Metabotropic Type 4 Is Involved in Autoinhibitory Cascade for Glucagon Secretion by ␣-Cells of Islet of Langerhans Shunsuke Uehara, Akiko Muroyama, Noriko Echigo, Riyo Morimoto, Masato Otsuka, Shouki Yatsushiro, and Yoshinori Moriyama

mGluR4, and its stimulation inhibits secretion of gluca- In islets of Langerhans, L-glutamate is stored in gluca- gon-containing secretory granules of ␣-cells and cose- gon through an inhibitory cAMP cascade. Thus, L-gluta- ␣ creted with glucagon under low-glucose conditions. The mate may directly interact with -cells and inhibit L-glutamate triggers secretion of ␥-aminobutyric acid glucagon secretion. Diabetes 53:998–1006, 2004 (GABA) from ␤-cells, which in turn inhibits glucagon secretion from ␣-cells through the GABAA receptor. In the present study, we tested the working hypothesis that L-glutamate functions as an autocrine/paracrine L-Glutamate, an excitatory in the cen- modulator and inhibits glucagon secretion through a tral nervous systems, plays important roles in fast ␣ glutamate receptor(s) on -cells. The addition of L- synaptic transmission and neuronal plasticity (1,2). For glutamate at 1 mmol/l; (R,S)-phosphonophenylglycine glutamate-evoked neurotransmission, neuronal cells de- (PPG) and (S)-3,4-dicarboxyphenylglycine (DCPG), specific for class III metabotropic glutamate velop systems: exocytosis of L-glutamate, receptor (mGluR), at 100 ␮mol/l; and (1S,3R,4S)-1- signal reception through various glutamate receptors, and aminocyclopentane-1,3,4-tricarboxylic acid (ACPT-I) at signal termination through plasma membrane glutamate 50 ␮mol/l inhibited the low-glucose–evoked glucagon transporters. The islet of Langerhans, a pancreatic minia- secretion by 87, 81, 73, and 87%, respectively. This ture endocrine organ for homeostasis of blood glucose, inhibition was dose dependent and was blocked by possesses glutamatergic systems similar to those in the (R,S)-cyclopropyl-4-phosphonophenylglycine (CPPG), a central (3): L-glutamate is stored in gluca- specific antagonist of class III mGluR. Agonists of other gon-containing secretory granules through vesicular gluta- glutamate receptors, including kainate and quisqualate, mate transporters, and it is cosecreted with glucagon from had little effectiveness. RT-PCR and immunological ␣-cells under low-glucose conditions (4,5). The released analyses indicated that mGluR4, a class III mGluR, was expressed and localized with ␣- and F cells, whereas no L-glutamate may bind to the glutamate receptor expressed evidence for expression of other mGluRs, including on islet cells, resulting in the occurrence of the glutamatergic mGluR8, was obtained. L-Glutamate, PPG, and ACPT-I response of islet cells, although the glutamatergic response is decreased the cAMP content in isolated islets, which not fully understood (3,6–12). Then, the glutamate signals was blocked by CPPG. Dibutylyl-cAMP, a nonhydrolyz- may be terminated by sequestration of L-glutamate through able cAMP analog, caused the recovery of secretion of Naϩ-dependent glutamate transporter (13). glucagon. Pertussis toxin, which uncouples adenylate In pioneering works, (RS)-␣-amino-3-hydroxy-5-methyl- cyclase and inhibitory G-protein, caused the recovery of 4-isoxazolepropionic acid (AMPA)-type ionotropic glu- both the cAMP content and secretion of glucagon. These ␣ tamate receptors, mainly GluR2/3, were found to be results indicate that - and F cells express functional expressed in ␣- and ␤-cells (6–8), and their stimulation was reported to enhance the secretion of insulin under high-glucose conditions (6,7,14). AMPA and kainate, but From the Department of Biochemistry, Faculty of Pharmaceutical Sciences, Okayama University, Okayama, Japan. not N-methyl-D-aspartate, were also shown to stimulate Address correspondence and reprint requests to Dr. Y. Moriyama, Depart- glucagon secretion from perfused pancreas (15). However, ment of Biochemistry, Faculty of Pharmaceutical Sciences, Okayama Univer- L sity, Okayama 700-8530, Japan. E-mail: [email protected] under low-glucose conditions, -glutamate secreted from u.ac.jp. ␣-cells stimulates an AMPA-type receptor and selectively Received for publication 22 September 2003 and accepted in revised form triggers ␥-aminobutyric acid (GABA) secretion, with a 5 January 2004. S.U. and A.M. contributed equally to the present study. lower effect on insulin secretion in isolated islets and ACPT-I, (1S,3R,4S)-1-aminocyclopentane-1,3,4-tricarboxylic acid; AMPA, clonal ␤-cells (4). Because GABA is stored in synaptic-like (RS)-␣-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; CPPG, (R,S)-cy- microvesicles, distinct secretory vesicles from insulin clopropyl-4-phosphonophenylglycine; DBcAMP, dibutyryl cAMP; DCPG, (S)-3,4-dicarboxyphenylglycine; DMEM, Dulbecco’s modified Eagle’s me- granules, in ␤-cells (16), the glutamatergic response was dium; dNTP, deoxynucleotide triphosphate; GABA, ␥-aminobutyric acid; explained as being caused by enhanced exocytosis of Gi, inhibitory G-protein; GST, S-transferase; L-CCG-I, (2S,3S,4S)- 2(carboxycyclopropyl); LY341495, (2S,1ЈS,2ЈS)-2-(2-carboxycyclopro- GABA-containing synaptic-like microvesicles (4). Because pyl)-2-(9H-xanthen-9-yl)glycine; mGluR, metabotropic glutamate receptor; ␣-cells express the GABAA receptor, a chloride channel, PPG, (R,S)-phosphonophenylglycine; PTX, pertussis toxin; (S)-3,5-DHPG, (S)- and its stimulation modulates the membrane potential, 3,5-; trans-ACPD, (Ϯ)-1-aminocyclopentane-trans-1,3- dicarboxylic acid. causing inhibition of glucagon secretion (17–19), glutama- © 2004 by the American Diabetes Association. tergic signaling from ␣-cells and subsequent GABAergic

998 DIABETES, VOL. 53, APRIL 2004 S. UEHARA AND ASSOCIATES signaling from ␤-cells seem to be involved in the inhibition of 35 temperature cycles were conducted, as follows: denaturation at 94°C for of glucagon secretion (3). 30 s, annealing at a specific temperature shown in Fig. 1A for 30 s, and ␣ extension at 72°C for 1 min. The amplification products were analyzed by Recently, Tong et al. (11) reported that islet -cells polyacrylamide gel electrophoresis. The sequences of the oligonucleotides express metabotropic glutamate receptor type 8 (mGluR8), used as primers were based on the published sequences of various mGluRs. and that its stimulation by agonists (R,S)-phosphono- DNA sequencing was performed by the chain-termination method (23). phenylglycine (PPG) and (S)-3,4-dicarboxyphenylglycine Expression of mGluR4. Rat mGluR4 cDNA, kindly provided by Dr. S. Nakanishi (24), was subcloned into the EcoRI-XhoI site of expression vector (DCPG; at around the nmol/l level) strongly inhibits glu- pcDNA3.1 (Invitrogen, San Diego, CA). The resultant construct was used to cagon secretion under low-glucose conditions. This report transfect COS7 cells by lipofection using Trans It reagent (Mirus, Madison, is of special interest because if mGluR8 actually functions WI). COS7 cells were cultured in Dulbecco’s modified Eagle’s medium in islet ␣-cells, the mGluR-mediated signaling cascade should (DMEM) containing 10% FCS as described previously (25). Membrane fraction be another signaling pathway for the inhibition of glucagon was prepared as described above. Antibodies. Site-specific rabbit polyclonal antibodies against rat mGluR4 secretion, which may support the involvement of L-glu- were produced as follows: DNA fragments encoding MSGKGGWAWWWARL tamate signaling in the inhibition of glucagon secretion. PLCLLLSLYAPWVPSSLGKPKGHPHMNS (24) were amplified by PCR and mGluRs belong to a big family of glutamate receptors cloned into the EcoR1 site of expression vector pGEX3X (Amersham Phar- and are classified into three distinct groups (classes I, II, macia Biotech) to form glutathione S-transferase (GST) fusion plasmids. DNA fragments encoding the NH2-terminal region of mGluR8, MVCEGKRLASCPCF and III) according to their sequence homology, second FLLTAKFYWILTMMQRTHSQEYAH (GenBank accession no. U63288) (26), messenger coupling, and selectivity (20). Class I were also amplified and cloned into pGEX3X as described above. After receptors consist of mGluR1 and -5, which are linked to transformation, the GST fusion proteins were purified on a glutathione- phosphatidylinositol hydrolysis/Ca2ϩ signal transmission. Sepharose 4B column (Amersham Pharmacia Biotech) and then injected into Class II receptors (mGluR2 and -3) and class III receptors a rabbit with complete adjuvant two times with a 2-week interval. The DNA fragments encoding the above-mentioned polypeptides of the NH2-terminal (mGluR4, -6, -7, and -8) are coupled with inhibitory G- region of mGluR4 and -8 were also cloned into the XhoI and EcoRI site of protein (Gi) protein, and thus their stimulation results in expression vector pBAD/myc-His B (Invitrogen). After transformation of the the inhibition of adenylate cyclase and a decreased cellular resultant plasmids, the six His-tagged polypeptides were purified through a cAMP level. A pioneering group has reported negative results Ni-NTA column (Qiagen, Tokyo) according to the manufacturer’s manuals. ␤ The mouse monoclonal antibodies against glucagon were from Progen. The as to the expression of mGluRs in clonal -cells (21). rat monoclonal antibodies against somatostatin, mGluR2/3, mGluR5, and However, Brice et al. (12) recently detected expression of the mGluR8 were from Chemicon. The guinea pig polyclonal antibodies against mGluR3, -4, -5, and -8 gene (using RT-PCR) and expression insulin and rat pancreatic polypeptide were from Biogenesis and Linco of the mGluR3 and 5 proteins (using immunoblotting) in Research, respectively. Alexa Fluor 568–labeled anti-mouse IgG and Alexa isolated islets and clonal ␣- and ␤-cells. Taking the obser- Fluor 488–labeled goat anti-rabbit IgG were from Molecular Probes. Western blotting. Membrane fraction was suspended in SDS sample buffer vation of Tong et al. (11) into consideration, it is clear that and then subjected to SDS gel electrophoresis. Immunoblotting was per- there is no consensus as to the expression of mGluRs in formed according to the published protocol, with visualization by enhanced islets. To further investigate the glutamatergic signaling chemiluminescence amplification (Amersham Pharmacia Biotech) (4). cascade through mGluR, it is necessary to establish which Immunohistochemistry. Indirect immunofluorescence microscopy was per- ␣ formed as described previously, using an Olympus FV-300 confocal laser type of mGluR is actually expressed in islet -cells. microscope (25). In the present study, we examined the expression of Assays. Isolated islets (20 pieces per assay) or cultured cells (4.0 ϫ 106 mGluRs by means of a combination of procedures, includ- cells/dish) were washed three times with DMEM and then incubated in ing pharmacological techniques, RT-PCR, immunoblotting, Ringer’s solution containing 10 mmol/l HEPES (pH 7.4), 0.2% BSA, and glucose and immunohistochemistry with pancreas and purified at the specified concentration for1hat37°C. Then, the islets or cultured cells were transferred to 2 ml of the above Ringer’s solution containing glucose at islets of rat. Here we report the evidence for functional the specified concentration. At the times indicated, samples (10 ␮l) were occurrence of mGluR4, but not other mGluRs (including taken, and the amount of glucagon and cAMP was quantified with enzyme mGluR8), in islets. mGluR4 is present in ␣-cells and F cells immunosorbent assay kits obtained from Amersham Pharmacia Biotech and that secrete pancreatic polypeptide, and its stimulation Yanaihara (Shizuoka, Japan) according to the manufacturers’ manuals. For measurement of cAMP, islets were extensively homogenized with 20 mmol/l inhibits the secretion of glucagon through an inhibitory MOPS-Tris (pH 7.0) containing 0.3 mol/l sucrose, 5 mmol/l EDTA, 5 ␮g/ml of cAMP cascade. The present results provide the molecular leupeptin, 5 ␮g/ml of chymostatin, 5 ␮g/ml of antipain, and 5 ␮g/ml of basis for the occurrence of an L-glutamate–mediated auto- pepstatin A and then centrifuged at 140.000g for 1 h. The resultant supernatant inhibitory mechanism for glucagon secretion by islet ␣-cells. was used for assay. Statistics. Statistical significance between sets of data were assessed using Student’s t test. RESEARCH DESIGN AND METHODS Materials. Agonists and antagonists for GluRs listed in Table 1 were Preparation. Islets of Langerhans were isolated from male Wistar rats at 7–8 purchased from Tocris Cookson (Avonmouth, U.K.). Other chemicals were of postnatal weeks by the collagenase digestion method combined with discon- the highest grade commercially available. tinuous Ficoll gradient centrifugation (22). Islets were then handpicked and suspended in a bicarbonate-buffered Hank’s solution supplemented with 0.2% BSA. In some experiments, islets (500 pieces) were washed with 20 mmol/l RESULTS MOPS-Tris (pH. 7.0) containing 0.3 mol/l sucrose, 5 mmol/l EDTA, 5 ␮g/ml leupeptin, 5 ␮g/ml chymostatin, 5 ␮g/ml antipain, and 5 ␮g/ml pepstatin A and Pharmacological evidence for involvement of class III then extensively homogenized. The homogenate was centrifuged at 800g for mGluR(s) in the inhibition of glucagon secretion. 10 min, and the resultant supernatant was centrifuged at 140.000g for1h. As the first step of this study, we investigated whether RT-PCR. Total RNA extracted from isolated islets (1 ␮g) was transcribed into cDNA in a final volume of 20 ␮l of a reaction buffer containing 0.2 mmol/l each mGluRs are involved in the regulation of glucagon se- deoxynucleotide triphosphate (dNTP), 10 mmol/l dithiothreitol, 25 pmol of cretion, using a pharmacological approach (Table 1). random octamers, and 200 units of Moloney murine leukemia virus reverse L-Glutamate at 1 mmol/l inhibited low-glucose–dependent transcriptase (Amersham). After 1 h incubation at 42°C, the reaction was glucagon secretion, as glucose secreted at 16.7 mmol/l terminated by heating at 90°C for 5 min. For PCR amplification, the 10-fold– diluted synthesized cDNA solution was added to the reaction buffer containing subtracted from that secreted at 3.3 mmol/l glucose, by 0.6 mmol/l total dNTP (150 ␮mol/l each dNTP), 25 pmol of primers (see Fig. 87%. L-Glutamate at 100 ␮mol/l was inhibitory, with 55% 1A), and 1.5 units of Ampli Taq-Gold DNA polymerase (Perkin Elmer). A total less effectiveness than that at 1 mmol/l. L-Glutamate is also

DIABETES, VOL. 53, APRIL 2004 999 mGluR4-MEDIATED INHIBITION OF GLUCAGON

FIG. 1. Expression of mGluRs genes in islets and exocrine pancreas. A: The probes used for RT-PCR analysis and the amplification conditions. B: Expression of various mGluR genes, as indicated, in brain (B), isolated islets (L), and exocrine pancreas (E) was examined by RT-PCR. No amplified products were obtained without the RT reaction (-RT). Tm, temperature. inhibitory to the arginine-evoked secretion of glucagon: (LY341495) is an antagonist with a Ki value of 57 nmol/l for arginine at 15 mmol/l enhanced secretion of glucagon mGluR8, which is a ϳ1,000-fold higher affinity than that for 2.3-fold at 3.3 mmol/l glucose, and arginine-enhanced mGluR4 (30). The compound moderately blocked the secretion of glucagon was inhibited by 79% upon treatment L-glutamate–evoked inhibition of glucagon secretion at 30 of L-glutamate at 1 mmol/l. PPG and DCPG (agonists for ␮mol/l, but it did not block inhibition at all at 300 nmol/l. class III mGluRs) (27) at 100 ␮mol/l also inhibited gluca- (S)-3,5-dihydroxyphenylglycine [(S)-3,5-DHPG] and quis- gon secretion by 81 and 73%, respectively. The inhibition qualate (specific agonists for class I mGluR) (31,32) at 1 was dose dependent: at 40 nmol/l both agonists were not mmol/l or (2S,3S,4S)-2(carboxycyclopropyl)glycine (L- effective, the effectiveness being 50% at 40 ␮mol/l. At Ͼ100 CCG-I) at 100 ␮mol/l and (Ϯ)-1-aminocyclopentane-trans- ␮mol/l, no further inhibition was observed. Moreover, it 1,3-dicarboxylic acid (trans-ACPD) at 1 mmol/l (agonists was found that (1S,3R,4S)-1-aminocyclopentane-1,3,4- for class II mGluRs) (20,33,34) were only slightly effective tricarboxylic acid (ACPT-I; another agonist for class III (Table 1). AMPA and kainate (agonists for ionotropic mGluR) (28) at 50 ␮mol/l inhibited glucagon secretion by glutamate receptors) at 0.5 mmol/l also slightly affected 87%. The L-glutamate–, PPG-, DCPG-, and ACPT-I–evoked glucagon secretion (Table 1). These results suggest that inhibition of glucagon secretion was blocked by (R,S)- class III mGluRs, most probably mGluR4, are involved in cyclopropyl-4-phosphonophenylglycine (CPPG), an antag- the inhibition of glucagon secretion. onist for class III mGluRs (29), at 100 ␮mol/l. Essentially Expression and localization of mGluR4 in islets. Sub- the same glucagon secretion was observed when CPPG sequently, we examined which types of mGluRs are ex- itself was included in the assay mixture. (2S,1ЈS,2ЈS)-2-(2- pressed in islets. Among the known mGluRs examined, carboxycyclopropyl)-2-(9H-xanthen-9-yl)glycine RT-PCR analysis indicated the amplification of RNA frag-

1000 DIABETES, VOL. 53, APRIL 2004 S. UEHARA AND ASSOCIATES

FIG. 1. Continued ments of mGluR4 (Fig. 1). The amplified products covered right panel). These results gave credence to the immuno- all predicted sequences. The sequence of the amplified logical specificity of the mGluR4 antibodies used in the RNA fragment was identical to the corresponding se- study. The mGluR4 antibodies recognized a single 100-kDa quence of mGluR4. On the other hand, no amplification of polypeptide in islet membranes as well as brain membrane RNA fragments for other mGluRs was observed, although (Fig. 2B). The immunoreactivity disappeared when anti- the same probes could amplify the predicted gene frag- genic polypeptides (1 mg) were included during the immu- ments from brain sources (Fig. 1). In particular, three noreaction. These results demonstrated the presence of the different sets of primers for mGluR8 did not give any mGluR4 protein in islets. Double-labeling indirect immuno- positive products for islet RNA, although these primers fluorescence microscopy revealed that mGluR4 immunore- gave amplified products with the expected sizes and activity is confined to islets but not to exocrine parts of sequences for brain RNA (Fig. 1). Essentially the same pancreas (Fig. 2C). mGluR4 immunoreactivity is colocal- results were obtained with the various amplification pro- ized with glucagon and with pancreatic polypeptide but not tocols used with an annealing temperature of 60 or 65°C with insulin or somatostatin (Fig. 2C). Treatment with (data not shown). These results led us to conclude that the mGluR4 gene is actually expressed in the islets, but antigenic peptides caused the mGluR4 immunoreactivity expression of the genes of other mGluRs is negative or to disappear (data not shown). These results indicate that ␣ ␤ ␦ under the detection limit of our assay. mGluR4 is present in - and F cells but not in - and -cells. Similar RT-PCR analyses were performed using RNA Consistent with the absence of mRNA of other mGluRs prepared from exocrine pancreas. Fragments of mGluR3 in islets, neither antibodies against mGluR2/3 nor those and -5 with the expected sizes and sequences were amplified, against mGluR5 recognized any polypeptides of islets on whereas no amplification of the mGluR4 gene was observed immunoblotting, whereas these antibodies recognized (Fig. 1). Thus, in contrast with isolated islets, the exocrine counterparts of brain membranes (Fig. 3A). Consistently, part of pancreas shows different expression of mGluRs. anti-mGluR2/3 and -mGluR5 antibodies did not immuno- To detect the expression of mGluR4 at the protein level, stain any islet cells, but they immunostained synaptic we prepared site-specific antibodies against mGluR4. The terminals and blood vessels in pancreas as well as brain mGluR4 antibodies recognized a 100-kDa polypeptide of sections (Fig. 3B). As for mGluR8, mGluR8 antibodies authentic mGluR4 expressed in COS7 cells (Fig. 2A, left did not recognize any polypeptides of islets on immuno- panel). The immunoreactivity disappeared when antigenic blotting (Fig. 3A). After immunohistochemistry, mGluR8 polypeptides (1 mg) were included during the immunore- antibodies slightly immunostained islet cells (Fig. 3B), action. Dot blot analysis indicated that the mGluR4 anti- although the immunoreactivity did not disappear even in bodies recognized the NH2-terminal region of mGluR4, but the presence of antigenic peptides (data not shown). they did not recognize the counterpart of mGluR8 (Fig. 2A, These results indicated that the level of mGluR2/3,

DIABETES, VOL. 53, APRIL 2004 1001 mGluR4-MEDIATED INHIBITION OF GLUCAGON

TABLE 1 The effects of L-glutamate and agonists of GluRs on low-glucose- dependent glucagon secretion Glucagon release Concentration (ng ⅐ 20 isletsϪ1 ⅐ Additions (␮mol/l) 30 minϪ1) 16.7316.7 (basal release) 1,000 0.51 Ϯ 0.23* 16.733.3 (100% control) 1,000 1.67 Ϯ 0.17 ϩ L-glutamate 1,000 0.66 Ϯ 0.19* ϩ L-glutamate ϩ 100 ␮mol/l CPPG 1,000 2.43 Ϯ 0.16* ϩ L-glutamate ϩ 300 nmol/l LY341495 1,000 0.69 Ϯ 0.12* ϩ L-glutamate ϩ 30 ␮mol/l LY341495 1,000 1.62 Ϯ 0.15 NS ϩ PPG 0.04 1.65 Ϯ 0.13 NS ϩ PPG 0.1 1.49 Ϯ 0.18 NS ϩ PPG 1 1.34 Ϯ 0.16† ϩ PPG 10 1.10 Ϯ 0.20‡ ϩ PPG 100 0.73 Ϯ 0.10* ϩ PPG ϩ 100 ␮mol/l CPPG 100 2.38 Ϯ 0.20‡ ϩ DCPG 0.02 1.72 Ϯ 0.13 NS ϩ DCPG 0.1 1.60 Ϯ 0.18 NS ϩ DCPG 1 1.48 Ϯ 0.15 NS ϩ DCPG 10 1.32 Ϯ 0.18† ϩ DCPG 100 0.82 Ϯ 0.12* ϩ DCPG ϩ 100 ␮mol/l CPPG 100 2.25 Ϯ 0.21‡ ϩ ACPT-I 1 1.61 Ϯ 0.13 NS ϩ ACPT-I 10 1.09 Ϯ 0.23‡ ϩ ACPT-I 20 0.92 Ϯ 0.17* ϩ ACPT-I 50 0.66 Ϯ 0.11* ϩ ACPT-I 100 0.68 Ϯ 0.20* ϩ ACPT-I ϩ 100 ␮mol/l FIG. 2. Expression and localization of mGluR4 protein in islets. A: Left CPPG 50 2.37 Ϯ 0.14* panel: The immunological specificity of mGluR4 antibodies was proven ϩ CPPG 100 2.28 Ϯ 0.19‡ with mGluR4-expressing COS7 cells. Membrane fractions of control ␮ ϩ (S)3,5DHPG 1,000 1.73 Ϯ 0.18 NS COS cells (lane 1) and mGluR4-expressing COS7 cells (100 g each) (lanes 2 and 3) were solubilized and subjected to SDS-PAGE, followed ϩ Quisqualate 1,000 1.36 Ϯ 0.15† by immunoblotting with anti-mGluR4 antibodies (lanes 1 and 2)orthe ϩ L-CCG-I 100 1.57 Ϯ 0.19 NS same antibodies plus 1 mg antigenic peptide (lane 3). Right panel: Dot ϩ trans-ACPD 1,000 1.54 Ϯ 0.13 NS blot analysis. Polypeptides of His-tagged NH2-terminal region of ␮ ϩ AMPA 500 1.37 Ϯ 0.21§ mGluR4 (lane 1) and mGluR8 (lane 2)at2 g each were spotted onto nitrocellulose sheets. One portion was stained with Amidoschwartz ϩ Kainate 500 1.33 Ϯ 0.15† 10B (Merck). Other portions were decorated with antibodies against mGluR4 or mGluR8, as indicated, and the immunological reaction was Data are means Ϯ SE, n ϭ 4. Isolated islets (20 pieces per assay) visualized with enhanced chemiluminescence. B: Western blotting of were first incubated with a solution containing 16.7 mmol/l glucose, brain membrane fraction (50 ␮g) (lanes 1 and 3) and islets (100 ␮g) and then transferred to a solution containing 3.3 mmol/l glucose to (lanes 2 and 4) was performed with mGluR4 antibodies (lanes 1 and 2) induce glucagon secretion in the presence or absence of the listed or the same antibodies plus 1 mg antigenic peptide (lanes 3 and 4). C: compounds at the indicated concentrations. After incubation for 30 Immunohistochemical localization of mGluR4 in islets. Sections of min, the medium was carefully sampled and the amount of glucagon pancreas were doubly immunostained with antibodies against mGluR4 was determined. Glucagon secretion with 16.7 mmol/l glucose and glucagon, mGluR4 and insulin, mGluR4 and somatostatin, or Ͻ Ͻ mGluR4 and pancreatic polypeptide, and then they were observed .␮m 10 ؍ throughout is also shown as basal release. *P 0.001; †P 0.05; under a confocal laser microscope. Bar ‡P Ͻ 0.01; §P Ͻ 0.1. NS, not significant vs. 100% control. PPG- or ACPT-I–evoked decreases in glucagon secretion mGluR5, or mGluR8 in islets is, if anything, very low or (Fig. 4). PTX also blocked the glutamate- and PPG- or under the detection limit of our assay. ACPT-I–dependent decreases in glucagon secretion and Intracellular signaling pathway for the inhibition of cAMP content (Figs. 4 and 5). Because it is known that ␣ glucagon secretion. Because mGluR4 is a class III islet -cells contain PTX-sensitive Gi protein (35,36), these mGluR, its stimulation may inhibit adenylate cyclase results strongly suggest that mGluR4 and adenylate cy- clase are functionally coupled through G in ␣-cells. through activation of Gi protein, followed by a decrease in i a cAMP level, resulting in the inhibition of glucagon secretion. To obtain convincing evidence for functional DISCUSSION coupling between mGluR4 and adenylate cyclase, the In this study, we showed that L-glutamate inhibits gluca- effects of dibutyryl cAMP (DBcAMP; a nonhydrolyzable gon secretion from isolated islets under low-glucose con- cAMP analog) and pertussis toxin (PTX; a specific ditions. AMPA and kainate did not stimulate glucagon uncoupler of adenylate cyclase and Gi) were investi- secretion under low-glucose conditions. These results gated. DBcAMP treatment abolished the glutamate- and support our proposal that glutamatergic signaling is

1002 DIABETES, VOL. 53, APRIL 2004 S. UEHARA AND ASSOCIATES

FIG. 2. Continued involved in the inhibition of glucagon secretion (3,4). mGluR4 after RT-PCR, although they did not perform Furthermore, L-glutamate–evoked inhibition of glucagon immunological identification. They also reported amplifi- secretion was blocked on treatment with CPPG, indicating cation of mRNA fragments of other mGluRs (mGluR3, -5, possible involvement of class III mGluR in this inhibitory and -8), although no expression of these mGluRs was process. We then investigated which mGluR is responsible observed in our case. At present, we do not know the for the inhibition. We found that mGluR4 is the major reason for the apparent discrepancy. It should be stressed, mGluR in the islet of Langerhans, and it is involved in however, that the mGluR3 and -5 genes are expressed in L-glutamate–evoked inhibition of glucagon secretion. Four the exocrine parts of pancreas (Figs. 1 and 3). Thus, the lines of evidence support this conclusion. At first, pharma- presence of acinar cells in an islet preparation may be cological analysis suggested the functional involvement of expected to lead to amplification of the gene fragments of class III mGluRs in the inhibition of low-glucose–stimu- these mGluRs. lated glucagon secretion. Second, RT-PCR analysis dem- More importantly, we could not confirm the expression onstrated the presence of mRNA of mGluR4. Third, we and localization of mGluR8 in islet ␣-cells. Even with the prepared antibodies specific to mGluR4 and demonstrated same DNA probes for RT-PCR analysis (Fig. 1A) (11), we the presence of the mGluR4 polypeptide in islets. Finally, could not detect any expression of the mGluR8 gene. immunohistochemical analysis indicated that mGluR4 is Although the mGluR8 antibodies slightly immunostained present in ␣- and F cells but not ␦- and ␤-cells, supporting islets, as judged by fluorescent microscopy, the immuno- the results of our pharmacological analysis. reactivity was observed throughout the islets, especially in Our results are partly consistent with previous observa- ␤-cells (Fig. 3), and did not disappear when excess tions reported by two groups (6,21): no expression of amounts of antigenic peptides were included during im- mGluRs was detected in ␤-cells and clonal ␤-cells. Brice et munoreaction. Consistent with these observations, West- al. (12) also detected amplification of a gene fragment of ern analysis with mGluR8 antibodies did not reveal any

DIABETES, VOL. 53, APRIL 2004 1003 mGluR4-MEDIATED INHIBITION OF GLUCAGON

FIG. 3. mGluR2/3, mGluR5, and mGluR8 were not detected in islets. A: Western blotting of brain membrane fraction (50 ␮g) (lane 1) and islets (100 ␮g) (lane 2) was performed with the listed antibodies. B: Sections of cerebrum (left panels) and pancreas (right panels) were immunostained with antibodies against mGluR2/3, mGluR5, and mGluR8 and then observed under a fluorescence microscope. Right panels:An islet is located at the center position. Immunoreactive blood vessels (middle right panel) and parts of nerve terminal of innervated .␮m 10 ؍ bottom right panel) were marked. Bar) polypeptide from isolated islets (Fig. 3). These results (30). These results further support the involvement of strongly suggest that immunoreaction by mGluR8 antibod- mGluR4 but not mGluR8 in the inhibition of glucagon se- ies is artificial in nature, and the concentration of mGluR8 cretion. Thus, the observation of Tong et al. (11) is incorrect. protein in ␣-cells is under the detection limit or very low as Stimulation of mGluR4 causes inhibition of glucagon compared with that of mGluR4. secretion. mGluR4 is a class III mGluR and is thus believed In addition, Tong et al. (11) reported that DCPG and to be negatively coupled with adenylate cyclase. As ex- PPG at 20 and 40 nmol/l had the maximum inhibitory effect pected, the addition of L-glutamate, PPG, or ACPT-I de- on glucagon secretion. The values are actually too low creased the cellular cAMP level, which was inhibited by compared with the above-mentioned abilities of authentic CPPG and PTX (Fig. 5). The involvement of DBcAMP also mGluR8 or even authentic mGluR4 of rat (27,28), and they led to the recovery of the PPG-evoked inhibition of secre- are far from the concentrations required for the inhibition tion of glucagon (Fig. 4). As for the cellular cAMP level, of glucagon secretion (Table 1). We showed that the essentially the same responses to L-glutamate and CPPG concentrations giving 50% effectiveness of DCPG and PPG were observed in ␣TC6 cells, which are clonal ␣-cells were ϳ28 and ϳ19 ␮mol/l, respectively, which are around (S.U., Y.M., unpublished observation). CPPG resulted in the Kd value of CPPG with respect to recombinant enhanced glucagon secretion and cAMP content (Table 1, mGluR4 expressed in cultured cells or rat tissues (37). Fig. 5), suggesting that endogenous L-glutamate released ACPT-I most potently inhibited glucagon secretion (Table from isolated islets tonically acts through an inhibitory 1). Moreover, LY341495 did not affect the L-glutamate– cascade. Although PTX is expected to show the same evoked inhibition of glucagon secretion at 300 nmol/l, a effect as CPPG, the compound did not enhance either several-fold higher concentration of Ki value for mGluR8 glucagon secretion or cAMP content (Figs. 4 and 5). The

1004 DIABETES, VOL. 53, APRIL 2004 S. UEHARA AND ASSOCIATES

FIG. 4. The effects of PTX and DB-cAMP on the L-glutamate–, PPG-, or ACPT-I–evoked inhibition of glucagon secretion. Isolated islets (20 pieces per assay) were incubated, and glucagon secretion was assayed as described in the legend to Table 1 in the presence of the listed compounds. The additions were made as follows: 100 ng/ml PTX, 1 mmol/l L-glutamate, 100 ␮mol/l PPG, 50 ␮mol/l ACTP-I, and 1 mmol/l DBcAMP. In the case of PTX treatment, islets were incubated with PTX for 21 h as described in (38), washed, and then treated with the listed compounds. After P < 0.001 when compared with the value# .(4 ؍ incubation for 30 min, the glucagon content in the medium was measured. Data are means ؎ SE (n of control (no addition); *P < 0.01; **P < 0.001. reason for the apparent ineffectiveness of PTX is unknown cyclase operate in ␣-cells. In vivo, the following signaling at present but seems to be caused by decreased viable cells cascade can be expected to operate: under low-glucose during incubation over a long period. Overall, we concluded conditions, L-glutamate is secreted by ␣- or F cells. Then, that signaling pathways from mGluR4 to Gi and adenylate the released L-glutamate may bind to the mGluR4 on

FIG. 5. The effects of PTX and CPPG on the L-glutamate–, PPG-, or ACPT-I–evoked change of the cAMP content. Isolated islets (20 pieces per assay) were incubated as described in the legend for Fig. 4 in the presence or absence of the listed compounds, and the cAMP content in islets was measured. The additions were made as follows: 100 ng/ml PTX, 100 ␮mol/l CPPG, 1 mmol/l L-glutamate, 100 ␮mol/l PPG, and 50 ␮mol/l ACTP-I. .P < 0.001 when compared with the value of control (no addition); *P < 0.01; **P < 0.001# .(4 ؍ Data are means ؎ SE (n

DIABETES, VOL. 53, APRIL 2004 1005 mGluR4-MEDIATED INHIBITION OF GLUCAGON

␣-cells so as to inhibit glucagon secretion through an Glutamate stimulates glucagon secretion via an excitatory re- inhibitory cAMP cascade. The mGluR4-mediated signaling ceptor of the AMPA subtype in rat pancreas. Eur J Pharmacol 237:45–50, 1993 16. Reetz A, Solimena M, Matteoli M, Folli F, Takei K, De Camilli P: GABA and pathway will provide the molecular basis for chemother- pancreatic ␤ cells: co-localization of decarboxylase (GAD) apeutics for hyperglycemia, one of the symptoms for type and GABA with synaptic-like microvesicles suggests their role in GABA 2 diabetes. storage and secretion. EMBO J 10:1275–1284, 1991 In conclusion, islet ␣-cells express functional mGluR4 17. Rorsman P, Berggren PO, Bokvist K, Ericson H, Mohler H, Otenson VG, as a major mGluR. L-Glutamate itself may function as an Smith PA: Glucose-inhibition of glucagon secretion involves activation of autocrine transmitter and inhibit glucagon secretion by GABAA- receptor chloride channels. Nature 341:233–236, 1989 18. Robbins MS, Grouse LH, Sorenson RL, Elde RP: Effect of on way of the mGluR4-mediated inhibitory cascade. The glucose-stimulated somatostatin and insulin release from the isolated, glutamatergic signaling also triggers GABA secretion from perfused rat pancreas. Diabetes 30:168–171, 1981 ␤-cells, and the released GABA in turn inhibits glucagon 19. Gilon P, Bertrand G, Loubatieres-Mariani MM, Remacle C, Henquin JC: The secretion by way of the GABAA receptor. Thus, the influence of ␥-aminobutyric acid on hormone release by the mouse and rat endocrine pancreas. Endocrinology 129:2521–2529, 1991 L-glutamatergic system may inhibit glucagon secretion 20. Nakanishi S: Molecular diversity of glutamate receptors and implications through two distinct signaling pathways, and it may con- for brain function. Science 258:597–603, 1992 stitute a novel regulatory mechanism of glucagon secre- 21. Wang Y, Small DL, Stanimirovic DB, Morley P, Durkin JP: AMPA receptor- tion in the islet of Langerhans. mediated regulation of a Gi-protein in cortical neurons. Nature 389:502– 504, 1997 ACKNOWLEDGMENTS 22. Gotoh M, Maki T, Satomi S, Porter J, Bonner-Weir S, O’Hara CJ, Monaco AP: Reproducible high yield preparation of rat islets by stationary in vitro This study was supported by a Grant-in-Aid for Scientific digestion following pancreatic ductal or portal venous collagenase injec- Research on Priority Areas–Advanced Brain Science tion. Transplantation 43:725–730, 1987 Project from the Ministry of Education, Science, Sports 23. Sambrook J, Fritsch EF, Maniatis T: A Laboratory Manual. 2nd ed. Cold and Culture of Japan, and the Umami Research Founda- Spring Harbor, NY, Cold Spring Harbor Laboratory, 1989 24. Tanabe Y, Masu M, Ishii T, Shigemoto R, Nakanishi S: A family of tions. S.Y. was supported by a Research Fellowship from metabotropic glutamate receptors. 8:169–179, 1992 the Japan Society for the Promotion of Science for Young 25. Hayashi M, Otsuka M, Morimoto R, Hirota S, Yatsushiro S, Takeda J, Scientists. Yamamoto A, Moriyama Y: Differentiation-associated Naϩ-dependent inor- We thank Prof. S. Nakanishi (Kyoto University School of ganic phosphate cotransporter (DNPI) is a vesicular glutamate transporter Medicine) for the kind supply of mGluR4 cDNA. in endocrine glutamatergic systems. J Biol Chem 276:43400–43406, 2001 26. Saugstad JA, Kinzie JM, Shinohara MM, Segerson TP, Westbrook GL: Cloning and expression of rat metabotropic glutamate receptor 8 reveals a REFERENCES distinct pharmacological profile. Mol Pharmacol 51:119–125, 1997 1. Fagg GE, Foster AC: Amino acid and their pathways in 27. Thomas NK, Wright RA, Howson PA, Kingston AE, Schoepp DD, Jane DE: the mammalian central nervous system. Neuroscience 9:701–719, 1983 (S)-3,4-DCPG, a potent and selective mGluR8a receptor agonist, activates 2. Mayer ML, Westbrook GL: The physiology of excitatory amino acids in the metabotropic glutamate receptors on primary afferent terminals in the vertebrate central nervous system. Prog Neurobiol 28:197–276, 1987 neonatal rat spinal cord. Neuropharmacol 40:311–318, 2001 3. Moriyama Y, Hayashi M: Glutamate-mediated signaling in the islets of 28. Acher FC, Tellier FJ, Azerad R, Brabet IN, Fagni L, Pin JP: Synthesis and Langerhans: a thread entangled. Trends Pharmacol Sci 24:511–517, 2003 pharmacological characterization of aminocyclopentanetricarboxylic ac- 4. Hayashi M, Yamada H, Uehara S, Morimoto R, Muroyama A, Takeda J, ids: new tools to discriminate between metabotropic glutamate receptor Yamamoto A, Moriyama Y: Secretory-granule-mediated co-secretion of subtypes. J Med Chem 40:3119–3129, 1997 L-glutamate and glucagon triggers the glutamatergic chemical transmis- 29. Toms NJ, Jane DE, Kemp MC, Bedingfield JS, Roberts PJ: The effect of sion in the islet of Langerhans. J Biol Chem 278:1966–1974, 2003 (RS)-a-cyclopropyl-4-phosphonophenylglycine ((RS)-CPPG), a potent and 5. Yamada H, Otsuka M, Hayashi M, Nakatsuka S, Hamaguchi K, Yamamoto selective metabotropic glutamate receptor antagonist. Br J Pharmacol ϩ A, Moriyama Y: Ca2 -dependent exocytosis of L-glutamate by ␣TC6 cells, 119:851–854, 1996 clonal mouse pancreatic ␣-cells. Diabetes 50:1012–1020, 2001 30. Howson PA, Jane DE: Actions of LY341495 on metabotropic glutamate 6. Gonoi T, Mizuno N, Inagaki N, Kuromi H, Seino Y, Miyazaki J, Seino S: receptor-mediated responses in the neonatal rat spinal cord. Br J Phar- Functional neuronal ionotropic glutamate receptors are expressed in the macol 139:147–155, 2003 non-neuronal cell line MIN6. J Biol Chem 269:16989–16992, 1994 31. Ito I, Kohda A, Tanabe S, Hirose E, Hayashi M, Mitsunaga S, Sugiyama H: 7. Inagaki N, Kuromi H, Gonoi T, Okamoto Y, Ishida H, Seino Y, Kaneko T, 3,5-Dihydroxyphenylglycine: a potent agonist of metabotropic glutamate Iwanaga T, Seino S: Expression and role of ionotropic glutamate receptors receptors. Neuroreport 3:1013–1016, 1992 in pancreatic islet cells. FASEB J 9:686–691, 1995 32. Porter RH, Roberts PJ, Jane DE, Watkins JC: (S)-homoquisqualate: a 8. Weaver CD, Yao TL, Powers AC, Verdoorn TA: Differential expression of potent agonist at the glutamate metabotropic receptor. Br J Pharmacol glutamate receptor subtypes in rat pancreatic islets. J Biol Chem 271: 106:509–510, 1992 12977–12984, 1996 33. Hayashi Y, Tanabe Y, Aramori I, Masu M, Shimamoto K, Ohfune Y, 9. Molnar E, Varadi A, McIIhinney RAJ, Ashcroft SJH: Identification of Nakanishi S: Agonist analysis of 2-(carboxycyclopropyl)glycine isomers functional ionotropic glutamate receptor proteins in pancreatic ␤-cells and for cloned metabotropic glutamate receptor subtypes expressed in Chi- in islets of Langerhans. FEBS Lett 371:253–257, 1995 nese hamster ovary cells. Br J Pharmacol 107:539–543, 1992 10. Morley P, MacLean S, Gendron TF, Small DL, Tremblay R, Durkin JP, 34. Monaghan DT, Bridges RJ, Cotman CW: The excitatory amino acid Meaing G: Pharmacological and molecular characterization of glutamate receptors: their classes, pharmacology, and distinct properties in the receptors in the MIN6 pancreatic beta-cell line. Neurol Res 22:379–385, 2000 function of the central nervous system. Annu Rev Pharmacol Toxicol 11. Tong Q, Ouedraogo R, Kirchgessner AL: Localization and function of group 29:365–402, 1989 III metabotropic glutamate receptors in rat pancreatic islets. Am J Physiol 35. Law SF, Manning D, Reisine T: Identification of the subunits of GTP- Endocrinol Metab 282:E1324–E1333, 2002 binding proteins coupled to somatostatin receptors. J Biol Chem 266: 12. Brice NJ, Varadi A, Ashcroft SJH, Molnar E: Metabotropic glutamate and 17885–17897, 1991 GABAB receptors contribute to the modulation of glucose-stimulated 36. Gromada J, Hoy M, Olsen HL, Gotfredsen CF, Buschard K, Rorsman P, insulin secretion in pancreatic beta cells. Diabetologia 45:242–252, 2002 Bokvist K: Gi2 proteins couple somatostatin receptors to low-conductance 13. Weaver CD, Gundersen V, Verdoorn TA: A high affinity glutamate/aspartate Kϩ channels in rat pancreatic ␣-cells. Pflugers Arch 442:19–26, 2001 transport system in pancreatic islets of Langerhans modulates glucose- 37. Schoepp DD, Jane DE, Monn JA: Pharmacological agents acting at sub- stimulated insulin secretion. J Biol Chem 273:1647–1653, 1996 types of metabotropic glutamate receptors. Neuropharmacol 38:1431– 14. Bertrand G, Gross R, Puech R, Loubatieres-Mariani MM, Bockaert J: 1476, 1999 Evidence for a glutamate receptor of the AMPA subtype which mediates 38. Gromada J, Hoy M, Buschard K, Salehi A, Rorsman P: Somatostatin insulin release from rat perfused pancreas. Br J Pharmacol 106:354–359, 1992 inhibits exocytosis in rat pancreatic ␣ cells by Gi2-dependent activation of 15. Bertrand G, Gross R, Puech R, Loubatieres-Mariani MM, Bockaert J: calcineurin and depriming of secretory granules. J Physiol 535:519–532, 2001

1006 DIABETES, VOL. 53, APRIL 2004