European Journal of Pharmacology - Molecular Pharmacology Section, 225 ( 19921253-262 © 1992 Elsevier Science Publishers B.V. All rights resewed 0922-4106/92/~05.00

253

FLJPMOL 90274

NMDA receptor activation stimulates phospholipase A z and somatostatin release from rat cortical neurons in primary cultures L u c i a T a p i a - A r a n c i b i a a, F l o r e n c e R a g e a, M a x R r c a s e n s b an d J e a n - P h i l i p p e Pin c Unit( de Neurobiologie Endocrinologique, CNRS URA 1197, Universit~ de Montpellier II, 34095 Montpellier Cedex 5, France, b INSERM U254. H~pital Saint Charles, Montpellier, France and c Centre CNRS-hVSERM de Pharmacologie-Endocrinologie, Montpellier, France Received 29 October 1991, accepted 3 December 1991

We have recently shown that glutamate exerts a stimulatory action on somatostatin secretion in cortical neurons essentially through NMDA receptor sites. Here, we investigated whether arachidonic acid release couid be modified after NMDA receptor activation in cortical neurons in primary culture. We also studied whether pharmacological manipulation of phospholipase Az could modify somatostatin release. We found that both glutamate and NMDA (N-methyl-D-aspartate) stimulated [3H]arachidonic acid release. NMDA-evoked arachidonic acid release was inhibited by MK-801 and TCP (two NMDA receptor-type antagonists), or by mepacrine, an inhibitor of phospholipase A 2. NMDA-induced somatostatin release was inhibited by MK-801, mepacrine and by another phospholipase A 2 inhibitor, p-bromophenacylbromide (pBPB). However, responses to NMDA were unaffected by H7, NDGA (nordihydroguaiaretic acid), indomethacin or by RHC 80267 (inhibitors of protein kinase C, lipooxygenase, cyclooxygenase and diacylglycerol lipasc, respectively). Mepacrine (~> 100 tzM) decreased NMDA-stimulated phosphatidylinositol (PI) hydrolysis and at higher concentrations (250 p.M) was also able to inhibit basal release whereas pBPB had no effect in the range of concentrations tested. Neomycin (which inhibits phosphatidylinositol metabolism by binding strongly and selectively to inositol phospholipids) reduced by 30% the NMDA-stimulated somatostatin release, although chronic treatment of neurons with the phorbol ester 12-myristate, 13-acetate (PMA) had no effect on this response. Melittin, an activator of phospholipase A2, was able to stimulate both arachidonic acid release and somatostatin secretion. High-performance liquid chromatography (HPLC) analysis of tritiated metabolites released from cortical neurons under basal or NMDA-stimulated conditions revealed that [3H]arachidonic acid was the only metabofite detectable. Furthermore, external addition of arachidonic acid increased somatostatin secretion. Our results show a correlation bev~,een the two parameters studied. NMDA (N-methyl-D-aspartate); Cortica! neurons; Arachidonic acid; Somatostatin; Phospholipase A 2

1. Introduction Somatostatin is widely distributed in neurons of the cerebral cortex (Epelbaum et al., 1977; Reichlin, 1983), although its role in cortical activity remains to be defined. Excitatory amino acid systems are also largely distributed in the telencephalon, where most cortical projections appear to use glutamate (Fagg and Foster, 1983; Fonnum, 1984; Cotman et al., 1987). We have already reported that primary cultures of rat cortical cells secrete somatostatin on stimulation by glutamate (Tapia-Arancibia and Astier, 1989) or by

Correspondence to: Dr. Lucia Tapia-Arancibia, Universit~ de Mont-

pel!ier I1, Place Eugene Bataillon, 34095 Montpellier Cedex 5, France. Tel. 67143815; Fax 67543079. A preliminary report of some of this work was presented at the Fidia SymposiumExcitatoryAminoAcidsConference in 1990,Padua, Italy.

depolarization of the plasma membrane (TapiaAranclbia ct al., 1988). N M D A receptor stimulation seems to prime the cells for glutamatergic stimulation since the receptors involved in somatostatin release fulfill all the characteristics of a typical N M D A receptor. The response is modulated by Mg 2+ and potentiated by glycine and serine through a strychnine-insensitive mechanism (Tapia-Arancibia and Astier, 1990). Furthermore the N M D A stimulato~ effect is decreased by Zn 2+, and strongly inhibited by DL-amino5-phosphonovaleric acid in a competitive manner or by thienybphencyclidine non-competitively. On the other hand, activation of NMDA-sensitive glutamat, receptors is able to stimulate arachidonic acid release in primary cultures of ceret~dtar granule cells (Lazarewicz et al., 1988), striatal neurons ~L~umuis et al., 1988), or hippocampal neurons (Sanfeliu et al., 1990), presumably by activation of phospholipase A 2 (Irvine, 1982).

254 In this study ceo investigated whether arachidonic acid release could be modified after NMDA receptor activation in cortical neurons in primary culture. We also attempted to determine whether pharmacological manipulation of phospholipase A 2 could modify somatostatin release.

2. Material and methods

2.1. Chemicals Monosodium L-glutamate, mepacrine (quinacrine), melittin, indomethacin, nordihydroguaiaretic acid (NDGA), 1-(5-isoquinolinyl sulfonyl) 2-methylpiperazine (H7), neomycin sulfate, phorbol 12-myristate 13acetate (PMA), and p-bromophenacylbromide (pBPB) were purchased from Sigma Chemical Co. (St. Louis, MO, USA). N-Methyl-D-aspartate (NMDA) was from Cambridge Research B'.qchemicals (UK); MK-801, (( + )-5-methyl-10,11-dihydro-5H-dibenzo(a.d.)cyclohept-5,10-imine hydrogen maleate) was from Research Biochemicals (Natick, MA, USA) and [3H]arachidonic acid (specific activity, 214 Ci/mmol) was purchased from Amersham France. Thienyl-phencyclidine (TCP) was kindly provided by Dr. J.M. Kamenka, Ecole Nationale Sup6rieure de Chimie (Montpellier, France) and RHC 80267 by Dr. J. Bockaert (CCIP, Montpellier, France). 2.2. Cell culture technique Primary cell cultures were prepared by mechanoenzymatic dissociation of fetal (day 17) rat telencephalon as previously described by Delfs et al. (1980) with some modifications (Par6s-Herbut6 et al., 1988). Briefly, cells were plated at 1.5-2.0 × 10 6 cells per culture dish (16 mm diameter, Costar, Cambridge, MA, USA) previously coated with poly-D-lysine (10 ~ g / m l ; 220,000 molecular weight) and pre-incubated for 1 h with 10% fetal calf serum in minimum essential medium (Grand Island Biological Co., Grand Island, NY, USA). After withdrawal of the last coating solution, cells were seeded in growth medium composed of minimum essential medium containing 10% Nu Serum (Collaborative Research, Lexington, MA, USA), glucose (0.6%), glutamine (2 mM), penicillin~streptomycin (2.5 U / m l ) adjusted to pH 7.3. The cultures were maintained at 37°C in a humid atmosphere of 95% air-5% CO 2. The preliferation of non-neuronal cells was inhibited by treatment with 10 ~M cytosine arabinoside for 48 h between days 3 and 5 after plating. These neuronal cultures did not contain more than 6% of glial cells, as revealed by counting cells positive for glial fibrillary acidic protein and vimentin (Par6s-Herbut6 et al., 1988). Thereafter, fresh

medium was added to the culture without the antimitotic agent. 2.3. Somatostatin release experiments After 8 days, cultures were removed from the incubator, randomized, and divided into treatment groups. The medium was quickly decanted, and the cells were washed with 3 ml prewarmed Earle's balanced salt solution (Grand Island Biological Co.), pregassed with 95% air-5% CO 2. Cells were washed twice for 20 min with 3 ml modified Locke medium without Mg z+ (raM): (NaCI, 154; KCI, 5.6; CaCI 2, 1.8; sucrose, 1; NaHCO3, 6; glucose, 10), buffered to pH 7.2 with 2 mM HEPES at 37°C in an atmosphere of 95% 02-5% CO 2. In most of the experiments Locke medium was added with 1 /~M tetrodoxin (qTX) to avoid any indirect effect mediated by TYX-sensitive action potentials via excitatory interneurons. The reactions were started by adding the same medium, containing bacitracin (30 p~g/ml) as a peptidase inhibitor, and then the test compounds. After the incubation period (15 rain) unless otherwise indicated, medium was aspirated and quickly decanted into tubes containing 0.1 times their volume of 10 N acetic acid to inhibit somatostatin degradation, and boiled for 10 min at 100°C. Aliquots (200 txl) of extracts were evaporated in a Speed Vac concentrator (Savant Instruments, Hicksville, NY, USA). Immunoreactive somatostatin was determined by a sensitive radioimmunoassay (RIA) (Patel and Reichlin, 1978) carried out in the same tube to minimize peptide loss. Neither salts nor other substances present in the dried samples interfered with the assay binding curves, which were run in tubes also containing 200/~1 evaporated medium. The antiserum (No. 693) was generously provided by Dr. Seymour Reichlin, New England Medical Center (Boston, MA, USA). The intra- and inter-assay coefficients of variation were 6.3% (n = 16) and 7.8% ( n - 2 0 ) , respectively. 2.4. Arachidonic acid release experiments In order to incorporate [3H]arachidonate into cellular lipids, neurons were incubated overnight with [3H]arachidonic acid (0.5/zCi per well). 80-85% of the added label was incorporated into cortical neurons. Before starting the experiments, cells were washed twice for 20 min with 3 ml of Locke medium containing 0.1% fatty acid-free bovine serum albumin (BSA). Neurons were then exposed for 15 rain at 37°C to the same medium plus the tested drugs. After incubation, the medium was decanted into vials and cells were dissolved in 1 ml of 1% sodium dodecyl sulfate (SDS), both fractions being assayed for radioactivity in a beta counter after addition of 6 ml ef scintillant (Amersham,

255 ACS II). Results are expressed as a percentage of released radioactivity during the 15 min incubation period over total radioactivity incorporated in the cells.

tions of [~H]inositol by other tritiated compounds as previously described (R4casens et al., 1988).

2. Z Statistical analysis 2.5. HPLC analysis of the [ 3H]arachidonic acid metaboiites Before being subjected to high-performance liquid chromatography (HPLC) analysis, arachidonic acid metabolites were extracted twice with 5 volumes of chloroform after acidification of the medium to pH 3.5 with formic acid. The chloroform phase was evaporated under vacuum with a Speed Vac concentrator (Savant). The [3H]arachidonic acid metabolites were dissolved in 200 /~1 aeetonitrile (30%)/water (pH 3.5 with 70% acetic acid) containing 5 mg of each of the unlabelled standards, and then injected into an Ultrasphere-ODS guard column connected to a reversephase column (Radial-pack, Resolve C18, 5 ram, Waters). The HPLC system consisted of a 421 controller, and two 114 Beckman pumps. Chromatographic conditions were adapted from Eling et al. (1982). Briefly, elution was performed at room temperature, at a flow rate of 3 m l / m i n , with a mixture of water (equilibrated at pH = 3.5 with acetic acid) and pure acetonitrile (Rathburn). A discontinuous step gradient was used, starting at 30% acetonitrile with increases to 54% at 45 min and 100% at 85 min. Fractions were collected every 0.5 min (1.5 ml) and counted in minMals containing 6 ml of scintillant (Amersham, ACS II). Elutions of unlabelled prostaglandins and HETEs were followed at 192 nm and 234 nm respectively (Merck spectrophotometer UVL 4000).

EC50 values were calculated using the 'GraphPAD' computer program (H. Motulsky) from the Departmeet of Pharmacology, University of California, San Diego (I.S.I. Software). The statistical significance of the results was calculated by one-way analysis of variance followed by Barlett's test.

3. Results

3.1. Glutamate and NMDA stimulate arachidonic acid release We have already shown that glutamate and NMDA are equally effective in eliciting somatostatin release from cortical neurons in primary culture (TapiaArancibia and Astier, 1989). In the present report, increasing concentrations of glutamate and NMDA in Mg2+-free medium evoked a dose-dependent stimulation of arachidonic acid release when added to cortical neurons (fig. 1). Glutamate and NMDA were equally potent with ECs0 values of 13.30 + 4.80 t~M (n = 8) and 8.40 _+ 1.76 /~M (n = 8) respectively. These EC50 values were closed to that found for somatostatin release (ECs0 = 20 p~M). Time course studies of [3H]araehidonic acid release are shown in fig. 1 (inset). 1.4

2. 6. Measurement o f inositol phospholipid hydrolysis

1.2

Cultures were labelled with 2 ~ C i / w e l l of myo-[23H]inositol and 1 mM cytidine for 24 h and then ex-enzNely washed with Locke solution (equilibrated with 5% C O z / 9 5 % O 2 to pH 7.4) prewarmed at 37°C. After 15 miw of preincubation with the antagonists in the presence of 10 mM LiC1, neurons were incubated for 15 mi,a at 37°C with 50 p~M NMDA or Locke solution for control samples. The reaction (release of [3H]inositol phosphates from membrane phospholipids) was stopped by replacement of the incubation medium by 5% PCA. Cells were harvested and [3H]IPs were extracted and separated on Dowex 1 x 8 formate columns according to the technique of Bone et al. (1984). The glycerophosphoinositides were eluted by 20 ml of 0.04 M ammonium formate/0.003 M sodium tetraborate. Then, the fraction containing IPs was eluted with 20 ml of 0.8 M ammonium formate/0.1 M formic acid. The measurements of levels of [3H]IPs were systematically corrected for possible contamina-

1.0

0.g

0.6

~1.-I

t

0.z I -8

log [ag~mist], M

Fig. 1. Glutamate and NMDA stimulationof aracbidonic acid release. Cortical neuronsmaintainedfor 8 days in vitrowere incubated overnightwit~0.5 p~Ci[3H]arachidonicacid. Cells,':z;c washed twice for 20 rain witt~ Mg2+-free Locke mediumand then exposed for 15 rain to the indicated concentrationsof agonists. Arachidwmcacid release is expressed as percent of total incorporated radioactivity. Results are meanszS.E.M, of four determinationsin one of three independentexperiments.Insetshows fl~etime-courseof arachidonic acid relee.se inducedby I ~ #M NMDA.

256 •

'"

NMDA (50 ~¢'i)

NMDA (50[tM)

i it a "0

.~

,~

150"

-~

I00

~.

50

-%.

0.8"

Control

0.6"

0.4"

0.2-

t

d

-9

-8

•

-7 log [MK

0.0 0

-9

-8

-7

-6

-5

-4

log [MK 801], M 2.0 NMDA(10011M)

,

-

,

-6

•

-5

801],

-4

M

Fig. 3. Inhibition by MK-801 of NMDA-induced somatostatin release. Antagonist was added 10 min before and throughout the 15 min incubation period with the agonist (50 ~ M NMDA) in Mg2+-free Locke medium. Results are means_+S,E.M, of four experiments. ** P < 0.01 vs. agonist stimulation.

!.6 ¸

NMDA-evoked release of somatostatin is shown in fig. 3. Half-maximal inhibition was obtained with 14.40 + 2.1 nM.

1.2

3.3. Metabolic pathways involved in NMDA stimulation o f arachidonic acid and somatostatin releases

•r- 0.8' r.