical activity significantly, leading to an increase in Ca2+, which causes glucagon exocytosis to accelerate through activation of L-type Ca2+ion channels27 (Kulkarni, Lecture).
■ Role in diabetes
Glucagon secretion exhibits characteristic abnormalities in type 2 diabetes (T2D). Very often patients have fasting hyperglucagonemia, and exaggerated responses to meal tests. Whereas in healthy controls intravenous and oral glucose brought about the same suppression of glucagon secretion, the diabetic subjects showed a paradoxical increase and lack of suppression after the oral challenge for the first 45 to 60min, but the suppression after intravenous isoglycemic glucose infusions was nearly normal.23 The same observation was made in patients with type 1 diabetes (T1D) and no residual β-cells,29 suggesting that the difference cannot result from disturbed intra-islet interaction between β- and α-cells. Recently, Lund et al30 demonstrated that the differential consequences of oral and intravenous glucose could be the result of the combined effects of 3 hormones: GLP-1 suppresses glucagon, GLP-2 reduces the suppression, while GIP reverses the suppression and causes clear stimulation for the first 30 min. Interestingly, it was recently shown that GLP-1-induced insulin stimulation and glucagon inhibition contribute equally to the glucose-lowering effect of GLP-1 in T2D patients31 (Holst, Lecture; Holst et al, unpublished).

Figure 5. Hypothetical model for the regulation of glucagon secretion by insulin and GLP-1 (from Kawamori et al, unpublished). (A) In a high glucose state, stimulated insulin secretion by β-cells acts on the insulin receptor on the surface of α-cells and then suppresses glucagon secretion in a paracrine manner. In a low glucose state, decreased insulin secretion by β-cells is recognized by α-cells as a reduction of insulin signaling in α-cells through insulin receptors, and α-cells then increase glucagon secretion in response. (B) GLP-1 directly suppresses glucagon secretion by α-cells through slight increase of cAMP followed by inhibition of N-type Ca2+ channels (27) GLP-1 also potentiates insulin secretion by β-cells and then suppresses glucagon secretion through insulin effects on α-cells. Glucose stimulates insulin secretion by β-cells and suppresses glucagon from α-cells through insulin effects, while glucose can stimulate glucagon secretion by α-cells.

 

2. Other factors

■ Uncoupling protein-2 (UCP2) is expressed in pancreatic β- and α-cells,32 as well asin other tissues, and has been shown to contribute to mitochondrial proton leakage. In the cell-cell, such activity may uncouple glucose metabolism from ATP synthesis thereby impeding the mechanism that triggers insulin secretion. The development of STZ-induced hyperglycemia is significantly less severe in UCP2KO than in wildtype mice. Interestingly, UCP2-deficient α- and β-cells had chronically higher cellular reactive oxygen species (ROS) levels than the wild-type prior to STZ application. Higher ROS in UCP2-deficient β-cells was associated with improved glucose-stimulated insulin secretion. However, higher ROS in UCP2-deficient α-cells was associated with impaired glucose-regulated glucagon secretion, leading to an attenuation of STZ-induced hyperglycemia33 (Wheeler, Lecture; Hardy et al, unpublished).
■ Somatostatin inhibits both glucagon and insulin secretion. Its effects on glucagon are most likely mediated by somatostatin receptor subtype 2, which is highly expressed in α-cells. Somatostatin activates K+ channels, which induce membrane hyperpolarization and reduce Ca2+-dependent exocytosis.34
■ GABA, a major neurotransmitter that inhibits neuronal firing in the central nervous system and is produced in the β-cells, also directly inhibits glucagon secretion by hyperpolarizing the α-cell plasma membrane.35
■ Glutamate, which has been shown to increase in parallel with glucagon, may be another important regulator of paracrine signaling to the α-cell. Glutamate, when secreted in response to low glucose, likely modulates plasma membrane potential and acts as a glucagon-positive regulator.36
■ Zinc may also act as a paracrine regulator of the α-cell. In β-cells, zinc is required for normal insulin crystallization, the formation of dense core granules, and is released along with insulin. In rat α-cells, zinc acts by increasing the activity of KATP channels, which leads to the inhibition of glucagon secretion. Conversely, in mouseα-cells, the inhibitory effects of zinc on glucagon secretion involves intracellular zinc transport by calcium channels and modulation of the cellular redox state. A recent report also indicated that a sudden decrease in zinc secretion during hypoglycemia triggers glucagon secretion, suggesting that zinc, like insulin, mediates a zinc switch-off signal for glucagon during glucose deprivation in perfused islets in mice.37 However, studies in human and mouse islets showed zinc to have no effects on glucagon secretion. Mice lacking ZnT8 zinc transporter in β-cells reduce first-phase insulin secretion but maintain normal glucagon secretion (less zinc is secreted from ZnT8βKO β-cells, leading to reduced extracellular zinc levels). Like wild-type islets, in Znt8βKO islets the cumulative inhibitory effects of other paracrine factors and potentially direct glucose sensing prevailed in the presence of high glucose. Overall, in this study, zinc secreted by β-cells does not appear to act as an inhibitor of glucagon secretion (Wheeler, Lecture; Hardy et al, unpublished).

3. Effects of glucose and role of KATP channels

In human and mouse islets, glucagon secretion is inhibited at glucose concentrations<6 mM (with a nearly maximal inhibition by 3 mM), but paradoxically is “stimulated” at higher glucose concentrations: 20 mM glucose is significantly less inhibitory than 5 mM (mouse islets) and 6 mM (human islets) (Figure 6). Interestingly, the effects of glucose on glucagon secretion, at least for the responses to concentrations up to 6 mM, seem not to be mediated by paracrine effects from β- or δ-cells since these glucose concentrations are not associated with any major stimulation of insulin or somatostatin secretion. Thus, and as already noted in this review, it seems likely that the α-cell is not only under paracrine control but is also equipped with an intrinsicregulation to respond to variations of ambient glucose. The nature of this intrinsicregulation remains poorly defined but ATP-sensitive potassium channels (KATP-channels) of the same type as those found in β-cells seem to be involved.38-39 Human α-cellscontain KATP-channels40 and their activation leads to repolarization. Rorsman et al proposed a unifying hypothesis that integrates both paracrine and intrinsic regulation of glucagon secretion. During hypoglycemia (ie, induced by exercise), glucagon secretion is enhanced due to slight activation of the KATP-channels with resultant firing of large-amplitude action potentials (Figure 7A). An increase in glucose (by ingestionof a glucose-rich meal) inhibits glucagon secretion by closure of the KATP-channels. The resulting membrane depolarization leads to a decreased amplitude of the α-cell action potential and reduced activation of voltage-gated P/Q-type Ca2+-channels, which culminates in suppression of exocytosis (Figure 7B). Following ingestion of amixed meal (rich in glucose, amino acids, and lipids), glucagon secretion is switched off by a combination of KATP-channel closure (as in Figure 7B) and activation of paracrineinhibitory signaling (indicated by the -) due to parallel stimulation of secretion in the neighboring β- and δ-cells. The latter supersedes the stimulatory effects of amino and free fatty acids (Figure 7C). In the fasted state (and starvation), however, when plasma concentrations of free fatty and amino acids are increased by mobilization of bodily depots, the low plasma glucose levels ensure minimal stimulation of exocytosis in β- and δ-cells. In the absence of inhibitory paracrine signals derived from these cells, the stimulatory effect of low glucose (mediated by activation of KATP-channels) is amplified by the presence of amino and free fatty acids (Figure 7D) (Rorsman,Lecture; Walker et al, unpublished).

Figure 6. Glucose dependence of pancreatic hormone secretion by mouse (A) and human (B) pancreatic islets (from Walker et al, unpublished). Data for insulin, glucagon and somatostatin were recorded from the same islets at 0-20 mM (mouse) and 1-20 mM glucose (man). Data are mean values± S.E.M. of 8 (A) and 7-12 (B) experiments. In (B), the individual experiments were conducted on islets from 3-4 different pancreases. Shaded area indicates range of glucose concentrations over which most of the regulation of the inhibition of glucagon secretion occurs. All values are statistically different (*P<0.05, or better) for all values measured at 3 mM glucose and above except for insulin secretion in mouse islets where the difference first becomes statistically different at 10 mM. †P<0.05 (or better) for glucagon secretion at 20 mM glucose vs that at 5 mM (mouse) or 6 mM (human).
Figure 7. Model for the regulation of glucagon secretion (from Walker et al, unpublished). The legend is detailed in the text (Chapter III.3.).

 

4. Role of the central and the autonomic nervous systems

■ Glucose-sensing cells are located at several anatomical sites: the mouth, gut, hepatoportal vein area, brainstem, and hypothalamus (mainly the arcuate nucleus and paraventricular nucleus). It is now well established that glucose-sensing cells present in the central nervous system are either excited by a rise in blood glucose (glucose-excited or GE neurons) or by fall in blood glucose concentrations (glucose-inhibited or GI neurons). These neurons are thought to be responsible for activation of the sympathetic and parasympathetic branches of the autonomic nervous system, which control glucagon and insulin secretion. On α-cells, the sympathetic neurotransmitter norepinephrine binds to β2-adrenergic receptor, which stimulates glucagon secretion, whereas on β-cells it binds to α2-adrenergic receptors, which inhibits insulin secretion.41 Activation of the sympatho adrenal system, which induces release of epinephrine into the blood by the adrenals, combines with nervous secretion of norepinephrine directly the islet cell level to stimulate glucagon secretion and to inhibit insulin secretion. Acetylcholine, the main neurotransmitter of the parasympathetic nervous system, stimulates secretion of both glucagon and insulin.
■ Central detection of hypoglycemia also controls counterregulation as demonstrated by the secretion of glucagon and catecholamines induced by intracerebroventricular injection of 2-DG. Interestingly, glucose-induced GABA secretion in the ventromedial hypothalamus blocks the hyperinsulinemic hypoglycemia-induced activation of glucagon secretion. Besides the ventromedial hypothalamus, the glucoregulatory response to hypoglycemia is also controlled by glucose-sensitive neurons of the brainstem.
■ The mechanisms of glucose sensing by GE and GI neurons are still incompletely defined.42 Some mechanisms involved in neuronal hypoglycemia detection and subsequent control of glucagon secretion can be mentioned:
• In a mouse model of the glucose transporter Glut2 gene inactivation, glucose-stimulated insulin secretion is prevented and glucagon secretion in response to insulin-induced hypoglycemia or 2-DG-induced neuroglucopenia is suppressed. Transgenic complementation studies indicate that GLUT2 expression in astrocytes but not in neurons is important for the counterregulatory response to hypoglycemia.43
• Glucokinase participates in both hyper- and hypoglycemia detection. Pharmacological intracerebroventricular inhibition of glucokinase prevents normal counterregulation to hyperinsulinemic hypoglycemia. Interestingly, recurrent hypoglycemia, which induces impaired counterregulation to subsequent hypoglycemic episodes (a model of hypoglycemia-associated autonomic failure), is associated with an increase in glucokinase activity in the hypothalamus.
• Central inhibition of KATP-channels prevents glucagon secretion in response to hyperinsulinemic hypoglycemia. In contrast, activation of this channel in the ventromedial hypothalamus amplifies the counterregulatory hormone responses to hypoglycemia in normal and recurrently hypoglycemic rats.44
• Inhibition of the metabolic sensor AMP-activated kinase in the arcuate nucleus and ventromedial hypothalamus impairs early glucagon and catecholamine responses to hypoglycemia. AMP-activated kinase may also be involved in the counterregulatory response to hypoglycemia or to neuroglucopenia45 (Thorens, Lecture; Thorens, unpublished).
■ Glucose-sensing cells are located at several anatomical sites: the mouth, gut, hepatoportal vein area, brainstem, and hypothalamus (mainly the arcuate nucleus and paraventricular nucleus). It is now well established that glucose-sensing cells present in the central nervous system are either excited by a rise in blood glucose (glucose-excited or GE neurons) or by fall in blood glucose concentrations (glucose-inhibitedor GI neurons). These neurons are thought to be responsible for activation of the sympathetic and parasympathetic branches of the autonomic nervous system, which control glucagon and insulin secretion. On α-cells, the sympathetic neurotransmitter norepinephrine binds to β2-adrenergic receptor, which stimulates glucagon secretion, whereas on β-cells it binds to α2-adrenergic receptors, which inhibits insulin secretion.41 Activation of the sympathoadrenal system, which induces release of epinephrine into the blood by the adrenals, combines with nervous secretion of norepinephrine directly the islet cell level to stimulate glucagon secretion and to inhibit insulin secretion. Acetylcholine, the main neurotransmitter of the parasympathetic nervous system, stimulates secretion of both glucagon and insulin.
■ Central detection of hypoglycemia also controls counterregulation as demonstrated by the secretion of glucagon and catecholamines induced by intracerebroventricular injection of 2-DG. Interestingly, glucose-induced GABA secretion in the ventromedialhypothalamus blocks the hyperinsulinemic hypoglycemia-induced activation of glucagon secretion. Besides the ventromedial hypothalamus, the glucoregulatory response to hypoglycemia is also controlled by glucose-sensitive neurons of the brainstem.
■ The mechanisms of glucose sensing by GE and GI neurons are still incompletely defined.42 Some mechanisms involved in neuronal hypoglycemia detection and subsequent control of glucagon secretion can be mentioned:
• In a mouse model of the glucose transporter Glut2 gene inactivation, glucose-stimulated insulin secretion is prevented and glucagon secretion in response to insulin-induced hypoglycemia or 2-DG-induced neuroglucopenia is suppressed. Transgenic complementation studies indicate that GLUT2 expression in astrocytes but not in neurons is important for the counterregulatory response to hypoglycemia.43
• Glucokinase participates in both hyper- and hypoglycemia detection. Pharmacological intracerebroventricular inhibition of glucokinase prevents normal counterregulation to hyperinsulinemic hypoglycemia. Interestingly, recurrent hypoglycemia, which induces impaired counterregulation to subsequent hypoglycemic episodes (a model of hypoglycemia-associated autonomic failure), is associated with an increase in glucokinase activity in the hypothalamus.
• Central inhibition of KATP-channels prevents glucagon secretion in response to hyperinsulinemic hypoglycemia. In contrast, activation of this channel in the ventromedial hypothalamus amplifies the counterregulatory hormone responses to hypoglycemia in normal and recurrently hypoglycemic rats.44
• Inhibition of the metabolic sensor AMP-activated kinase in the arcuate nucleus and ventromedial hypothalamus impairs early glucagon and catecholamine responses to hypoglycemia. AMP-activated kinase may also be involved in the counterregulatory response to hypoglycemia or to neuroglucopenia45 (Thorens,Lecture; Thorens, unpublished).

Editorial
“Pancreatic α-cells and glucagon—neglected metabolic actors”
I- Birth and death of the α-cell
II- Regulation of glucagon expression
III- Regulation of glucagon secretion
IV- Role of glucagon in metabolism
V- Therapeutic perspectives
VI- Conclusion
Lectures during IGIS meeting and unpublished reviews
References