Ghrelin Uses Gα_i2 and Activates Voltage-Dependent K⁺ Channels to Attenuate Glucose-Induced Ca²⁺ Signaling and Insulin Release in Islet β-Cells

RESEARCH DESIGN AND METHODS

Male Wistar rats (Japan SLC, Hamamatsu, Japan), Ghr-KO mice (17), and wild-type C57BL/6J mice (Charles River Laboratories Japan, Yokohama, Japan) were housed in accordance with our institutional guidelines and with the Japanese Physiological Society's guidelines for animal care. PTX (5 μg/kg) (Sigma-Aldrich, St. Louis, MO) or saline was intravenously injected. Three days after injections, animals were used for in vivo and in vitro studies.

Measurements of plasma insulin and GH concentrations.

Ghrelin (Peptide Institute, Osaka, Japan) and [d-Lys³]-GHRP-6 (Sigma-Aldrich) were administered intraperitoneally to male Wistar rats (8 weeks old) after overnight fasting. U-73122 (Wako Pure Chemical, Osaka, Japan) and U-73343 (Wako) were administered intraperitoneally 15 min before ghrelin administration. In glucose tolerance tests (GTTs), 2 g/kg glucose with or without ghrelin (10 nmol/kg) was injected intraperitoneally into rats. Blood was obtained from tails, and plasma insulin and GH concentrations were measured using ELISA kits (Morinaga Institute of Biological Science, Yokohama, Japan) and enzyme immunoassay kits (GE Healthcare Bio-Science, Piscataway, NJ), respectively.

Preparation of pancreatic islets and single β-cells from rats and mice.

Islets of Langerhans were isolated by collagenase digestion from male Wistar rats aged 8–12 weeks and male Ghr-KO and wild-type mice aged 10 weeks, as reported previously (4,25) with slight modification. Animals were anesthetized by intraperitoneal injection of pentobarbitone at 80 mg/kg, followed by injection of collagenase at 1.05 mg/ml (Sigma-Aldrich) into the common bile duct. Collagenase was dissolved in 5 mmol/l Ca²⁺-containing HEPES-added Krebs-Ringer bicarbonate buffer (HKRB) solution (129 mmol/l NaCl, 5 mmol/l NaHCO₃, 4.7 mmol/l KCl, 1.2 mmol/l KH₂PO₄, 2 mmol/l CaCl₂, 1.2 mmol/l MgSO₄, and 10 mmol/l HEPES, at pH 7.4 with NaOH). HKRB was added with 0.1% BSA except for patch-clamp experiments. The pancreas was dissected out and incubated at 37°C for 16 min. Islets were collected and either used for insulin release experiments or dispersed into single cells in Ca²⁺-free HKRB. The single cells were plated sparsely on coverslips or 12-well plates and maintained in short-term culture for up to 2 days at 37°C in an atmosphere of 5% CO₂ and 95% air in Eagle's minimal essential medium containing 5.6 mmol/l glucose supplemented with 10% fetal bovine serum, 100 μg/ml streptomycin, and 100 units/ml penicillin.

Measurements of insulin release in rat and mouse islets.

Groups of 12–15 islets were incubated for 1 h at 37°C in HKRB with 2.8 mmol/l glucose for stabilization, followed by test incubation for 1 h in HKRB with 2.8 or 8.3 mmol/l glucose. Ghrelin, [d-Lys³]-GHRP-6, tetraethylammonium (TEA; Wako), or stromatoxin (ScTx; Sigma-Aldrich) was present throughout the incubation. As a positive control, islets were stimulated by addition of 25 mmol/l KCl (final concentration 29.7 mmol/l) in HKRB in the presence of 8.3 mmol/l glucose. Insulin concentrations were determined by ELISA (Morinaga Institute of Biological Science).

Measurements of [Ca²⁺]_i in rat single β-cells.

Isolated and dissociated single β-cells on coverslips were mounted in an open chamber and superfused in HKRB. [Ca²⁺]_i in single β-cells were measured by dual-wavelength fura-2 microfluorometry with excitation at 340/380 nm and emission at 510 nm using a cooled charge-coupled device camera. The ratio image was produced on an Aquacosmos system (Hamamatsu Photonics, Hamamatsu, Japan).

Patch-clamp experiments in rat single β-cells.

Perforated whole-cell currents were recorded using a pipette solution containing nystatin (150 μg/ml) dissolved in 0.1% DMSO, as previously reported (26). Membrane currents, recorded using an amplifier (200B; Axopatch, Foster, CA), were stored online in a computer using pCLAMP9 software. Voltage clamp in perforated mode was considered to be adequate when the series resistance was <20 MΩ. Patch pipettes were pulled from glass tubings (Narishige, Tokyo, Japan); the resistances of the pipettes ranged from 4 to 7 MΩ when filled with pipette solution that contained 40 mmol/l K₂SO₄, 50 mmol/l KCl, 5 mmol/l MgCl₂, 0.5 mmol/l EGTA, and 10 mmol/l HEPES at pH 7.2 with KOH. Cell capacitance recorded from single β-cells was 7.7 ± 0.22 pF (n = 64). β-Cells were voltage-clamped at a holding potential of −70 mV and then shifted to test potentials from −60 to 40 mV in 10-mV steps with pulses of 100-ms durations at 5- to 8-s intervals. Because the ATP-sensitive K⁺ (K_ATP) channel has voltage-independent kinetics and is a major determinant of resting potential of β-cells, the currents elicited by voltages around the resting potential are those through the K_ATP channels. Thus, to record the K_ATP channel current, the β-cells were voltage-clamped to the holding potential of −70 mV, stepped to −100 mV to apply voltage ramp from −100 to −50 mV at a speed of 25 mV/100 ms, and stepped back to −70 mV every 10 s. The currents evoked by the ramp pulses during superfusion with 2.8 mmol/l glucose were substantially inhibited by exposure to tolbutamide or by increasing glucose concentration to 8.3 mmol/l. These properties are consistent with features of K_ATP channels. The membrane potentials were recorded by switching to current-clamp mode from whole-cell clamp mode on the amplifier. Electrophysiological experiments were performed at room temperature (25°C).

Antisense study.

Antisense oligonucleotides specific for the Gα_i1-, Gα_i2-, and Gα_i3-subunits of G-proteins were obtained from BIOGNOSTIK (Göttingen, Germany). Dissociated single β-cells were incubated for 48 h with 5 μmol/l antisense or control oligonucleotide in tissue culture medium at 37°C. Cells were used for [Ca²⁺]_i and insulin release studies. Knockdown of each protein was determined by immunocytochemical analysis using monoclonal antibodies as follows. Cells were fixed in 4% paraformaldehyde and then incubated overnight with mouse monoclonal anti-insulin antibodies (Sigma-Aldrich) at dilutions of 1:1,000 at 4°C. Samples were then incubated in Alexa Fluor 488-labeled goat anti-mouse IgG (Molecular Probes, Eugene, OR). Preparations were then incubated overnight with monoclonal antibodies against the Gα_i1-, Gα_i2-, or Gα_i3-subunit of G-proteins (Calbiochem, San Diego, CA) at a dilution of 1:1,000 at 4°C. Antibody binding was detected with Alexa Fluor 568–labeled goat anti-rabbit IgG (Molecular Probes). Control experiments were carried out without the primary antibodies. Immunofluorescences for insulin and G-protein were observed with photomultipliers of a multiphoton laser-scanning microscope (FluoView FV300-TP; Olympus, Tokyo, Japan).

Statistical analysis.

Data represent the means ± SE. Statistical analyses were performed using the Student's t test or one-way ANOVA. P values <0.05 were considered statistically significant.

RESULTS

Ghrelin decreases plasma insulin concentrations in a PTX-sensitive manner.

Intraperitoneal administration of ghrelin (10 nmol/kg) decreased plasma insulin concentrations at 30 min in rats fasted overnight (Fig. 1A). Conversely, intraperitoneal administration of a specific GHS-R antagonist, [d-Lys³]-GHRP-6 (10 μmol/kg), significantly increased plasma insulin levels, reflecting counteraction of endogenous ghrelin activity to inhibit insulin release. On the other hand, ghrelin (10 nmol/kg i.p.) increased plasma GH concentrations at 15 min (Fig. 1B), and this effect was completely blocked by treatment with a PLC inhibitor U-73122 (2 μmol/kg i.p.), but not with its inactive analog U-73343 (2 μmol/kg i.p.). These effects are consistent with the current concept that the effects of ghrelin are mediated by G₁₁-PLC signaling (2). In contrast, the ghrelin action to inhibit insulin release was unaffected by the PLC inhibitor (Fig. 1A). The insulinostatic effects of both endogenous and exogenous ghrelin were blunted in rats treated with PTX (5 μg/kg), whereas ghrelin-induced GH release was unaltered (Fig. 1A and B). Thus, the G-protein coupling for the action of ghrelin on insulin release is distinct from that on GH release.

To further confirm the PTX sensitivity of the insulinostatic function of ghrelin, we examined GTTs in rats fasted overnight. In control experiments without PTX, ghrelin (10 nmol/kg i.p.) injected together with glucose (2 g/kg) markedly attenuated plasma insulin responses (Fig. 1C and D) and enhanced blood glucose responses (Fig. 1E and F) to GTT in comparison with control values without ghrelin. In rats treated with PTX (5 μg/kg), insulin responses during glucose challenge were higher and glucose responses were lower than those in control rats, and the actions of ghrelin to attenuate insulin responses and to enhance glucose responses were not observed (Fig. 1C–F).

Insulinostatic effects of endogenous ghrelin in islets.

In isolated rat islets, 8.3 mmol/l glucose-induced insulin release was augmented by GHS-R blockade, whereas it was inhibited by exogenous ghrelin (Fig. 2A), as previously reported. These effects of endogenous and exogenous ghrelin were blunted in islets isolated from PTX-treated rats, whereas addition of 25 mmol/l KCl enhanced insulin release from these islets (Fig. 2B).

Glucose (8.3 mmol/l)-induced insulin release from isolated islets of Ghr-KO mice was significantly greater than that of control mice (Fig. 2C), as already reported. The enhancement of glucose-induced insulin release from islets of Ghr-KO mice was undetectable after treatment with PTX (Fig. 2C), indicating that endogenous ghrelin attenuates insulin release via PTX-sensitive G-proteins.

Ghrelin decreases electrical activities in β-cells.

We next electrophysiologically examined the effects of ghrelin in β-cells. Under the condition of perforated whole-cell current-clamp mode, glucose (8.3 mmol/l) elicited firings of action potentials in β-cells. The firings were characterized by spike-like and repetitively occurring action potentials overlapping on slowly depolarized wave as depicted in the time-expanded scale (Fig. 3, control inset). These electrical firings were attenuated by ghrelin administration in a reversible manner (Fig. 3). Ghrelin decreased both the frequency and amplitude of the firings (Fig. 3, right inset). Mean membrane potentials measured at most repolarized levels between slow-wave potentials at 8.3 mmol/l glucose were not significantly changed by ghrelin (−38.1 ± 0.9 mV for control vs. −36.9 ± 4.5 mV for ghrelin treatment, n = 3). These findings suggest that the ghrelin administration does not hyperpolarize the membrane potential but decreases the activity of action potentials in β-cells.

Ghrelin PTX sensitively activates Kv channels but not K_ATP channels in β-cells.

In the presence of 8.3 mmol/l glucose and 100 μmol/l tolbutamide, outward K⁺ currents evoked by depolarizing pulses to +20 mV were increased by exposure to 10 nmol/l ghrelin (Fig. 4A), confirming a previous report (4). The current-voltage relationship demonstrated that ghrelin increased the amplitude of current densities through Kv channels (Fig. 4B). At holding potentials of −20 and −10 mV, which are in the physiological range of membrane potentials in β-cells stimulated with high glucose, ghrelin significantly enhanced Kv currents (Fig. 4C). We examined whether activation of Kv currents is involved in ghrelin-induced suppression of insulin release. The Kv-channel current was largely inhibited by Kv-channel blocker TEA, confirming a previous report (4). In the presence of TEA (10 mmol/l), glucose (8.3 mmol/l)-induced insulin release was potentiated, whereas basal insulin release at 2.8 mmol/l glucose was unchanged (Fig. 4D). Ghrelin failed to attenuate glucose (8.3 mmol/l)-induced insulin release in the presence of TEA (Fig. 4D), suggesting that the enhancement of Kv channel conductance is linked to insulinostatic ability of ghrelin. Furthermore, ScTx (0.1 μmol/l), an inhibitor of the Kv2.1 channel (27), potentiated glucose (8.3 mmol/l)-induced insulin release, and in the presence of ScTx ghrelin, it failed to attenuate glucose-induced insulin release (Fig. 4D). This result suggests that ghrelin inhibits insulin release by increasing the Kv2.1 channel conductance.

We next examined whether the ghrelin-induced Kv channel potentiation is sensitive to PTX. As shown in Fig. 4E, in β-cells treated with PTX (1 μg/ml), Kv currents evoked by depolarizing pulses to 20 mV were not increased by exposure to 10 nmol/l ghrelin. The ghrelin action to enhance Kv currents was blunted in the entire range of potentials in β-cells treated with PTX (Fig. 4F). These results indicate that PTX-sensitive G-protein is required for ghrelin activation of Kv currents. Moreover, in the presence of membrane-permeable cyclic AMP analog dibutylyl-cyclic AMP (db-cAMP) at 1 mmol/l, ghrelin did not potentiate Kv currents (Fig. 4G), suggesting that ghrelin activation of Kv channels is mediated by cAMP signaling pathway.

Effects of ghrelin on the K_ATP channel currents, which were evoked by ramp pulses, were determined in rat single β-cells. K_ATP channel currents were decreased by changing glucose concentrations from 2.8 to 8.3 mmol/l. The conductance density of the K_ATP channel current measured at a slope during ramp pulses was not altered by ghrelin (Fig. 4H and I), indicating that K_ATP channels may not contribute to the inhibition of β-cell membrane excitability by ghrelin.

Ghrelin inhibits [Ca²⁺]_i and insulin release via Gα_i2-dependent pathway.

In islet β-cells, [Ca²⁺]_i is considered the major regulator of insulin secretion. Repetitive glucose (8.3 mmol/l) stimulation induced repeated [Ca²⁺]_i increases in rat single β-cells. Ghrelin (10 nmol/l), added 5 min before the second glucose stimulation, suppressed [Ca²⁺]_i responses (Fig. 5A and C), and this effect was abolished by treatment with PTX (Fig. 5B and C). Thus, the effects of ghrelin on insulin release, Kv channels, and [Ca²⁺]_i were all inhibited by PTX, suggesting that certain subtype(s) of PTX-sensitive G-proteins are crucial in the signal transduction of ghrelin in β-cells. Treatment of primary cultured rat β-cells with antisense oligonucleotide specific for the Gα_i2-subunit of G-proteins markedly decreased Gα_i2-subunit expression (Fig. 5F). In the antisense-treated, but not control oligonucleotide-treated, β-cells, ghrelin failed to attenuate [Ca²⁺]_i responses to glucose (Fig. 5D and E). The inhibitory effect of ghrelin on [Ca²⁺]_i responses was unaltered in β-cells treated with antisense oligonucleotides specific for Gα_i1 and Gα_i3 (Fig. 5E). Furthermore, in primary cultured β-cells treated with antisense oligonucleotide for Gα_i2, ghrelin failed to suppress glucose (8.3 mmol/l)-induced insulin release, whereas significant suppression was observed in the control oligonucleotide-treated and untreated cells (Fig. 5G). These results suggest that Gα_i2-mediated signaling is crucial for the action of ghrelin to suppress glucose-induced [Ca²⁺]_i increase and insulin release and that the ghrelin-induced attenuation of [Ca²⁺]_i increase is linked to that of insulin release.

DISCUSSION

In this study, we present the novel ghrelin signaling in β-cells, which is distinct from that for GH release. The function of endogenous ghrelin was assessed by the effects of GHS-R antagonist in vivo and in islets. In rats, ghrelin of both endogenous and exogenous origins resulted in PTX-sensitive decrease in plasma insulin concentrations, contrasting with PTX-insensitive increase in GH levels by ghrelin. In rat isolated islets, endogenous and exogenous ghrelin suppressed glucose-induced insulin release. Moreover, isolated islets from Ghr-KO mice displayed PTX-sensitive enhancement of glucose-induced insulin release. Ghrelin also increased Kv currents and inhibited glucose-induced [Ca²⁺]_i increases in β-cells. Thus, all the effects of endogenous and exogenous ghrelin on Kv, [Ca²⁺]_i, and insulin release were blunted in the presence of PTX. Moreover, the effects of ghrelin on [Ca²⁺]_i and insulin release were abolished in the islet β-cells treated with the antisense oligonucleotide specific for Gα_i2-subunit of G-proteins. TEA, a Kv channel blocker, eliminated the ability of ghrelin to suppress insulin release. These findings demonstrate that ghrelin suppresses glucose-induced insulin release via Gα_i2- and Kv channel–mediated attenuation of Ca²⁺ signaling in β-cells.

PTX, an inhibitor of G_i/o subtypes of trimeric G-proteins, was formally recognized as the islet-activating protein (IAP) because this agent increased insulin release from isolated islets (28). The effect of IAP has suggested the presence of intra-islet substance that activates G_i/o-protein–coupled receptors to decrease insulin release. However, the endogenous islet substance whose insulinostatic action is antagonized by PTX/IAP is as yet unidentified. This study demonstrated that the insulinostatic effect of endogenous ghrelin in islets, as evidenced by enhancement of insulin release due to GHS-R antagonist and ghrelin gene knockout, was blunted by PTX treatment. The enhancement of insulin release by GHS-R antagonist and that by ghrelin gene knockout were as large as 70–80% of the enhancement by PTX/IAP treatment (Fig. 2A vs. B and C). These findings indicate that a large part, but not all, of the PTX/IAP effect is exerted by blocking the action of endogenous ghrelin in islets. The apparently ghrelin-independent portion of the PTX/IAP effect could be due to inhibition of the signaling of somatostatin, an islet hormone that is released from δ-cells and inhibits insulin release via PTX-sensitive G-proteins (29), although a paracrine role for somatostatin in islets has long been controversial (30).

This study revealed a novel ghrelin signaling coupled to inhibition of glucose-induced insulin release in β-cells. The effects of ghrelin on Kv, [Ca²⁺]_i, and insulin release were all completely blocked by PTX. Moreover, the effects of ghrelin on [Ca²⁺]_i and insulin release were blunted in the islet β-cells treated with the antisense oligonucleotide specific for Gα_i2-subunit of G-proteins. In contrast, neither the antisense oligonucleotide for Gα_i1 nor that for Gα_i3 affected the ghrelin effects on [Ca²⁺]_i. The data indicate that the insulinostatic ghrelin signaling is produced via PTX-sensitive Gα_i2 in β-cells. The mediation by the G-protein Gα_i2-subtype is surprising in light of the current concept that GHS-R signaling is mediated primarily by the G₁₁ subtype (2). However, our result is consistent with the reports that Gα_i2 is expressed in β-cells (29,31), whereas Gα₁₁ is expressed primarily in non–β-cells in islets (32–34).

Voltage-gated potassium channels are involved in repolarization of excitable cells. In pancreatic β-cells, activation of delayed rectifier K⁺ (Kv) channels repolarizes cells and attenuates glucose-stimulated action potentials, limiting Ca²⁺ entry through voltage-dependent Ca²⁺ channels to suppress insulin secretion (35). Blockade of Kv channels can promote glucose-dependent insulin secretion (36–38). In the present study, outward K⁺ currents were increased by exposure to 10 nmol/l ghrelin, and this effect was blunted in β-cells treated with PTX (Fig. 4). We previously reported that Kv current was largely inhibited by a Kv channel blocker TEA and that the presence of TEA diminished the ability of ghrelin to suppress glucose-induced [Ca²⁺]_i increases in β-cells (4). Moreover, the present study showed that attenuation of glucose-induced insulin release by ghrelin was also abolished in the presence of a Kv2.1 channel blocker ScTx (Fig. 4D). Kv2.1 is expressed at high levels in islets of various species (36,37), and immunohistochemical analysis indicates that expression of Kv2.1 is restricted to islet β-cells (39). Taken together, an increase in Kv2.1 channel conductance is linked to ghrelin-induced inhibition of insulin release and Ca²⁺ signaling. K_ATP channel is well known as a key molecule that determines resting membrane potentials and converts the glucose metabolism to the membrane excitation in β-cells (40,41). This channel, however, may not be the target for ghrelin signaling for the following reasons; membrane potentials, which are mainly controlled by K_ATP channels, were not significantly changed by ghrelin, and K_ATP channel currents at 8.3 mmol/l glucose conditions were not altered by ghrelin application. Therefore, the ghrelin-induced decrease in the action potential firing may be caused largely, if not completely, by activation of Kv channels. It should also be noted that ghrelin significantly activated Kv currents at the membrane potentials of −20 and −10 mV (Fig. 4C) that correspond to the peak values of bursting action potentials in glucose-stimulated β-cells (Fig. 3). Collectively, our results corroborate a signaling cascade that ghrelin activates Kv channels via PTX-sensitive mechanisms to cause a rapid repolarization and shortening of bursting action potentials, leading to attenuation of glucose-induced [Ca²⁺]_i increases and insulin secretion. Moreover, ghrelin did not potentiate the Kv currents in the presence of cyclic AMP analog db-cAMP (Fig. 4G), suggesting that ghrelin activation of Kv channels is mediated by cAMP signaling pathway. It is also known that members of the Gα_i-family are involved in transducing the information of inhibitory hormones and neurotransmitters to cyclic AMP productions (42–44), voltage-dependent Ca²⁺ channels (45,46), and insulin exocytosis (29). Whether the ghrelin action could also involve these processes is unknown and definitely requires further studies. Possible differences in the coupling of G-proteins to signaling pathways between ghrelin and other inhibitory hormones, norepinephrine and somatostatin, also remain to be studied.

Low plasma ghrelin levels are associated with elevated fasting insulin levels, insulin resistance, and type 2 diabetes (16). The inverse relationship between plasma levels of ghrelin and insulin may be explained in part by the inhibition of insulin release by ghrelin. Our present and previous (4,17) studies revealed that ghrelin is a physiological autocrine/paracrine hormone in islets and that its antagonism could promote insulin release and prevent high-fat diet–induced glucose intolerance. Moreover, the present study provided evidence to support both endogenous and exogenous ghrelin attenuating glucose-induced insulin release via Gα_i2-mediated activation of Kv channels and suppression of action potential firings and [Ca²⁺]_i increases in β-cells. These unique signaling mechanisms and molecules mediating the insulinostatic action of ghrelin in β-cells provide potential therapeutic targets for the prevention and treatment of type 2 diabetes and hyperinsulinemia.