- Made available online as an Accepted Preprint 31 July 2008
- Accepted Preprint first posted online on 31 July 2008
Physiology of the pancreatic α-cell and glucagon secretion: role in glucose homeostasis and diabetes
- CIBER de Diabetes y Enfermedades Metabólicas Asociadas (CIBERDEM) and Instituto de Bioingeniería, Universidad Miguel Hernández, Avenida de la Universidad s/n, 03202 Elche, Spain
- (Correspondence should be addressed to I Quesada; Email: ivanq{at}umh.es)
Abstract
The secretion of glucagon by pancreatic α-cells plays a critical role in the regulation of glycaemia. This hormone counteracts hypoglycaemia and opposes insulin actions by stimulating hepatic glucose synthesis and mobilization, thereby increasing blood glucose concentrations. During the last decade, knowledge of α-cell physiology has greatly improved, especially concerning molecular and cellular mechanisms. In this review, we have addressed recent findings on α-cell physiology and the regulation of ion channels, electrical activity, calcium signals and glucagon release. Our focus in this review has been the multiple control levels that modulate glucagon secretion from glucose and nutrients to paracrine and neural inputs. Additionally, we have described the glucagon actions on glycaemia and energy metabolism, and discussed their involvement in the pathophysiology of diabetes. Finally, some of the present approaches for diabetes therapy related to α-cell function are also discussed in this review. A better understanding of the α-cell physiology is necessary for an integral comprehension of the regulation of glucose homeostasis and the development of diabetes.
Introduction
The principal level of control on glycaemia by the islet of Langerhans depends largely on the coordinated secretion of glucagon and insulin by α- and β-cells respectively. Both cell types respond oppositely to changes in blood glucose concentration: while hypoglycaemic conditions induce α-cell secretion, β-cells release insulin when glucose levels increase (Nadal et al. 1999, Quesada et al. 2006a). Insulin and glucagon have opposite effects on glycaemia as well as on the metabolism of nutrients. Insulin acts mainly on muscle, liver and adipose tissue with an anabolic effect, inducing the incorporation of glucose into these tissues and its accumulation as glycogen and fat. By contrast, glucagon induces a catabolic effect, mainly by activating liver glycogenolysis and gluconeogenesis, which results in the release of glucose to the bloodstream. An abnormal function of these cells can generate failures in the control of glycaemia, which can lead to the development of diabetes (Dunning et al. 2005). Actually, diabetes is associated with disorders in the normal levels of both insulin and glucagon. An excess of glucagon plasma levels relative to those of insulin can be determinant in the higher rate of hepatic glucose output, which seems to be critical in maintaining hyperglycaemia in diabetic patients (Dunning et al. 2005).
Despite the importance of the α-cell and glucagon secretion in the regulation of glycaemia and nutrient homeostasis, little is known about the physiology of these cells compared with the overwhelming information about β-cells. Several factors may explain this lack of information about glucagon secretion. First, the scarcity of this cell population in islets of animal models such as mice and rats along with several technical limitations of conventional methods have made it more difficult to study α-cells than β-cells (Quoix et al. 2007). Second, the lack of functional identification patterns has also been an important limitation in α-cell research. However, in recent years notable progress has been made in the study of α-cell function at the cellular and molecular levels. This review attempts to describe recent advances in α-cell physiology and the regulation of glucagon secretion. Additionally, it focuses on the pathophysiology of these cells, their role in diabetes, as well as potential therapeutic strategies.
Islet of Langerhans: cell architecture and function
Glucagon-secreting α-cells are one of the main endocrine cell populations that coexist in the islet of Langerhans along with insulin-secreting β-cells. The islet is further composed by other scarce secretory populations such as δ- and poly-peptide releasing (PP)-cells, which release somatostatin and pancreatic polypeptide respectively. This multicellular structure constitutes the endocrine unit of the pancreas and is responsible for the regulation of blood glucose homeostasis. Approximately one million islets are distributed throughout a healthy adult human pancreas, representing 1 and 2% of the total mass of the organ. Each islet, with sizes varying from 100 to 500 μm, is made up of 1000–3000 cells. In mouse and rat islets, β-cells are the main population accounting for 60–80% of the total number of cells, while 15–20% are α-cells, <10% are δ-cells and less than 1% correspond to the PP-cell population (Brelje et al. 1989, Brissova et al. 2005). The architecture of rodent islets is characterized by the location of β-cells in the core and the non-β cells distributed in a mantle around the insulin-secreting cell population. This cellular distribution along with several studies on microcirculation within the islet suggests that the order of paracrine interactions is from β- to α- and δ-cells (Bonner-Weir 1991). The rich vascularization within the islet ensures a rapid sensing of plasma glucose levels by these endocrine cells, allowing an appropriate secretory response. In human islets, however, there are important differences in composition and spatial organization compared with rodents (Cabrera et al. 2006). While the proportion of δ- and PP-cells are similar in the human islet, β-cells are less abundant (48–59%) and the α-cell population reaches a 33–46%, suggesting that glucagon secretion plays a major role in humans (Cabrera et al. 2006). These islet cell populations show a random distribution pattern, where the majority of β-cells are in contact with non-β-cells, suggesting that paracrine interactions among different populations may be more active (Cabrera et al. 2006). Another divergence between human and rodent islets is the intercellular communication among the different populations. In mice, β-cells work as a syncytium in terms of electrical activity and Ca2+ signalling due to the high level of coupling mediated by gap junctions of connexin36 (Gopel et al. 1999, Nadal et al. 1999, Quesada et al. 2003). This coupling favours a more vigorous insulin secretion (Vozzi et al. 1995). By contrast, coupling can be found between several human β-cells in clusters within the same islet but not in the whole β-cell population (Quesada et al. 2006b). This kind of intercellular communication is probably the result of the human islet cytoarchitecture and its functional meaning is still unknown (Cabrera et al. 2006). Unlike β-cells, α- and δ-cells from rodents and humans are not functionally coupled and work as independent units. In addition to nutrients and paracrine signals, islet function is further regulated by sympathetic, parasympathetic and sensory nerves that go deeply into the islet (Ahren 2000). Thus, multiple regulation levels determine hormone release from pancreatic islets.
Glucagon secretion by pancreatic α-cells
Stimulus-secretion coupling in α-cells: from ion channel activity to exocytosis
Pancreatic α-cells are equipped with a specific set of channels that generate action potentials of Na+ and Ca2+ in the absence or at low levels of glucose (Gromada et al. 1997). This electrical activity triggers Ca2+ signals and glucagon secretion. Elevated glucose concentrations inhibit all these events. ATP-dependent K+ (KATP) channels play a fundamental role in α-cells, such as they do in β-cells, since they couple variations in extracellular glucose concentrations to changes in membrane potential and electrical activity. In intact rat α-cells, KATP channels have a lower ATP sensitivity (Ki=0.94 mM) than the one observed in excised patches (Ki=16 μM; Bokvist et al. 1999, Gromada et al. 2007), but a very similar one to the values recorded in mouse and rat intact β-cells. However, KATP channels exhibit a higher ATP sensitivity in intact mouse α-cells (Ki=0.16 mM; Leung et al. 2005). Consequently, lower ATP concentrations are required to obtain the maximal inhibition of KATP conductance compared with mouse β-cells. Recent evidence has indicated that the densities of these channels are similar in mouse α- and β-cells (Leung et al. 2005). The repolarization of action potentials is mediated by voltage-dependent K+ channels. While delayed rectifying K+ channels have been demonstrated in rat, mouse and guinea pig α-cells, a tetraethylammonium-resistant voltage-dependent K+ current (A-current) has only been identified in mice (Barg et al. 2000, Leung et al. 2005). Furthermore, tetrodotoxin-sensitive Na+ currents are fundamental for the generation of action potentials in these cells. Na+ channels are activated at voltages above −30 to −20 mV (Gopel et al. 2000), and their blockade by tetradotoxin leads to the inhibition of glucagon secretion (Gromada et al. 2004, Olsen et al. 2005, MacDonald et al. 2007). Additionally, α-cells have a heterogeneous presence of Ca2+ channel subtypes with different roles. While L and N channels have been reported in rat α-cells (Gromada et al. 1997), mouse α-cells express L-, T-, N- and probably R-type Ca2+ channels (Gopel et al. 2000, Gromada et al. 2004, Pereverzev et al. 2005, MacDonald et al. 2007). The low voltage-activated T-type channels work as pacemakers in the initiation of action potentials in mice (Gopel et al. 2000). They open around −60 mV, the action potential initiation threshold in α-cells. The high voltage-activated L and N channels open during action potentials when the membrane potential exceeds −40 to −30 mV. Although most of the Ca2+ current goes through L-type channels in α-cells, the Ca2+ required for exocytosis at low-glucose levels is mediated by N-type channels, and their blockade by ω-conotoxin-GVIA inhibits glucagon secretion in these conditions (Gromada et al. 1997, 2004, Olsen et al. 2005, MacDonald et al. 2007). However, in the presence of cAMP-elevating agents, L channels are the major conduit for Ca2+ (Gromada et al. 1997).
A model to explain the glucose regulation of electrical activity in mouse α-cells has been postulated in the light of recent studies (Fig. 1). At low-glucose levels, the activity of KATP channels renders a membrane potential of about −60 mV. At this voltage, T-type channels open, which depolarize the membrane potential to levels where Na+ and N-type Ca2+ channels are activated, leading to regenerative action potentials (Gromada et al. 2004, MacDonald et al. 2007). Ca2+ entry through N-type channels induces glucagon secretion. The repolarization of action potentials is mediated by the flowing of K+A-currents. At low-glucose concentrations, this electrical activity triggers oscillatory Ca2+ signals in both human and mouse α-cells in intact islets (Nadal et al. 1999, Quesada et al. 1999, 2006b; Fig. 2). However, the increase in extracellular glucose levels rises the cytosolic ATP/ADP ratio which blocks KATP channels, depolarizing α-cells to a membrane potential range where the channels involved in action potentials become inactivated (Gromada et al. 2004, MacDonald et al. 2007). As a consequence, electrical activity, Ca2+ signals and glucagon secretion are inhibited (Figs 1 and 2). Thus, glucagon release from α-cells is mainly supported by an intermediate KATP channel activity that maintains a membrane potential range able to sustain regenerative electrical activity (MacDonald et al. 2007). A similar model has been also proposed for human α-cells (MacDonald et al. 2007). Nevertheless, this scheme has been argued by some reports indicating that glucose may be hyperpolarizing rather than depolarizing (Liu et al. 2004, Manning Fox et al. 2006). It has also been proposed that glucose would inhibit glucagon secretion by suppressing a depolarizing Ca2+ store-operated current independent of KATP channels (Liu et al. 2004, Vieira et al. 2007).
Schematic model for glucose-dependent regulation of glucagon secretion in the mouse α-cell. Glucose is incorporated into the α-cell by the transporter SLC2A1. At low-glucose concentrations, the moderate activity of KATP channels situates the α-cell membrane potential in a range that allows the opening of voltage-dependent T- and N-type Ca2+ channels and voltage-dependent Na+ channels. Their activation triggers action potentials, Ca2+ influx and exocytosis of glucagon granules. The opening of A-type K+ channels is necessary for action potential repolarization. However, high-glucose concentrations elevate the intracellular ATP/ADP ratio, blocking KATP channels and depolarizing the membrane potential to a range where the inactivation of voltage-dependent channels takes place. This results in the inhibition of electrical activity, Ca2+ influx and glucagon secretion. The function of L-type channels predominates when cAMP levels are elevated. See text for further details.
Opposite Ca2+ signalling patterns in α- and β-cells in response to glucose. At low-glucose concentrations (0.5 mM), electrical activity triggers oscillatory Ca2+ signals in α-cells that lead to glucagon release. Elevation of glucose levels (11 mM) inhibits all these events. By contrast, 11 mM glucose stimulate Ca2+ signalling and insulin secretion in β-cells. Both fluorescence records were obtained by confocal microscopy from two cells within an intact mouse islet. Inset shows a thin optical section (∼6 μm) of a mouse islet loaded with the Ca2+-sensitive fluorescent probe Fluo-3.
In rat α-cells, the activity of KATP channels at low-glucose concentrations also keeps the membrane potential at about −60 mV, where spontaneous Na+ and Ca2+ action potentials are produced (Gromada et al. 1997). However, in contrast to the situation in mice, the stimulus-secretion coupling in rat α-cells is similar to that of β-cells. That is, elevations of extracellular glucose levels increase the intracellular ATP/ADP ratio, blocking KATP channels and depolarizing the membrane potential, which stimulates Ca2+ influx through N channels and glucagon secretion (Franklin et al. 2005, Olsen et al. 2005). Accordingly, the pharmacological inhibition of glucose metabolism increases KATP channel activity in rat α-cells (Olsen et al. 2005). This model indicating a β-cell-like stimulus-secretion coupling is based on recent studies that have used isolated rat α-cells. However, these results contrast with the observations showing that glucose inhibits α-cell electrical activity and glucagon secretion in intact rat islets (Franklin et al. 2005, Manning Fox et al. 2006). Therefore, the blocking effect observed in rat islets at high-glucose concentrations is most likely the result of paracrine signalling by β-cell activation (Wendt et al. 2004).
Regulation of α-cell function by glucose: direct or paracrine effect?
Whether glucose inhibits α-cells directly or by paracrine mechanisms has been a matter of debate, and, probably, the predominant level of control may depend on the physiological situation. Part of this controversy is also due to the divergences found in the stimulus-secretion coupling of different animal models. Although paracrine signalling may be critical for the glucose inhibition of glucagon secretion in rats (Wendt et al. 2004, Franklin et al. 2005, Olsen et al. 2005), a direct effect has been observed in mice and humans (Asplin et al. 1983, MacDonald et al. 2007). In mice and humans, a glucose direct action on α-cells has been proven in isolated cells under conditions where paracrine effects are negligible, and in intact islets incubated with different paracrine signalling inhibitors (Gromada et al. 2004, Ravier & Rutter 2005, Shiota et al. 2005, MacDonald et al. 2007, Vieira et al. 2007). Moreover, secretion studies prove that glucose inhibits glucagon release at concentrations below the threshold for β-cell activation and insulin release (MacDonald et al. 2007, Vieira et al. 2007).
Several reports on experiments using genetic mouse models support the role of glucose-modulated KATP channels in α-cell function. The regulation of glucagon secretion by glucose is impaired in ABCC8-deficient mice lacking functional KATP channels (Gromada et al. 2004, Munoz et al. 2005). A similar situation occurs in KCNJ11Y12X mouse with a KCNJ11 mutation in the KATP channel (MacDonald et al. 2007). In humans, the Glu23Lys polymorphism in the KCNJ11 subunit of these channels is associated with diminished suppression of glucagon release in response to hyperglycaemia (Tschritter et al. 2002). Nevertheless, since KATP channels seem to be essential for the α-cell regulation in the proposed models, some considerations on glucose metabolism should be taken into account. Although α-cells possess the high-affinity, low-capacity glucose transporter SLC2A1, instead of the high-capacity SLC2A2 present in the β-cell, it has been demonstrated that glucose transport is not a limiting factor in α-cell glucose metabolism (Gorus et al. 1984, Heimberg et al. 1995, 1996). However, several studies indicate that important biochemical differences exist between both cell types. While the ratio of lactate dehydrogenase/mitochondrial glycerol phosphate dehydrogenase is low in the β-cell, this ratio is higher in non-β cells (Sekine et al. 1994). Additionally, α-cells may express higher levels of the lactate/monocarboxylate transporter than β-cells but lower ones of pyruvate carboxylase (Sekine et al. 1994, Zhao et al. 2001). These biochemical differences indicate that β-cells are more efficient in the mitochondrial oxidation of glucose, while α-cells rely more on anaerobic glycolysis (Schuit et al. 1997, Quesada et al. 2006a). This lower coupling between glycolytic events in the cytosol and ATP synthesis in mitochondrial respiration of α-cells would explain the fact that, in response to glucose, cytosolic ATP increases are small in these cells (Ishihara et al. 2003, Ravier & Rutter 2005) and that ATP/ADP changes are almost invariable (Detimary et al. 1998). Therefore, some aspects at the above-mentioned models for α-cell stimulus-secretion coupling deserve more attention, especially those concerning the modulation of KATP channel activity by glucose metabolism and ATP production. Other mechanisms regulating KATP channels may also have an important role.
Regulation of glucagon secretion by fatty acids and amino acids
Although the lipotoxicity theory and its role in obesity-induced diabetes have increased the interest in the interactions between fatty acids and islet functions, little is known about their effect on the regulation of the α-cell compared with those on β-cells. While initial studies suggested an inhibitory effect on glucagon secretion (Andrews et al. 1975), more recent investigations have indicated that short-term exposure to fatty acids stimulates the release of this hormone (Bollheimer et al. 2004, Olofsson et al. 2004, Hong et al. 2005). The short-term stimulatory action depends on the chain length, spatial configuration and degree of saturation of the fatty acid (Hong et al. 2005). The action of palmitate has been studied in mice at the cell level. This fatty acid increases α-cell exocytosis by enhancing Ca2+ entry through L-type Ca2+ channels and also, by relief of the inhibitory paracrine action of the somatostatin secreted from δ-cells (Olofsson et al. 2004). A study using clonal α-cells on the long-term effect of palmitate and oleate concluded that they also enhance glucagon secretion and triglyceride accumulation in a time- and dose-dependent manner but inhibit cell proliferation (Hong et al. 2007). In agreement with this, the long-term exposure of rat islets to fatty acids induces a marked increase in glucagon release, a decrease in glucagon content and no changes in glucagon gene expression (Gremlich et al. 1997, Dumonteil et al. 2000). In addition to fatty acids, amino acids are also relevant in the modulation of the α-cell function. Amino acids such as arginine, alanine and glutamine are potent stimulators of glucagon secretion (Pipeleers et al. 1985, Kuhara et al. 1991, Dumonteil et al. 2000). However, a few amino acids such as isoleucine can also inhibit α-cell secretion while leucine has a dual effect: it is a positive stimulus at physiological concentrations but becomes a negative one at elevated levels (15 mM; Leclercq-Meyer et al. 1985). In any case, the function of amino acids and fatty acids in the α-cell requires further investigation at the cellular and molecular levels.
Autocrine, paracrine, endocrine and neural regulation of glucagon secretion
Autocrine, paracrine and endocrine signalling
The spatial distribution of α-cells and the vascular organization within the islet sustain an important intercellular communication through autocrine and paracrine mechanisms (Fig. 3). In addition to insulin, glucagon or somatostatin, secretory granules from islet cells contain other molecules with biological activity, which are released to the extracellular space by exocytosis, activating surface receptors in the same cell, in neighbouring islet cells, or in distant cells within the islet via the vascular system. Several paracrine mechanisms are activated at high-glucose concentrations as a result of β- and δ-cell stimulations, and thus, they may participate in the glucose-induced inhibition of glucagon release.
Paracrine signalling in the α-cell. See text for details. ADCY, adenylate cyclase; AMPA-R, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor; GABA, γ-aminobutyric acid; GLP1, glucagon-like peptide-1; GRM, metabotrophic glutamate receptor; PKA, protein kinase A; SSTR2, somatostatin receptor-2.
Insulin and zinc
One of the most important paracrine mechanisms responsible for inhibiting glucagon release is conducted by insulin, acting via several pathways. An appropriate expression of the insulin receptor in mouse α-cells seems to be essential for glucose-regulated glucagon secretion (Diao et al. 2005). In INR1-G9 clonal α-cells, insulin has been found to inhibit glucagon release through the activation of phosphatidylinositol 3-kinase (PIK3; Kaneko et al. 1999). The insulin receptor–PIK3 signalling pathway is also involved in the modification of the sensitivity of KATP channels to ATP in mouse α-cells, which may affect the secretory response (Leung et al. 2006). Furthermore, insulin increases KATP channel activity in isolated rat α-cells, inducing an inhibitory effect on glucagon release via membrane hyperpolarization (Franklin et al. 2005). In addition to the effects on KATP channels, insulin can translocate A-type GABA receptors to the cell membrane, which increases the response to GABA secreted by β-cells, favouring membrane hyperpolarization and suppression of glucagon secretion (Xu et al. 2006). Since α-cell Ca2+ signals rely on electrical activity, insulin also inhibits Ca2+ signals induced by low-glucose concentrations (Ravier & Rutter 2005). Therefore, several pieces of evidence indicate that insulin inhibits glucagon release mainly by altering α-cell membrane potential.
Insulin is stored within the secretory granule forming stable hexamers around two atoms of Zn2+. After exocytosis, these hexameric crystals are exposed to a change in pH from 5.5 to 7.4, become dissociated and release both atoms of Zn2+. Recent studies have claimed that zinc atoms can also work as modulators of the α-cell function (Gyulkhandanyan et al. 2008), although their role remains controversial. On one hand, it has been found that Zn2+ activates KATP channels and decreases glucagon release in isolated rat α-cells (Ishihara et al. 2003, Franklin et al. 2005). Zn2+ seems to be the switch-off signal to initiate glucagon secretion during hypoglycaemia in streptozotocin-induced diabetic rats (Zhou et al. 2007a). However, these results contrast with the absence of effects on mouse α-cells (Ravier & Rutter 2005).
Somatostatin and glucagon
Somatostatin is produced and secreted by several tissues in addition to the δ-cell population of the islet and works as an inhibitor of both glucagon and insulin release (Fehmann et al. 1995). Immunocytochemical studies in human islets have demonstrated that, among the five identified somatostatin receptor (SSTR) subtypes, SSTR2 is highly expressed in α-cells while SSTR1 and SSTR5 are expressed in β-cells (Kumar et al. 1999). In mice and rats, SSTR2 also predominates in the α-cell and SSTR5 in the β-cell population (Hunyady et al. 1997, Strowski et al. 2000). These receptors are coupled to G-proteins and induce multiple intracellular effects. Electrophysiological studies have shown that somatostatin activates K+ channels in α-cells, inducing membrane hyperpolarization and suppressing electrical activity, which affects Ca2+-dependent exocytosis (Yoshimoto et al. 1999, Gromada et al. 2001). Capacitance measurements have further elucidated that somatostatin directly decreases exocytosis by depriming secretory granules through the activation of the serine/threonine protein phosphatase calcineurin pathway (Gromada et al. 2001). Also, a negative interaction of somatostatin with adenylate cyclase and cAMP levels has been reported in rat α-cells (Schuit et al. 1989). In addition to the effects of insulin and somatostatin on α-cells, glucagon itself works as an extracellular messenger. It exerts an autocrine positive feedback that stimulates secretion in both isolated rat and mouse α-cells by an increase in exocytosis associated to a rise in cAMP levels (Ma et al. 2005).
GLP1
The incretin hormone glucagon-like peptide 1 (GLP1) is released from the L-cells of the small intestine after food intake, stimulating insulin production and inhibiting glucagon release. Because of this dual effect, GLP1 is a potential therapeutic agent in the treatment of diabetic patients that manifest insulin deficiency as well as hyperglucagonaemia (Dunning et al. 2005). The observed suppressing effect of GLP1 on glucagon secretion in vivo and in perfused pancreas contrasts with those effects found in single α-cells (Dunning et al. 2005). In isolated rat α-cells, GLP1 stimulates glucagon secretion by interacting with specific receptors coupled to G-proteins that activate adenylate cyclase, which increases cAMP levels (Ding et al. 1997, Ma et al. 2005). Thus, it seems that paracrine mechanisms may be responsible for the GLP1 suppressing action (Dunning et al. 2005). This possibility has been underscored by the findings in experiments using β-cell-specific knock-out mice for the transcription factor Pdx1. In these mice, the lack of effect of GLP1 on β-cells is also accompanied by its inability to induce an inhibitory action on glucagon plasma levels (Li et al. 2005). Moreover, GLP1 may also affect the α-cell function by interacting with the autonomic nervous system (Balkan & Li 2000).
Other extracellular messengers
The neurotransmitter γ-aminobutyric acid (GABA) is another α-cell modulator. GABA accumulates in β-cell vesicles and is released by Ca2+-dependent exocytosis, stimulating A-type GABA receptors in neighbouring α-cells. Activation of these receptors is coupled to inward Cl− currents that hyperpolarize the α-cell plasma membrane, decreasing glucagon release in rats and guinea pigs (Rorsman et al. 1989, Wendt et al. 2004). Similar conclusions were obtained in mouse islets and clonal αTC1–9 cells (Xu et al. 2006, Bailey et al. 2007). The neurotransmitter l-glutamate also accumulates in the α-cell secretory granules because of vesicular glutamate transporters 1 and 2 found in these cells (Hayashi et al. 2003a). In low-glucose conditions, l-glutamate is cosecreted with glucagon, triggering GABA release from neighbouring β-cells and, subsequently, inhibiting the α-cell function as previously described (Hayashi et al. 2003b). Additionally, glutamate can activate autocrine signalling pathways in α-cells through the multiple glutamate receptors expressed in these cells, which include ionotrophic AMPA and kainate subtypes and metabotrophic receptors (Inagaki et al. 1995, Uehara et al. 2004, Cabrera et al. 2008). Although activation of ionotrophic receptors may stimulate glucagon release (Bertrand et al. 1993), metabotrophic glutamate receptors inhibit rat glucagon secretion at low-glucose concentrations through a down-regulation of cAMP levels (Uehara et al. 2004). Another α-cell regulator is amylin or islet amyloid pancreatic polypeptide (Iapp). This polypeptide is a 37 amino acid hormone mainly synthesized in β-cells, although it can be produced in δ-cells as well. This peptide is cosecreted with insulin by exocytosis and has an inhibitory effect on glucagon basal concentrations as well as on those levels observed after arginine stimulation (Akesson et al. 2003, Gedulin et al. 2006). This glucagonostatic effect has been reported in the plasma levels of mice and rats as well as in perfused pancreas or intact islets. Since amylin also reduces somatostatin and insulin release, some authors have proposed that endogen amylin within the islet may establish a negative feedback to avoid excessive secretion from α-, β- and δ-cells (Wang et al. 1999). Also, the purinergic messenger ATP is highly accumulated in β-cell secretory granules and in nerve terminals. It has recently been reported that ATP inhibits Ca2+ signalling and glucagon secretion in mouse α-cells, indicating that purinergic receptors are involved in α-cell function (Tudurí et al. 2008). Purinergic regulation of glucagon release has also been described in rat islets (Grapengiesser et al. 2006).
Neural regulation
As previously stated, the islet of Langerhans is highly innervated by parasympathetic and sympathetic nerves that ensure a rapid response to hypoglycaemia and protection from potential brain damage (Ahren 2000). Some terminals of these nerves store and release classical neurotransmitters, such as acetylcholine and noradrenaline, as well as several neuropeptides, which stimulate or inhibit glucagon secretion depending on the neural messenger released. Cholinergic stimulation involving muscarinic receptors and intracellular Ca2+ mobilization enhances α-cell function both in vivo and in isolated cells (Ahren & Lundquist 1982, Berts et al. 1997). Noradrenaline increases glucagon secretion as well (Ahren et al. 1987). Sympathetic activation can also induce adrenaline release from the adrenal medulla, which potently stimulates glucagon secretion by enhancing Ca2+ influx through L-type Ca2+ channels and accelerating granule mobilization (Gromada et al. 1997). In addition to classical neurotransmitters, several neuropeptides such as vasoactive intestinal polypeptide, pituitary adenylate cyclase-activating polypeptide and gastrin-releasing peptide, which may stimulate glucagon release from pancreas, can be accumulated in parasympathetic nerves, while galanin and neuropeptide Y can be stored in sympathetic nerve terminals (Ahren 2000). Multiple actions have been reported for the latter neuropeptides. The effects and mechanisms involved in neural regulation of α-cells have yet to be established at the cellular and molecular levels. These systems are mainly regulated by glucose-sensing neurons of the ventromedial hypothalamus, which respond to plasma glucose levels with mechanisms very similar to those of the β-cell, including the activity of glucose-regulated KATP channels (Borg et al. 1994, Miki et al. 2001, Song et al. 2001). Actually, it has been observed that the α-cell response to hypoglycaemia is also impaired in KCNJ11-deficient mice whose neurons of the ventromedial hypothalamus lack functional KATP channels and glucose responsiveness (Miki et al. 2001).
Glucagon physiological and pathophysiological actions and its role in diabetes
Glucagon synthesis
The preproglucagon-derived peptides glucagon, GLP1 and GLP2, are encoded by the preproglucagon gene, which is expressed in the central nervous system, intestinal L-cells and pancreatic α-cells. A post-translational cleavage by prohormone convertases (PC) is responsible for the maturation of the preproglucagon hormone that generates all these peptides (Mojsov et al. 1986). The different expression of PC subtypes in each tissue mediates the production of each different peptide. In α-cells, the predominance of PCSK2 leads to a major production of glucagon together with the products glicentin, glicentin-related pancreatic polypeptide, intervening peptide 1 and the major proglucagon fragment (Dey et al. 2005). The absence of PCSK2 in knock-out mice leads to a lack of mature glucagon (Furuta et al. 2001). In enteroendocrine cells, PCSK1/3 enzymes cleave the preproglucagon hormone to generate GLP1 and GLP2 along with glicentin, intervening peptide 2 and oxyntomodulin (Mojsov et al. 1986).
The regulation of glucagon gene expression has not been studied as extensively as the insulin gene. The inhibitory effect of insulin on glucagon secretion has also been confirmed in gene expression and it occurs at the transcriptional level (Philippe et al. 1995). In diabetic rats, glucagon gene expression is augmented and is accompanied by hyperglucagonaemia in conditions of hyperglycaemia and insulin deficiency. Insulin treatment normalized glucagon expression and plasma levels in these rats, an effect that was not attributed to the restoration of normal glucose levels (Dumonteil et al. 1998). It was concluded that insulin, unlike glucose, modulates glucagon expression. The lack of response to glucose was further confirmed in isolated rat islets (Gremlich et al. 1997, Dumonteil et al. 2000), contrasting with the up-regulation of glucagon expression observed in clonal α-cells (Dumonteil et al. 1999, McGirr et al. 2005). The effect of amino acids on glucagon gene regulation has also been studied. While arginine increases glucagon expression in isolated rat islets: a process that is mediated by protein kinase C (PKA; Yamato et al. 1990, Dumonteil et al. 2000), histidine plays a fundamental role in clonal αTC1–6 cells (Paul et al. 1998). Other nutrients, such as the fatty acid palmitate, produces a down-regulated glucagon expression at short term in rat islets in a dose-dependent manner (Bollheimer et al. 2004). By contrast, no effect with palmitate has been observed in other long-term studies (Gremlich et al. 1997, Dumonteil et al. 2000). Like insulin, somatostatin also inhibits glucagon expression. It has been reported that somatostatin down-regulates glucagon expression basal levels as well as those produced by forskolin stimulation in clonal INR1G9 cells (Fehmann et al. 1995, Kendall et al. 1995).
Glucagon receptor
The rat and mouse glucagon receptor is a 485 amino acid protein, belonging to the secretin–glucagon receptor II class family of G protein-coupled receptors (Mayo et al. 2003). Glucagon binding to this receptor is coupled to GTP-binding heterotrimeric G proteins of the Gαs type that leads to the activation of adenylate cyclase, cAMP production and PKA. This receptor can also activate the phospholipase C/inositol phosphate pathway via Gq proteins, resulting in Ca2+ release from intracellular stores (Fig. 4; Wakelam et al. 1986, Mayo et al. 2003). The glucagon receptor is present in multiple tissues including the liver, pancreas, heart, kidney, brain and smooth muscle. Thus, it modulates multiple responses in these tissues, including effects on ion transport and glomerular filtration rate in kidney among others (Ahloulay et al. 1992). In any case, the regulation of glucose homeostasis is the major function of glucagon and its receptor. This role will be described in the next paragraph.
The role of glucagon and the glucagon receptor in the liver. Glucagon signalling regulates positively (+) or negatively (−) the multiple steps of the hepatic glucose metabolism. This modulation affects the expression and/or the activity of several enzymes of glucose metabolism. See text for details. ADCY, Adenylate cyclase; CREB, cAMP response element binding; F(1,6)P2, fructose-1,6-bisphosphate; F(2,6)P2, fructose-2,6-bisphosphate; F-6-P, fructose 6-phosphate; FBP1, fructose-1,6-bisphosphatase; FBP2, fructose-2,6-bisphosphatase; G-1-P, glucose 1-phosphate; G-6-P, glucose 6-phosphate; G6PC, glucose-6-phosphatase; GP, glycogen phosphorylase; GS, glycogen synthase; IP3, inositol 1,4,5-trisphosphate; OAA, oxaloacetate; PC, pyruvate carboxylase; PEP, phosphoenolpyruvate; PCK2, phosphoenolpyruvate carboxykinase; PFKM, phosphofructokinase-1; PPARGC1A, peroxisome proliferators-activated receptor-γ coactivator-1; PIP2, phosphatidylinositol 4,5-bisphosphate; PKLR, pyruvate kinase; PLC, phospholipase C; Pyr, pyruvate. Dashed lines: red, inhibition; blue, stimulation.
Glucagon control of glucose homeostasis and metabolism
Several lines of defence protect the organism against hypoglycaemia and its potential damaging effects, especially in the brain, which depends on a continuous supply of glucose, its principal metabolic fuel. These defences include decreased insulin release and increased secretion of adrenaline and glucagon. Additionally, glucose-sensing neurons of the ventromedial hypothalamus further control responses to glycaemia changes, as previously mentioned. Among all these regulatory systems, glucagon plays a central role in the response to hypoglycaemia and also opposes to insulin effects. The main action of glucagon occurs in the liver where the insulin/glucagon ratio controls multiple steps of hepatic metabolism. Glucagon stimulates gluconeogenesis and glycogenolysis, which increases hepatic glucose output, ensuring an appropriate supply of glucose to body and brain, and at the same time, it decreases glycogenesis and glycolysis. The glucagon receptor in the liver is highly selective for glucagon, but it exhibits a modest affinity for glucagon-like peptides (Hjorth et al. 1994). Its main action on the liver is mediated by the activation of adenylyl cyclase and the PKA pathway. Glucagon regulates gluconeogenesis mainly by the up-regulation of key enzymes such as glucose-6-phosphatase (G6PC) and phosphoenolpyruvate carboxykinase (PCK2) through the activation of the cAMP response element-binding protein (CREB) and peroxisome proliferator-activated receptor γ-coactivator-1 (PPARGC1A; Herzig et al. 2001, Yoon et al. 2001; Fig. 4). PCK2 and G6PC, along with fructose-1,6-biphosphatase (FBP1) have a key role in the rate of gluconeogenesis (Fig. 4). PCK2 mediates the conversion of oxalacetate into phosphoenolpyruvate while G6PC regulates glucose production from glucose-6-phosphate. FBP1 is responsible for the conversion of fructose-1,6-biphosphate (F(1,6)P2) into fructose-6-phosphate (F6P). Its activity is regulated by glucagon since this hormone decreases the intracellular levels of fructose-2,6-biphosphate (F(2,6)P2), an allosteric inhibitor of FBP1 (Kurland & Pilkis 1995). Additionally, this decrease in F(2,6)P2 also reduces the activity of phosphofructokinase-1 (PFKM), down-regulating glycolysis. The glycolytic pathway is further inhibited by glucagon at the pyruvate kinase (PKLR) level (Slavin et al. 1994). Glycogen metabolism is mainly determined by the activity of glycogen synthase (GS) and glycogen phosphorylase (GP). While glucagon is important for GP phosphorylation and activation, it inhibits GS function by phosphorylation and its conversion into an inactive form of the enzyme (Band & Jones 1980, Ciudad et al. 1984, Andersen et al. 1999).
Glucagon can also stimulate the uptake of amino acids for gluconeogenesis in the liver. Indeed, subjects with hyperglucagonaemia can develop plasma hypoaminoacidaemia, especially of amino acids involved in gluconeogenesis, such as alanine, glycine and proline (Cynober 2002). Glucagon is also involved in the regulation of fatty acids in adipocytes. Hormone-sensitive lipase mediates the lipolysis of triacylglycerol into the non-esterified fatty acids and glycerol, which are released from adipocytes. It has been reported that although glucagon does not modify the transcriptional levels of this enzyme, it increases the release of glycerol from adipocytes (Slavin et al. 1994). This mobilization of glycerol from adipose tissue can further be used in the liver during gluconeogenesis. However, the existence of a lipolytic action of glucagon observed in several animal models is still controversial in humans. While a positive effect of glucagon on lipolysis has been reported in human subjects (Carlson et al. 1993), several recent studies have indicated that it lacks a role in a physiological context (Gravholt et al. 2001). An elevated glucagon to insulin ratio accelerates gluconeogenesis as well as fatty acid β-oxidation and ketone bodies formation (Vons et al. 1991). Thus, glucagon may also be involved in diabetic ketoacidosis, a medical complication in diabetes derived from the overproduction of ketone bodies (Eledrisi et al. 2006).
The role of α-cell function in diabetes
More than 30 years ago, Unger & Orci (1975) proposed the bihormonal hypothesis to explain the pathophysiology of diabetes. According to this hypothesis, this metabolic disease is the result of an insulin deficiency or resistance along with an absolute or relative excess of glucagon, which can cause a higher rate of hepatic glucose production than glucose utilization, favouring hyperglycaemia. At present, there exists multiple clinical and experimental evidence that support this hypothesis. The rate of hepatic glucose output has been correlated with the hyperglycaemia found in animal models of diabetes as well as in human diabetes, and the maintenance of this abnormality has also been associated with hyperglucagonaemia (Baron et al. 1987, Consoli et al. 1989, Gastaldelli et al. 2000, Dunning & Gerich 2007, Li et al. 2008). In type 2 diabetes, the impairment of insulin release and development of insulin resistance is often accompanied by absolute or relative increased levels of glucagon in the fasting and postprandial states (Reaven et al. 1987, Larsson & Ahren 2000). In this situation, insulin is not effective as a negative feedback for hepatic glucose output while glucagon potentiates glucose mobilization from the liver, thus contributing to hyperglycaemia. Another malfunction reported in diabetic patients is the lack of suppression of glucagon release in hyperglycaemic conditions, which would contribute further to postprandial hyperglycaemia in both type 1 and type 2 diabetes (Dinneen et al. 1995, Shah et al. 2000). However, this irregular α-cell behaviour does not occur when insulin levels are adequate, suggesting that abnormalities in glucagon release are relevant for hyperglycaemia in the context of diabetes or impairment of insulin secretion or action (Shah et al. 1999). Hyperglucagonaemia is also responsible for the development of hyperglycaemia and diabetes in patients with the glucagonoma syndrome, a paraneoplastic phenomenon characterized by an islet α-cell pancreatic tumour (Chastain 2001).
Another defect in normal glucagon secretion has important consequences in the management of hypoglycaemia. The secretory response of α-cells to low-glucose concentrations is impaired in type 1 and long-lasting type 2 diabetes, increasing the risk of episodes of severe hypoglycaemia, especially in patients treated with insulin (Cryer 2002). In this regard, iatrogenic hypoglycaemia is a situation that implies insulin excess and compromised glucose counter-regulation, and it is responsible for a major complication in diabetes treatment, increasing the morbidity and mortality of this disease (Cryer 2002). This lack of glucagon response to hypoglycaemia has been associated with multiple failures in α-cell regulation; yet, the mechanisms are still under study (Bolli et al. 1984, Cryer 2002, Zhou et al. 2007b). Even though islet allotransplantation can provide prolonged insulin independence in patients with type 1 diabetes, the lack of α-cell response to hypoglycaemia usually persists after transplantation, indicating that this procedure does not restore the physiological behaviour of α-cells (Paty et al. 2002).
All these problems in the glucagon secretory response observed in diabetes have been attributed to several defects in α-cell regulation including defective glucose sensing, loss of β-cell function, insulin resistance or autonomic malfunction. However, the mechanisms involved in α-cell pathophysiology still remain largely unknown and deserve more investigation for better design of therapeutic strategies. In this regard, although direct therapeutic approaches to correct the lack of α-cell response to hypoglycaemia are missing, several proposals have been developed to amend glucagon excess, as we will see in the next section.
Molecular pharmacology of glucagon release and action: therapeutic potential in diabetes treatment
Given that absolute or relative glucagon excess seems to be critical in the development and/or maintenance of hyperglycaemia in diabetes by increasing hepatic glucose output, the strategies targeted to correct this malfunction are suitable for the improvement of glucose levels. In this respect, several experimental and therapeutic approaches have been developed (for a further review, see Dunning & Gerich 2007). The specific control of glucagon secretion by pharmacological modulation is complex since several components of the α-cell stimulus-secretion coupling are also present in β- and δ-cells. Thus, the manipulation of glucagon action by modulating the glucagon receptor signalling seems to be an effective alternative (Li et al. 2008). This strategy has been supported by several studies. Glucagon receptor knock-out mice have hyperglucagonaemia and α-cell hyperplasia, but their glucose tolerance is improved and they develop only a mild fasting hypoglycaemia (Gelling et al. 2003). These mice have a normal body weight, food intake and energy expenditure although less adiposity and lower leptin levels. These results are consistent with the experiments with anti-sense oligonucleotides for the glucagon receptor. Diabetic db/db mice treated with these oligonucleotides had lower glucose, triglyceride and free fatty acids blood levels, as well as improved glucose tolerance, and they developed hyperglucagonaemia without apparent effects on α-cell size or number (Liang et al. 2004). This approach is also accompanied by an increase in GLP1 and insulin levels in Zucker diabetic fatty rats and db/db and ob/ob mice (Sloop et al. 2004). Furthermore, the use of high affinity, neutralizing glucagon monoclonal antibodies improved glycaemic control and reduced hepatic glucose production in diabetic ob/ob mice (Sorensen et al. 2006). Therefore, these experimental results are a further support that glucagon antagonism may be beneficial for diabetes treatment.
Modulation of glucagon secretion
Sulphonylureas
Sulphonylureas are efficient KATP channel blockers that have been extensively used for the clinical treatment of diabetes. In rat α-cells, sulphonylureas stimulate electrical activity, Ca2+ signalling and glucagon release (Franklin et al. 2005). In mice, however, tolbutamide produces membrane depolarization, but a decrease in Ca2+ signalling and glucagon release (Gromada et al. 2004). Recent experiments in mouse α-cells have shown that, in the absence of glucose, this drug increases glucagon secretion at concentrations up to 1 μM, but higher doses are inhibitory (MacDonald et al. 2007). This biphasic effect is due to the mouse α-cell electrical behaviour (Fig. 1): glucagon release takes place within a narrow window of intermediate KATP channel activity (and membrane potential), and thus it is inhibited when the cell is hyperpolarized or depolarized beyond this membrane potential range (MacDonald et al. 2007). Accordingly, with this scheme, the KATP channel opener diazoxide can also have a biphasic effect on glucagon secretion. These effects will change depending on the extracellular glucose concentrations that necessarily influence KATP channel activity (MacDonald et al. 2007). This biphasic behaviour may explain the disparity of effects found for sulphonylureas (Loubatieres et al. 1974, Ostenson et al. 1986). In humans, sulphonylureas are associated to a glucagon secretion decrease in healthy and type 2 diabetic subjects (Landstedt-Hallin et al. 1999), while they stimulate glucagon levels in type 1 diabetic patients (Bohannon et al. 1982). Since sulphonylureas also induce insulin and somatostatin secretion, which affect α-cells, these drugs offer a poor specific control of glucagon secretion.
GLP1 mimetics and DPP4 inhibitors
In addition to stimulating insulin release, GLP1 can suppress glucagon secretion in humans, perfused rat pancreas and isolated rat islets in a glucose-dependent manner (Guenifi et al. 2001, Nauck et al. 2002). Because GLP1 is rapidly cleaved and inactivated by the enzyme dipeptidyl peptidase-IV (DPP4), a good alternative would be to design either GLP1 derivatives with higher resistance to DPP4 or agents that increase GLP1 endogenous levels. Among the GLP1 mimetics, exenatide is a synthetic polypeptide with high resistance to DPP4 cleavage that decreases glucagon levels in normal and diabetic subjects (Degn et al. 2004). Liraglutide, another GLP1 derivative with long-lasting actions, can reduce glucagon release after a meal in patients with type 2 diabetes (Juhl et al. 2002). Alternatively, DPP4 inhibitors like sitagliptin and vildagliptin increase the endogen effects of GLP1, reducing glucagon plasma concentrations in diabetic individuals (Rosenstock et al. 2007). Since all these alternatives produce opposing actions on insulin and glucagon, they generate promising expectations for diabetes treatment.
Imidazolines
The insulinotrophic effects of imidazoline compounds are mediated by KATP channel blockade, which leads to depolarization, Ca2+ influx and secretion, and by direct interactions with the exocytotic machinery (Zaitsev et al. 1996). Remarkably, imidazoline compounds such as phentolamine also suppress glucagon secretion in both rat and mouse islets, an action mediated by a direct effect on α-cell exocytosis via activation of phosphatase calcineurin proteins, and independent of KATP channels or Ca2+ currents (Hoy et al. 2001). Given that imidazoline compounds stimulate insulin release while inhibiting glucagon secretion, these drugs are potentially valuable in diabetes.
Somatostatin analogues
Because of the different expression of SSTR in the islet (Kumar et al. 1999), several studies have explored the modulation of glucagon secretion by subtype-specific somatostatin analogues (Strowski et al. 2006). It has been shown that SSTR2 is the subtype receptor predominantly expressed in rodent α-cells, and that SSTR2-deficient mice develop hyperglycaemia and non-fasting hyperglucagonaemia (Singh et al. 2007). In mice, the use of a highly SSTR2-selective non-peptide agonist inhibited glucagon release without affecting insulin release (Strowski et al. 2006). However, there is some overlapping in human islets between the different SSTR subtypes in α- and β-cells that limit, at present, the use of subtype-specific somatostatin analogues (Singh et al. 2007).
Amylin and pramlintide
Amylin, which is cosecreted with insulin from β-cells, inhibits glucagon secretion stimulated by amino acids but does not affect hypoglycaemia-induced glucagon release (Young 2005). Since α-cell response to amino acids is often exaggerated in diabetic patients, amylin or amylinomimetic compounds such as pramlintide are used as an effective alternative for the treatment of postprandial and amino acid-induced excess of glucagon secretion (Dunning et al. 2005, Young 2005).
Modulation of glucagon action and glucagon receptor signalling
Peptide-based glucagon receptor antagonists
Several linear and cyclic glucagon analogues have been developed to work as glucagon receptor antagonists. Essentially, they impair the ability of glucagon to stimulate adenylate cyclase activity in liver, thus reducing hepatic glucose output and improving plasma glucose levels. This is the case of [des-His1, des-Phe6, Glu9] glucagon-NH2, which reduces glucose levels in streptozotocin-induced diabetic rats (Van Tine et al. 1996). Recent investigations have demonstrated that the antagonist des-His-glucagon binds preferentially to the hepatic glucagon receptor in vivo, and this correlates with the glucose lowering effects (Dallas-Yang et al. 2004).
Non-peptide glucagon receptor antagonists
Multiple, competitive and non-competitive, non-peptide antagonists have been reported to act on glucagon binding and/or function. For instance, a novel competitive antagonist (N-[3-cyano-6-(1, 1-dimethylpropyl)-4, 5, 6, 7-tetrahydro-1-benzothien-2-yl]-2-ethylbutanamide) was recently shown to inhibit glucagon-mediated glycogenolysis in primary human hepatocytes and to block the increase in glucose levels after the administration of exogenous glucagon in mice (Qureshi et al. 2004). The information about the effect of these antagonists on humans is, however, scarce. In this respect, Bay 27-9955 is an oral glucagon receptor antagonist that has been tested in humans, demonstrating its efficacy in reducing glucose levels induced by exogenous glucagon (Petersen & Sullivan 2001).
Despite the success of several approaches to modulate glucagon secretion or action and improve glucose control in animal models or in humans, more information is still required. Long-standing studies should address whether the utilization of these agents could lead to undesired hypoglycaemia in humans, accumulation of lipids or compensatory mechanisms that decrease the benefits of these therapies in the long term. In this aspect, the results obtained in animal models are positive: although the glucagon receptor knock-out mouse develops hyperglucagonaemia, it is not hypoglycaemic and does not have an abnormal accumulation of lipids (Gelling et al. 2003). Additionally, recent long-term studies in mice further prove the viability of glucagon antagonism (Winzell et al. 2007). Thus, present data are promising and indicate that several therapeutic agents targeted to glucagon signalling and α-cell secretion may be useful for the management of diabetes.
Conclusions
Pancreatic α-cells and glucagon secretion are fundamental components of the regulatory mechanisms that control glucose homeostasis. However, α-cell physiology has remained elusive compared with the overwhelming information about insulin secretion and the β-cell. In recent years, however, several groups have initiated intensive efforts to understand α-cell physiology and identified essential pieces of its stimulus-secretion coupling. Additionally, important aspects of the regulation of α-cell metabolism and the control of glucagon expression are being elucidated. All of this information will favour an overall comprehension of the α-cell function and its role in glucose homeostasis. Nevertheless, more research is required to understand the α-cell behaviour, not only in healthy subjects but in pathological conditions as well. In conclusion, since the malfunction of the glucagon secretory response is involved in diabetes and its complications, a complete understanding of the α-cell will allow for a better design of therapeutic approaches for the treatment of this disease.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by grants from the Ministerio de Educación y Ciencia (BFU2007-67607 and PCI2005-A7-0131 to I Q; BFU2005-01052 to A N). CIBERDEM is an initiative of the Instituto de Salud Carlos III.
- Received in final form 30 July 2008
- Accepted 31 July 2008
- © 2008 Society for Endocrinology
References
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
