glutamatergic innervation of neuropeptide y and pro-opiomelanocortin-containing neurons in the...

Glutamatergic innervation of neuropeptide Y andpro-opiomelanocortin-containing neurons in thehypothalamic arcuate nucleus of the rat

Jozsef Kiss, Zsolt Csaba, Agnes Csaki and Bela HalaszNeuroendocrine Research Laboratory, Hungarian Academy of Sciences and Semmelweis University, Department of HumanMorphology and Developmental Biology, Tuzolto u. 58. Budapest, Hungary

Keywords: asymmetric synapse, body weight regulation, excitatory amino acids, glutamatergic innervation, neuromorphology

Abstract

The hypothalamic arcuate nucleus contains a number of neurochemically different cell populations, among others neuropeptide Y(NPY)- and pro-opiomelanocortin (POMC)-derived peptide-expressing neurons; both are involved in the regulation of feeding andenergy homeostasis, NPY neurons also in the release of hypophysiotropic hormones, sexual behaviour and thermogenesis. Recentobservations indicate that there is a dense plexus of glutamatergic fibres in the arcuate nucleus. The aim of the present studies wasto examine the relationship of these fibres to the NPY and POMC neurons in the arcuate nucleus. Double-label immunoelectronmicroscopy was used. Glutamatergic elements were identified by the presence of vesicular glutamate transporter 1 (VGluT1) or 2(VGluT2) (selective markers of glutamatergic elements) immunoreactivity. A significant number of VGluT2-immunoreactive terminalswas observed to make asymmetric type of synapses with NPY and with b-endorphin (a marker of POMC neurons)-immunostainednerve cells of the arcuate nucleus. About 15% of VGluT2 synapsing terminals established asymmetric synapses with NPY-positivecells and more than 40% of VGlut2-positive terminals formed synapse on b-endorphin-positive neurons. VGluT2-positive perikaryawere also observed, part of them also contained b-endorphin. Nerve terminals containing both VGluT2 and b-endorphin weredemonstrated in the cell group. Only very few VGluT1 fibres were detected. Our observations provide the first directneuromorphological evidence for the existence of glutamatergic innervation of NPY and POMC neurons of the arcuate nucleus.

Introduction

The hypothalamic arcuate nucleus contains a number of neurochem-ically different cell populations, among others neuropeptide Y(NPY)- and pro-opiomelanocortin (POMC)-derived peptide-expres-sing neurons (Everitt et al., 1986; Meister et al., 1989). NPY, a 36-amino acid peptide of the pancreatic polypeptide family, stimulatesfood intake and body weight gain (Kalra et al., 1999), it is involvedin the release of hypophysiotropic hormones (Kalra & Crowley,1992), inhibits sexual behaviour (Clark et al., 1984) and decreasesthermogenesis (Billington et al., 1994). POMC is the polypeptideprecursor of b-endorphin, a-melanocyte-stimulating hormone(a-MSH) and adrenocorticotropin. Both b-endorphin and a-MSHin the arcuate neurons participate in the regulation of feeding andenergy homeostasis (Elmquist et al., 1999; Kalra et al., 1999; Kalra& Kalra, 2003).

Several data suggest that glutamate, which has been identified as thedominant excitatory transmitter in hypothalamic neuroendocrineregulation (van den Pol et al., 1990; van den Pol, 1991; van denPol & Trombley, 1993), is involved in the innervation of arcuateneurons. Both ionotropic (Petralia & Wenthold, 1992; Petralia et al.,1994a,b,c; van den Pol et al., 1994; Eyigor et al., 2001) andmetabotropic glutamate receptors (Ghosh et al., 1997; Kiss et al.,1997) are expressed in this cell group. It has been reported (van den

Pol et al., 1990) that arcuate neurons receive synaptic contact fromaxons exhibiting glutamate immunoreactivity. However, glutamateexists as four different pools in the CNS: transmitter pool, metabolicpool, glial pool and c-aminobutyric acid (GABA) precursor pool. Thismakes it difficult and requires special care to differentiate those cellsthat secrete glutamate from those that are simply immunoreactive formetabolic glutamate that is not destined for release (Storm-Mathisenet al., 1983). There is now fairly strong evidence for the view that therecently discovered vesicular glutamate transporters [vesicular glu-tamate transporter type 1 (VGluT1), vesicular glutamate transportertype 2 (VGluT2), vesicular glutamate transporter type 3 (VGluT3)] arespecific to presumed glutamatergic neuronal elements, and comprise anovel and valuable tool for neuromorphological studies to identifyglutamatergic neuronal structures. These proteins transport glutamateinto secretory vesicles with the same kinetics as previously describedfor synaptic vesicles (Bellocchio et al., 2000; Takamori et al., 2000;Fremeau et al., 2001; Gras et al., 2002). It has been reported (Collinet al., 2003; Lin et al., 2003) that there are single VGluT1 fibres and avery dense plexus of VGluT2-immunoreactive fibres in the arcuatenucleus. Further, it has been observed by Collin et al. (2003) thatVGluT2 fibres surround NPY- and POMC-immunoreactive cell bodiesin the arcuate nucleus. However, we still do not know whetherglutamatergic terminals synapse on such neurons or not? The aim ofthe present investigations was to study this question. Single- anddouble-label immunocytochemistry was used, and the brain sectionswere examined by confocal laser scanning microscopy and under theelectron microscope.

Correspondence: Dr B. Halasz, as above.E-mail: [email protected]

Received 15 November 2004, revised 10 January 2005, accepted 20 January 2005

European Journal of Neuroscience, Vol. 21, pp. 2111–2119, 2005 ª Federation of European Neuroscience Societies

doi:10.1111/j.1460-9568.2005.04012.x

2112 J. Kiss et al.

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 2111–2119

Materials and methods

Animals

A total of 14 adult male (250–300 g body weight) Sprague–Dawleyrats (Charles River, Budapest, Hungary) were used. Animals were keptunder standard laboratory conditions, with tap water and regular ratchow ad libitum, in a 12 h light : dark cycle. Animal treatment was inaccordance with the rules set for experimentation on animals by theRegulation for Animal Care and Use in Hungary (Parliamentary Act of1998.XXVIII). All efforts were made to minimize number of animalsused. The animal treatments were made with special care.

Tissue preparation and immunohistochemistry

Three kinds of investigations were made. (i) Double-immunofluores-cence histochemistry for VGluT1 and NPY or b-endorphin, and forVGluT2 and NPY or b-endorphin in the arcuate nucleus for lightmicroscopy. (ii) Single-immunogold labelling of VGluT1 and VGluT2in the arcuate nucleus for electron microscopy. (iii) Double-immuno-labelling of VGluT1 and NPY or b-endorphin, and of VGluT2 andNPY or b-endorphin in the arcuate nucleus for electron microscopy.

Double immunofluorescence histochemistry for VGluT1 and NPYor b-endorphin, and for VGluT2 and NPY or b-endorphinin the arcuate nucleus for light microscopy

Six animals were used for double-labelling immunofluorescenceexperiments. To increase NPY and b-endorphin immunoreactivity inthe arcuate neurons, colchicine was injected into the lateral ventricle72 h before perfusion in three rats. The animals were anaesthetizedwith Equithesin (chlornembutal 0.3 mL ⁄ 100 g body weight, injectedintraperitoneally), immobilized in a stereotaxic head frame (Stoelting,Wood Dale, IL, USA), and an intraventricular injection of colchicine(75 lg in 10 lL of a 0.9% solution of sodium chloride) wasadministered into the ventricle. Additionally three rats were processedwithout colchicine treatment in order to examine double-labelling innerve fibres and axon terminals. Three days later the animals were re-anaesthetized with intraperitoneally administered chloral hydrate(0.35 g ⁄ kg body weight) dissolved in 0.1 m phosphate buffer (Merck,Darmstadt, Germany), pH 7.4 (PB), then they were rapidly perfusedtranscardially with 50 mLTyrode’s solution, followed by 400 mL of amixture of freshly depolymerized 4% paraformaldehyde (Merck),0.1% glutaraldehyde (TAAB Laboratories, Aldermaston, Berks, UK)and 10% (v ⁄ v) saturated picric acid (Merck) in PB. This wasimmediately followed by 200 mL of glutaraldehyde-free fixative. Thebrains were then removed from the skull and postfixed by immersionin the same glutaraldehyde-free fixative for an additional 3 h at 4 �C.After rigorous washing in PB, a coronal segment including the entirerostrocaudal extent of the arcuate nucleus region was cut out andplaced into PB containing 30% sucrose overnight at 4 �C. Thirty-micrometre-thick coronal sections were cut on a freezing microtome(type 1206, Reichert-Jung, Nussloch, Germany) and were collected in

four groups. Sections were washed three times in PB, then incubatedin PB containing 0.5% (v ⁄ v) Triton X-100 (SIGMA, Steinheim,Germany) for 30 min, washed in PB, then preincubated 1 h at roomtemperature in PB containing 5% normal goat serum (NGS) and 2%bovine serum albumin (BSA, SIGMA).Double-immunofluorescence histochemistry was performed for

VGluT1 plus NPY; VGluT1 plus b-endorphin; VGluT2 plus NPY;and VGluT2 plus b-endorphin on free-floating 30-lm-thick sections.Primary antibodies were applied in cocktails prepared by the followingcombinations: (i) guinea-pig anti-VGluT1 polyclonal antibody(diluted 1 : 6000, Chemicon International, Temecula, CA, USA) plusrabbit anti-NPY polyclonal antibody (diluted 1 : 2000, ChemiconInternational) or rabbit polyclonal antibody to b-endorphin, a markerfor POMC-containing neurons (diluted 1 : 2000, a kind gift from DrD.T. Krieger; Tsong et al., 1982); (ii) guinea-pig anti-VGluT2polyclonal antibody (diluted 1 : 3000, Chemicon International) plusrabbit anti-NPY polyclonal antibody (diluted 1 : 2000) or rabbit anti-b-endorphin antibody (diluted 1 : 2000). The cocktails containingprimary antibodies were all prepared in PB containing 2% NGS, 1%BSA and 0.05% sodium azide. Incubations for the primary antibodieswere processed for 48 h at 4 �C. After washing in PB, the sectionswere incubated with biotinylated goat anti-guinea-pig IgG (F(ab¢)2fragment specific), as second antibody to the VGluT1 and VGluT2antibodies (diluted 1 : 500 in PB, Jackson ImmunoResearch, WestGrove, PA, USA) for 3 h at room temperature, then after washing thesections were incubated in a cocktail containing Alexa 488-labelledgoat anti-rabbit IgG antibody (diluted 1 : 100, Molecular Probes,Eugene, OR, USA) plus Streptavidin Alexa FluorR 594 (diluted1 : 200 in PB, Molecular Probes) in the dark, overnight, at 4 �C.Subsequently, the immunofluorescence-labelled sections were rinsedin PB, mounted onto glass slides and coverslipped with AquaPoly ⁄Mount (Polysciences, Warrington, PA, USA). Sections wereanalysed by confocal laser scanning microscopy using a Bio-RadRadiance2100 Rainbow system equipped with Ar 488-nm and HeNe543-nm lasers (Bio-Rad Laboratories, Hercules, CA, USA). Digitalimages were collected from a single optical plane using a 40 · PlanFluor oil-immersion lens (numerical aperture 1.3). Pinhole setting was1 Airy unit for all images. For each optical section, double-fluorescence images were acquired in sequential mode to avoidpotential contamination by linkage-specific fluorescence ‘cross-talk’.To build and label the composite illustrations, Adobe Photoshop

(version 7.0; Adobe Systems, San Jose, CA, USA) was used.

Single-immunogold labelling of VGluT1 and VGluT2 for electronmicroscopy

Four rats were anaesthetized with chloral hydrate and transcardiallyperfused first with 50 mL saline, then with 400 mL of fixativecontaining 4% paraformaldehyde (Merck), 0.05% glutaraldehyde(TAAB Laboratories) and 10% (v ⁄ v) saturated picric acid in PB,followed by 200 mL of the same fixative, but without glutaraldehyde.After perfusion fixation, brains were removed and a coronal segment

Fig. 1. Comparative distribution of vesicular glutamate transporter 1 (VGluT1; A and C, D and F; red) or vesicular glutamate transporter 2 (VGluT2; G and I, J andL, M and O; red) and neuropeptide Y (NPY; B and C, H and I; green) or b-endorphin (E and F, K and L, N and O; green) in the arcuate nucleus of colchicine-treatedrats (A–L) and of animals not receiving colchicine (M–O) as visualized in single optical sections by confocal microscopy. (A–C) In the ventromedial part of thearcuate nucleus, no close contact is evident between the scant VGluT1-immunoreactive puncta and the numerous NPY-immunoreactive cell bodies. (D–F) In theventrolateral arcuate nucleus, only few close appositions are formed by the scant VGluT1-immunoreactive fibres on b-endorphin-immunoreactive cells. (G–I)VGluT2-immunoreactive puncta are densely distributed in the ventromedial arcuate nucleus, where they surround and closely appose NPY-immunoreactive cells(arrowheads). (J–L) In the ventrolateral arcuate nucleus, VGluT2 immunoreactivity is present in some b-endorphin-immunoreactive neurons (arrows). VGluT2-immunoreactive fibres are densely distributed, surrounding and forming close contacts with b-endorphin-immunoreactive neurons (arrowheads), either VGluT2-immunopositive or not. (M–O) Punctate VGluT2 immunoreactivity is present in some of the b-endorphin-immunolabelled varicose fibres (arrows) in theventrolateral arcuate nucleus. Note that in a b-endorphin-positive fibre, few varicosities display VGluT2 immunoreactivity. Scale bar, 20 lm.

VGluT2 terminals and NPY and POMC neurons 2113


Fig. 2. (A–D) Electron micrographs showing vesicular glutamate transporter 2 (VGluT2) immunogold labelling in axon terminals synapsing with unidentifieddendritic portions localized in various areas of the arcuate nucleus. The synaptic connections are all formed by the asymmetric type of synapse. The silver–gold-labelled bouton (T) shown in (A) forms synaptic connection (arrow) with a distal dendritic portion of medium calibre (D). Synaptic connections (arrows) formed byVGluT2-positive axon terminal (T) with small calibre terminal dendrite branch (D) are demonstrated in (B). The silver–gold grains cover the axoplasmic areacontaining accumulation of numerous round clear synaptic vesicles. The large-sized labelled bouton (T) demonstrated in (C) forms two separate synapses with thesame very thin calibre terminal dendrite branchlet (D, arrows). (D) Synaptic connections (arrow) made by VGluT2-immunopositive axon terminal (T) on the headof dendritic spine (Sp). (E) VGluT2 immunogold labelling in the perikaryon (Pk) of an unidentified neuron located in the ventromedial area of the arcuate nucleus.Silver–gold granules (arrowheads), the end-products of the immunogold reaction for detecting the appearance of VGluT2 on ultrathin sections, are accumulatedselectively in the cytoplasm of the cell body, covering the cytoplasmic areas containing cisternae of the endoplasmic reticulum (ER). No silver granules areaccumulated on the cytoplasmic area involving the organelles of the Golgi-complex (Go). No silver–gold granules can be seen on the cell nucleus. (F) A cross-section profile of an unidentified large calibre proximal dendrite (D), which is immunopositive for VGluT2. Silver–gold granules (some are marked by arrowheads)fill up the entire areas containing endoplasmic reticulum cisternae (ERc). Scale bars, 0.25 lm (A); 0.4 lm (B–D); 0.6 lm (E and F).

2114 J. Kiss et al.


Fig. 3. (A and B) Synaptic contacts made by vesicular glutamate transporter 2 (VGluT2) immunogold-labelled axon terminal with DAB-stained target elementsimmunoreactive for neuropeptide Y (NPY) or for b-endorphin (one of the markers of POMC-containing neurons) are demonstrated. (A) VGluT2-silver–gold-labelled bouton (T) making asymmetric synapse (arrows) with an NPY-containing immunoperoxidase-positive DAB-stained dendrite (D). (B) An asymmetric typeof synaptic contact established by VGluT2-immunogold positive bouton (T) with b-endorphin-immunoreactive, DAB-stained distal dendrite (D). (C) The co-localization of the presence of VGluT2-immunolabelling in the perikaryon (Pk) of a b-endorphin-immunopositive neuron. Arrowheads point to some VGluT2 silver–gold-labelled granules. Note that the cell nucleus (N) is free of silver–gold granules. (D) Double-labelled b-endorphin-containing axon terminal is demonstrated.VGluT2 silver–gold-labelled granules (arrowheads) are accumulated in a bouton immunopositive for b-endorphin (T). Note that in the DAB-stained bouton thesilver–gold granules are associated only with synaptic vesicles (thin arrows), while the axoplasm of the axon terminal is free from these granules. Silver–goldgranules preferentially cover axoplasmic areas including high density of synaptic vesicles. (E) A b-endorphin-immunopositive axon terminal (T) lacking VGluT2silver–gold labelling. Thin arrows point to synaptic vesicles DAB-stained for b-endorphin, an arrowhead indicates a silver–gold granule in a different axon terminal(T) not showing b-endorphin immunoreactivity. Scale bars, 0.25 lm (A); 0.30 lm (B); 0.8 lm (C); 0.35 lm (D and E).



containing the entire rostrocaudal extent of the arcuate nucleus regionwas cut out and postfixed in fresh glutaraldehyde-free fixative at 5 �Covernight. Serial vibratome sections (50 lm thick) in the frontal planewere cut and immersed for 20 min in 1% sodium borohydride in PB.After rigorous washing, the sections for cryoprotection were placed inPB containing 30% sucrose until sinking, then they were freeze-thawedby using liquid nitrogen to enhance the penetration of antibodies.For electron microscopic visualization of VGluT1 or VGluT2

immunoreactivity, silver intensification of pre-embedding immunogoldlabelling was used as described in detail earlier (Kiss et al., 2003).

Double-immunolabelling of VGluT1 and NPY or b-endorphin,and of VGluT2 and NPY or b-endorphin for electron microscopy

Four rats were used for electron microscopic double-immunolabelexperiments using immunoperoxidase reaction for NPY or b-endor-phin neurons and pre-embedding immunogold labelling for VGluT1-and VGluT2-containing axon terminals on the same sections forexamining the synaptic contacts between NPYor b-endorphin neuronsand glutamatergic axon terminals in the arcuate nucleus.Each animal in this experimental group, anaesthetized with

Equithesin and immobilized in a stereotaxic head frame, waspretreated with an intraventricular injection of colchicine. Thirty-sixto 48 h after the colchicine treatment, the animals were re-anaesthet-ized with chloral hydrate and then perfused through the heart first withTyrode’s solution, followed by the fixatives described above. Afterperfusion, the brains were left in situ for 15–20 min, then they wereremoved from the skull. Blocks from the entire rostrocaudal extent ofthe arcuate nucleus were dissected and washed in PB, followed bysectioning on a vibratome at 40 lm thickness.One series of the sections was incubated for 48–72 h at 4 �C in a

cocktail containing primary antibodies specific to VGluT1 (dil.1 : 6000) or VGluT2 (dil. 1 : 3000) and NPY (dil. 1 : 1000), and thesections in the other of the two series were incubated in a cocktailcontaining primary antibodies specific to VGluT1 or VGluT2 andb-endorphin (dil. 1 : 2500). This was followed by incubation in acocktail of the second antibodies for VGluT1 or VGluT2 and NPY, orVGluT1 or VGluT2 and b-endorphin, then silver-intensified withIntenSETMM reagent kit (Amersham, Biosciences, Buckinghamshire,UK), as described (Kiss et al., 2003). The sections were then incubatedin ABC-Elite (Vector Laboratories, Burlingame, CA, USA; 1 : 1000),then the tissue-bound peroxidase was visualized by a diaminobenzidine(DAB) reaction (15 mg DAB, 12 mg NH4Cl, 0.12 mg glucose oxidaseand 60 mg b-d-glucose in 30 mL PB, for 5–20 min). Sections forelectron microscopy were further processed in the same way asdescribed (Kiss et al., 2003), and examined under a transmissionelectron microscope (H-7600, Hitachi, Tokyo, Japan).

Specificity controls

Although the primary antisera (anti-VGluT1, VGluT2, NPY,b-endorphin) used in these experiments are well characterized,controls were performed to examine the specificity of the immuno-reagents under our laboratory conditions. Controls for double-labelling immunofluorescence included omission of the primaryantibodies to test the non-specific binding of the secondary antibodiesand incubation with one primary, but both secondary, antibodies todemonstrate the absence of cross-labelling. For immunogold-detectionspecificity of VGluT1 or VGluT2 antigens, or for double-immuno-cytochemical labelling of VGluT1 or 2 and NPY or b-endorphin, lightmicroscopic control was performed on 50-lm-thick vibratomesections flat-embedded into Durcupan. To control the immunogold

labelling of the vesicular glutamate transporters, the primary antibod-ies were omitted or replaced with normal immunoglobulin in theprocedure. In controls for double-label experiments the antibodies toNPY or to b-endorphin were omitted or replaced with normal serum.In these cases no immunogold labelling to VGluT1 or VGluT2, or noDAB-staining of NPY or b-endorphin cells could be detected on theDurcupan-embedded sections. In addition to guinea-pig VGluT2polyclonal antibody, rabbit anti-VGluT2 polyclonal antibody (Synap-tic Systems GmbH, Gottingen, Germany) was also used for specificitycontrol of our guinea-pig anti-VGluT2 antibody. Rabbit and guinea-pig anti-VGluT2 antisera have given similar results.

Results

Light microscopic observations

On sections double-immunolabelled for VGluT1 ⁄NPY and forVGluT1 ⁄ b-endorphin, an extremely low density of VGluT1-immu-noreactive rod-like fibres and punctate appearance of nerve terminalswas detected (Fig. 1A and D). Very rarely immunolabelled VGluT1elements could be seen in close apposition with NPY- or b-endorphin-positive arcuate neurons in colchicine-treated rats (Fig. 1C and F).In contrast, on sections double-immunolabelled for VGluT2 and

NPY or b-endorphin, a dense network of VGluT2-immunoreactivefibres and terminals was observed in the arcuate nucleus (Fig. 1G, Jand M), both in its ventromedial and ventrolateral division. Confocalmicroscopic examination of single optical sections revealed thatVGluT2-immunoreactive terminals surrounded and apposed NPY-immunolabelled cells (Fig. 1H and I) localized mainly in theventromedial division of the nucleus and b-endorphin-immunoreactiveneurons (Fig. 1K and L) localized mainly in the ventral andventrolateral part of the nucleus of colchicine-treated rats. Thelabelled terminals seemed to make close contacts with immunola-belled cell bodies and dendrites (Fig. 1I and L), suggesting thatVGluT2 (but not VGluT1)-containing fibres presumably form synapticcontacts with NPY- and b-endorphin-immunoreactive neurons in thearcuate nucleus.VGluT1-immunolabelled cell bodies were not detected in the

arcuate nucleus. By contrast, in colchicine-treated rats, numerousVGluT2-immunoreactive perikarya were present in the nucleus, themajority of them located in the ventrolateral part and only few werefound in the ventromedial part. NPY-immunoreactive neurons werenever found to be immunopositive for VGluT2 in the ventromedialarcuate nucleus (Fig. 1G–I). By contrast, a subpopulation of b-endor-phin-immunoreactive neurons displayed VGluT2 immunoreactivity inthe ventrolateral part of the arcuate nucleus (Fig. 1J–L); however, notall VGluT2-immunoreactive cells were b-endorphin-positive in thearea. A network of b-endorphin-immunoreactive varicose fibres wasalso observed embedded in the dense VGluT2-immunoreactive punctain the arcuate nucleus of rats not treated with colchicine. Confocalmicroscopic examination of single optical sections revealed that someof the b-endorphin-immunoreactive varicosities were also labelledwith VGluT2 immunoreactivity (Fig. 1M–O).

Electron microscopic observations

Lack of VGluT1 immunogold-labelled axon terminalsand VGluT1-positive neurons

Approximately 200 sections on 40 grids from blocks of fourexperimental animals were examined under the electron microscopeto reveal the occurrence and localization of silver–gold-labelled

2116 J. Kiss et al.


VGluT1 neuronal elements. About half of the sections were single-labelled only for VGluT1 and the others were dual-labelled forVGluT1 plus NPY and for VGluT1 plus b-endorphin. Very fewVGluT1 immunogold-labelled axons were detected, pre-terminalaxonal profiles or terminals were not seen, presumably due to thepresence of extremely low density of VGluT1-immunoreactiveelements (Fig. 1A and D). Cell bodies or dendritic parts showingimmunogold labelling for VGluT1 were not detected.

Synaptic contacts of VGluT2 immunogold-labelled axons.VGluT2-immunoreactive neurons

Approximately 200 ultrathin sections on 40 grids from four experi-mental animals were analysed for studying the appearance of VGluT2-containing axons and terminals localized in different divisions of thenucleus. In all sections cut from the three divisions examined, VGluT2immunogold-labelled axon terminals and pre-terminal axonal profileswere found. Labelled terminals had usually round or oval shape, andvery rarely they seemed to be axonal varicosity with irregular ellipticalshape. They were enriched in small clear round synaptic vesicles andoften contained one or more mitochondria. The silver-intensifiedimmunogold particles (black rounded granules of various size)accumulated over clusters of synaptic vesicles in axon terminals andless frequently in pre-terminal axonal profiles. Numerous VGluT2immunogold-labelled boutons made synaptic contacts with unlabelleddendritic shaft and spine-like profiles (Fig. 2A–D) forming exclusivelythe asymmetric type of synapses.

To estimate the relative occurrence of axon terminals silver–gold-labelled for VGluT2 and forming synaptic contacts with neuronalprofiles, they were counted on 100 sections selected randomly fromfour experimental animals. Among 1050 gold-labelled boutonscounted, 607 were in synaptic contact with unidentified neuronaltarget element. The labelled terminals clearly formed the asymmetrictype of synapse. Investigating approximately 560 synapticallycontacted immunoreactive boutons, no labelled terminals containingpleomorphic vesicles (which characterize GABAergic boutons) and nosynapses of the symmetric type established by the labelled terminalswere found. Of the 560 silver–gold-labelled axon terminals, 319formed synaptic contact with dendritic shaft, 124 with terminal thinbranch of a dendrite, 94 with dendritic spine and 22 with the origin ofa proximal part of a dendrite or with perikaryon.

VGluT2-immunoreactive cells were seen to be distributed randomlyin the entire extent of the nucleus on single VGluT2-immunolabelledultrathin sections. The number of cells containing silver–gold particlesin their perikaryon was relatively small, but the labelling in thecytoplasm was clearly specific. The immunogold labelling wasrevealed in areas of perikarya (Fig. 2E) and medium to large dendriteprofiles (Fig. 2F).

VGluT2 synapses onto NPY- or b-endorphin-immunoreactive neurons

In sections double-immunolabelled for VGluT2 and NPY orb-endorphin, a relatively high number of VGluT2 terminals madeasymmetric type of synapses with various neuronal compartmentsimmunostained for NPYor b-endorphin (Fig. 3A and B). The majorityof these synapses were in contact with dendritic shafts. Thinnerdendrite branches and spine-like processes could also be observed tobe synaptically contacted by VGluT2-labelled terminals. Silver–gold-labelled axon terminals were also found to form asymmetric type ofsynaptic contact with b-endorphin-containing DAB-stained perikar-yon (not shown). Boutons that established synaptic contact andboutons among these that synapsed with NPY- or b-endorphin-immunostained neuronal elements were counted. Among 2250

VGluT2-labelled boutons (the total number of labelled boutonsobserved in 100 sections double-labelled with NPY) counted, 1175established synaptic contact with neuronal target structure. Onehundred and eighty of the 1175 labelled boutons formed asymmetricalsynapse on NPY-immunostained neuronal elements. Of the 1675VGluT2-labelled boutons counted in 100 sections containingb-endorphin immunolabelling, 1135 formed synapse with neuronalelements. Four hundred and seventy-five of the 1135 terminalsestablished asymmetric synapses with b-endorphin-immunostainedelements. No symmetric type of synapse was found to be formedby VGluT2-immunolabelled axon terminal either with NPY- orb-endorphin-containing neurons.

Colocalization of VGluT2 in b-endorphin-containing neurons

In some neurons the coexpression of VGluT2 protein with immuno-reactive b-endorphin in cell body (Fig. 3C) and ⁄ or dendrite profilewas revealed. Moreover, similarly to the light microscopic immuno-fluorescent observation (Fig. 1M–O), electron microscopic examina-tion also showed immunoreactive axon terminals in ultrathin sectionsdouble-labelled for VGluT2 and b-endorphin immunoreactivities(Fig. 3D). Colocalization of VGluT2 and NPY in neuronal cell bodiesand axons was not observed.

Discussion

The present electron microscopic observations demonstrate thatglutamatergic terminals form asymmetric type of synapses on NPYand POMC-peptide-containing neurons of the arcuate nucleus. Thisfinding provides the first direct neuromorphological evidence for theexistence of glutamatergic innervation of these nerve cells. Accordingto our rough estimate, about 15% of VGluT2 asymmetrical synapseswere on NPY-immunostained neuronal elements and more than 40%on b-endorphin-positive elements. This suggests that glutamatergicinnervation of both neuron types of the arcuate nucleus is presumablysignificant. We detected also at the light and electron microscopic levelVGluT2-immunoreactive cell bodies in the arcuate nucleus. This latterfinding is in line with the observations of Collin et al. (2003) and Linet al. (2003), and indicates that besides the already known largenumber of neurochemically different cell populations (Everitt et al.,1986; Meister et al., 1989), there are also glutamatergic neurons in thisnucleus.Concerning the origin of the VGluT2-immunoreactive fibres

synapsing on NPY and POMC neurons in the arcuate nucleus, someof them may arise from glutamatergic neurons within the nucleus.Another source of such glutamatergic fibres could be the orexin(hypocretin)-containing nerve cells implicated in feeding (Kalra et al.,1999) and residing primarily in the lateral hypothalamus (Nambuet al., 1999). About half of orexin-immunoreactive neurons containVGluT2 mRNA (Rosin et al., 2003), and orexin terminals contact anddirectly interact with NPY and POMC neurons in the arcuate nucleus(Muroya et al., 2004). Glutamatergic neurons projecting to the NPYand POMC neurons of the arcuate nucleus may also be situated invarious other regions of the brain. Further studies are needed to mapthe location of glutamatergic neurons terminating in this nucleus.Our findings indicate that glutamate acts directly on NPY- and

POMC-derived peptide-containing nerve cells. This assumption isconsistent with the observations that endogenous glutamate stimulatesrelease of a-MSH (Wayman & Wilson, 1992) and that microinjectionof N-methyl-d-aspartate (NMDA) into the arcuate nucleus releasedb-endorphin immunoreactivity into perfusate (lateral ventricle-cisternamagna perfusion) and the release was blocked by systemic pretreat-



ment with the NMDA antagonist dizocilpine (MK-801; Bach &Yaksh, 1995). Glutamate presumably stimulates NPY and POMCneurons involved in the control of food intake and energy homeostasis(Kalra et al., 1999). It has to be mentioned that the glutamate receptoragonist, NMDA stimulated immediate and transient feeding lastingfor about 10 min only when microinjected into the lateral hypotha-lamus (Stanley et al., 1993a,b). Glutamate may also stimulate NPYneurons participating in the regulation of pituitary tropic hormonesecretion (Kalra & Crowley, 1992) or in the control of sexualbehaviour (Clark et al., 1984) and ⁄ or in thermogenesis (Billingtonet al., 1994). The situation is complicated by the fact that, forexample, NPY neurons of the arcuate nucleus project to manydifferent hypothalamic cell groups and various extrahypothalamicstructures (Broberger et al., 1998b), and that these neurons areinvolved in the control of various functions. NPY neurons of thearcuate nucleus terminate among others on corticotropin-releasinghormone neurons (Liposits et al., 1988), thyrotropin-releasinghormone neurons in the paraventricular nucleus (Toni et al., 1990;Legradi & Lechan, 1998), gonadotropin-releasing hormone neurons inthe medial preoptic area (Tsuruo et al., 1990; Turi et al., 2003) and onlateral hypothalamic orexin-positive cells (Broberger et al., 1998a).POMC neurons project to the medial preoptic area (Horvath et al.,1992a), to gonadotropin-releasing hormone neurons (Leranth et al.,1988) and to histamine neurons of the tuberomamillary nucleus(Fekete & Liposits, 2003).It should also be mentioned that the control of the activity of NPY

and POMC-peptide-containing neurons in the arcuate nucleus isextremely complex. Both cells are innervated by GABAergic fibres(Horvath et al., 1992a; Backberg et al., 2004) and orexin-containingneurons (Horvath et al., 1999; Muroya et al., 2004). b-endorphin-immunoreactive nerve cells also receive catecholaminergic- (Horvathet al., 1992a), enkephalin- (Zhang et al., 1987) and oxytocin-containing nerve fibres (Csiffary et al., 1992), and NPY neurons aresynaptically connected to b-endorphin-immunoreactive cells (Horvathet al., 1992b). Arginine vasopressin, corticotropin-releasing hormoneand serotonin release hypothalamic b-endorphin (Bornstein & Akil,1990). There are corticotropin-releasing hormone receptor 1 typereceptors in the NPY neurons of the arcuate nucleus (Campbell et al.,2003). In addition, it is well established that both leptin (producedmainly by adipocytes) and ghrelin (product of the stomach), act onNPY and POMC neurons. Leptin inhibits NPY and stimulates a-MSHneurons, and ghrelin exerts opposite effects on these cells (Elmquistet al., 1999; Kalra et al., 1999; Kalra & Kalra, 2003). Thus, it appearsthat the glutamatergic innervation of the NPY and POMC neuronsdemonstrated by us means a further member of the system controllingthe activity of these neurons.Taking into account that the NPY and POMC neurons of the

arcuate nucleus receive a vast amount of neurochemically differentafferents, project to several different structures and are involved inthe control of various functions, the functional significance of theglutamatergic innervation of NPY and POMC neurons needs furtherelucidation.Our findings confirm previous observations of Collin et al. (2003)

showing that glutamate is colocalized with b-endorphin in asubpopulation of b-endorphin-positive neurons. In addition, wedemonstrated that there are nerve terminals in the arcuate nucleuscontaining both b-endorphin and VGluT2, indicating that suchelements terminate in the cell group. It has been reported (Hentgeset al., 2004) that in situ RNA hybridization for POMC and glutamicacid decarboxylase, the GABA synthetic enzyme, revealed colocal-ization of these two mRNAs in approximately one-third of POMCneurons in the arcuate nucleus. The results of the mentioned authors

suggest that these neurons project to extrahypothalamic structures. Itcould be assumed that those POMC neurons, which contain alsoglutamate, terminate, at least partly, within the arcuate nucleus andmay belong to the intrinsic circuitry of the cell group, while thosePOMC neurons in which GABA is present may project primarily todistant areas.

Acknowledgements

This research was supported by the Hungarian National Research Fund (OTKAT-034481 to J.K., T-042516 to B.H.), and the Hungarian Academy of Sciencesand the Ministry of Health (ETT 054 ⁄ 2003 to B.H.). Z.C. is a recipient ofBolyai Fellowship.

Abbreviations

a-MSH, a-melanocyte-stimulating hormone; BSA, bovine serum albumin;DAB, diaminobenzidine; GABA, c-aminobutyric acid; NGS, normal goatserum; NMDA, N-methyl-d-aspartate; NPY, neuropeptide Y; PB, phosphatebuffer; POMC, pro-opiomelanocortin; VGluT1, vesicular glutamate transporter1; VGluT2, vesicular glutamate transporter 2.

References

Bach, F.W. & Yaksh, T.L. (1995) Release of b-endorphin immunoreactivityfrom brain by activation of a hypothalamic N-methyl-D-aspartate receptor.Neuroscience, 65, 775–783.

Backberg, M., Ultenius, C., Fritschyt, J.-M. & Meister, B. (2004) Cellularlocalization of GABAA receptor a subunit immunoreactivity in the rathypothalamus: relationship with neurons containing orexigenic or anorexi-genic peptides. J. Neuroendocrinol., 16, 589–604.

Bellocchio, E.E., Reimer, R.J., Fremeau, R.T. Jr & Edwards, R.H. (2000)Uptake of glutamate into synaptic vesicles by an inorganic phosphatetransporter. Science, 289, 957–960.

Billington, C.J., Briggs, J.E., Harker, S., Grace, M. & Levine, A.S. (1994)Neuropeptide Y in hypothalamic paraventricular nucleus: a center coordinat-ing energy metabolism. Am. J. Physiol., 266, R1765–R1770.

Bornstein, D.M. & Akil, H. (1990) In vitro release of hypothalamicb-endorphin (bE) by arginine vasopressin, corticotropin-releasing hormoneand 5-hydroxytryptamine: evidence for release of opioid active and inactivebE forms. Neuropeptides, 16, 33–40.

Broberger, C., De Sutcliff, J.G., Lecea, L. & Hokfelt, T. (1998a)Hypocretin ⁄ orexin- and melanin-concentrating hormone-expressing cellsfrom distinct populations in the rodent lateral hypothalamus: relationship tothe neuropeptide Y and agouti gene-related protein systems. J. Comp.Neurol., 402, 460–474.

Broberger, C., Johansen, J., Johansson, C., Schalling, M. & Hokfelt, T. (1998b)The neuropeptide Y ⁄agouti gene-related protein (AGRP) brain circuitry innormal, anorectic, and monosodium glutamate-treated mice. Proc. Natl.Acad. Sci. USA (Neurobiology), 95, 15043–15048.

Campbell, R.E., Grove, K.L. & Smith, M.S. (2003) Distribution ofcorticotropin-releasing hormone receptor immunoreactivity in the rathypothalamus: coexpression in neuropeptide Y and dopamine neurons inthe arcuate nucleus. Brain Res., 973, 223–232.

Clark, J.T., Kalra, P.S., Crowley, W.R. & Kalra, S.P. (1984) Neuropeptide Yandhuman pancreatic polypeptide stimulate feeding behavior in rats. Endocri-nology, 115, 427–429.

Collin, M., Backberg, M., Ovesjo, M.-L., Fisone, G., Edwards, R.H., Fujiyama,F. & Meister, B. (2003) Plasma membrane and vesicular glutamatetransporter mRNAs ⁄ proteins in hypothalamic neurons that regulate bodyweight. Eur. J. Neurosci., 18, 1265–1278.

Csiffary, A., Ruttner, Z., Toth, Zs. & Palkovits, M. (1992) Oxytocin nervefibers innervate b-endorphin neurons in the arcuate nucleus of the rathypothalamus. Neuroendocrinology, 56, 429–435.

Elmquist, J.K., Elias, C.F. & Saper, C.B. (1999) From lesions to leptin:hypothalamic control of food intake and body weight. Neuron, 22, 221–232.

Everitt, B.J., Meister, B., Hokfelt, T., Melander, T., Terenius, L., Rokaeus, A.,Theodorsson-Norheim, E., Dockray, G., Edwardson, J., Cuello, C., Elde, R.,Goldstein, M., Hemmings, H., Ouimet, C., Walaas, I., Greengard, P., Vale,

2118 J. Kiss et al.


W., Weber, E., Wu, J.-Y. & Chang, K.-J. (1986) The hypothalamic arcuatenucleus median eminence complex: immunohistochemistry of transmitters,peptides and DARPP-32 with special reference to coexistence in dopamineneurons. Brain Res. Rev., 11, 97–155.

Eyigor, O., Centers, A. & Jennes, L. (2001) Distribution of ionotropicglutamate receptor subunit mRNAs in the rat hypothalamus. J. Comp.Neurol., 434, 101–124.

Fekete, C. & Liposits, Z. (2003) Histamine-immunoreactive neurons of thetuberomammillary nucleus are innervated by a-melanocyte stimulatinghormone-containing axons. Generation of a new histamine antiserum forultrastructural studies. Brain Res., 969, 70–77.

Fremeau, R.T. Jr, Troyer, M.D., Pahner, I., Nygaard, G.O., Tran, C.H., Reimer,R.J., Bellocchio, E.E., Fortin, D., Storm-Mathisen, J. & Edwards, R.H.(2001) The expression of vesicular glutamate transporters defines two classesof excitatory synapse. Neuron, 31, 247–260.

Ghosh, P.K., Baskaran, N. & van den Pol, A.N. (1997) Developmentallyregulated gene expression of all eight metabotropic glutamate receptors inhypothalamic suprachiasmatic and arcuate nuclei – a PCR analysis. Dev.Brain Res., 102, 1–12.

Gras, C., Herzog, E., Bellenchi, G.C., Bernard, V., Ravassard, P., Pohl, M.,Gasnier, B., Giros, B. & El Mestikawy, S. (2002) A third vesicular glutamatetransporter expressed by cholinergic and serotoninergic neurons.J. Neurosci., 22, 5442–5451.

Hentges, S.T., Nishiyama, M., Overstreet, L.S., Stenzel-Poore, M., Williams,J.T. & Low, M.J. (2004) GABA release from proopiomelanocortin neurons.J. Neurosci., 24, 1578–1583.

Horvath, T.L., Diano, S. & van den Pol, A.N. (1999) Synaptic interactionbetween hypocretin (orexin) and neuropeptide Y cells in the rodent andprimate hypothalamus: a novel circuit implicated in metabolic and endocrineregulations. J. Neurosci., 19, 1072–1087.

Horvath, T.L., Naftolin, F., Kalra, S.P. & Leranth, C. (1992b) Neuropeptide Yinnervation of b-endorphin-containing cells in the rat mediobasal hypotha-lamus. A light- and electron-microscopic double-immunostaining analysis.Endocrinology, 131, 2461–2467.

Horvath, T.L., Naftolin, F. & Leranth, C. (1992a) GABAergic andcatecholaminergic innervation of mediobasal hypothalamic b-endorphincells projecting to the medial preoptic area. Neuroscience, 51, 391–399.

Kalra, S.P. & Crowley, W.R. (1992) Neuropeptide Y: a novel neuroendocrinepeptide in the control of pituitary hormone secretion, and its relation toluteinizing hormone. Front. Neuroendocrinol., 13, 1–36.

Kalra, S.P., Dube, M.G., Pu, S., Xu, B., Horvath, T.L. & Kalra, P.S. (1999)Interacting appetite-regulating pathways in the hypothalamic regulation ofbody weight. Endocrine Rev., 20, 68–100.

Kalra, S.P. & Kalra, P.S. (2003) Neuropeptide Y. A physiological orexigenmodulated by the feedback action of ghrelin and leptin. Endocrine, 22,49–55.

Kiss, J., Kocsis, K., Csaki, A., Gorcs, T.J. & Halasz, B. (1997) Metabotropicglutamate receptor in GHRH and b-endorphin neurones of the hypothalamicarcuate nucleus. Neuroreport, 8, 3703–3707.

Kiss, J., Kocsis, K., Csaki, A. & Halasz, B. (2003) Evidence for vesicularglutamate transporter synapses onto gonadotropin-releasing hormone andother neurons in the rat medial preoptic area. Eur. J. Neurosci., 18, 3267–3287.

Legradi, G. & Lechan, R.M. (1998) The arcuate nucleus is the major source forneuropeptide Y-innervation of thyrotropin-releasing hormone neurons in thehypothalamic paraventricular nucleus. Endocrinology, 139, 3262–3270.

Leranth, C., MacLusky, N.J., Shanabrough, M. & Naftolin, F. (1988)Immunohistochemical evidence for synaptic connection between pro-opiomelanocortin-immunoreactive axons and LH-RH neurons in the preopticarea of the rat. Brain Res., 449, 167–176.

Lin, W., McKinney, K., Liu, L., Lakhlani, S. & Jennes, L. (2003)Distribution of vesicular glutamate transporter-2 messenger ribonucleicacid and protein in the septum-hypothalamus of the rat. Endocrinology,144, 662–670.

Liposits, Z., Sievers, L. & Paull, W.K. (1988) Neuropeptide-Y and ACTH-immunoreactive innervation of corticotropin releasing factor (CRF)-synthe-sizing neurons in the hypothalamus of the rat. An immunocytochemicalanalysis at the light and electron microscopic levels. Histochemistry, 88,227–234.

Meister, B., Ceccatelli, S., Hokfelt, T., Anden, N.-E., Anden, M. &Theodorsson, E. (1989) Neurotransmitters, neuropeptides and binding sitesin the rat mediobasal hypothalamus: effects of monosodium glutamate(MSG) lesions. Exp. Brain Res., 76, 343–368.

Muroya, S., Funahashi, H., Yamanaka, A., Kohno, D., Uramura, K., Nambu,T., Shibahara, M., Kuramochi, M., Takigawa, M., Yanagisawa, M.,Sakurai, T., Shioda, S. & Yada, T. (2004) Orexins (hypocretins) directlyinteract with neuropeptide Y, POMC and glucose-responsive neurons toregulate Ca2+ signaling in a reciprocal manner to leptin: orexigenicneuronal pathways in the mediobasal hypothalamus. Eur. J. Neurosci., 19,1524–1534.

Nambu, T., Sakurai, T., Mizukami, K., Hosoya, Y., Yanagisawa, M. & Goto, K.(1999) Distribution of orexin neurons in the adult rat brain. Brain Res., 827,243–260.

Petralia, R.S., Wang, Y.X. & Wenthold, R.J. (1994a) Histological andultrastructural localization of the kainate receptor subunits KA2 andGluR6 ⁄ 7 in the rat nervous system using selective antipeptide antibodies.J. Comp. Neurol., 349, 85–110.

Petralia, R.S., Wang, Y.X. & Wenthold, R.J. (1994b) The NMDA receptorsubunits NR2A and NR2B show histological and ultrastructural localizationpatterns similar to those of NR1. J. Neurosci., 14, 6102–6120.

Petralia, R.S. & Wenthold, R.J. (1992) Light and electron immunocytochemicallocalization of AMPA-selective glutamate receptors in the rat brain. J. Comp.Neurol., 318, 329–354.

Petralia, R.S., Yokotani, N. & Wenthold, R.J. (1994c) Light and electronmicroscope distribution of the NMDA receptor subunit NMDAR1 in the ratnervous system using a selective anti-peptide antibody. J. Neurosci., 14,667–696.

van den Pol, A.N. (1991) Glutamate and aspartate immunoreactivity inhypothalamic presynaptic axons. J. Neurosci., 11, 2087–2101.

van den Pol, A.N., Hermans-Borgmeyer, I., Hofer, M., Ghosh, P. &Heinemann, S. (1994) Ionotropic glutamate-receptor gene expression inhypothalamus: localization of AMPA, kainate and NMDA receptor RNAwith in situ hybridization. J. Comp. Neurol., 338, 377–390.

van den Pol, A.N. & Trombley, P.Q. (1993) Glutamate neurons inhypothalamus regulate excitatory transmission. J. Neurosci., 13, 2829–2836.

van den Pol, A.N., Wuarin, J.-P. & Dudek, F.E. (1990) Glutamate, the dominantexcitatory transmitter in neuroendocrine regulation. Science, 250, 1276–1278.

Rosin, D.L., Weston, M.C., Sevigny, C.P., Stornetta, R.L. & Guyenet, P.G.(2003) Hypothalamic orexin (hypocretin) neurons express vesicularglutamate transporters VGluT1 or VGluT2. J. Comp. Neurol., 465, 593–603.

Stanley, B.G., Ha, L.H., Spears, L.C. & Dee II, M.G. (1993a) Lateralhypothalamic injections of glutamate, kainic acid, D.L-a-amino-3-hydroxy-5-methyl-isoxazole propionic acid or N-methyl-D-aspartic acid rapidly elicitintense transient eating in rats. Brain Res., 613, 88–95.

Stanley, B.G., Willett III, V.L., Donias, H.W., Ha, L.H. & Spears, L.C. (1993b)The lateral hypothalamus: a primary site mediating excitatory amino acidelicited eating. Brain Res., 630, 41–49.

Storm-Mathisen, J., Leknes, A.K., Bore, A.T., Waaland, J.L., Edminson, P.,Haug, F.M. & Ottersen, O.P. (1983) First visualization of glutamate andGABA in neurones by immunocytochemistry. Nature, 301, 517–520.

Takamori, S., Rhee, J.S., Rosenmund, C. & Jahn, R. (2000) Identification of avesicular glutamate transporter that defines a glutamatergic phenotype inneurons. Nature, 407, 189–194.

Toni, R., Jackson, I.M.D. & Lechan, R.M. (1990) Neuropeptide Y-immuno-reactive innervations of thyrotropin-releasing hormone-synthesizing neuronsin the rat hypothalamic paraventricular nucleus. Endocrinology, 118, 2444–2453.

Tsong, S.D., Phillips, D., Halmi, N., Krieger, D.T. & Bardin, C.W. (1982)b-Endorphin is present in the male reproductive tract of five species. Biol.Reprod., 27, 755–764.

Tsuruo, Y., Kawano, H., Kagotani, Y., Hisano, S., Daikoku, S., Chihara, K.,Zhang, T. & Yanaihara, N. (1990) Morphological evidence for neuronalregulation of luteinizing hormone-releasing hormone-containing neuronsby neuropeptide Y in the rat septo-preoptic area. Neurosci. Lett., 110, 261–266.

Turi, G.F., Liposits, Z., Moenter, S.M., Fekete, C. & Hrabovszky, E.(2003) Origin of neuropeptide Y-containing afferents to gonadotropin-releasing hormone neurons in male mice. Endocrinology, 144, 4967–4974.

Wayman, C.P. & Wilson, J.F. (1992) Endogenous glutamate stimulates releaseof a-melanocyte-stimulating hormone from the rat hypothalamus. Neuro-peptides, 23, 93–97.

Zhang, R., Hisano, S., Chikamori-Aoyama, M. & Daikoku, S. (1987) Synapticassociation between enkephalin-containing axon terminals and proopiome-lanocortin-containing neurons in the arcuate nucleus of rat hypothalamus.Neurosci. Lett., 82, 151–156.



glutamatergic innervation of neuropeptide y and pro-opiomelanocortin-containing neurons in the...

Documents