The expression of GLP-1 receptor mRNA and ... - Semantic Scholar

Journal of Neurochemistry, 2005, 92, 798–806

doi:10.1111/j.1471-4159.2004.02914.x

The expression of GLP-1 receptor mRNA and protein allows the effect of GLP-1 on glucose metabolism in the human hypothalamus and brainstem Elvira Alvarez,* M. Dolores Martıńez, Isabel Roncero,* Julie A. Chowen,à Beatriz Garcıá-Cuartero,* Juan D. Gispert,§ Carmen Sanz,* Patricia Va´zquez,* Antonio Maldonado,¶ Javier de Caćeres, Manuel Desco,§ Miguel Angel Pozo¶ and Enrique Bla´zquez* *Department of Biochemistry and Molecular Biology, Faculty of Medicine, Complutense University, Madrid, Spain Clinical Biochemistry Service, Hospital Clıńico San Carlos-Pabelloń 8, Madrid, Spain àDepartment of Endocrinology, Laboratory of Investigation, Hospital Ninõ Jesu´s, Madrid, Spain §Service of Medical Image, Hospital Gregorio Maranõń, Madrid, Spain ¶PET Complutense, Madrid, Spain

Abstract In the present work, several experimental approaches were used to determine the presence of the glucagon-like peptide-1 receptor (GLP-1R) and the biological actions of its ligand in the human brain. In situ hybridization histochemistry revealed specific labelling for GLP-1 receptor mRNA in several brain areas. In addition, GLP-1R, glucose transporter isoform (GLUT-2) and glucokinase (GK) mRNAs were identified in the same cells, especially in areas of the hypothalamus involved in feeding behaviour. GLP-1R gene expression in the human brain gave rise to a protein of 56 kDa as determined by affinity cross-linking assays. Specific binding of 125I-GLP-1(7–36) amide to the GLP-1R was detected in several brain areas and was inhibited by unlabelled GLP-1(7–36) amide, exendin-4 and exendin (9–39). A further aim of this work was to evaluate

cerebral-glucose metabolism in control subjects by positron emission tomography (PET), using 2-[F-18] deoxy-D-glucose (FDG). Statistical analysis of the PET studies revealed that the administration of GLP-1(7–36) amide significantly reduced (p < 0.001) cerebral glucose metabolism in hypothalamus and brainstem. Because FDG-6-phosphate is not a substrate for subsequent metabolic reactions, the lower activity observed in these areas after peptide administration may be due to reduction of the glucose transport and/or glucose phosphorylation, which should modulate the glucose sensing process in the GLUT-2- and GK-containing cells. Keywords: biological effects, gene expression, glucagon-like peptide-1 receptor, glucose sensing, human brain. J. Neurochem. (2005) 92, 798–806.

The existence of specific subpopulations of neurones involved in energy homeostasis, and located in the so-called ‘satiety and hunger centres’ of the hypothalamus, is well established. These neuronal pathways, containing both orexigenic and anorexigenic peptides, generate integrated responses to afferent stimuli that are related to modifications in metabolites or in the storage of fuels. Feeding behaviour is controlled by the antagonist effects of both classes of molecules, glucagon-like peptide-1 (GLP-1) being one of the components of the numerous groups of anorexigenic peptides. GLP-1(7–36) amide is a member of

the glucagon-related peptide family. It is produced by posttranslational modification of GLP-1, which is encoded by the

798

Received July 29, 2004; revised manuscript received October 5, 2004; accepted October 7, 2004. Address correspondence and reprint requests to Enrique Bla´zquez, Departamento de Bioquı´mica y Biologıá Molecular, Facultad de Medicina, Universidad Complutense, 28040-Madrid, Spain. E-mail: [email protected] Abbreviations used: FDG, 2-[F-18] deoxy-D-glucose; GK, glucokinase; GKRP, glucokinase regulatory protein; GLP-1R, glucagon-like peptide-1 receptor; PET, positron emission tomography.

2005 International Society for Neurochemistry, J. Neurochem. (2005) 92, 798–806

GLP-1 receptor expression in human brain 799

proglucagon gene (Mojsov et al. 1986), and it exerts several biological effects on peripheral tissues and the central nervous system (CNS). In both humans and experimental animals, this peptide stimulates insulin secretion (Kreymann et al. 1987) in a glucose-dependent manner, and inhibits gastric acid secretion and gastric emptying (Schjoldager et al. 1989). In addition, GLP-1(7–36) amide increases arterial blood pressure and heart rate in rats (Barragań et al. 1994, 1999) and stimulates surfactant secretion by human and rat type II pneumocytes (Benito et al. 1998; Vara et al. 2001). The proglucagon gene is expressed in the brain, the mRNA transcript being identical to that produced in the endocrine pancreas and intestine. In rats, it has been reported that the expression of the GLP-1 receptor gene gives rise to a protein (Wei and Mojsov 1995; Alvarez et al. 1996) with effects on the selective release of neurotransmitters (Mora et al. 1992), appetite and fluid homeostasis (Navarro et al. 1996; Turton et al. 1996). In men, GLP-1(7–36) amide has been shown to exert an anti-diabetogenic effect (Gutniak et al. 1992) and serves as a signal to reduce food intake. Here, we present experimental evidence of the presence of GLP-1 receptor mRNA and protein in human brain, as well as the coexpression of this gene with those of GLUT-2 and GK, proteins that may play a role in glucose sensing. Positron emission tomography (PET) is an imaging technology (Phelps 2000) in which compounds labelled with positron-emitting radioisotopes serve as molecular probes to identify and determine biochemical processes in vivo. Here, we used intravenously injected 2-[F-18] fluoro-2-deoxy-Dglucose (FDG) to trace the transport and phosphorylation of glucose in brain. Because of the inhibitory effect of deoxyglucose, FDG-6-phosphate is the end product of the process of glucose metabolism. Furthermore, since it is not a substrate for subsequent metabolic reactions, it is retained in the cell proportionately to the rate of glycolysis. The activity measured reflects FDG transport and phosphorylation by the cells and, in the case of the hypothalamus and brainstem, which contain cells expressing GLUT-2 and GK, may provide some information about brain glucose sensing. Accordingly, we report experimental evidence not only of the existence of the GLP-1 receptor mRNA and protein, and of GLP-1 receptor co-expression with GLUT-2 and GK, but also of the biological effects of GLP-1(7–36) amide on brain glucose metabolism in areas involved in energy homeostasis and glucose sensing.

Experimental procedures Materials GLP-1(7–36) amide was obtained from Bachem (St. Helens, UK) and exendin peptides were a generous gift from Dr John Eng, Veterans Hospital, New York, USA. Na125I (580–600 MBq/lg)

was supplied by Amersham Biotech (Barcelona, Spain). Disuccinimidyl suberate was from Pierce (Rockford, IL, USA) and reagents for sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) were obtained from Bio-Rad Laboratories S.A. (Madrid, Spain). All other reagents were from SigmaAldrich (Madrid, Spain) or Merck Biosciences (Madrid, Spain). Donors Human brain specimens were obtained at our School of Medicine from the Tissue Bank for Neurological Research. The donors of these organs had given written consent while alive, or their families had signed their consent after their death. Specimens were from six men and four women between 31 and 85 years of age. The study was approved by the Ethical Committee of our Institution in accordance with the Declaration of Helsinki-II. We only used human brain tissue from donors previously shown to have no neurological or psychiatric disease, or when brain histochemistry, as determined post-mortem, was normal. Post-mortem processing of tissue was performed according to the Tissue Bank protocol. Immediately after brain extraction, two symmetrical halves were obtained through a mid-sagittal incision. Thereafter, one half-brain was immersed in buffered 4% formaldehyde for routine neuropathological processing and diagnosis. The other half-brain was sectioned in coronal slices, fresh-frozen in liquid nitrogen and stored at ) 80C. For this study, frozen tissue samples were obtained from several brain regions.

In situ hybridization histochemistry In situ hybridization was performed as previously described (Chowen et al. 1993). Antisense or sense cRNA probes were generated with SP6 or T7 RNA polymerase in pGLPR-1, generously provided by Dr B. Thorens (Lausanne, Switzerland), in pGEM4z containing fragment 1–820 of human GLUT-2 cDNA (Fukumoto et al. 1988), a gift from Dr G. I. Bell (University of Chicago, Illinois, USA), or pGEMT-hGKi containing the coding region of the human pancreatic glucokinase gene (Alvarez et al. 2002). In situ hybridization double-labelling experiments were performed using an antisense digoxigenin-labelled cRNA probe complementary to GLP-1 receptor mRNA and 35S-labelled cRNA probes for the location of GLUT-2 or GK mRNA transcripts. The GLPR-1 riboprobe was synthesized using SP6 polymerase in the GLPR-1 and Dig RNA labelling kit (Roche Molecular Biochemicals, Barcelona, Spain). GLUT-2 and GK riboprobes were synthesized using T7 polymerase in a standard transcription reaction containing 10 lM 35S-UTP. This resulted in probes with specific activities of approximately 1.8 · 109 dpm/lg. The probes were hydrolysed in bicarbonate buffer to an average length of 150 bases. Sense cRNA probes were used as specificity controls and, under identical conditions, showed no detectable labelling. Detection of the digoxigenin-labelled probe was performed, according to the manufacturer’s directions (Roche Molecular Biochemicals, Barcelona, Spain), through the colour reaction to detect the digoxigenin. The slides were then rinsed in distilled water, rapidly dehydrated and allowed to air-dry. They were then dipped in Pyroxilin (3% in isoamyl acetate) and dried before coating with photographic emulsion (LM1, Amersham Pharmacia Biotech, Barcelona, Spain) and exposed at 4C for 6 weeks.


800 E. Alvarez et al.

Binding assays and cross-linking of 125I-GLP-1(7–36) amide to human brain particulate fractions Binding assays were performed by incubating aliquots of brain homogenates (200 lg protein) with 0.1–0.2 nM 125I GLP-1(7–36) amide (specific activity 3500–4000 dpm/fol) as described previously (Calvo et al. 1995). The amount of 125I-GLP-1(7–36) amide bound in the presence of 1 lM unlabelled GLP-1(7–36) was considered nonspecific binding. Protein contents were determined by the method of Lowry et al. (1951) with bovine serum albumin as the standard. For cross-linking experiments, 125I-GLP-1(7–36) amide was bound to brain homogenates as described above. The resulting sediments were then washed in Krebs Ringer HEPES [KRH; 118 mM NaCl, 5 mM KCL, 2.5 mM CaCl2, 1.2 mM KH2PO4, 1.2 mM MgSO4, 25 mM HEPES (pH 7.4)] buffer without bovine serum albumin and resuspended in the same medium. The crosslinking reaction was initiated by addition of 0.5 mM disuccinimidyl suberate (DSS) dissolved in dimethyl sulfoxide and was carried out for 15 min at 4C. The reaction was quenched with 10 mM Tris-HCl plus 1 mM EDTA, pH 7.4. Membranes were then recovered by centrifugation and subsequently solubilized and/or subjected to SDS-PAGE. Positron emission tomography (PET) studies Subjects Positron emission tomography with 2-[F-18] fluoro-2-deoxy-Dglucose (FDG) was used to evaluate the effect of GLP-1(7–36) amide on cerebral glucose metabolism in five men and five women aged between 20 and 24 years. After the volunteers had been informed about the nature and content of the test, they signed the written consent. The study was approved by the Ethical Committee of our Institution in accordance with the Declaration of Helsinki-II. On the test day, overnight-fasted subjects arrived at the Department and then rested on a bed in the supine position with slight elevation of the head. A catheter was inserted in an antecubital arm vein and saline (0.9% NaCl) was continuously infused during the test. After a period of rest, a basal blood sample was taken and then saline or GLP-1(7–36) amide was infused, providing a steadystate infusion rate of 0.75 pmol/kg body weight/min for 30 min. Blood samples were taken at 0 (basal sample), 5, 15, 30 and 90 min after saline or GLP-1(7–36) amide infusion from the antecubital vein of the other arm. Five minutes after starting the test, FDG was administered intravenously. With a lapse of least 6 months, each subject was subjected to two tests to verify the differences between the i.v. administration of saline alone or with GLP-1(7–36) amide. Immediately after i.v. perfusion had ended, PET scans provided measures of the tissue concentrations of the probe. More accurate information about the anatomical localization of the probe in the tissue was obtained with the help of nuclear magnetic resonance (NMR) of the brain. For GLP-1(7–36) amide infusion, the synthetic human peptide was dissolved in a 0.9% saline solution containing 3% of the volunteer’s own blood serum, subjected to sterile filtration and checked for sterility. PET procedures Data acquisition 18-Fluoride was produced in a 12 megaelectron volt (Mev) cyclotron (Oxford Instruments, Oxford, UK) by proton bombard-

ment of a high-pressure water target. [18F]fluoro-deoxy-D-glucose was synthesized by the method of nucleophilic substitution of a precursor by 18-F (Hamacher et al. 1986). PET scans were performed on a POSICAN HZL-R full ring tomograph (Positron Corporation, Houston, TX, USA) with an axial full field of view of 15 cm. The measured axial resolution was 5 mm FWHM; estimated in-plane resolution was 5.8 mm, with a measured system sensitivity of 140 Kcps/lCi/cc. PET scans were performed in the fasting state (more than 10 h) and the blood glucose concentration was < 100 mg/dL in all cases. Static PET imaging was performed 35 min after intravenous injection of 185 MBq of FDG/70 kg body weight, in a quiet, low-light room with the subject’s eyes closed. Data acquisition consisted of a 20 min emission scan and a 5 min transmission scan for attenuation correction, with a germanium 68 source. Image analysis PET images were analysed with the SPM 99 software package (Friston et al. 1995) [proportional scaling 12 · 12 · 12 mm full width half maximum (FWHM) smoothing]. The grey-level threshold was set at 0.8, i.e. only voxels with an intensity level above 0.8 of the mean value for that scan were included in the statistical analysis. Studies were transformed into a Talairach stereotaxic space (Talairach and Tournoux 1988), matching each scan to a reference template image. The metabolic effect of GLP-1(7–36) amide was assessed by means of longitudinal comparisons using the one-tailed paired Student’s t-test to check separately for hyper- and hypoactivations between both study conditions. The one-tailed significance threshold was set at p ¼ 0.001, resulting in an overall significance level of p ¼ 0.002 for the two-tail model. The significance of the metabolic changes was assessed and labelled in our results according to three different decreasing levels of evidence (LOE): L1, peak-height corrected p-value below 0.05; L2, an extent-corrected p-value below 0.05; and L3, uncorrected p-value below 0.001 and the area has already been previously reported as relevant. Areas with an uncorrected p-value below 0.001 that did not match any of the above criteria were not reported in the results, since they were likely to represent false positives. Assays Blood samples were collected in plastic tubes containing aprotinin (Trasylol with 1000 IU of kallikrein inhibitor; Bayer, Leverkusen, Germany). The samples were centrifuged and the plasma was frozen at ) 20C. Plasma glucose levels and other metabolic parameters were measured using a Hitachi (Tokyo, Japan) 737 automatic analyser. Circulating insulin was determined by radioimmunoassay (RIA) using a human insulin-specific kit in which the hormone does not cross-react with human proinsulin (< 0.2%) (Linco Research Inc., St Charles, MO, USA). This is a completely homologous assay since the antibody was raised against purified human insulin and both the standard and the tracer were prepared with human insulin. Glucagon was also determined by RIA (Faloona and Unger 1974) using an antibody obtained against the carboxy terminus of the glucagon molecule, and therefore mainly measured glucagon of pancreatic origin. GLP-1(7–36) amide concentrations were measured (Orskov et al. 1991) after extraction



of serum with 70% ethanol (v/v, final concentration) against standards of this human synthetic peptide and an antiserum that cross-reacts with equal strength with all peptides containing the GLP-1 sequence, regardless of NH2- or COOH-terminal extensions. The intra- and interassay coefficients of variation were < 5 and 15%, respectively. All samples were analysed in the same assay, and the coefficient of variation of the triplicate samples did not exceed 5%.

Results

In situ hybridization histochemistry of GLP-1 receptor, GLUT-2 and glucokinase mRNAs By in situ hybridization histochemistry with the specific antisense probes, labelled cells were found throughout the entire brain, whereas no labelling was detected with the sense probes. Cells positive for GLP-1 receptor mRNA were found to be distributed throughout the cerebral cortex, hypothalamus (mainly the ventromedial and arcuate nuclei), hippocampus, thalamus, caudate-putamen and globus pallidum. Dark-field photomicrographs of labelled cells of the hypothalamus (arcuate nucleus) and cerebral cortex are shown in Fig. 1, with clusters of white grains representing cells positive for GLP-1 receptor mRNA. Double-labelling assays revealed that the GLP-1 receptor mRNA, as identified by a blue reaction product, and GLUT-2 and GK mRNAs, as localized by silver grains, were present in the hypothalamus, caudate-putamen, mesencephalon and bulb (Fig. 2). In the hypothalamus, GLP-1R was widely

Fig. 2 Double-labelling in situ hybridization histochemistry of GLP-1 receptors and GLUT-2 or glucokinase mRNAs. (a) Co-localization of GLP-1 receptor and GLUT-2 mRNAs in the hypothalamus (ventromedial nucleus). (b) Co-localization of GLP-1 receptor and glucokinase mRNAs in the hypothalamus (ventromedial nucleus). (c) Bulb shows co-localization of GLP-1 receptor and GLUT-2 mRNAs. (d) Co-localization of GLP-1 receptor and GLUT-2 mRNAs in the mesencephalon. (e) Co-localization of GLP-1 receptor and GK mRNAs in the caudate putamen. Blue reaction product indicates labelling of GLP-1 receptor mRNA, whereas silver grains indicate the localization of either GLUT-2 or glucokinase mRNAs. Filled arrows are representative of double-labelled cells, while the unfilled arrows represent single labelled cells. Scale bars: 20 lm.

expressed and many cells positive for both GLP-1R and GLUT-2, GLP-1R and GK, or GLUT-2 and GK were found, as well as a few areas with only a GLUT-2-positive signal. Positive cells were quite evenly dispersed throughout the caudate-putamen, all three combinations of double-labelled cells being found. In the occipital cerebral cortex and bulb, the three types of double-labelled cells appeared in epithelial cells around blood vessels and in specific cell layers. In the mesencephalon and cerebellum, more cells expressed GLP1R than GLUT-2 or GK, but double-labelled cells of all combinations were also found. Fig. 1 In situ hybridization histochemistry of GLP-1 receptor mRNA in human brain. Dark-field photomicrographs of labelled cells in hypothalamus (arcuate nucleus) (a) and cerebral cortex (b). Clusters of white grains represent cells positive for GLP-1 receptor mRNA. Scale bar: 30 lm.

Characterization of 125I-GLP-1(7–36) amide binding sites in human brain Binding assays were performed in brain homogenates rather than membrane-enriched fractions because in preliminary



Table 1 Binding of human brain

125

I-GLP-1(7–36) amide to different regions in the

Region

Specific binding (fmol/mg protein)

Bulb Frontal cortex Parietal cortex Occipital cortex Temporal cortex Globus pallidum Hypothalamus Caudate-putamen Thalamus

0.98 2.90 1.55 4.10 3.10 2.70 2.70 3.80 2.70

± ± ± ± ± ± ± ± ±

0.30 0.84 0.40 0.75 0.94 0.97 0.85 0.80 0.85

Data are means ± SE (n ¼ 4–6).

studies, no differences in GLP-1(7–36) amide binding between these preparations were observed (Calvo et al. 1995). The same observations have been reported for the binding of glucagon to rat brain (Hoosein and Gurd 1984). The distribution of 125I-GLP-1(7–36) amide-binding sites in several regions of the human brain is indicated in Table 1. Specific binding was detected in several human brain areas, the highest values (fmol/mg of protein) being found in cerebral cortex (4.1 ± 0.7), hypothalamus (2.7 ± 0.8) and caudate putamen (3.2 ± 0.6). Because these regions were the richest in GLP-1(7–36) amide-binding sites, they were selected for further characterization by transient-state and steady-state binding sites, together with chemical crosslinking and SDS-PAGE procedures. The binding of 125I-GLP-1(7–36) amide to hypothalamus and cerebral cortex was time- and temperature-dependent. The specificity of GLP-1(7–36) amide binding to brain tissue was estimated from competition studies, using the unlabelled peptides GLP-1(7–36) amide, exendin-4 and exendin (9–39). As shown in Fig. 3, specific tracer binding was inhibited in a concentration-dependent manner by the unlabelled peptides, the pattern of displacement being similar in both hypothalamus and cerebral cortex. The data obtained with increasing concentrations of native GLP-1(7–36) amide afforded a curvilinear plot in both hypothalamus and cerebral cortex (Fig. 4b), suggesting a two-site model of ligand binding. Cross-linking of 125I-GLP-1(7–36) amide to human brain homogenates To identify structurally and characterize the brain GLP-1 (7–36) amide-binding protein, we used the homobifunctional reagent DSS to attach covalently-radiolabelled GLP-1(7–36) amide to the binding component. The resulting 125I-GLP1(7–36) amide-cross-linked samples were then analysed by SDS-PAGE, followed by autoradiography. As shown in the representative autoradiograph of Fig. 4(a), a single labelled band centred at Mr 59 kDa was identified in cerebral cortex.

Fig. 3 Effect of unlabelled GLP-1(7–36) amide and related peptides exendin (9–39) and exendin-4 on specific 125I-GLP-1(7–36) amide binding to occipital cerebral cortex (a) and hypothalamus (b) homogenates. Binding assays were carried out by incubating aliquots of tissue homogenates with 125I-GLP-1(7–36) amide in the presence of increasing amounts of the different unlabelled GLP-1(7–36) amide (j), exendin (9–39) (.) and exendin-4 (d) for 2 h at 25C. Specific binding was determined as outlined in Experimental procedures and is expressed as the percentage of specific binding measured in the presence of tracer alone (maximum bound). Data means ± SE values (n ¼ 3–4).

(a)

(b)

Fig. 4 Cross-linking of 125I-GLP-1(7–36) amide to homogenates from occipital cerebral cortex in the presence of increasing amounts of unlabelled GLP-1(7–36) amide, followed by treatment with 0.5 mM DSS as outlined in Experimental procedures. Cross-linked preparations were then analysed by SDS-PAGE. Autoradiograms of representative dried gels are shown (a). Scatchard analysis of 125I-GLP-1(7–36) amide binding to occipital cerebral cortex (d) and hypothalamus (s) homogenates. Displacement curves obtained with increasing concentrations of unlabelled GLP-1(7–36) amide as in Fig. 3 were transformed to give Scatchard plots (b). Data are means (n ¼ 3).



73 ± 5 mg/mL) or insulin concentration levels (ranging from 7 ± 1 to 11 ± 1 lU/mL). By contrast, plasma GLP-1(7–36) amide increased significantly after the administration of the peptide (from 197 ± 32 at 0 time to 345 ± 73 pg/mL 30 min after starting peptide infusion), whereas circulating glucagon concentrations decreased (from 100 ± 8 pg/mL at 0 time to 82 ± 12 and 85 ± 17 pg/mL at 5 and 30 min of peptide infusion, respectively). In addition, PET imaging was sensitive to GLP-1(7–36) amide administration (Fig. 5). Thus, this peptide significantly reduced (p < 0.001, uncorrected) cerebral glucose metabolism in the hypothalamus and brainstem as compared with the data obtained in normal control subjects without GLP-1(7–36) amide administration. Discussion

Fig. 5 Positron emission tomography of glucose metabolism in brains of control subjects after i.v. perfusion with GLP-1(7–36) amide. The procedure followed is described in Experimental procedures. SPM99 group analysis of 10 subjects in two conditions: with GLP-1(7–36) amide (n ¼ 10) and with saline (n ¼ 10). Differences were projected on transverse, sagittal and coronal normalized brain MRI. Hypometabolism in hypothalamus area (x )2; y )16; z )2 in red) was identified in GLP-1(7–36) amide compared with saline group (p < 0.0001 uncorrected; Z ¼ 4.73).

Labelling of the brain 125I-GLP-1(7–36) amide-binding protein complex was abolished when 1 lM GLP-1(7–36) amide was added during the binding step. Effect of GLP-1(7–36) amide on brain glucose metabolism as assessed by positron emission tomography Positron emission tomography with FDG was used to evaluate cerebral glucose metabolism in normal control subjects. Because glucose metabolism provides most of the chemical energy required for the brain to function properly, FDG represents a good general molecular imaging probe to evaluate basal and GLP-1(7–36) amide on ATP-dependent functions. Intravenous administration of GLP-1(7–36) amide did not modify blood plasma glucose (ranging from 71 ± 5 to

Here, we report experimental evidence indicating that the GLP-1 receptor gene is expressed in human brain as a protein, with binding properties similar to those of the GLP1R located in peripheral tissues. In addition, we present data indicating that GLP-1(7–36) amide may act through this receptor to modify glucose metabolism in selective areas of the human brain involved in glucose sensing. In situ hybridization histochemistry was used to identify the mRNA of the GLP-1 receptor in several cerebral areas such as cerebral cortex, hypothalamus, caudate putamen or cerebellum. The GLP-1 receptor cDNAs from human and rat brain have been cloned and sequenced (Wei and Mojsov 1995; Alvarez et al. 1996) and in both species, the deduced amino acid sequences are the same as those found in pancreatic islets. The GLP-1R gene gives rise to a protein with a regional distribution in well defined areas of the human brain, with the cerebral cortex, caudate putamen, hypothalamus including the ventromedial and arcuate nuclei, thalamus and globus pallidum being the richest locations of GLP-1 binding sites. These results are partly in agreement with previous findings reporting that in the rat brain, a high degree of specific binding of labelled GLP-1 is mainly concentrated in the hypothalamus, thalamus and basal nuclei. However, they differ in that a low degree of binding is found in the cerebral cortex of rats. The high density of GLP-1 binding sites observed in the rodent and human hypothalamus suggests a functional role for this peptide in certain physiological processes, such as central control of food intake and pancreatic hormone release. In addition, the higher content of GLP-1 receptors in the human cerebral cortex, especially in the occipital and frontal cortex, suggests some unknown functions of this peptide. The interaction of GLP-1(7–36) amide with human brain homogenates matches the generally accepted criteria for peptide–receptor interactions, such as high affinity, saturability and specificity. Steady-state binding was present in high- and low-affinity forms in brain homogenates. These results are comparable with those obtained in membrane fractions from rat brain



(Calvo et al. 1995) and rat adipose tissue (Valverde et al. 1993). However, only one population of high-affinity GLP1(7–36) amide-binding sites was identified in membranes from rat insulinoma-derived (RIN) cells (Go¨ke and Conlon 1988), rat lung membranes (Kause et al. 1988) and rat gastric glands (Uttenthal and Bla´zquez 1990). Whether these differences are due to the experimental conditions used or to tissue-dependent differences in primary receptor structure, or whether they might be due to environmentally determined effects on receptor conformation, remains to be determined. Nevertheless, two kinds of affinities have been found for many members of the G-protein-coupled receptors, and the high-affinity sites are believed to be due to the interaction of the receptor with G-proteins (Lefkowitz et al. 1983; Spiegel et al. 1992). The evidence strongly suggests that the covalent 125I-GLP-1 (7–36) amide-binding protein complex identified in human brain homogenates represents the structural correlate of the interaction of the 125I-GLP-1(7–36) amide with nervous tissue GLP-1 receptor, as detected in binding assays. Increasing amounts of unlabelled GLP-1(7–36) amide resulted in a decrease in radiolabelled-peptide binding, which was paralleled by a concomitant reduction in the intensity of the crosslinked band, the dose–response curves for both effects being superimposable. Furthermore, unlabelled exendin-4 and exendin (9–39), an agonist and an antagonist, respectively, for the GLP-1 receptor, inhibited complex labelling with an order of potency that was in very good agreement with the ligand specificity exhibited by the GLP-1 receptor in competitive binding assays. The presence of this GLP-1 receptor in the human brain, with high affinity and specificity for its ligand, suggests that it may function to mediate the biological actions of GLP-1 (7–36) amide in the CNS. Our findings indicate that GLUT-2, GK and GLP-1 receptors are expressed in many of the same cells of the human hypothalamus, and are located in areas involved in the regulation of energy homeostasis, feeding behaviour and glucose metabolism. It is noteworthy that in the brain, GLP-1(7–36) amide contributes to reducing food intake (Navarro et al. 1996; Turton et al. 1996), and the co-localization of those three components in hypothalamic neurones suggests that a glucose sensor system may be involved in the transduction of signals required to produce a state of satiety. Accordingly, increased glycaemia after meals may be recognized by these hypothalamic cells due to the high-Km glucose transporter activity of GLUT-2 and the highKm glucose phosphorylation of GK. GK functions as a true glucose sensor (Matschinsky 1990; Jetton et al. 1994) since it is not inhibited by glucose-6-phosphate, has a low affinity for glucose, and its kinetic co-operativity with glucose allows the rate of glucose phosphorylation to be directly proportional to blood glucose concentrations. Although GLUT-2 also plays a role in the discrimination of glucose concentrations, its ability in this process is limited. GK activity may also be regulated by

the effects of glucokinase regulatory protein (GKRP), which functions as a metabolic sensor, acting in accordance with the metabolic needs of the cells (Van Schaftingen et al. 1984; Shiota et al. 1999; Roncero et al. 2004). Thus, GK and GKRP interact in rat brain and may respond to fructose esters (Alvarez et al. 2002). To understand further the biological effects induced through the GLP-1 receptor in human brain, the effect of GLP-1(7–36) amide on cerebral glucose metabolism was analysed using PET technology. As expected, intravenous administration of GLP-1(7–36) amide increased the circulating levels of this peptide and reduced plasma glucagon concentrations. In contrast, plasma glucose and insulin levels remained unmodified during the study. Peripheral administration of GLP-1(7–36) amide induced changes in the carbohydrate metabolism of the brain; this can be explained in terms of the fact that this peptide can enter the brain by binding to blood barrier-free organs such as the subfornical organ and the area postrema (Orskov et al. 1996). In addition, it could be transported into the brain through the choroid plexus, which has a high density of GLP-1 receptors (Alvarez et al. 1996). These findings explain the satiety-induced effects of peripheral or central administration of GLP-1 (7–36) or, as shown here, the central alterations produced by i.v. administration of the peptide. PET allows the molecular imaging of biological actions with an in vivo integrative view of such processes. Using this technology, we observed that i.v. administration of GLP-1(7–36) amide produced a significant reduction in carbohydrate metabolism in selective areas of the brain, including the hypothalamus and brainstem, both areas that are involved in feeding behaviour. Since we used the FDG as a probe, the end product was FDG-6-phosphate, which is not an appropriate substrate for subsequent reactions and is retained in the cell proportionately to the rate of glycolysis. Thus, the accumulation of FDG-6-phosphate into the cells serves to assess the facilitated transport and hexokinase phosphorylation of glucose, which, in the case of hypothalamus and brainstem cells containing GLUT-2 and GK, might facilitate the glucose-sensing process. These findings are of interest because glucoreceptive sites controlling food intake and blood glucose have been found in the medulla oblongata and mesencephalon of the rat (Ritter et al. 2000), in addition to those reported in the hypothalamus. Our results indicate that the GLP-1 receptor is also expressed in the human brain, and that it most likely mediates the effect of GLP-1(7–36) on glucose metabolism in selective areas of the hypothalamus and brainstem; it may also facilitate the process of glucose sensing in these areas. Because of the reduced number of neurones involved in glucose sensing (Oomura et al. 1969; Ashford et al. 1990), approximately 40% of cells in the ventromedial and 30% of cells in the lateral areas of the hypothalamus, PET imaging offers a good procedure for identifying these cell signals in vivo more accurately compared with the more commonly used in vitro procedures.



These findings open new doors for studying the effects of other regulatory peptides in subjects under control and pathophysiological situations. Acknowledgements The authors are indebted to Dr G. I. Bell (Chicago, IL, USA) and Dr B. Thorens (Lausanne, Switzerland) for the generous gift of the human GLUT-2 cDNA and the GLP-1 receptor cDNA probes, respectively, and to the Tissue Bank for Neurological Research in our School of Medicine for the human brain specimens used in this study. This work was supported by grants from the Fondo de Investigacioń Sanitaria, Instituto Carlos III, RGDM (603/212) and RCMN (C03/08), the Ministerio de Ciencia y Tecnologıá and the Comunidad de Madrid, Spain.

References Alvarez E., Roncero I., Chowen J. A., Thorens B. and Bla´zquez E. (1996) Expression of the glucagon-like peptide-1 receptor gene in rat brain. J. Neurochem. 66, 920–927. Alvarez E., Roncero I., Chowen J. A., Va´zquez P. and Bla´zquez E. (2002) Evidence that glucokinase regulatory protein is expressed and interacts with glucokinase in rat brain. J. Neurochem. 80, 45–53. Ashford M. L. J., Borden R. and Treherne J. M. (1990) Glucose-induced excitation of hypothalamic neurons is mediated by ATP-sensitive K+ channels. Pflu¨gers Arch. 415, 479–483. Barragań J. M., Rodriguez R. and Bla´zquez E. (1994) Changes in arterial blood pressure and heart rate induced by glucagon-like peptide-1 (7–36) amide in rats. Am. J. Physiol. 266, E459–E466. Barragań J. M., Eng J., Rodriguez R. and Bla´zquez E. (1999) Neural contribution to the effect of glucagon-like peptide-1 (7–36) amide on arterial blood pressure in rats. Am. J. Physiol. 277, E784– E791. Benito E., Bla´zquez E. and Bosch M. A. (1998) Glucagon-like peptide-1 (7–36) amide increases pulmonary surfactant secretion through a cyclic adenosine 3¢, 5¢-monophosphate-dependent protein kinase mechanism in rat type II pneumocytes. Endocrinology 139, 2363– 2368. Calvo J. C., Yusta B., Mora F. and Bla´zquez E. (1995) Structural characterization by affinity cross-linking of glucagon-like peptide1 (7–36) amide receptor in rat brain. J. Neurochem. 64, 299–306. Chowen J. A., Argente J., Gonza´lez-Parra S. and Garcıá Segura L. M. (1993) Differential effects of the neonatal and adult sex steroid environments on the organization and activation of hypothalamic growth hormone-releasing hormone and somatostatin neurons. Endocrinology 133, 2792–2280. Faloona G. R. and Unger R. H. (1974) Glucagon, in Methods of Hormone Radioimmunoassay (Saffe B. and Behrman H. R., eds), pp. 310–330. Academic Press, New York. Friston K. J., Holmes A. P., Worsley K. J., Poline J. P., Frith C. D. and Frackowiak R. S. J. (1995) Statistical parametric maps in functional imaging: a general linear approach. Hum. Brain Mapping 2, 189–210. Fukumoto H., Seino S., Imura H., Seino Y., Eddy R. L., Fukushima Y., Byers M. G., Shows T. B. and Bell G. I. (1988) Sequence, tissue distribution, and chromosomal localization of mRNA encoding a human glucose transport-like protein. Proc. Natl Acad. Sci. USA 85, 5434–5438. Go¨ke R. and Conlon J. M. (1988) Receptors for glucagon-like peptide-1 (7–36) amide on rat insulinoma-derived cells. J. Endocrinol. 116, 357–362.

Gutniak M., Orskov C., Holst J. J., Alvień B. and Efendic S. (1992) Antidiabetogenic effect of glucagon-like peptide-1 (7–36) amide in normal subjects and patients with diabetes mellitus. N. Engl. J. Med. 326, 1316–1322. Hamacher K., Coenen H. H. and Stocklin G. (1986) Efficient stereospecific synthesis of non-carried-added 2-[18 F]flouro-deoxy-Dglucose using aminopolyether-supported nucleophilic substitution. J. Nucl. Med. 27, 235–238. Hoosein N. M. and Gurd R. S. (1984) Identification of glucagon receptors in rat brain. Proc. Natl Acad. Sci. USA 81, 4368–4437. Jetton T. L., Liang Y., Pettepher C. C., Zimmerman E. C., Cox F. G., Horvath K., Matschinsky F. M. and Magnuson M. A. (1994) Analysis of upstream glucokinase promoter activity in transgenic mice and identification of glucokinase in rare neuroendocrine cells in the brain and gut. J. Biol. Chem. 269, 3641–3654. Kause S. M., Kreymann B., Ghatei M. A. and Bloom S. R. (1988) Identification and characterization of glucagon-like peptide-1 (7–36) amide-binding sites in the brain and lung. FEBS Lett. 241, 209–212. Kreymann B., Williams G., Ghatei M. A. and Bloom S. R. (1987) Glucagon-like peptide-1 (7–36): a physiological incretin in man. Lancet 2, 1300–1304. Lefkowitz R. J., Stadel J. M. and Caron M. G. (1983) Adenylate cyclasecoupled beta-adrenergic receptors: structure and mechanisms of activation and desensitization. Annu. Rev. Biochem. 52, 159–186. Lowry O. H., Rosebrough N. J., Farr A. L. and Randall R. J. (1951) Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193, 265–275. Matschinsky F. M. (1990) Perspectives in diabetes: glucokinase as glucose sensor and metabolic signal generator pancreatic, b-cells and hepatocytes. Diabetes 39, 647–652. Mojsov S., Heinrich G., Wilson I. B., Ravazzola M., Orci L. and Habener J. F. (1986) Preproglucagon gene expression in pancreas and intestine diversifies at level of posttranslational processing. J. Biol. Chem. 261, 11 880–11 889. Mora F., Expo´sito I., Sanz B. and Bla´zquez E. (1992) Selective release of glutamine and glutamic acid produced by perfusion of glucagonlike peptide-1 (7–36) amide in the basal ganglia of the conscious rat. Brain Res. Bull. 29, 359–361. Navarro M., Rodriguez-Fonseca F., Alvarez E., Chowen J. A., Zueco J. A., Go´mez R., Eng J. and Bla´zquez E. (1996) Colocalization of glucagon-like peptide-1 (GLP-1) receptors, glucose transporter GLUT-2, and glucokinase mRNAs in rat hypothalamic cells: Evidence for a role of GLP-1 receptor agonists as an inhibitory signal for food and water intake. J. Neurochem. 67, 1982–1991. Oomura Y., Ono T., Ooyama H. and Waynes M. J. (1969) Glucose and osmosensitive neurons of the rat hypothalamus. Nature 222, 282– 284. Orskov C., Seppensen J., Madsbad S. and Holst J. J. (1991) Proglucagon products in plasma of non-insulin-dependent diabetics and nondiabetic controls in the fasting state and after oral glucose and intravenous arginine. J. Clin. Invest. 85, 415–423. Orskov C., Poulsen S. S., Moller M. and Holst J. J. (1996) Glucagon-like peptide-1 receptors in the subfornical organ and the area postrema are accessible to circulating glucagon-like peptide-1. Diabetes 45, 832–835. Phelps M. E. (2000) Positron emission tomography provides molecular imaging of biological processes. Proc. Natl Acad. Sci. USA 97, 9226–9233. Ritter S., Dinh T. T. and Zhang Y. (2000) Localization of hindbrain glucoreceptive sites controlling food intake and blood glucose. Brain Res. 856, 37–47. Roncero I., Alvarez E., Chowen J. A., Sanz C., Ra´bano A., Va´zquez P. and Bla´zquez E. (2004) Expression of glucose transporter isoform



GLUT-2 and glucokinase genes in human brain. J. Neurochem. 88, 1203–1210. Schjoldager B. T. G., Mortesen P. E., Christiansen C., Orskov C. and Holst J. J. (1989) GLP-1 (glucagon-like peptide-1) and truncated GLP-1 fragments of human proglucagon inhibit gastric acid secretion in humans. Dig. Diseases Sci. 34, 703–707. Shiota C., Coffey J., Grippo J. F. and Magnuson M. A. (1999) Nuclear import of hepatic glucokinase depends upon glucokinase regulatory protein, whereas export is due to a nuclear export signal sequence in glucokinase. J. Biol. Chem. 274, 37 125–37 130. Spiegel A. M., Shenker A. and Weinstein L. S. (1992) Receptor-effector coupling by G proteins: Implication for normal and abnormal signal transduction. Endocr. Rev. 13, 536–565. Talairach J. and Tournoux P. (1988) A Co-Planar Stereotactic Atlas of the Human Brain. Thieme, New York. Turton M. D., O’Sea D., Gunn I. et al. (1996) A role for glucagon-like peptide-1 in the central regulation of feeding. Nature 379, 69–72.

Uttenthal L. O. and Bla´zquez E. (1990) Characterization of high affinity receptors for truncated glucagon-like peptide-1 in rat gastric glands. FEBS Lett. 262, 139–141. Valverde I., Me´rida E., Delgado E., Trapote M. A. and VillanuevaPenãcarrillo M. L. (1993) Presence and characterization of glucagon-like peptide-1 (7–36) amide receptors in solubilized membranes of rat adipose tissue. Endocrinology 132, 75–79. Van Schaftingen E., Detheux M. and Veiga da Cunha M. (1984) Shortterm control of glucokinase activity: role of a regulatory protein. FASEB J. 8, 414–419. Vara E., Arias-Dıáz J., Balibrea J. L. and Bla´zquez E. (2001) Glucagon-like peptide-1 (7–36) amide stimulates surfactant secretion in human type II pneumocytes. Am. J. Respir. Crit. Care Med. 163, 840–846. Wei Y. and Mojsov S. (1995) Tissue-specific expression of the human receptor for glucagon-like peptide-1: brain, heart and pancreatic forms have the same deduced amino acid sequences. FEBS Lett. 358, 219–224..
