Journal of Neuroendocrinology, 1998, Vol. 10, 377–381

Glutamate Pathways Mediate Somatostatin Responses to Glucose inNormal and Diabetic Rat Hypothalamus

B. G. Issa, B. M. Lewis, J. Ham, J. R. Peters and M. F. ScanlonDepartment of Medicine, University of Wales College of Medicine, Cardiff, UK.

Key words: glutamate, ionotropic, NMDA, somatostatin, hypothalamus, Goto-Kakizaki, diabetes.

Abstract

We investigated the role of hypothalamic glutamate receptors in mediating the stimulatory effect oflow glucose (<5 mM) on somatostatin release. We also studied whether alteration in glutamaterelease might contribute to the reduced hypothalamic somatostatin response to low glucoseobserved in diabetic (Goto-Kakizaki) rat hypothalami. Hypothalamic somatostatin release in responseto incubation with 1 mM d-glucose was inhibited by the ionotropic glutamate receptor antagonistsMK801, D-AP5 and DNQX but not by the metabotropic antagonists L-AP3 or MCPG. The release ofsomatostatin was increased by the ionotropic agonists NMDA, AMPA and kainate but not bymetabotropic agonists t-ACPD or L-AP4. Basal and peak glutamate release in response toincubation with 1 mM glucose, were significantly lower from GK hypothalami There were nosignificant differences in the basal or stimulated release of serine and GABA. These data indicatethat ionotropic NMDA/AMPA/kainate receptors and not metabotropic receptors mediate the effectsof glucose on rat hypothalamic somatostatin release. Reduced hypothalamic somatostatin release inresponse to low glucose in diabetic (Goto-Kakizaki) rats may well be secondary, at least in part, toreduced glutamate release.

We and others have demonstrated the specific release of synaptic activity between hypothalamic neurons (9).Application of glutamate to hypothalamic neurons increasessomatostatin, growth hormone releasing hormone and thyro-their firing rate (10) and can stimulate hypothalamic somatos-tropin releasing hormone but not luteinising hormone releas-tatin release (11, 12). Therefore, we investigated the role ofing hormone from normal Wistar rat hypothalami in responseglutamate and the type of glutamate receptor mediatingto a decrease in the ambient glucose concentration below 5hypothalamic somatostatin release in response to low glucose.mM (1–4). These effects are markedly reduced in diabeticIn addition, we investigated whether the reduced somatostatinGoto-Kakizaki (GK) rats, a model of nonobese diabetesresponses to low glucose in diabetic GK hypothalami mightproduced by inbreeding of Wistar rats using glucose tolerancebe mediated by alterations in glutamate release.as a selection index (5). This may reflect reduced sensitivity

This is important because understanding the mechanismsto glucose of these hypothalamic neurons or other mediatingunderlying the reduced sensitivity of hypothalamic neuronspathways, secondary to the diabetic state. Because the hypo-to alterations in glucose concentrations may facilitate thethalamus is vital for the integration of neuroendocrine anddevelopment of specific pharmacological protection againstautonomic functions in normal and diabetic states, we usedhypoglycemic neuronal injury and may also contribute to anthis experimental model to investigate further the mechanismsimproved understanding of the so-called ‘hypoglycemiaunderlying the neuropeptide releasing effects of decreasingunawareness’ phenomenon in diabetes.glucose concentrations.

Glutamate is the major excitatory amino acid neuro-transmitter in the nervous system of mammals, and brain Resultsglutamatergic systems are activated by hypoglycemia (6).The hypothalamus shows strong immunoreactivity for glu- There was no significant difference between body weight

(mean SEM, GK vs normal, 249±3 vs 254±20 g) or hypo-tamate (7, 8) which accounts for almost all the excitatory

Correspondence to: Professor M. F. Scanlon, Department of Medicine, University of Wales College of Medicine, Heath Park, Cardiff,CF4 4XN, UK.

© 1998 Blackwell Science Ltd

378 Glutamate and hypothalamic somatostatin in normal and Goto-Kakizaki rats

thalamic weight (GK vs normals 41±2 vs 47±3 mg) of GK antagonist (294.1±15.4% over basal in the absence of drugsvs 146.2±21.1%, P<0.01, for D-AP5 and 115.8±26.3%,and normal rats. Fasting plasma glucose concentrations

(mean±SEM, mmol/l ) were significantly higher in all GK P<0.05 for DNQX. There was no effect of EAA antagonistson the basal (5 mM glucose) release of somatostatin (datarats (GK vs normals, 9.4±0.4 vs 6.1±0.3, P<0.0001). After

an oral glucose load (2 g/kg), there was an increase in plasma not shown). The effect of ionotropic agonists was consistentwith the data using antagonists showing significant stimula-glucose concentrations in both GK (peak 17.5±1.0 at 30 min)

and normal rats (peak 9.9±0.6 at 30 min). The mean area tion of somatostatin release in the presence of 5 mM glucose(249.2±55.3%, P<0.05, for NMDA, 524.0±96.7%. P<0.01,under curve was significantly higher in GK rats (P<0.001).

Incubation of hypothalami with 1 mM glucose caused signi- for AMPA and 294.5±53.8%, P<0.05 for kainate) (Fig. 1).Metabotropic antagonists and agonists had no significantficant release of somatostatin in normals (101.2±3.4% vs

294.1±15.3%, P<0.01, Fig. 1) but not in GK rats effect on somatostatin release (Table 1).(100.4±12.0% vs 126.4±21.5%, NS).

Amino acid release (normal and GK rats) (Fig. 2)Effect of excitatory amino acids antagonists/agonists (normal

Release of glutamate and aspartate was significantly less fromrats) (Fig. 1; Table 1)GK than from normal rat hypothalami [GK, 48.4 (8 mMglucose) and 56.2 (5 mM glucose) vs normal, 120.0 (5 mMMK801, a noncompetitive NMDA antagonist caused a dose-

related inhibition of somatostatin release in response to glucose), P<0.02 (in both comparisons) for glutamate; GK,14.9 (8 mM glucose) and 20.1 (5 mM glucose) vs normal,incubation with 1 mM glucose (294.1±15.4% over basal in

the absence of drug vs 10−6 M: 228.0±35.8%, 10−5 M: 44.6 (5 mM glucose), P<0.02 (in both comparisons) foraspartate].225.1±32.6%, 10−4 M: 121.1±11.7%, 10−3 M: 159.5±

21.5%) with a Bmax dose of 10−4 M (P<0.01). Subsequent Incubation with 1 mM glucose caused significant release ofboth glutamate (120.0 vs 257.0, P<0.01, for normal; 48.4 vsstudies were then performed using drugs at a concentration

of 10−4 M. Similar results were obtained using D-AP5, a 96.9, P<0.02 for GK) and aspartate (44.6 vs 86.6, P<0.02for normal; 14.8 vs 48.4, P<0.01 for GK ) but there were nocompetitive antagonist and DNQX, a non-NMDA ionotropicdifferences for GABA or serine. Peak release of both aminoacids in response to incubation with 1 mM glucose was alsosignificantly less in GK than normal rats (96.9 vs 257.0,P<0.01 for glutamate and 48.4 vs 86.7, P<0.02 for aspart-ate). Basal and peak release of GABA and serine were notsignificantly different between GK and normal rats.

Discussion

These results demonstrate that rat hypothalamic somatostatinrelease is evoked by both ionotropic NMDA and non-NMDAreceptors but not by metabotropic receptors. In particular,ionotropic receptors mediate somatostatin release in responseto incubation with low glucose concentrations. This is inkeeping with the fact that stimulation of ionotropic receptorslinked to ion channels results in rapid glutamate-mediatedexcitatory transmission in the central nervous sytem (CNS),in contrast to metabotropic receptors which mediate pro-longed synaptic modulation via activation of phosphatidyli-nositol pathways (13).

Glutamate receptors are present in the hypothalamus (8)with particularly high levels of NMDA R1, AMPA andkainate receptors (14–16). Glutamate and glutamate agonistsstimulate the release of somatostatin from a variety of brainregions (11, 12, 17–20) and there is evidence that the releaseof several hypothalamic neuropeptides, including somatosta-tin release from hypothalamic neurons in primary culture(21–23), can be evoked via ionotropic NMDA receptors.

5 mM G 1 mM G 5 mMNMDA

5 mMKainate

5 mMAMPA

B700

600

500

400

300

200

100

0

SS

% r

elea

se

5 mM G 1 mM G 1 mMMK801

1 mMD-AP5

1 mMDNQX

A400

300

200

100

0

SS

% r

elea

se

Despite this, we must be cautious about the interpretationF. 1. Incubation of normal rat hypothalami with 1 mM glucose (n=of our data. In vitro incubation of intact hypothalamic tissue24) vs 5 mM glucose (n=24) causes significant release of somatostatin

(P<0.01; ,). () Shows the effects of ionotropic excitatory amino acids is a relatively crude experimental model, which, at best, canantagonists on somatostatin release in response to 1 mM glucose (n=12 provide an overall picture of control systems that can sub-treatment group) and () the effect of ionotropic excitatory amino acids sequently be analysed in more detail using alternative tech-agonists on somatostatin release in 5 mM glucose (n=12 treatment

niques. First, the high concentrations of drugs required togroup). Data are presented as mean±SEM. () *P<0.05, **P<0.01 vs1 mM G. () *P<0.05, **P<0.01 vs 5 mM G. produce an effect could argue against any physiological

© 1998 Blackwell Science Ltd, Journal of Neuroendocrinology, 10, 377–381

Glutamate and hypothalamic somatostatin in normal and Goto-Kakizaki rats 379

T 1. Incubation with 1 mM Glucose Caused a Significant Release of Somatostatin (% Control ) Compared with 5 mMGlucose (P<0.01) which was Unaffected by Metabotropic Antagonists. Metabotropic Agonists had no Effect onSomatostatin Release.

Glucose concentration Antagonists Agonists

5 mM 1 mM 1 mM+L-AP3 1 mM+MCPG 5 mM+t-ACPD 5 mM+D-AP4

101.2±3.3 294.1±15.3 395.7±85.5 184.5±12.2 166.5±14.8 129.0±26.4

L-AP3=-2-amino-3-phosphonopropionic acid, MCPG=a-methyl-4-carboxyphenylglycine, t-ACPD=trans-(1S,3R)-amino-1,3-cyclopentanedicar-boxylic acid, L-AP4=-2-amino-4-phosphonobutyrate.

Second, it is not possible, using this experimental model,to localize accurately the sites of, or neurotransmitter releaseinto, the incubation medium. Our incubate comprised com-plete hypothalamic halves and it is likely that release occursfrom both intact cells and severed nerve terminals.

Despite the known effects of magnesium ions and glycineon the NMDA receptor (magnesium produces a voltage-dependent block of the NMDA ion channel and glycinepotentiates the action of agonists at the NMDA receptor),we have not measured these levels in the incubation mediumof the relevant experiments. However, all our experimentswere performed under the same incubation conditions.

We have also demonstrated a marked reduction in therelease of the excitatory amino acids, glutamate and aspartate,in response to incubation with low glucose in GK, as opposedto normal hypothalami. The effect is specific because therewere no differences in basal or stimulated release of GABAand serine. This ‘attenuated activation’ of the hypothalamicglutamatergic system has not been described previously andmay explain, at least in part, the reduced somatostatinresponse to low glucose which we have previously demon-strated in GK rats (5). It is also possible, however, thatreduced responsiveness of somatostatin neurons to glutamatemay contribute to this phenomenon in GK rats. Because ofthe low somatostatin response to 1 mM glucose in GKhypothalami, it is not possible to undertake mechanisticstudies of the mediation of this response by glutamate recep-tors in these animals. It is difficult to be certain that excitatoryamino acids levels in the incubation media accurately reflectglutamate transmission in the hypothalamic block. However,the effects of low glucose on excitatory amino acids release

5/8mM

B250

200

150

100

50

0

pm

ol/m

g t

issu

e

1 /1mM

5/8mM

1/1mM

GABA Serine

NSNS

NS

NS

5/8mM

A300

200

150

100

50

0

pm

ol/m

g t

issu

e

1 /1mM

5/8mM

1/1mM

Glutamate Aspartate

250

Normal

GK

correlate with previously demonstrated in vivo responses (6)F. 2. Incubation with 1 mM glucose causes significant release of glu- and both normal and GK rat tissues were exposed to thetamate and aspartate, but not serine and GABA from both normal

same incubation conditions.(P<0.01 for glutamate and P<0.02 for aspartate) and GK hypothalami[P<0.02 for glutamate (n=12) and P<0.01 for aspartate (n=12)]. Previous studies of glutamate in the CNS of diabetic animalGlutamate and aspartate release from GK rats is significantly less both models have concentrated on cerebral cortical function. Inbasally and in the presence of 1 mM glucose compared with normal rats rats with streptozotocin-induced diabetes, the specific activity(). Similar data were obtained when GK hypothalami were incubated

of glutamic acid decarboxylase was decreased by 18% inin 5 mM glucose rather than 8 mM glucose (see text). There were nosignificant differences for the release of GABA or serine (). Data are cortical tissue, -glutamate uptake was decreased by 48% inpresented as median values. *P<0.02, **P<0.01. cortical slices but not synaptosomes and the glutamine (but

not glutamate or aspartate) content of cortical tissue wasincreased by 75% (28, 29). These disparate data are difficultrelevance. However, such concentrations have been usedto interpret and cannot be extrapolated to the hypothalamus.previously by ourselves and other investigators using similarIt is interesting, however, that some workers (29), usingexperimental models, in particular in relation to the controlnormal rat cortical slices, showed a considerable reductionof LHRH release (24–27). It has been argued that such(84%) of glutamate in the incubating medium in the absenceconcentrations may be required to produce adequate tissue

and synaptic penetration. of glucose, in contrast to our findings in the hypothalamus.

© 1998 Blackwell Science Ltd, Journal of Neuroendocrinology, 10, 377–381

380 Glutamate and hypothalamic somatostatin in normal and Goto-Kakizaki rats

receptor antagonist [Dizocilpine (MK801) and -2-amino-5-phosphopen-Although GABA neurons in the medial basal hypothalamustanoic acid (D-AP5) for ionotropic noncompetitive and competitive (N-are innervated by glutamate neurons (8, 30), we did not findmethyl--aspartate) NMDA receptors, respectively; 6,7-dinitroquioxaline-2,3-

any differences in the release of GABA between normal and dione (DNQX ) for ionotropic non-NMDA receptors and -2-amino-3-GK hypothalami. phosphonopropionic acid (L-AP3) and a-methyl-4-carboxyphenylglycine

(MCPG) for metabotropic receptors]. Dose response studies for MK 801The cause and significance of the ‘attenuated activation’ ofwere performed using MK 801 at concentrations 10−6−10−3 M.the hypothalamic excitatory amino acids system in GK rats

In separate experiments, hypothalami were incubated in 5 mM glucoseis unclear. The cerebral glutamate pool of amino acids is an with a glutamate receptor agonist [NMDA, AMPA and kainate for ionotropicimportant source of oxidizable substrates (31) and it is receptors and trans-(1S,3R)-amino-1,3-cyclopentanedicarboxylic acid (t-

ACPD) and -2-amino-4-phosphonobutyrate (L-AP4) for metabotropicpossible that in diabetic states, exposure to high levels ofreceptors]. These drugs show specificity, as indicated, for ionotropic receptorsglucose or reduced levels of insulin decreases the glutamatethat are linked to ion channels or metabotropic receptors, which are linkedcontent in the CNS. However, this view is not supported byto second message pathways (cAMP and phospholipase C). Group I and

the findings of others (28) who found no significant differ- group II metabotropic receptors can both be stimulated by glutamate and t-ences in the cortical content of glutamate and aspartate in ACPD, but differ in that group I receptors activate phospholipase C whereas

group II receptors inhibit adenyl cyclase. The third group of metabotropicdiabetic rats. Alternatively, there may be alterations in thereceptors is activated by L-AP4, but not by t-ACPD, and this action isfunction and/or expression of glucose transporters, particu-mediated by the inhibition of adenyl cyclase. L-AP3 has been reported to belarly Glut 3, on glutamate neurons. a putative antagonist of metabotropic receptors whereas MCPG has some

It is established that glucose deprivation causes an influx selectivity for group II receptors (33). All drugs were obtained from SemetTechnical Ltd. (St Albans, UK) and were used at a final concentration ofof calcium into neurons and release of excitatory amino acids.10−4 M. These concentrations of drugs were used previously with similarThe subsequent excitotoxic effect of glutamate and aspartateexperimental models and are required to ensure adequate tissue and synapticis also mediated by calcium influx into the cell throughpenetration (24–27).

ionotropic glutamate, particularly NMDA receptors (6, 32). In the second series of experiments on excitatory amino acids release,However, if this becomes excessive, neurotoxic damage can normal and diabetic (GK) hypothalami were incubated in KRB containing

5 mM glucose (normal ) or 8 mM glucose (GK) and 1 mM glucose KRB.result. It is possible that the reduced excitatory amino acidsThe concentrations of 5 mM and 8 mM were chosen to simulate ambientrelease in diabetic animals serves as a protective mechanismfasting glucose concentrations in each group of animals. However, experimentsagainst hypoglycemia and excitatory amino acids induced were also performed on GK hypothalami incubated in 5 mM glucose after

neuronal damage. removal.At the end of all experimental incubations, media were removed, centrifuged

and aliquoted in tubes (containing 1 mM acetic acid in somatostatin tubes toensure peptidase inhibition) for assay of somatostatin, glutamate, aspartate,Materials and methodsGABA and serine.

AnimalsRadioimmunoassayAll rats used were from the University of Wales College of Medicine breedingThe radioimmunoassay procedure for somatostatin has been described previ-colony. GK rats ( Wistar strain) were initially given as a gift by Professorously (1). Briefly, somatostatin was measured by standard RIA using a fullyY. Goto of the Tohoku University School of Medicine, Sendai, Japan. Thecharacterized antiserum (34). Appropriate dilutions of labelled (125I-Tyrl )-colony was started in 1989 with six breeding pairs. The diabetic offspringsomatostatin and antiserum were used to obtain 10 000 cpm and 30% binding.were raised in parallel to the non-diabetic Wistar rats, and all rats were24 h preincubation and incubation times were used at 4 °C and nonspecificmaintained six to a cage in a thermo-statically controlled room at 24 °C withbinding was 3–4%. Assay sensitivity was less than 1 pg/tube and somatostatin-a 24 h lighting cycle ( light periods being from 07.00 h to 19.00 h). The rats14 and somatostatin-28 compete on an equimolar basis for binding of labelwere fed on a standard diet (Pilsbury’s modified rat and mouse breeding diet)to this particular antibody (35, 36).and allowed free access to tap water. Rats were weaned at 28 days after birth.

To assess the glucose tolerance of the rats, oral glucose (2 g/kg) was given toPlasma glucose measurementsnormal and diabetic animals after a 12 h overnight fast and tail vein samplesPlasma glucose was measured using the glucose oxidase method on a YSIwere taken for plasma glucose measurements at 0, 30, 60, 90 and 120 minModel 2300 STAT Plus Glucose Analyser (Yellow Springs Instrument Ltd,

In vitro methodology Yellow Springs, OH, USA).The methods for in vitro hypothalamic incubation have been described

Assay of amino acidspreviously in detail (1). Briefly, the experiments were performed on hypothal-Glutamate, aspartate, GABA and serine were determined by reverse phaseamic tissue removed from 16-week-old male GK and normal Wistar rats.HPLC. The mobile phase consisted of a linear gradient between sodiumRats were fasted for 12–16 h and weighed before tissue removal. The ratsacetate-tetrahydroflurane and methanol; a 25 cm long reverse C18 columnwere rendered unconscious in CO2 chambers then decapitated and the brainswas used to separate the amino acids. 200 ml of sample was added to 200 mlremoved rapidly. Hypothalami, defined by the posterior margin of the opticof homoserine (internal standard) and vortex mixed. Following centrifugation,chiasm and the anterior margins of the mamillary bodies to a depth of 2 mm,200 ml was taken and mixed with 50 ml of O–phthaldialdehyde. 100 ml of thewere dissected out, weighed, halved longitudinally and the two halves placedmixture were subsequently injected onto the column and the resultingtogether in each well of a multiwell tissue culture plate containing 500 ml offluorescence measured.Krebs-Ringer bicarbonate solution ( KRB; 4.75 mM KCl, 1.18 mM KH2PO4,

1.18 mM MgSO4, 2.52 mM CaCl2, 25 mM NaHCO3 and 118.5 mM NaCl,Data presentation and statistical analysispH 7.4) with varying concentrations of -glucose, 0.1% (w/v) bovine serum

albumin and bacitracin (300 mg/ml ). A minimum of four hypothalami per Statistical analysis of weight and glucose status in normal vs GK rats wasperformed using the nonpaired Students ‘t’ test. The areas under the glucosetreatment group were used in each experiment and each experiment was

repeated at least three times. tolerance test curves (AUC) were calculated using the trapezoidal rule.All data for the glutamate agonist/antagonist experiments are expressed asWe have previously demonstrated that the basal release of somatostatin

stabilizes after incubation for 60 min (5). Hypothalami were therefore incub- mean±SEM. Data for somatostatin were calculated as pg somatostatin perhypothalamus per 20 min incubation period, which produces similar resultsated for three successive 20 min periods and the media discarded. The

experimental treatment was then applied for the final 20 min incubation. All to expression of the data as pg somatostatin per wet weight or protein content(4). Somatostatin release was expressed as a percentage of the mean of theincubations were performed at 37 °C gassed with 95% O2 and 5% CO2. In

the first series of experiments, normal hypothalami were incubated in KRB control (i.e. somatostatin release in the presence of 5 mM G) in order toreduce variability in the results between experiments. Control values werecontaining 5 mM glucose, 1 mM glucose and 1 mM glucose with a glutamate

© 1998 Blackwell Science Ltd, Journal of Neuroendocrinology, 10, 377–381

Glutamate and hypothalamic somatostatin in normal and Goto-Kakizaki rats 381

calculated as a percentage of the mean control and expressed as mean±SEM. Anatomical organisation of excitatory amino acid receptors and theirpathways. Trends Neurosci 1987; 10: 273–279.To assess the efficacy of various glutamate receptor antagonists, somatostatin

17 Tapia-Arancibia L, Astier H. Actions of excitatory amino acids onrelease was expressed as a percentage of the mean of peak somatostatinsomatostatin release from cortical neurons in primary cultures. Jrelease in the presence of 1 mM glucose. Statistical analysis was performedNeurochem 1989; 53: 1134–1141.using one way analysis of variance (ANOVA) with post hoc comparisons by

18 Tapia-Arancibia L, Rage F, Recasens M, Pin J. NMDA receptorDunnett multiple comparisons test.activation stimulates phospholipase A2 and somatostatin release fromExcitatory amino acids data are expressed as median pmol/mg tissue.rat cortical neurons in primary cultures. Eur J Pharmacol 1992; 225:Amino acid release from normal and GK rats was compared using a253–262.nonparametric test (Mann–Whitney test) because the data did not satisfy the

19 Fontana G, De Bernardi R, Ferro F, Gemignani A, Raiteri M.criteria for a normal distribution.Characterization of the glutamate receptors mediating release of somatos-tatin from cultured hippocampal neurons. J Neurochem 1996; 66: 161–168.Acknowledgements 20 Ham J, Rickards C, Scanlon M. Prolonged exposure to N-methyl-D-aspartate increases intracellular and secreted somatostatin in rat corticalThis work was supported by grants from the British Diabetic Associationcells. Neurosc Lett 1991; 129: 262–264.and the Welsh Scheme for the Development of Health and Social Research.

21 Cocilovo L, de Gennaro Colonna V, Zoli M, Biagini G, Settembrini BP,We are grateful to Miss N. Knott and Dr J. Srinivasan for help with theMuller EE, Cocchi D. Central mechanisms subserving the impairedassays of amino acids.growth hormone secretion induced by persistent blockade of NMDAreceptors in immature male rats. Neuroendocrinology 1992; 55: 416–221.

Accepted 14 January 1998 22 Bourguignon J, Gerard A, Mathieu J, Simons J, Franchimont P. Pulsatilerelease of gonadotropin-releaseing hormone from hypothalamic explantsis restrained by blockade of of N-methyl-D,L-aspartate receptors.Endocrinology 1989; 125: 1090–1096.References

23 Urbanski H. A role of N-methyl-D-aspartate receptors in the control ofseasonal breeding. Endocrinology 1990; 127: 2223–2228.1 Lewis BM, Dieguez C, Ham J, Page MD, Creagh FM, Peters JR,

24 Mahachoklertwatdana P, Black SM, Kaplan SL, Bristow JD, GrumbachScanlon MF. Effects of glucose on thyrotropin-releasing hormone, growthMM. Nitric oxide synthesised by gonadotropin releasing hormone neu-hormone-releasing hormone, somatostatin and leutinizing hormone-rons is a mediator of N-methyl-D-aspartate (NMDA)-induced GnRHreleasing hormone release from rat hypothalamus in vitro. Jsecretion. Endocrinology 1994; 135: 1709–1712.Neuroendocrinol 1989; 1: 437–441.

25 Wayman CP, Pike N, Wilson J. N-methyl-D-aspartate (NMDA) stimu-2 Baes M, Vale WW. Characterisation of the glucose-dependent release oflates release of alpha-MSH from the rat hypothalamus through releasegrowth hormone-releasing factor and somatostatin from superfused ratof nitric oxide. Brain Res 1994; 666: 201–206.hypothalami. Neuroendocrinology 1990; 51: 202–207.

26 Giri M, Kaufman JM. Involvement of neuroexcitatory amino acids in3 Lengyel AM, Grossman A, Nieuwenhuyzen-Kruseman A, Ackland J,the control of gonadotropin release hormone from the hypothalamus ofRees LH, Besser M. Glucose modulation of somatostatin and LHRHthe adult male guinea: predominantly inhibitory action of N-methyl-D-release from rat hypothalamic fragments in vitro. Neuroendocrinologyaspartate mediated neurotransmission and its reversal after orchidectomy.1988; 39: 31–38.Endocrinology 1996; 136: 2404–2407.4 Berelowitz M, Dudlak D, Frohman LA. Release of somatostatin-like

27 Giri M, Changqing G, Kaufman JM. The N-methyl-D-aspartate medi-immunoreactivity from incubated rat hypothalamus and cerebral cortex.ated inhibitory control of gonadotropin-releasing hormone release in theEffects of glucose and glucoregulatory hormones. J Clin Invest 1982; 69:hypothalamus of the adult male guinea pig is expressed through opioid-1293–1301.ergic systems. Endocrinology 1996; 137: 1468–1473.5 Lewis BM, Ismail IS, Issa B, Peters JR, Scanlon MF. Desensitisation of

28 Galanopoulos E, Lellos V, Papadakis M, Philippidis H, Palaiologos G.somatostatin, TRH, and GHRH responses to glucose in the diabetic Effect of fasting and diabetes on some enzymes and transport of(GK) rat hypothalamus. J Endocrinol 1996; 151: 13–17. glutamate in cortex slices or synaptosomes from rat brain. Neurochem

6 Weiloch T. Hypoglycemia induce neuronal damage prevented by an N- Res 1988; 13: 243–248.methyl-D-aspartate antagonist. Science 1985; 230: 681–683. 29 Palaiologos G, Philippidis H, Chomatas H, Iakovou D, Linardou A.

7 van den Pol AN. Glutamate and aspartate immunoreactivity in hypothal- Effect of branched chained amino acids, pyruvate, or ketone bodies onamic presynaptic neurons. J Neurosci 1991; 11: 2087–2101. the free amino acid pool and release from brain cortex slices of normal

8 van den Pol AN, Waurin JP, Dudek FE. Glutamate the dominant and streptozotocin-diabetic rats. Neurochem Res 1987; 12: 1–7.excitatory transmitter in neuroendocrine regulation. Science 1990; 250: 30 Decavel C, van den Pol AN. Converging GABA and glutamate-1276–1278. immunoreactive axons make synaptic contact with identified hypothal-

9 van den Pol AN, Trombley PQ. Glutamate neurons in hypothalamus amic neurosecretory neurons. J Comp Neurol 1992; 316: 104–116.regulate excitatory transmission. J Neurosci 1993; 13: 2829–2836. 31 Lewis LD, Ljunggren B, Norberg K, Siesjo BK. Changes in carbohydrate

10 Arnauld E, Cirino M, Layton BS, Renaud LP. Contrasting actions of substrates, amino acids and ammonia in the brain during insulin-inducedamino acids, acetyl choline, adrenaline, and leucine enkephalin on the hypoglycemia. J Neurochem 1974; 23: 659–671.excitability of supraoptic vasopressin-secreting neurons. 32 Tasker CT, Coyle JT, Vornov JJ. The regional vulnerability to hypoglyce-Neuroendocrinology 1983; 36: 187–196. mia-induced neurotoxicity in organotypic hippocampal cultue: protection

11 Rage F, Benyassi A, Arancibia S, Tapia-Arancibia L. Gamma amino by early tetrodotoxin or delayed MK-801. J Neurosci 1992; 12: 4298–butyric acid–glutamate interaction in the control of somatostatin release 4308.from hypothalamic neurons in primary culture: in vivo corroboration. 33 Pin JP, Duvoisin R. Neurotransmitter receptors I. The metabotropicEndocrinology 1992; 130: 1056–1062. glutamate receptors: structure and functions. Neuropharmacology 1995;

12 Tapia-Arancibia L, Astier H. Glutamate stimulates somatostatin release 34: 1–26.from diencephalic neurons in primary culture. Endocrinology 1988; 123: 34 Patel YC, Reichlin S. Somatostatin in hypothalamus, extrahypothalamic2360–2366. brain and peripheral tissues of the rat. Endocrinology 1978; 102: 523–530.

13 Brann DW. Glutamate: a major excitatory transmitter in neuroendocrine 35 Scanlon MF, Robbins RJ, Bolaffi JL, Jackson IMD, Reichlin S.regulation. Neuroendocrinology 1995; 61: 213–225. Characterisation of somatostatin and TRH release by rat hypopthalamic

14 Meeker RB, Greenwood RS, Hayward JN. Glutamate receptors in the and cerebral cortical neurons maintained on a capillary membranerat hypothalamus and pituitary. Endocrinology 1994; 134: 621–629. perfusion system. Neuroendocrinology 1983; 37: 269–274.

15 Kus L, Beitz AJ, Kerr JE, Handa RJ. NMDA R1 receptor mRNA 36 Robbins RJ, Leidy JWJ, Landon RM. The effects of growth hormone,expression in the hypothalamus of intact, castrate and DHT-treated male prolactin, corticotophin and thyrotrophin on the production and secre-rats. Soc Neuroscience 1993; 19: 919. tion of somatostatin by hypothalamic cells in vitro. Endocrinology 1985;

117: 538–543.16 Cotman CW, Monaghan DT, Ottersen OP, Strom-Mathisen J.

© 1998 Blackwell Science Ltd, Journal of Neuroendocrinology, 10, 377–381