G_o2 G protein mediates galanin inhibitory effects on insulin release from pancreatic β cells

Abstract

The neuropeptide galanin regulates numerous physiological activities in the body, including feeding and metabolism, learning and memory, nociception and spinal reflexes, and anxiety and related behaviors. Modulation of blood glucose levels by suppressing insulin release was the first reported activity for galanin. This inhibition was mediated by one or more pertussis toxin-sensitive G proteins of the G_i/o subfamily. However, the molecular identities of the specific G protein(s) and intracellular effectors have not been fully revealed. Recently, we demonstrated that mice lacking G_o2, but not other members of the G_i/o protein family, secrete more insulin than controls upon glucose challenge, indicating that G_o2 is a major transducer for the inhibitory regulation of insulin secretion. In this study, we investigated galanin signaling mechanisms in β cells using cell biological and electrophysiological approaches. We found that islets lacking G_o2, but not other G_i/o proteins, lose the inhibitory effect of galanin on insulin release. Potentiation of ATP-sensitive potassium (K_ATP) and inhibition of calcium currents by galanin were disrupted by anti-G_o2α antibodies. Galanin actions on K_ATP and calcium currents were completely lost in G_o2^−/− β cells. Furthermore, the hyperglycemic effect of galanin is also blunted in G_o2^−/− mice. Our results demonstrate that G_o2 mediates the inhibition of insulin release by galanin by regulating both K_ATP and Ca²⁺ channels in mice. Our findings provide insight into galanin's action in glucose homeostasis. The results may also be relevant to the understanding of galanin signaling in other biological systems, especially the central nervous system.

Galanin, a 29- to 30-residue neuropeptide hormone initially discovered in porcine intestine (1), is distributed throughout the central and peripheral nervous systems and the intestinal neuroendocrine system of many mammalian species (2, 3). The first 15 N-terminal amino acids, which retain the biological activity of the full-length peptide hormone, are highly conserved, underscoring a physiological importance across species. Galanin coexpresses and colocalizes with many neurotransmitters (4) and functions as an inhibitory modulator.

The biological effects of galanin are mediated by galanin receptors. Three types of galanin receptors (GalR1, GalR2, and GalR3) have been identified by molecular cloning and characterized pharmacologically in various species (5–8). All three subtypes of galanin receptors are members of the GTP-binding protein-coupled receptor (GPCR) superfamily. The three galanin receptors exhibit overlapping but distinctive patterns of expression in the central nervous system and periphery. The distinct distribution patterns of receptors support the notion that each receptor mediates some unique physiological function in the body.

Physiological Effects of Galanin.

The biological activity of galanin has been studied intensively in the central and peripheral nervous systems, as well as in the pituitary and the endocrine system. Galanin signaling is involved in feeding and metabolism, learning and memory, nociception and spinal reflexes, anxiety, neuron regeneration, and the pathogenesis of Alzheimer's disease (9–12). The precise role of galanin signaling in these states has not been elucidated. The first reported biological activity of galanin was its effect on plasma glucose levels in dogs and rats (1). The pancreas is highly innervated, and galanin localizes to autonomic nerve terminals in the endocrine pancreas (3) as well as to nerve cell bodies in the celiac ganglion (13). Infusion of galanin into animals results in a significant increase of blood glucose levels by inhibition of insulin secretion from pancreatic β cells (1, 14). Infusion of galanin directly into animals through the pancreatic artery at a concentration that is similar to that released from stimulated pancreatic nerve termini is sufficient to inhibit insulin secretion (15). Conversely, a galanin antagonist can block galanin-mediated inhibition of insulin release from islets (16). In galanin-deficient animals, the inhibition of insulin secretion induced by the chemical activation of sympathetic nerves was observed to be impaired (17). Interestingly, genetically obese ob/ob mice have severely decreased pancreatic content of galanin of less than 10% of levels found in control animals (18). In addition, the number of galanin immunoreactive cells is dramatically reduced in diabetic animals (19). Reduced islet innervation has been associated with impaired insulin secretion in type II diabetic hamsters (20). These results suggest that pancreatic galaninergic nerve dysfunction may contribute to the development of type II diabetes, which is an increasing worldwide public health problem.

Galanin-Receptor Signal Transduction.

The signaling mechanisms of GalR1 and GalR3, which are coupled to their effectors by pertussis toxin (PTX)-sensitive G_i/o G proteins, have been pharmacologically studied. Stimulation of receptors expressed in transfected cell lines or oocytes can inhibit forskolin-stimulated cAMP production, or activate G protein-regulated inwardly rectifying K⁺ channels in a PTX-sensitive manner (5, 7, 21, 22). Multiple subclasses of G proteins may be involved in GalR2 signaling (6, 22–24). GalR2 can activate phospholipase C and protein kinase C via G_q/11 and activate MAPK and/or inhibit forskolin-stimulated cAMP production through PTX-sensitive G_i/o G proteins. Like the galanin receptors, all five nonsensory PTX-sensitive G_i/o members are expressed in neurons and endocrine tissues. The signal transduction mechanisms mediating galanin's physiological effects have not been studied in native tissue or primary cells. Our previous studies demonstrated that all nonsensory members of the PTX-sensitive G_i/G_o protein family, namely G_i1, G_i2, G_i3, G_o1, and G_o2, are expressed in pancreatic islets, and that only G_o2, and not G_i1–3 or G_o1, G protein-deficient mice show an “improved” glucose tolerance test and enhanced insulin release from pancreatic β cells (25). This suggests that G_o2 is the major transducer mediating inhibitory effects on insulin release that prevent oversecretion. In this study, using G_i/G_o α-subunit gene knockout animals established in our laboratory (26, 27), we demonstrate that the G_o2 G protein mediates galanin's inhibitory effect on insulin release. We also identify potentiation of ATP-sensitive potassium (K_ATP) currents and inhibition of Ca²⁺ channels as possible molecular mechanisms mediating galanin-GalR-G_o2 signaling.

Results and Discussion

Inhibition of Insulin Release by Galanin Is Lost in G_o2α^−/− Islets.

In the present study, we addressed the possible molecular mechanisms by which galanin inhibits insulin release. Galanin has been shown to elevate blood glucose levels by inhibiting insulin release. PTX blocks the inhibitory effects of galanin, suggesting the involvement of G_i/G_o proteins in the process. PTX ADP ribosylates a cysteine at the carboxyl termini (−4 position) of the α subunits of G_i/o G proteins and collectively blocks all G_i/G_o signaling. Therefore, PTX assays cannot delineate through which specific G protein(s) galanin signals. To investigate the mechanism and to screen the responsible G_i/G_o protein(s) for mediating the inhibition of insulin release by galanin, we established a perifusion assay of isolated islets from G_i/G_o-deficient animals aimed at identifying which G_i/G_o protein(s) mediates galanin's inhibitory effects. A pool of 30–50 islets isolated shortly after a recovery period were placed onto a nylon membrane in a perifusion chamber. After 60 min of perifusion with 1.8 mM glucose in Krebs Ringer bicarbonate (KRB) solution to establish a stable rate of insulin secretion, the glucose concentration in the buffer was elevated to 16.2 mM and islets were continually perifused for 120 min at this high-glucose concentration. The perifusion solution was then switched back to 1.8 mM (low) glucose for 30 min, followed by perifusion with KCl (30 mM) for an additional 30 min. During the high-glucose (16.2 mM) perifusion period, galanin (100 nM) was included transiently (from minute 60 to 90) in the buffer. Fig. 1 shows insulin secretion profiles of islets isolated from WT and knockout mice lacking G_i2α, G_i1/G_i3α, G_o1α, and G_o2α. High glucose (16.2 mM) induced insulin release from pancreatic islets effectively from all of the genotypes. When the lower-glucose (1.8 mM) buffer was applied, insulin release from islets declined to basal levels. KCl (30 mM), which directly depolarizes the β-cell membrane, triggered insulin release from islets. Galanin (perifused together with high glucose) suppressed insulin secretion in islets from WT, G_i1α^−/− G_i3α^−/− double KO, G_i2α^−/−, and G_o1α^−/− mice. Upon removal of galanin from the buffer, insulin secretion resumed at the elevated rates. In contrast, the inhibitory effect of galanin on insulin release was absent in islets from G_o2α^−/− mice. The results indicate that G_o2 mediates the inhibitory effect of galanin on insulin release in pancreatic β cells. The levels of insulin released under basal, glucose-stimulated, and K⁺-induced conditions were comparable for islets from WT, G_i2α^−/−, G_i1α^−/− G_i3α^−/− double knockout, and G_o1α^−/− mice, but were significantly higher for islets from G_o2α^−/− (P < 0.01). This suggests that, in addition to mediating galanin inhibition of insulin release, G_o2 modulates the sensitivity of islets to glucose and maybe directly or indirectly regulates secretory machinery.

Fig. 1.

Lack of G_o2α in the islets impairs the inhibitory effect of galanin on insulin secretion. Insulin release profiles of perifusion assays on islets from G_o2α^−/−, WT, and other G_i/G_oα knockout mice are shown. Hand-picked islets (30–50) were packed in a perifusion chamber and perifused with low (1.8 mM) and high (16.2 mM) concentrations of glucose solutions and KCl as indicated. Galanin was applied during a 60–90 min period as indicated in the first shaded area in the charts. Insulin contents were measured by ELISA with a mouse insulin standard. Perifusion rate was 0.25 mL/min; fractions were collected every 2 min. Shown are traces of mean values ± SEM of insulin secretion from WT, G_o2^−/−, G_o1^−/−, G_i1/3 double knockout, and G_i2^−/− islets’ response to glucose, galanin, and KCl.

K_ATP Channels in G_o2α^−/− Islets Show Normal Glucose Sensitivity.

It is well-established that glucose and other nutrients promote the depolarization of the β-cell membrane and Ca²⁺ influx through voltage-dependent channels. This constitutes the principal stimulus for insulin exocytosis (28). A glucose transporter (Glut-2) in the β cell facilitates entry of glucose into the cell (29). The enzyme glucokinase phosphorylates glucose to glucose-6-phosphate. Glucokinase appears to function as the fundamental glucose sensor that controls the subsequent β-cell response (30). Glucose is oxidized, leading to a rapid increase of the concentration of intracellular ATP and a decrease of the concentration of ADP. This in turn changes the ATP:ADP ratio in the β cell. ATP is a potent inhibitor of the ATP-sensitive K⁺ channel, whereas ADP is a stimulator of the channel (31, 32). Inhibition of K⁺ channels by ATP decreases K⁺ efflux from the β cell, thus leading to membrane depolarization (33). Depolarization opens voltage-dependent Ca²⁺ channels, causing a rapid increase in the concentration of intracellular Ca²⁺. This triggers exocytosis of pancreatic β-cell insulin granules (34). As observed, galanin can inhibit insulin secretion from pancreatic β cells in a PTX-sensitive manner.

To investigate whether the G_o2-mediated signaling pathway regulates K_ATP and Ca²⁺ channels, we studied channel properties via patch-clamp recording. First, the integrity of K_ATP channels was investigated in β cells from G_o2α knockout mice. Specifically, K_ATP channels were studied in response to glucose (Fig. S1). Freshly isolated islets were digested briefly with trypsin to generate single β cells. Short-term cultured β cells were used in the studies. Glucose inhibited K_ATP currents in β cells from both wild-type and G_o2α^−/− (Fig. S1 A–C). High glucose (16.2 mM) inhibited K_ATP currents in isolated β cells from −58.0 ± 15.6 to −22 ± 6.1 pA/pF for WT (n = 10, P < 0.01) and −49.8 ± 13.3 to −27.1 ± 5.1 pA/pF for G_o2α^−/− (n = 12, P < 0.05) mice (Fig. S1D). Furthermore, glucose deprivation stimulated K_ATP currents in both WT (−26.5 ± 5.0 to −62.6 ± 13.5 pA/pF; n = 9, P < 0.01) and G_o2α^−/− mice (−22.26 ± 4.0 to −65.34 ± 14 pA/pF; n = 6, P < 0.01). This result demonstrates that K_ATP channels in β cells isolated from G_o2α^−/− mice respond to glucose similarly to WT controls and suggests that lack of G_o2α does not alter the responsiveness of K_ATP channels to glucose in β cells.

Galanin Fails to Potentiate K_ATP Channels in G_o2α^−/− Mice.

We next investigated the effects of galanin on the modulation of K_ATP channel activity by glucose. K_ATP channel currents in the β cells of WT and G_o2α^−/− mice were recorded in the absence of galanin, presence of 100 nM galanin, or 50 μM diazoxide. As shown in Fig. 2, the averages of glucose-sensitive K_ATP currents in β cells from WT mice were −35.4 ± 6.3 and −58.0 ± 10.7 pA/pF (n = 12) before and after application of 100 nM galanin to the bath solution, respectively. The averages of K_ATP currents in G_o2α^−/− β cells were −46.5 ± 6.7 and −45.3 ± 6.9 pA/pF (n = 11) before and after application of 100 nM galanin to the bath solution, respectively (Fig. 2). Diazoxide, a selective K_ATP channel opener, was able to further potentiate the channels in both WT and G_o2α^−/− β cells in the presence of galanin. Galanin significantly enhanced (P < 0.01) K_ATP currents in β cells from WT control mice; however, this current augmentation was completely absent in the β cells from G_o2α^−/− mice. The results suggest that G_o2 protein mediates the inhibitory effect of galanin on K_ATP channels in mouse β cells, and that the potentiation of K_ATP by diazoxide is independent of the G_o2-mediated pathway. These data are consistent with the existence of a specific signaling mechanism whereby the inhibitory neuropeptide galanin binds to its receptors and activates K_ATP channels through G_o2 to inhibit insulin secretion from β cells.

Fig. 2.

Galanin-stimulated K_ATP channel activity is missing in G_o2^−/− β cells. (A) Representative K_ATP current traces recorded from β cells potentiated by 100 nM galanin. Testing potential (TP) = −100 to +50 mV, holding potential (HP) = −20 mV. (B) I–V (current-voltage) relationship curves show the effect of galanin and diazoxide (50 μM) on K_ATP current from WT and G_o2^−/− β cells. (C) Summary of K_ATP currents in β cells stimulated by galanin in WT and the diminished galanin stimulation in G_o2^−/− in the presence of 10 mM glucose (in normal culture medium-glucose condition) (n = 11–12, **P < 0.01, TP = −100 mV to +50 mV, HP = −20 mV).

Anti-G_o2α Antibodies Suppress Basal K_ATP Currents and Abolish the Potentiating Effects of Galanin on K_ATP in WT β Cells.

To further validate that G_o2 G protein is a critical component in the inhibitory pathway mediating galanin signaling, we used affinity-purified anti-G_o2α antibodies to titrate the endogenous G_o2α protein and tested K_ATP current activity in islet β cells. WT β cells dialyzed with anti-G_o2α antibodies showed a reduced basal K_ATP channel activity. After rupture of the patch, antibodies diffused into the cells from the holding pipettes, and the K_ATP current decreased from −43.7 ± 0.7 to −15.7 ± 0.1 pA/pF (n = 5, P < 0.01) within 10 min and reached a new steady state. K_ATP currents were stable when the holding pipettes contained nonspecific IgG or no antibodies. Galanin (100 nM) was perifused into the cell chamber during the steady phase. As expected, K_ATP currents were not enhanced by galanin in β cells dialyzed with anti-G_o2α antibodies. Notably, the antibodies had no effects of the K_ATP channel opener diazoxide on β cells (Fig. 3). These findings are in agreement with studies on isolated β cells from G_o2α^−/− mice (Fig. 2), and confirm that G_o2 is a critical downstream signal transducer of galanin receptors in β cells. Through modulation of K_ATP channel activity, G_o2 regulates insulin secretion. The stimulation of K_ATP channels by galanin was abolished by dialyzing anti-G_o2α antibodies into β cells (Fig. 3 A and B). Thus, antibodies dialyzed into the cytoplasm appear to titrate G_o2 away from their effectors and change the dynamic regulation of effectors. The reduction of the basal K_ATP channel activity by anti-G_o2α antibodies suggests that under basal conditions, G_o2 potentiates K_ATP channel activity in β cells. The K_ATP channels in β cells are involved in maintaining membrane potential, and activation of K_ATP channels is required for retaining β cells in the hyperpolarized state to suppress insulin secretion. In addition, the observation of dynamic changes in K_ATP activity by dialyzed anti-G_o2α antibodies suggests that the α subunit may be involved directly. Our results from native channel studies are also in agreement with previous findings that demonstrated that the activity of cloned SUR1-Kir6.2 channels (pancreatic form) can be enhanced by a G_i/oα protein during inside-out patch recordings in a reconstitution assay (35). In addition, recombinant G_oα protein can potentiate neuronal (brain) K_ATP channel activity in inside-out patches (36). Taken together, these results support that the α subunit of G_o also participates in the regulation of K_ATP channel activity.

Fig. 3.

Anti-G_o2 antibodies reduce basal K_ATP currents and abolish galanin-potentiated K_ATP currents in WT β cells. (A) Representative K_ATP current traces in WT β cells dialyzed with anti-G_o2α antibody and perifused with 100 nM galanin or 50 μM diazoxide. TP = −100 to +50 mV, HP = −20 mV. Basal K_ATP currents were blocked by the dialyzed anti-G_o2 antibodies, and the anti-G_o2 antibodies abolished galanin-potentiated K_ATP currents but not diazoxide-stimulated K_ATP currents. (B) Corresponding I–V (current-voltage) relationship curves. (C) Summary of time course of basal K_ATP currents blocked by anti-G_o2 antibodies (n = 5, TP = −100 mV, HP = −20 mV). (D) Summary of abolishment of galanin-potentiated K_ATP currents by anti-G_o2 antibody dialysis. Compare K_ATP currents before and 10 min after dialysis of anti-G_o2 antibodies (n = 5, **P < 0.01, TP = −100 mV, HP = −20 mV).

Inhibitory Effect of Galanin on Ca²⁺ Currents Is Missing in Pancreatic β Cells from G_o2α^−/− Mice.

The L-type Ca²⁺ channel is a major route for elevating intracellular Ca²⁺ concentration, which in turn triggers insulin exocytosis. Modulation of Ca²⁺ channel activity would effectively regulate intracellular Ca²⁺ concentrations and insulin release in β cells. It has been shown that mouse β cells express L-type Ca²⁺ channels (both Ca_v1.2 and Ca_v1.3) (37, 38). To address whether the L-type Ca²⁺ channel is another intracellular effector in the galanin G_o2-mediated pathway, we investigated galanin's effect on the regulation of Ca²⁺ currents. First, we verified the nature of the recorded Ca²⁺ currents in β cells using a Ca²⁺ channel potentiator and blocker. As shown in Fig. 4A, Bay K-8644 (0.2 μM) potentiated current, whereas nifedipine (1 μM) inhibited current, confirming the existence of classical L-type Ca²⁺ channels in β cells. The low-voltage T-type Ca²⁺ channels were not detectable in our mouse pancreatic β cells. We next tested the effect of galanin on regulating Ca²⁺ currents in β cells. Galanin caused a reduction in Ca²⁺ flow into β cells by inhibiting Ca²⁺ currents. Galanin (100 nM) significantly reduced Ca²⁺ current in β cells isolated from WT mice, with a decrease of 55% of peak currents from −108.93 ± 25.68 to −60.26 ± 17.96 pA/pF (n = 8, P < 0.01) (Fig. 4 B and C). When the cells were dialyzed with anti-G_o2α antibodies, the average peak currents were −85.13 ± 21.58 and −72.07 ± 26.41 pA/pF (n = 8) before and after galanin (100 nM) application, respectively. No significant difference was observed before and after application of galanin in the presence of anti-G_o2α antibodies. This demonstrates that anti-G_o2α antibodies can block the inhibitory effect of galanin on Ca²⁺ currents. These observations suggest that in addition to modulating K_ATP channel activities, G_o2 protein could also mediate effects of galanin by regulating the activity of the pancreatic L-type Ca²⁺ channel. To confirm this hypothesis, we performed patch-clamp recordings of Ca²⁺ channel activity of β cells isolated from G_o2α^−/− animals. In G_o2α^−/− β cells, average peak currents before and after application of galanin were −51.87 ± 11.9 and −38.54 ± 15.6 pA/pF (n = 6), respectively. There was also no statistically significant difference before and after application of galanin. These results show that the lack of G_o2α in β cells by gene knockout or titration of G_o2α by antibodies leads to the loss of galanin's inhibitory effect on Ca²⁺ channels. It is noteworthy that there is a statistically significant difference (P < 0.05) in the Ca²⁺ currents in WT and G_o2α^−/− β cells. The cause for this reduction of Ca²⁺ channel activity in G_o2α knockout is currently unknown. Taken together, our results suggest that G_o2 mediates the potentiation of K_ATP channels and inhibition of Ca²⁺ channels induced by galanin. Hence, G_o2 is a major transducer for galanin receptor signaling in the pancreatic β cell.

Fig. 4.

The inhibition of Ca²⁺ current by galanin is missing in G_o2α^−/− β cells. (A) Current traces of whole-cell Ca²⁺ currents recorded from WT β cells under a control condition or the presence of 1 μM nifedipine or 0.2 μM Bay K-8644. TP = 0 mV, HP = −70 mV. (B) Representative Ca²⁺ current traces before and after 100 nM galanin application to WT β cells without or with dialyzed anti-G_o2 antibodies and to G_o2α^−/− β cells (TP = 0 mV, HP = −70 mV). (C) Summary of galanin effects on Ca²⁺ currents in β cells from WT without or with dialyzed anti-G_o2α antibodies and from G_o2α^−/−. The results suggest that G_o2 mediates the inhibition of Ca²⁺ currents by galanin (TP = 0 mV, HP = −70 mV, *P < 0.05, **P < 0.01, n = 5–8). Values are means ± SEM.

Loss of G_o2α Blunts the Hyperglycemic Effect of Galanin in Vivo.

To confirm our findings that G_o2 mediates galanin receptor signaling in pancreatic β cells and galanin's inhibitory effect on insulin secretion, we tested the effect of galanin administration to G_o2α^−/− and WT mice on their blood glucose levels. Galanin can effectively suppress insulin release from β cells; therefore, administration of galanin into the bloodstream will suppress insulin levels, and in turn increase blood glucose levels. Galanin (2 μg/kg) was delivered via the retroorbital vein in overnight-fasted mice. Blood glucose levels were monitored before and after administration of galanin. As shown in Fig. 5, galanin elevates blood glucose levels rapidly, reaching a maximum after 10 min in WT mice and declining slowly thereafter. In WT mice, blood glucose levels were raised from a basal level of 74.5 ± 3.4 mg/dL to 130.1 ± 7.4 mg/dL 10 min postinjection and 121.8 ± 5.7 mg/dL 20 min postinjection. These correspond to a 77.8 and 64.5% increase after administration of galanin, respectively. In contrast, galanin only elevated glucose levels slightly over basal levels in G_o2α-null mice. The basal blood glucose level of G_o2-null mice was 60.4 ± 1.7 mg/dL. The glucose levels were raised moderately after galanin injection to 85.5 ± 3.1 mg/dL 10 min postinjection and 81.8 ± 3.7 mg/dL 20 min postinjection, corresponding to a 45.7 and 36.9% increase over basal, respectively. The results demonstrate that galanin can effectively elevate blood glucose levels in WT mice, and that this effect is significantly reduced (P < 0.01) in mice lacking G_o2α. This suggests that G_o2 mediates a critical signaling pathway for galanin in the modulation of glucose homeostasis in the body. We did not observe a complete loss of galanin's hyperglycemic effect in G_o2α-null mice, suggesting that galanin may also use other, G_o2-independent mechanisms to augment glucose levels in β cells. In fact, galanin has been shown to enhance plasma glucagon levels secreted from pancreatic α cells. During the systemic infusion of galanin, pancreatic glucagon output rapidly doubled in dogs and mice (3, 39). Therefore, galanin-induced hyperglycemia is the additive action of lowering basal plasma insulin levels, increasing basal plasma glucagon levels, and possibly other mechanisms. Increasing plasma glucagon by galanin is believed to be mediated by G_11/q proteins, but not by G_i/o proteins. Therefore, elevation of blood glucose levels by the action of glucagon should not be affected in G_o2α^−/− mice. Physiologically, the body responds to hypoglycemia by increasing glucose production and decreasing glucose sequestration (by reducing insulin levels) simultaneously.

Fig. 5.

Effects of galanin on blood glucose levels in G_o2 knockout mice. Galanin at 2 μg/kg bodyweight was given at time 0. Blood glucose levels were monitored. (Left) Actual glucose levels. (Right) Percentage changes over the basal. WT: n = 9, 74.5 ± 3.4 (basal at 0 min), 130.1 ± 7.4 (10 min, increased 77.8% ± 14.3), 121.8 ± 5.7 (20 min, increased 64.5% ± 8.1); G_o2α^−/−: n = 18, 60.4 ± 1.7 (basal at 0 min), 85.5 ± 3.1 (10 min, increased 45.7% ± 5.7), 81.8 ± 3.7 (20 min, increased 36.9% ± 6.7). For the control no-galanin injection experiment, n = 6 and 8 for WT and G_o2α^−/− mice, respectively. *P < 0.01; statistical significance was determined using the Student's t test from unpaired data or analysis of variance.

The distribution of galanin suggests that it is a neuropeptide. Its activity is exerted locally, rather than systemically as a humoral agent (40). Pancreatic islets are highly innervated, and activation of pancreatic nerves is sufficient to influence islet function (15). G_o was originally identified as the “other” PTX-sensitive G protein in the brain and is highly expressed in the CNS and endocrine cells (41). Logically, antagonizing the G_s pathway by suppressing adenylyl cyclase (AC) activity might be a mechanism for galanin receptor-G_o protein-mediated signaling in cells. Galanin can inhibit AC activity through G_i3 (42) or G_i2/G_i3 proteins in the RINm5F cell line (43). However, galanin can still suppress insulin release from insulin-secreting cells in the presence of cAMP in the media (44). Using static islet incubation assays, we also confirmed that the suppression of insulin release by galanin is not dependent on the AC-cAMP pathway. Islets were incubated with or without galanin in the presence of 0.1 mM dibutyryl-cAMP and 16.2 mM glucose, and released insulin was measured. Galanin can inhibit insulin release from wild-type islets even in the presence of cAMP, as observed previously (45). This observation suggests that inhibiting AC and reducing cAMP levels are unlikely to be a mode for suppressing insulin release by galanin. In fact, G_o is at best a poor AC inhibitor in vitro compared with G_i proteins (46). In addition, cellular cAMP levels are not elevated in G_o^−/− islets (47). Taken together, reducing cAMP levels by inhibiting AC activity in β cells does not appear to constitute a major route for the inhibitory effect of galanin on insulin release.

Our findings can be summarized as follows: (i) The inhibitory effects of galanin on insulin release are lost in β cells from G_o2α^−/− mice, but are intact in the other G_i/o knockouts; (ii) the potentiation by galanin of K_ATP currents is missing in G_o2α^−/− β cells; (iii) the inhibition of β-cell L-type Ca²⁺ currents by galanin is lost in G_o2^−/− β cells; (iv) antibodies specific against G_o2α abolish galanin-mediated potentiation of K_ATP and inhibition of Ca²⁺ currents in WT β cells; and (v) in vivo, the galanin-induced hyperglycemic effect is blunted in mice lacking G_o2α.

Galanin inhibits Ca²⁺-induced insulin release from permeabilized RINm5F cells in a PTX-sensitive manner (48), suggesting that direct inhibition of exocytosis is one mechanism of action for galanin. Constitutively active α subunits of G_i/o protein can inhibit the exocytosis process, suggesting α subunits can directly influence the process (49). In addition, enhanced potassium-induced insulin release in G_o2^−/− islets shown in Fig. 1 and our previous study (25) supports the notion that G_o2 may act on the insulin exocytosis process directly. Interestingly, Gβγ can also regulate the exocytosis process. Gβγ can bind SNAP-25 directly, as demonstrated in in vitro assays (50). The C terminus of SNAP-25 may be a target of Gβγ for presynaptic inhibition (51). Recently, it has been reported that norepinephrine can block the exocytosis process via the Gβγ subunit, and that this process is mediated specifically by Gα_i1/2 (52). Through study of G-protein knockout mice, we and others have observed that the expression levels of Gβγ correlate well with α-subunit levels in cells; Gβγ are down-regulated in α-subunit knockout and heterozygous mice (53). This suggests that signaling through Gα and Gβγ is well-coordinated in the cells. The mechanism for G-protein inhibition of the exocytosis process remains to be further elucidated.

Galanin has been shown to be an important neuropeptide that regulates glucose homeostasis in the body (17). Our study demonstrates that galanin inhibition of insulin release is mediated mainly by G_o2, not other G_i/G_o proteins in pancreatic β cells. Other G_i or G_o proteins cannot compensate for the loss of G_o2 in mediating galanin's modulation of insulin secretion. This indicates that galanin receptors preferentially couple to effectors by G_o2 protein in β cells in vivo. At the molecular level, loss of G_o2 protein in β cells results in impairment of galanin's modulation of K_ATP and Ca²⁺ channels. This suggests in the insulin secretion pathway that galanin achieves its inhibitory effects on insulin release by at least at three different action sites: by potentiating K_ATP channels, inhibiting L-type Ca²⁺ currents, and regulating the exocytosis process. Inhibitory peptide hormones play a critical role in maintaining euglycemia by preventing oversecretion of insulin into the bloodstream. Oversecretion of insulin could result in hypoglycemia that may impair the normal biological function of the endocrine and nervous systems. Furthermore, prolonged oversecretion of insulin may desensitize peripheral tissues and lead to the development of insulin resistance characteristic of type II diabetes.

Materials and Methods

All procedures were designed and performed in accordance with the generally accepted ethical standards for animal experimentation and approved by the Chancellor's Animal Research Committee of the University of California, Los Angeles. Detailed methods for islets and β-cell isolations, islet perifusion, electrophysiological recordings, and in vivo tests for galanin effects on glucose levels are described in SI Materials and Methods. Statistical analyses were done using paired and unpaired Student's t tests and analyses of variance in conjunction with Newman–Keuls tests where appropriate. Group differences at the level of P < 0.05 were considered statistically significant.