Directory UMM :Data Elmu:jurnal:B:Brain Research:Vol887.Issue2.Dec2000:

(1)

www.elsevier.com / locate / bres

Research report

N-Methyl-

D

-aspartate receptor activation results in regulation of

extracellular signal-regulated kinases by protein kinases and

phosphatases in glutamate-induced neuronal apototic-like death

a,b a a ,

_*

Qian Jiang

, Zhenglin Gu , Guangyi Zhang

, Guozhang Jing

Research Center for Biochemistry and Molecular Biology, Xuzhou Medical College, 84 West Huai-hai Road, Xuzhou, Jiangsu 221002, PR China

Shanghai Institute of Materia Medica, Chinese Academy of Sciences, 294 Tai-yuan Road, Shanghai 200031, PR China Accepted 19 September 2000

Abstract

Extracellular signal-regulated kinases (ERK1 / ERK2) have been shown transiently activated and involved in excitotoxicity. We searched for upstream molecules responsible for the regulation of glutamate-induced ERK1 / ERK2 activation and ERK1 / ERK2-mediated apototic-like death in cultured rat cortical neurons. ERK1 / ERK2 activation (monitored by anti-active ERK1 / ERK2 antibody) was almost

completely prevented by blockage of NMDA receptor (NMDA-R) or elimination of extracellular Ca , but not any other glutamate 21

receptor or L-type voltage-gated Ca channel. It was prevented largely by inhibition of protein kinase C (PKC), protein-tyrosine kinases (PTK), respectively, but mildly by that of CaM kinase II. Combined inhibition of CaM kinase II (but not PTK) and PKC had an additive effect. Reversion of ERK1 / ERK2 activation was largely prevented by inhibition of protein phosphatase (PP) 1 or protein tyrosine phosphatase (PTP). Combined inhibition of PP 1 and PTP had no additive effect. Glutamate-induced apoptotic-like death (determined by DAPI staining) was largely prevented by inhibition of NMDA-R, PKC, CaM kinase II, PTK and MEK1 / MEK2 (ERK1 / ERK2 kinase), respectively. Combined inhibition of CaM kinase II (but not PKC or PTK) and MEK1 / MEK2 had an additive effect. Glutamate-induced apoptotic-like death was promoted by inhibition of PP1 and PTP, respectively. The above results suggested that in glutamate-induced cortical neurotoxicity ERK1 / ERK2 activation be mainly mediated by NMDA-R. Subsequently, a pathway dependent on both PKC and PTK was mainly involved, which was also mainly responsible for ERK1 / ERK2-mediated apoptotic-like death, and a CaM kinase II-dependent pathway was relatively mildly involved. Reversion of ERK1 / ERK2 activation was mainly mediated by a pathway dependent on both PP1 and PTP, which might be involved in the restrain of glutamate-induced neurotoxicity.  2000 Elsevier Science B.V. All rights reserved.

Theme: Neurotransmitters, modulators, transporters, and receptors Topic: Excitatory amino acids: excitotoxicity

Keywords: Extracellular signal-regulated kinase; Excitotoxicity; Glutamate receptor; Protein kinase; Protein phosphatase; Cortical neuron

1. Introduction intracellular cascades leading to excitotoxicity are largely

unknown, potential Ca targets include nitric oxide Glutamate-induced excitotoxicity, with certain charac- synthase (NOS) [2], protein phosphatases and protein teristics of apoptosis [5,25,48], has been implicated in kinases [9,28]. Recently, several molecules including NOS many neuronal degenerative diseases and considered pre- [49,50], protein kinase C (PKC), CaM kinase II [18] and

dominantly mediated by an overload of intracellular Ca , some members of protein tyrosine kinases (PTK), such as mainly as a result of influx via N-methyl-D-aspartate PYK2 (proline-rich tyrosine kinase 2) and c-Src

21 21

receptor (NMDA-R) [30,51]. Although the Ca -activated [16,29,39], have been shown to transduce Ca signaling to ERK1 / ERK2 (extracellular signal-regulated kinases) cascade.

*Corresponding author. Tel.:186-516-574-8423; fax: 1

86-516-574-ERK1 / ERK2, with molecular masses of 44 and 42 kDa,

8429.

E-mail address: gyzhang@xzmc.edu.cn (G. Zhang). respectively, are classical members of mitogen-activated

(2)

protein kinase (MAPK) superfamily. Both require specific glutamine, Gibco-BRL). Cultures were used after 13 days diphosphorylation of both threonine and tyrosine residues in vitro when the cells were vulnerable to glutamate insult at the regulatory sites by MEK1 / MEK2 (ERK1 / ERK2 and verified .95% neurofilaments positive by immuno-kinase) for activation [33,42]. ERK1 / ERK2 cascades play staining.

important roles in signal transduction from cell surface to

nucleus. The well-documented neurotropic growth factor 2.2. Drug treatment receptor-mediated activation cascade (Ras / Raf / MEK /

ERK) has been thought to play important roles in cell Excitotoxicity was induced by 50 mM glutamate expo-growth, proliferation and survival [10,18,20,22,32]. sure for 15 min, during which the medium was changed

Recently, ERK1 / ERK2 have been found activated after into modified EBSS (Mg -free Earle’s balanced salt relatively mild stimulation of glutamate receptors and solution (EBSS) (Gibco-BRL) supplemented with 5 mM involved in some activity-dependent functions [18,20]. glycine). For restoration, at the end of glutamate exposure, Furthermore, ERK1 / ERK2 have also been found activated cultures were rinsed twice with EBSS, and the original in some excitotoxicity-associated events, such as stroke, feeding medium was restored. For drug treatments, seizure and Alzheimer’s disease [3,19,24]. We previously PD98059 (Calbiochem, San Diego, CA, USA), MK-801 observed that ERK1 / ERK2 were transiently activated in (dizocilpine maleate, RBI, Natick, MA, USA.), L-AP-3 glutamate-induced apoptotic-like death in cultured rat (L-(1)-2-amino-3-phosphono-propionic acid, RBI), cortical neurons, and PD98059, a specific inhibitor for DNQX (6,7,-dinitroquinoxaline-2,3 (1H,4H)-dione, MEK1 / MEK2, completely inhibited such activation and Sigma), nifedipine (Sigma), EGTA (Sigma), genistein partially prevented the glutamate-induced apoptotic-like (Sigma), H-89 (Calbiochem), KN-62 (RBI), sphingosine

death [26]. Therefore, ERK1 / ERK2 might be excessively (RBI),L-NNA (N -nitro-L-argine, Sigma) were added form activated transiently and involved in the glutamate-induced 20 min before till the end of glutamate exposure,

respec-cortical neurotoxicity. tively, or in combination. In some other cases, sodium

However, little is known about the upstream cascade of orthovanadate (Sigma), okadaic acid (RBI) and cyclos-the variation of ERK1 / ERK2 in excitotoxicity. In this porin A (Sigma) were added from 20 min before until 3 h study, we searched for upstream molecules responsible for after glutamate exposure, respectively. PD98059, DNQX, the variation of ERK1 / ERK2 and the ERK1 / ERK2-me- nifedipine, genistein, sphingosine, okadaic acid and cyclos-diated apoptotic-like death in the glutamate-induced neuro- porin A were made as 5003 stocks in dimethyl sulfoxide toxicity in cultured rat cortical neurons. Several molecules (DMSO), respectively. KN-62 was made as 3003 stocks were investigated, including three subtypes of glutamate in methanol. Other drugs were made as 2003 stocks in

receptors, L-type voltage-gated Ca channel (L-VGCC), water. Vehicle controls were treated only with 50 mM NOS, some protein kinases and protein phosphatases, each glutamate and vehicle (0.5% water or 0.2% DMSO or of which has been implicated involved either in excitotox- 0.3% methanol) in modified EBSS. Sham controls were icity or in the regulation of ERK1 / ERK2 activation in treated only with modified EBSS.

other cases.

2.3. Cell extracts preparation and Western immunoblot

2. Materials and methods Cultured cells were rinsed with PBS, scraped off the

wells. Each sample was pooled from two wells

(approxi-7

2.1. Neuronal cultures mately 1.2310 cells) and homogenized in 160ml ice-cold

buffer (50 mM 3-(N-morpholino) propane-sulfonic acid, Cortical neuronal cultures were prepared from 17-day- MOPS, pH 7.4), 0.5 mM dithiothreitol, 2 mM sodium old Sprague–Dawley rat embryos as previously described orthovanadate, 0.5 mM EDTA, 1 mM EGTA, 0.5 mM [6]. Briefly, neocortex was meticulously isolated in ice- ouabain, 1 mM phenylmethylsulfonyl fluoride, 0.5 mM cold high glucose Dulbecco’s modified Eagle medium leupeptin and 0.5 mM pepstatin A, and centrifuged at (h-DMEM, Gibco-BRL, Grand Island, NY, USA). Cortical 15 0003g for 15 min at 48C. Ten ml were removed for cells were dissociated by trypsinization (0.25% (w / v) in protein concentration determination by Lowry method

21 21

Ca - and Mg -free Hank’s balanced salt solution (Gib- [31], the remaining supernatant was incubated in sample co-BRL), at 378C for 15 min), followed by gentle triturat- buffer (2% sodium dodecyl sulfate, 20% glycerol, 5% ing in plating medium (h-DMEM supplemented with 10% b-mercaptoethanol, 62.5 mM Tris–HCl, pH 6.8, and fetal bovine serum and 10% horse serum, Gibco-BRL). 0.01% bromphenol blue) at 968C for 5 min. Equal amount Cells were seeded onto poly-L-lysine (Sigma, St. Louis, of proteins (40mg) were separated by 10% SDS–PAGE by MO, USA)-coated wells or coverslips at a density of the method of Laemmli [27] and electrotransferred onto

5 2

2310 cells per cm , incubated at 378C in 5% carbon nitrocellulose filter (pore size, 0.45mm, Amersham, Buc-dioxide atmosphere, and fed by feeding medium (Neuro- kingham, UK) by the method of Towbin [46]. The filter basal Medium supplemented with 2% B-27 and 0.5 mM was probed with anti-ERK1 / ERK2 antibody (Sigma,

(3)

polyclonal, 1:10 000) or anti-active (diphosphorylated) ERK1 / ERK2 antibody (Sigma, monoclonal, 1:5000) at 48C overnight. Detection was carried out by alkaline phosphatase-conjugated goat anti-rabbit IgG (Sigma, 1:20 000) or goat anti-mouse IgG (Sigma, 1: 40 000) and developed with NBT / BCIP color substrate (Sigma). After immunoblot, bands on filter were scanned, quantitative analyzed and printed by an image analyzer affiliated with digital graphic printer (LabWorks Software, UVP, Upland, CA, USA). Protein level and activation (diphosphorylation) level of ERK1 / ERK2, based on immunoreactivities of ERK1 / ERK2 and active ERK1 / ERK2, respectively, were expressed as fold versus sham control of optical density of certain band from Western immunoblot.

2.4. Assessment of apoptotic-like cell death

Live cells grown on each coverslip (2310 cells) were incubated with 10 mg / ml fluorescent DNA binding dye DAPI (49,6-diamidino-2-phenylindole, Sigma) at 378C for 30 min, washed with PBS and excited with vertical fluorescent at 400 nm. With fluorescence collected at 455 nm, apoptotic-like cells were characterized by the presence

of condensed and fragmented nuclei, as opposed to the Fig. 1. Effects of antagonists of glutamate receptors, L-VGCC and

extracellular Ca elimination on ERK1 / ERK2 activation in

glutamate-diffuse staining observed in nonapoptotic cells. Each

induced apoptotic-like death in cultured rat cortical neurons. Thirteen

sample was pooled from three coverslips. The proportion

days in vitro cortical neurons were exposed to 50mM glutamate for 15

of apoptotic-like cells was calculated as a percentage of

min. MK-801 (20mM), L-AP-3 (1 mM), DNQX (20mM), nifedipine

total cells counted in 10 microscopic fields (3400). _{(ND, 20}_m_{M) and EGTA (5 mM) were added, respectively, 20 min}

before and during the glutamate exposure. (A) Western immunoblot at 15 min of glutamate exposure with anti-ERK1 / ERK2 antibody (top) or

2.5. Statistics

anti-diphosphorylated ERK1 / ERK2 (p-ERK1 / ERK2) antibody (bottom). (B) Quantitative representations expressed as fold versus sham control

Values were expressed as mean6S.D. from five

in-(Sham, treated with no drug) of optical density (O.D.) ERK2 or of

dependent cultures. One-way ANOVA was used. Com- _p-ERK2 _band _{from Western} _immunoblot. _Each _point _represents

a b

parisons of each group to control were by LSD (least mean6S.D. of five independent cultures. P,0.05 versus sham, P,0.05 versus vehicle control (Vehi, treated only with glutamate and vehicle).

significant difference) test. Others were by q test (New-man–Keuls test). A P value of ,0.05 was considered significant.

(1 mM) [41], or KA /AMPA-R with DNQX (20mM) [23], or L-VGCC with nifedipine (20 mM) [44]. Since the

3. Results alterations of p-ERK1 and p-ERK2 were much the same,

we directed attention to p-ERK2, which displayed a 3.1. The role of glutamate receptors,L-VGCC and relatively stronger immunoreactivity.

extracellular Ca in ERK1 /ERK2 activation in

glutamate-induced apoptotic-like death in cultured rat 3.2. The role of protein kinases and NOS in ERK1 /

cortical neurons ERK2 activation in glutamate-induced apoptotic-like

death in cultured rat cortical neurons

Protein level of ERK1 / ERK2, as indicated by ERK1 /

ERK2 immunoreactivities, was unaffected (Fig. 1A, top, As shown in Fig. 2, ERK1 / ERK2 activation was largely and B). While activation level of ERK1 / ERK2, as indi- prevented by inhibition of either PTK with genistein (100 cated by diphosphorylated ERK1 / ERK2 (p-ERK1 / ERK2), mM, also as an inhibitor of topoisomerase II kinase) [1] or was increased to 4.9-fold versus sham control at 15 min of PKC with sphingosine (10 mM) [12]. It was weakly but glutamate exposure (Fig. 1A, bottom, and B). Such still significantly prevented by inhibition of CaM kinase II activation was almost totally prevented by either blockage with KN-62 (40mM) [45], but not significantly affected by of NMDA-R with MK-801 (20 mM) or elimination of inhibition of PKA with H-89 (40mM, also as an inhibitor

extracellular Ca with EGTA (5 mM). It was not of PKG) [8], or NOS with L-NNA (100 mM) [11]. The significantly affected by blockage of mGlu-R withL-AP-3 inhibitory effect of KN-62 was significantly weaker than

(4)

Fig. 3. Effects of NMDA-R antagonist, extracellular Ca eliminator, inhibitors of NOS and protein kinases on glutamate-induced apoptotic-like death in cultured rat cortical neurons. Thirteen days in vitro cortical neurons were exposed to 50mM glutamate for 15 min. MK-801 (20mM), EGTA (5 mM),L-NNA (20mM), genistein (Gen, 100mM), KN-62 (40

mM), sphingosine (SS, 10mM) and PD98059 (PD, 50mM) were added into the medium 20 min before and during the glutamate exposure, respectively. DAPI staining at 18 h after exposure were quantitative represented as percentage of total cells counted in 10 microscopic fields (3400). Each point represents mean6S.D. of five independent cultures.

a b

P,0.05 versus sham control (Sham, with no drug treatment), P,0.05 Fig. 2. Effects of inhibitors of protein kinase and NOS on ERK1 / ERK2

versus vehicle control (Vehi, treated only with glutamate and vehicle), activation in glutamate-induced apoptotic-like death in cultured rat _c

P,0.05 versus single drug treatment. cortical neurons. Thirteen days in vitro cortical neurons were exposed to

50mM glutamate for 15 min. Genistein (Gen, 100mM), H-89 (40mM), KN-62 (40mM), sphingosine (SS, 10 mM) andL-NNA (20 mM) were

added, respectively, 20 min before and during the glutamate exposure. (A) PD98059 (50mM) [26], or NOS withL-NNA (100 mM).

Western immunoblot at 15 min of glutamate exposure with anti-ERK1 / _{A significant additive effect was observed in combined} ERK2 antibody (top) or anti-diphosphorylated ERK1 / ERK2 (p-ERK1 /

inhibition of CaM kinase II (but not PKC or PTK) and

ERK2) antibody (bottom). (B) Quantitative representations expressed as

ERK1 / ERK2. The number of total cells counted in 10

fold versus sham control (Sham, treated with no drug) of optical density

microscopic fields (3400) is about 34006360 cells.

(O.D.) of ERK2 or p-ERK2 band from Western immunoblot. Each point

represents mean6S.D. of five independent cultures. P,0.05 versus sham,

P,0.05 versus vehicle control (Vehi, treated only with glutamate and

vehicle), P,0.05 versus single drug treatment. _{3.4. The role of protein phosphatases in the reversion of}

ERK1 /ERK2 activation in glutamate-induced apoptotic-that of either sphingosine or genistein. Combined inhibi- like death in cultured rat cortical neurons

tion of CaM kinase II and PKC had a significantly additive

and completely preventive effect. Combined inhibition of As shown in Fig. 4, at 3 h after glutamate exposure, PTK and PKC had no significant additive effect. ERK1 / ERK2 activation reverted to sham control level. Such reversion was largely prevented by inhibition of both

3.3. The role of NMDA-R, extracellular Ca , PKC, protein phosphatase (PP) 1 and PP2A with 300 nM

PTK, CaM kinase II and NOS in ERK1 /ERK2-mediated okadaic acid [4], or protein tyrosine phosphatase (PTP)

apoptotic-like death in glutamate-induced excitotoxicity with sodium orthovanadate (200 mM, also as an inhibitor

in cultured rat cortical neurons of ATPase and alkaline phosphatase) [15], but not

sig-nificantly affected by PP2A with 5 nM okadaic acid [4], or As determined by DAPI staining (Fig. 3), at 18 h after PP 2B with cyclosporin A (2 mM, also as an immuno-glutamate exposure, the number of apoptotic-like cells was suppressant) [17]. Combined use of okadaic acid (300 nM) significantly increased to 82% versus sham control (11%). and sodium orthovanadate (200mM) had no significantly Such increase was significantly prevented by elimination additive effect. The concentration of each drug shown in

of extracellular Ca with EGTA (5 mM), or inhibition of Figs. 1–5, except for okadaic acid (5 nM), was responsible NMDA-R with MK-801 (20 mM), PTK with genistein for the maximal effect of the drug (data not shown). (100 mM), or PKC with sphingosine (10 mM), or CaM Modified EBSS and vehicles had little effect on ERK1 / kinase II with KN-62 (40 mM), or MEK1 / MEK2 with ERK2 activation (data not shown).

(5)

Fig. 5. Effects of inhibitors of protein phosphatases on glutamate-induced apoptotic-like death in cultured rat cortical neurons. Thirteen days in vitro cortical neurons were exposed to 50mM glutamate for 15 min. Okadaic acid (OA, 300 nM) and sodium orthovanadate (Van, 200mM) were added from 20 min before until 3 h after glutamate exposure, respectively. DAPI staining at 18 h after exposure were quantitative represented as per-centage of total cells counted in 10 microscopic fields (3400). Each point

represents mean6S.D. of five independent cultures. P,0.05 versus sham

control (Sham, with no drug treatment), P,0.05 versus vehicle control

(Vehi, treated only with glutamate and vehicle), P,0.05 versus single drug treatment.

Fig. 4. Effects of inhibitors of protein phosphatases on the reversion of _{death in cultured rat cortical neurons, and such activation} ERK1 / ERK2 activation in glutamate-induced apoptotic-like death in

was almost completely prevented by either blockage of

cultured rat cortical neurons. Thirteen days in vitro cortical neurons were 21

NMDA-R or elimination of extracellular Ca [26]. L

-exposed to 50mM glutamate for 15 min. Sodium orthovanadate (Van,

VGCC and glutamate receptors, including NMDA-R, a

-200 mM), okadaic acid (OA, 5 nM), okadaic acid (OA, 300 nM) and

cyclosporin A (Cys A, 2 mM) were added, respectively, from 20 min amino-3-hydroxy-5-methyl-4-isoxazolepropionate / kainate

before until 3 h after the glutamate exposure. (A) Western immunoblot at _{receptor (AMPA / KA-R) and metabotropic glutamate} re-3 h after glutamate exposure with anti-diphosphorylated anti-ERK1 /

ceptor (mGlu-R), have all been shown involved in ERK1 /

ERK2 antibody (top) or ERK1 / ERK2 (p-ERK1 / ERK2) antibody

(bot-ERK2 activation in some other cases [18,20]. However, in

tom). (B) Quantitative representations expressed as fold versus sham

this study, blockage of none of them, except for NMDA-R,

control (Sham, treated with no drug) of optical density (O.D.) of ERK2 or

p-ERK2 band from Western immunoblot. Each point represents significantly affected ERK1 / ERK2 activation. Therefore,

a b

mean6S.D. of five independent cultures. P,0.05 versus sham, P,0.05 _{in glutamate-induced cortical neurotoxicity ERK1 / ERK2} versus vehicle control (Vehi, treated only with glutamate and vehicle).

activation might be mainly mediated by NMDA-R-induced

influx of extracellular Ca .

3.5. The role of PP1 and PTP in glutamate-induced Stimulation of NMDA-R has been shown to mediate

excitotoxicity in cultured rat cortical neurons ERK1 / ERK2 activation, which might involve PKC and CaM kinase II in hippocampal neurons [18] and PTK in As shown in Fig. 5, inhibition of PP1 with 300 nM striatal neurons [47]. However, the upstream cascade(s) is okadaic acid and PTP with sodium orthovanadate (200 unknown in excitotoxicity. In this study, in

glutamate-mM) can promote glutamate-induced apoptotic-like death induced cortical neurotoxicity ERK1 / ERK2 activation was from 82 to 93 and 96%, respectively. Modified EBSS and largely prevented by inhibition of PKC, relatively mildly vehicles had little effect on apoptotic-like death (data not prevented by that of CaM kinase II, and combined

shown). inhibition of these two had an additive and complete

inhibitory effect. These results strongly indicate that in ERK1 / ERK2 activation a PKC-dependent pathway was mainly involved, a CaM kinase II-dependent pathway was

4. Discussion relatively mildly involved, and these two pathways were

relatively independent on each other. Moreover, such We have previously shown that ERK1 / ERK2 were activation was largely prevented by inhibition of PTK and transiently activated in glutamate-induced apoptotic-like no significant additive effect was observed in combined

(6)

inhibition of PTK and PKC. Therefore, PKC and the PTK significantly prevented the glutamate-induced apoptotic-might act on the same pathway largely contributing to like cell death and combined inhibition of CaM kinase II ERK1 / ERK2 activation. The exact relationship between (but not PKC or PTK) and ERK1 / ERK2 had a significant PTK and PKC is unclear. However, two pathways may additive effect. These results suggest that the ERK1 / ER-exist. One may be PKC–c-Src pathway. Schlaepfer et al. K2-mediated apoptotic-like death was mainly a down-has shown that PKC-mediated c-Src activity played a role stream event of PKC and PTK, but not CaM kinase II, in in integrin-induced ERK1 / ERK2 activation in NIH 3T3 glutamate-induced cortical neurotoxicity.

fibroblasts [40]. And the other may be PKC–PYK2 We have previously observed that ERK1 / ERK2 activa-pathway. Treatment of PC12 cells with agents that in- tion reverted to basal level at 3 h after glutamate exposure

creased intracellular Ca - or activated PKC-stimulated in cortical excitotoxicity [26]. Other studies also revealed PYK2 activity, and overexpression of PYK2 increased that PP1, PP2A, PP2B and PTP were involved in the ERK1 / ERK2 activity in these cells [29]. Thus, an im- reversion of ERK1 / ERK2 activation in some other cases

portant connection linking Ca , PKC, PYK2 and ERK1 / [34,35,38,43]. However, little is known about the upstream ERK2 may be present in some neuronal cells. cascade(s) of the reversion of ERK1 / ERK2 activation in CAMP-dependent protein kinase (PKA) has been shown excitotoxicity. In the present study, reversion of ERK1 / activated after NMDA-R [7], although there are numerous ERK2 activation was largely prevented by 300 nM okadaic

examples of antagonism and synergism between Ca and acid, supposed to inhibit both PP1 and PP2A [4], but not cAMP signaling. PKA has also been shown to interfere by 5 nM okadaic acid, supposed to inhibit only PP2A [4]. with ERK1 / ERK2 activation in many cell types, although Thus, PP1, but not PP2A, was mainly involved in rever-the interactions between cAMP and ERK1 / ERK2 sig- sion of ERK1 / ERK2 activation. Moreover, such reversion naling system are complex and variable [20]. However, our was also largely prevented by inhibition of PTP and results suggest that PKA be not involved in ERK1 / ERK2 combined inhibition of PP1 and PTP had no significant

activation in excitotoxicity. additive effect. Therefore, PP1 and PTP might act on the

Another possible mediator of ERK1 / ERK2 activation in same signaling pathway largely contributing to the rever-excitotoxicity is NOS. NOS-produced nitric oxide (NO) sion of ERK1 / ERK2 activation in excitotoxicity.

has been shown involved not only in glutamate-induced Since we have also showed that either PKC or PTK was ERK1 / ERK2 activation [49,50] but also in excitotoxicity mainly involved in ERK1 / ERK2-mediated apoptotic-like [2,36]. Unexpectedly, we did not observed any effect of death in glutamate-induced neurotoxicity, we further de-L-NNA, a NOS inhibitor, on ERK1 / ERK2 activation in termined the role of protein phosphatases in glutamate-glutamate-induced cortical neurotoxicity, although we induced apoptotic-like death. Our results showed that found that L-NNA provided significant neuroprotection inhibition of either PP 1 or PTP could not only prevent the against the glutamate-induced excitotoxicity. Our finding is reversion of ERK1 / ERK2 activation but also promote contrary to that by Yun et al. [50]. The basis for this glutamate-induced apoptosis. This would suggest that discrepancy is unclear, but presumably reflects differences either PP 1 or PTP be involved in the restrain of gluta-in culture conditions. We used a serum-free medium mate-induced neurotoxicity. Therefore, they might have an specific to neuron culture, while they used a serum-con- opposite effect to PKC or PTK on glutamate-induced taining medium. In fact, the time course of ERK1 / ERK2 apoptotic-like death through regulating ERK1 / ERK2 acti-activation in our study is also different from theirs. This vation. However, it is still unclear whether these phosphat-may also reflects the different mechanism in ERK1 / ERK2 ases act just the reverse action(s) of those kinases. activation. They found that ERK1 / ERK2 is not fully In conclusion, we have clearly shown that in glutamate-activated until 10 min after either NO, or NMDA, or induced cortical neurotoxicity ERK1 / ERK2 activation was glutamate treatment, while we found that ERK1 / ERK2 mainly mediated by NMDA-R-induced influx of

extracel-21

activation peaked during glutamate treatment and was lular Ca . Subsequently, a pathway dependent on both rapidly recovered after glutamate treatment. The time PKC and PTK was mainly involved, which was also course of ERK1 / ERK2 activation in our study is con- mainly responsible for ERK1 / ERK2-mediated apoptotic-sistent with that reported by Fukunaga and Miyamoto [18]. like death, and a CaM kinase II-dependent pathway was We have previously observed that ERK1 / ERK2 activa- relatively mildly involved. Reversion of ERK1 / ERK2 tion was involved in the glutamate-induced cortical neuro- activation was mainly mediated by a pathway dependent toxicity [26]. Some other studies have also shown in- on both PP1 and PTP, which might be involved in the volvements of PKC, PTK and CaM kinase II in excitotox- restrain of glutamate-induced apoptotic-like death. Further icity [13,14,21,37]. Therefore, we determined whether study is ongoing to clarify the exact relationship between protein kinases, which has been shown involved in ERK1 / PKC and PTK and between PP1 and PTP in regulating ERK2 activation in excitotoxicity, are also responsible for ERK1 / ERK2-mediated apoptosis, the exact PTK mem-ERK1 / ERK2-mediated apoptotic-like death. Inhibition of ber(s) involved, and target molecules of those protein ERK1 / ERK2, PKC, PTK and CaM kinase II, respectively, kinases and phosphatases.

(7)

mechanisms for neuronal survival, differentiation, and plasticity?,

Acknowledgements

Neuron 16 (1996) 233–236.

[17] D.A. Fruman, C.B. Klee, B.E. Bierer, S.J. Burakoff, Calcineurin

The authors thank Professor Zhida Xu for excellent _{phosphatase activity in T lymphocytes is inhibited by FK 506 and} assistance in assessment of the apoptotic-like death. cyclosporin A, Proc. Natl. Acad. Sci. USA 89 (1992) 3686–3690. [18] K. Fukunaga, E. Miyamoto, Role of MAP kinase in neurons, Mol.

Neurobiol. 16 (1998) 79–95.

[19] P. Gass, M. Kiessling, H. Bading, Regionally selective stimulation

References _{of mitogen activated protein (MAP) kinase tyrosine phosphorylation}

after generalized seizures in the rat brain, Neurosci. Lett. 162 (1993) [1] T. Akiyama, J. Ishida, S. Nakagawa, H. Ogawara, S. Watanabe, N. 39–42.

Itoh, M. Shibuya, Y. Fukami, Genistein, a specific inhibitor of [20] S.S. Grewal, R.D. York, P.J.S. Stork, Extracellular-signal-regulated tyrosine-specific protein kinases, J. Biol. Chem. 262 (1987) 5592– kinase signalling in neurons, Curr. Opin. Neurobiol. 9 (1999)

5595. 544–553.

[2] A. Almeida, J.P. Golanos, J.M. Medina, Nitric oxide mediates [21] I. Hajimohammadreza, A.W. Probert, L.L. Coughenour, S.A. glutamate induced mitochondrial depolarization in rat cortical Borosky, F.W. Marcoux, P.A. Boxer, K.K. Wang, A specific inhibitor neurons, Brain Res. 816 (1999) 580–586. of calcium / calmodulin-dependent protein kinase-II provides

neuro-¨

[3] T. Arendt, M. Holzer, A. Grobmann, D. Zedlick, M.K. Bruckner, protection against NMDA- and hypoxia / hypoglycemia-induced cell Increased expression and subcellular translocation of the mitogen death, J. Neurosci. 15 (1995) 4093–4101.

activated protein kinase kinase and mitogen-activated protein kinase [22] C.S. Hill, R. Treisman, Transcriptional regulation by extracellular in Alzheimer’s disease, Neuroscience 68 (1995) 5–18. signals: mechanisms and specificity, Cell 80 (1995) 199–211. [4] C. Bialojan, A. Takai, Inhibitory effect of a marine-sponge toxin, [23] T. Honore, S.N. Davies, J. Drejer, E.J. Fletcher, P. Jacobsen, D.

okadaic acid, on protein phosphatases, specificity and kinetics, Lodge, F.E. Nielsen, Quinoxalinediones: potent competitive non-Biochem. J. 256 (1988) 283–290. NMDA glutamate receptor antagonists, Science 241 (1988) 701– [5] E. Bonfoco, D. Krainc, M. Ankarcrona, P. Nicotera, S.A. Lipton, 703.

Apoptosis and necrosis: two distinct events induced, respectively, by [24] B.R. Hu, T. Wieloch, Tyrosine phosphorylation and activation of mild and intense insults with N-methyl-D-aspartate or nitric oxide / mitogen-activated protein kinase in the rat brain following transient superoxide in cortical cell cultures, Proc. Natl. Acad. Sci. USA 76 cerebral ischemia, J. Neurochem. 62 (1994) 1357–1367.

(1995) 7162–7166. [25] J. Ikeda, S. Terakawa, S. Murota, I. Morita, K. Hirakawa, Nuclear [6] G.J. Brewer, Serum-free B27 / neurobasal medium supports differen- disintegration as a leading step of glutamate excitotoxicity in brain

tiated growth of neurons from the striatum, substantia nigra, septum, neurons, J. Neurosci. Res. 43 (1996) 613–622.

cerebral cortex, cerebellum, and dentate gyrus, J. Neurosci. Res. 42 [26] Q. Jiang, Z. Gu, G. Zhang, G. Jin, Diphosphorylation and in-(1995) 674–683. volvement of extracellular signal-regulated kinases (ERK1 / 2) in [7] D.M. Chetkovich, J.D. Sweatt, NMDA receptor activation increases glutamate-induced apoptotic-like death in cultured rat cortical

cyclic AMP in area CA1 of the hippocampus via calcium / cal- neurons, Brain Res. 857 (2000) 71–77.

modulin stimulation of adenylyl cyclase, J. Neurochem. 61 (1993) [27] U.K. Laemmli, Cleavage of structural proteins during the assembly 1933–1942. of the head of bacteriophage T4, Nature 227 (1970) 680–685. [8] T. Chijiwa, A. Mishima, M. Hagiwara, M. Sano, K. Hayashi, T. [28] M. Leist, P. Nicotera, Calcium and neuronal death, Rev. Physiol.

Inoue, K. Naito, T. Toshioka, H. Hidaka, Inhibition of forskolin- Pharmacol. Biochem. 132 (1997) 79–125.

induced neurite outgrowth and protein phosphorylation by a newly [29] S. Lev, H. Moreno, R. Martinez, P. Canoll, E. Peles, J.M. Musac-synthesized selective inhibitor of cyclic AMP-dependent protein chio, G.D. Plowman, B. Rudy, J. Schlessinger, Protein tyrosine kinase, N-[2-( p-bromocinnamylamino)ethyl]-5-isoquinolinesul- kinase PYK2 involved in Ca (21)-induced regulation of ion channel fonamide (H-89), of PC12D pheochromocytoma cells, J. Biol. and MAP kinase functions, Nature 376 (1995) 737–745.

Chem. 265 (1990) 5267–5272. [30] S.A. Lipton, P.A. Rosenberg, Excitatory amino acids as a final [9] D.W. Choi, Calcium: still center-stage in hypoxic-ischemic neuronal common pathway for neurologic disorders, New Engl. J. Med. 330

death, Trends Neurosci. 18 (1995) 58–60. (1994) 613–622.

[10] R.J. Davis, The mitogen-activated protein kinase signal transduction [31] O.H. Lowry, H.J. Rosebrough, A.L. Farr, R.J. Randall, Protein pathway, J. Biol. Chem. 268 (1993) 14553–14556. measurement with Folin-phenol reagent, J. Biol. Chem. 193 (1951) [11] M.A. Dwyer, D.S. Bredt, S.H. Snyder, Nitric oxide synthase: 265–275.

irreversible inhibition byL-NG-nitroarginine in brain in vitro and in [32] C.J. Marshall, Specificity of receptor tyrosine kinase signaling: vivo, Biochem. Biophys. Res. Commun. 176 (1991) 1136–1141. transient versus sustained extracellular signal-regulated kinase acti-[12] M. Faucher, N. Girones, Y.A. Hannun, R.M. Bell, R.J. Davis, vation, Cell 80 (1995) 179–185.

Regulation of the epidermal growth factor receptor phosphorylation [33] S. Nakielny, P. Cohen, J. Wu, T. Sturgill, MAP kinase activator form state by sphingosine in A431 human epidermoid carcinoma cells, J. insulin-stimulated skeletal muscle is a protein threonine / tyrosine Biol. Chem. 263 (1988) 5319–5327. kinase, EMBO J. 11 (1992) 2123–2129.

[13] M. Favaron, H. Manev, R. Siman, M. Bertolino, A.M. Szekely, G. [34] E.D. Norman, E. Thiels, G. Barrionuevo, E. Klann, Long-term DeErausquin, A. Guiditti, E. Costa, Down-regulation of protein depression in the hippocampus in vivo is associated with protein kinase C protects cerebellar granule neurons in primary culture from phosphatase-dependent alterations in extracellular signal-regulated glutamate-induced neuronal death, Proc. Natl. Acad. Sci. USA 87 kinase, J. Neurochem. 74 (2000) 192–198.

(1990) 1983–1987. [35] M. Ogata, M. Oh-hora, A. Kosugi, T. Hamaoka, Inactivation of [14] V. Felipo, M.D. Minana, S. Grisolia, Inhibitors of protein kinase C mitogen-activated protein kinases by a mammalian tyrosine-specific prevent the toxicity of glutamate in primary neuronal cultures, Brain phosphatase, PTPBR7, Biochem. Biophys. Res. Commun. 256

Res. 604 (1993) 192–196. (1999) 52–56.

[15] I. Figiel, L. Kaczmarek, Orthovanadate induces cell death in rat [36] T. Ohmori, Y. Hirashima, M. Kurimoto, S. Endo, A. Takaku, In vitro dentate gyrus primary culture, Neuroreport 8 (1997) 2465–2470. hypoxia of cortical and hippocampal CA1 neurons: glutamate, nitric

(8)

selective neural death in CA1 neurons, Brain Res. 743 (1996) [43] J. Tian, M. Karin, Stimulation of Elk1 transcriptional activity by 109–115. mitogen-activated protein kinases is negatively regulated by protein [37] T. Ohtsuki, M. Matsumoto, K. Kitagawa, T. Mabuchi, K. Mandai, phosphatase 2B (calcineurin), J. Biol. Chem. 274 (1999) 15173–

K. Matsushita, K. Kuwabara, M. Tagaya, S. Ogawa, H. Ueda, T. 15180.

Kamada, T. Yanagihara, Delayed neuronal death in ischemic hip- [44] E.C. Toescu, Activity of voltage-operated calcium channels in rat pocampus involves stimulation of protein tyrosine phosphorylation, cerebellar granule neurons and neuronal survival, Neuroscience 94 Am. J. Physiol. 271 (1996) C1085–1097. (1999) 561–570.

[38] N. Parameswaran, P. Nambi, C.S. Hall, D.P. Brooks, W.S. Spielman, [45] H. Tokumitsu, T. Chijiwa, M. Hagiwara, A. Mizutani, M. Terasawa, Adrenomedullin decreases extracellular signal-regulated kinase ac- H. Hidaka, KN-62, 1-[N,O-bis(5-isoquinolinesulfonyl)-N-methyl-L -tivity through an increase in protein phosphatase-2A ac-tivity in tyrosyl]-4-phenylpiperazi, J. Biol. Chem. 265 (1990) 4315–4320. mesangial cells, Eur. J. Pharmacol. 388 (2000) 133–138. [46] H. Towbin, T. Stachelin, J. Gordon, Electrophoretic transfer of [39] G. Rusanescu, H. Qi, S.M. Thomas, J.S. Brugge, S. Halegoua, proteins from polyacrylamide gels to nitrocellulose sheets. Pro-Calcium influx induces neurite growth through a Src-Ras signaling cedure and some applications, Proc. Natl. Acad. Sci. USA 76 (1979) cassette, Neuron 15 (1995) 1415–1425. 4350–4354.

[40] D.D. Schlaepfer, K.C. Jones, T. Hunter, Multiple Grb2-mediated [47] S.R. Vincent, M. Sebben, A. Dumuis, J. Bockaert, Neurotransmitter integrin-stimulated signaling pathways to ERK2 / mitogen-activated regulation of MAP kinase signaling in striatal neurons in primary protein kinase: summation of both c-Src- and focal adhesion kinase- culture, Synapse 29 (1998) 29–36.

initiated tyrosine phosphorylation events, Mol. Cell. Biol. 18 (1998) [48] A.M. Wood, K.R. Bristow, N-Methyl-D-aspartate receptor desensiti-2571–2585. sation is neuroprotective by inhibiting glutamate-induced apoptotic-[41] D.D. Schoepp, B.G. Johnson, E.C. Smith, L.A. McQuaid, like death, J. Neurochem. 70 (1998) 677–687.

Stereoselectivity and mode of inhibition of phosphoinositide-cou- [49] H.Y. Yun, V.L. Dawson, T.M. Dawson, Glutamate-stimulated cal-pled excitatory amino acid receptors by 2-amino-3-phosphonop- cium activation of Ras / Erk pathway mediated by nitric oxide, ropionic acid, Mol. Pharmacol. 38 (1990) 222–228. Diabetes Res. Clin. Pract. 45 (1999) 113–115.

[42] R. Seger, N.G. Ahn, J. Posada, E.S. Munar, A.M. Jensen, J.A. [50] H.Y. Yun, M. Gonzalez-Zulueta, V.L. Dawson, T.M. Dawson, Nitric Cooper, M.H. Cobb, E.G. Krebs, Purification and characterization of oxide mediates N-methyl-D-aspartate receptor-induced activation of mitogen-activated protein kinase activator(s) from epidermal growth p21ras, Proc. Natl. Acad. Sci. USA 95 (1998) 5773–5778. factor-stimulated A431 cells, J. Biol. Chem. 267 (1992) 14373– [51] C.F. Zorumski, J.W. Olney, Excitotoxic neuronal damage and

(1)

polyclonal, 1:10 000) or anti-active (diphosphorylated)

ERK1 / ERK2 antibody (Sigma, monoclonal, 1:5000) at

4

8 C overnight. Detection was carried out by alkaline

phosphatase-conjugated

goat

anti-rabbit

IgG

(Sigma,

1:20 000) or goat anti-mouse IgG (Sigma, 1: 40 000) and

developed with NBT / BCIP color substrate (Sigma). After

immunoblot, bands on filter were scanned, quantitative

analyzed and printed by an image analyzer affiliated with

digital graphic printer (LabWorks Software, UVP, Upland,

CA, USA). Protein level and activation (diphosphorylation)

level of ERK1 / ERK2, based on immunoreactivities of

ERK1 / ERK2 and active ERK1 / ERK2, respectively, were

expressed as fold versus sham control of optical density of

certain band from Western immunoblot.

2.4. Assessment of apoptotic-like cell death

Live cells grown on each coverslip (2

3 10 cells) were

incubated with 10

m

g / ml fluorescent DNA binding dye

DAPI (4

9 ,6-diamidino-2-phenylindole, Sigma) at 37

8 C for

30 min, washed with PBS and excited with vertical

fluorescent at 400 nm. With fluorescence collected at 455

nm, apoptotic-like cells were characterized by the presence

of condensed and fragmented nuclei, as opposed to the

Fig. 1. Effects of antagonists of glutamate receptors, L-VGCC and

extracellular Ca elimination on ERK1 / ERK2 activation in

glutamate-diffuse staining observed in nonapoptotic cells. Each

induced apoptotic-like death in cultured rat cortical neurons. Thirteen

sample was pooled from three coverslips. The proportion

days in vitro cortical neurons were exposed to 50mM glutamate for 15

of apoptotic-like cells was calculated as a percentage of

min. MK-801 (20mM), L-AP-3 (1 mM), DNQX (20mM), nifedipine

total cells counted in 10 microscopic fields (

3 400).

_{(ND, 20}_m_{M) and EGTA (5 mM) were added, respectively, 20 min} before and during the glutamate exposure. (A) Western immunoblot at 15 min of glutamate exposure with anti-ERK1 / ERK2 antibody (top) or

2.5. Statistics

anti-diphosphorylated ERK1 / ERK2 (p-ERK1 / ERK2) antibody (bottom). (B) Quantitative representations expressed as fold versus sham control

Values were expressed as mean

6 S.D. from five

in-(Sham, treated with no drug) of optical density (O.D.) ERK2 or of

dependent cultures. One-way ANOVA was used. Com-

_p-ERK2 _band _{from Western} _immunoblot. _Each _point _represents

a b

parisons of each group to control were by LSD (least

mean6S.D. of five independent cultures. P,0.05 versus sham, P,0.05 versus vehicle control (Vehi, treated only with glutamate and vehicle).

significant difference) test. Others were by q test

(New-man–Keuls test). A P value of

,

0.05 was considered

significant.

(1 mM) [41], or KA /AMPA-R with DNQX (20

m

M) [23],

or

-VGCC with nifedipine (20

m

M) [44]. Since the

3. Results

alterations of p-ERK1 and p-ERK2 were much the same,

we directed attention to p-ERK2, which displayed a

3.1. The role of glutamate receptors,

-VGCC and

relatively stronger immunoreactivity.

extracellular Ca

in ERK1 /ERK2 activation in

glutamate-induced apoptotic-like death in cultured rat

3.2. The role of protein kinases and NOS in ERK1 /

cortical neurons

ERK2 activation in glutamate-induced apoptotic-like

death in cultured rat cortical neurons

Protein level of ERK1 / ERK2, as indicated by ERK1 /

ERK2 immunoreactivities, was unaffected (Fig. 1A, top,

As shown in Fig. 2, ERK1 / ERK2 activation was largely

and B). While activation level of ERK1 / ERK2, as indi-

prevented by inhibition of either PTK with genistein (100

cated by diphosphorylated ERK1 / ERK2 (p-ERK1 / ERK2),

m

M, also as an inhibitor of topoisomerase II kinase) [1] or

was increased to 4.9-fold versus sham control at 15 min of

PKC with sphingosine (10

m

M) [12]. It was weakly but

glutamate exposure (Fig. 1A, bottom, and B). Such

still significantly prevented by inhibition of CaM kinase II

activation was almost totally prevented by either blockage

with KN-62 (40

m

M) [45], but not significantly affected by

of NMDA-R with MK-801 (20

m

M) or elimination of

inhibition of PKA with H-89 (40

m

M, also as an inhibitor

extracellular Ca

with EGTA (5 mM). It was not

of PKG) [8], or NOS with

-NNA (100

m

M) [11]. The

(2)

a b

P,0.05 versus sham control (Sham, with no drug treatment), P,0.05 Fig. 2. Effects of inhibitors of protein kinase and NOS on ERK1 / ERK2

versus vehicle control (Vehi, treated only with glutamate and vehicle), activation in glutamate-induced apoptotic-like death in cultured rat _c

P,0.05 versus single drug treatment. cortical neurons. Thirteen days in vitro cortical neurons were exposed to

50mM glutamate for 15 min. Genistein (Gen, 100mM), H-89 (40mM), KN-62 (40mM), sphingosine (SS, 10 mM) andL-NNA (20 mM) were

added, respectively, 20 min before and during the glutamate exposure. (A)

PD98059 (50

m

M) [26], or NOS with

-NNA (100

m

M).

Western immunoblot at 15 min of glutamate exposure with anti-ERK1 /

_{A significant additive effect was observed in combined}

ERK2 antibody (top) or anti-diphosphorylated ERK1 / ERK2 (p-ERK1 /

inhibition of CaM kinase II (but not PKC or PTK) and

ERK2) antibody (bottom). (B) Quantitative representations expressed as

ERK1 / ERK2. The number of total cells counted in 10

fold versus sham control (Sham, treated with no drug) of optical density

microscopic fields (

3 400) is about 3400

6 360 cells.

(O.D.) of ERK2 or p-ERK2 band from Western immunoblot. Each point

represents mean6S.D. of five independent cultures. P,0.05 versus sham,

P,0.05 versus vehicle control (Vehi, treated only with glutamate and

vehicle), P,0.05 versus single drug treatment.

_{3.4. The role of protein phosphatases in the reversion of}

ERK1 /ERK2 activation in glutamate-induced

apoptotic-that of either sphingosine or genistein. Combined inhibi-

like death in cultured rat cortical neurons

tion of CaM kinase II and PKC had a significantly additive

and completely preventive effect. Combined inhibition of

As shown in Fig. 4, at 3 h after glutamate exposure,

PTK and PKC had no significant additive effect.

ERK1 / ERK2 activation reverted to sham control level.

Such reversion was largely prevented by inhibition of both

3.3. The role of NMDA-R, extracellular Ca

, PKC,

protein phosphatase (PP) 1 and PP2A with 300 nM

PTK, CaM kinase II and NOS in ERK1 /ERK2-mediated

okadaic acid [4], or protein tyrosine phosphatase (PTP)

apoptotic-like death in glutamate-induced excitotoxicity

with sodium orthovanadate (200

m

M, also as an inhibitor

in cultured rat cortical neurons

of ATPase and alkaline phosphatase) [15], but not

sig-nificantly affected by PP2A with 5 nM okadaic acid [4], or

As determined by DAPI staining (Fig. 3), at 18 h after

PP 2B with cyclosporin A (2

m

M, also as an

immuno-glutamate exposure, the number of apoptotic-like cells was

suppressant) [17]. Combined use of okadaic acid (300 nM)

significantly increased to 82% versus sham control (11%).

and sodium orthovanadate (200

m

M) had no significantly

Such increase was significantly prevented by elimination

additive effect. The concentration of each drug shown in

of extracellular Ca

with EGTA (5 mM), or inhibition of

Figs. 1–5, except for okadaic acid (5 nM), was responsible

NMDA-R with MK-801 (20

m

M), PTK with genistein

for the maximal effect of the drug (data not shown).

(100

m

M), or PKC with sphingosine (10

m

M), or CaM

Modified EBSS and vehicles had little effect on ERK1 /

kinase II with KN-62 (40

m

M), or MEK1 / MEK2 with

ERK2 activation (data not shown).

(3)

represents mean6S.D. of five independent cultures. P,0.05 versus sham

control (Sham, with no drug treatment), P,0.05 versus vehicle control

(Vehi, treated only with glutamate and vehicle), P,0.05 versus single drug treatment.

Fig. 4. Effects of inhibitors of protein phosphatases on the reversion of

_{death in cultured rat cortical neurons, and such activation}

ERK1 / ERK2 activation in glutamate-induced apoptotic-like death in

was almost completely prevented by either blockage of

cultured rat cortical neurons. Thirteen days in vitro cortical neurons were 21

NMDA-R or elimination of extracellular Ca

[26].

-exposed to 50mM glutamate for 15 min. Sodium orthovanadate (Van,

VGCC and glutamate receptors, including NMDA-R,

a

-200 mM), okadaic acid (OA, 5 nM), okadaic acid (OA, 300 nM) and

cyclosporin A (Cys A, 2 mM) were added, respectively, from 20 min

amino-3-hydroxy-5-methyl-4-isoxazolepropionate / kainate

before until 3 h after the glutamate exposure. (A) Western immunoblot at

_{receptor (AMPA / KA-R) and metabotropic glutamate}

re-3 h after glutamate exposure with anti-diphosphorylated anti-ERK1 /

ceptor (mGlu-R), have all been shown involved in ERK1 /

ERK2 antibody (top) or ERK1 / ERK2 (p-ERK1 / ERK2) antibody

(bot-ERK2 activation in some other cases [18,20]. However, in

tom). (B) Quantitative representations expressed as fold versus sham

this study, blockage of none of them, except for NMDA-R,

control (Sham, treated with no drug) of optical density (O.D.) of ERK2 or

p-ERK2 band from Western immunoblot. Each point represents

significantly affected ERK1 / ERK2 activation. Therefore,

a b

mean6S.D. of five independent cultures. P,0.05 versus sham, P,0.05

_{in glutamate-induced cortical neurotoxicity ERK1 / ERK2}

versus vehicle control (Vehi, treated only with glutamate and vehicle).

activation might be mainly mediated by NMDA-R-induced

influx of extracellular Ca

.

3.5. The role of PP1 and PTP in glutamate-induced

Stimulation of NMDA-R has been shown to mediate

excitotoxicity in cultured rat cortical neurons

ERK1 / ERK2 activation, which might involve PKC and

CaM kinase II in hippocampal neurons [18] and PTK in

As shown in Fig. 5, inhibition of PP1 with 300 nM

striatal neurons [47]. However, the upstream cascade(s) is

okadaic acid and PTP with sodium orthovanadate (200

unknown in excitotoxicity. In this study, in

glutamate-m

M) can promote glutamate-induced apoptotic-like death

induced cortical neurotoxicity ERK1 / ERK2 activation was

from 82 to 93 and 96%, respectively. Modified EBSS and

largely prevented by inhibition of PKC, relatively mildly

vehicles had little effect on apoptotic-like death (data not

prevented by that of CaM kinase II, and combined

shown).

inhibition of these two had an additive and complete

inhibitory effect. These results strongly indicate that in

ERK1 / ERK2 activation a PKC-dependent pathway was

mainly involved, a CaM kinase II-dependent pathway was

4. Discussion

relatively mildly involved, and these two pathways were

relatively independent on each other. Moreover, such

We have previously shown that ERK1 / ERK2 were

activation was largely prevented by inhibition of PTK and

transiently activated in glutamate-induced apoptotic-like

no significant additive effect was observed in combined

(4)

ERK1 / ERK2 activation. The exact relationship between

(but not PKC or PTK) and ERK1 / ERK2 had a significant

PTK and PKC is unclear. However, two pathways may

additive effect. These results suggest that the ERK1 /

ER-exist. One may be PKC–c-Src pathway. Schlaepfer et al.

K2-mediated apoptotic-like death was mainly a

down-has shown that PKC-mediated c-Src activity played a role

stream event of PKC and PTK, but not CaM kinase II, in

in integrin-induced ERK1 / ERK2 activation in NIH 3T3

glutamate-induced cortical neurotoxicity.

fibroblasts [40]. And the other may be PKC–PYK2

We have previously observed that ERK1 / ERK2

activa-pathway. Treatment of PC12 cells with agents that in-

tion reverted to basal level at 3 h after glutamate exposure

creased intracellular Ca

- or activated PKC-stimulated

in cortical excitotoxicity [26]. Other studies also revealed

PYK2 activity, and overexpression of PYK2 increased

that PP1, PP2A, PP2B and PTP were involved in the

ERK1 / ERK2 activity in these cells [29]. Thus, an im-

reversion of ERK1 / ERK2 activation in some other cases

portant connection linking Ca

, PKC, PYK2 and ERK1 /

[34,35,38,43]. However, little is known about the upstream

ERK2 may be present in some neuronal cells.

cascade(s) of the reversion of ERK1 / ERK2 activation in

CAMP-dependent protein kinase (PKA) has been shown

excitotoxicity. In the present study, reversion of ERK1 /

activated after NMDA-R [7], although there are numerous

ERK2 activation was largely prevented by 300 nM okadaic

examples of antagonism and synergism between Ca

and

acid, supposed to inhibit both PP1 and PP2A [4], but not

cAMP signaling. PKA has also been shown to interfere

by 5 nM okadaic acid, supposed to inhibit only PP2A [4].

with ERK1 / ERK2 activation in many cell types, although

Thus, PP1, but not PP2A, was mainly involved in

rever-the interactions between cAMP and ERK1 / ERK2 sig-

sion of ERK1 / ERK2 activation. Moreover, such reversion

naling system are complex and variable [20]. However, our

was also largely prevented by inhibition of PTP and

results suggest that PKA be not involved in ERK1 / ERK2

combined inhibition of PP1 and PTP had no significant

activation in excitotoxicity.

additive effect. Therefore, PP1 and PTP might act on the

Another possible mediator of ERK1 / ERK2 activation in

same signaling pathway largely contributing to the

rever-excitotoxicity is NOS. NOS-produced nitric oxide (NO)

sion of ERK1 / ERK2 activation in excitotoxicity.

has been shown involved not only in glutamate-induced

Since we have also showed that either PKC or PTK was

ERK1 / ERK2 activation [49,50] but also in excitotoxicity

mainly involved in ERK1 / ERK2-mediated apoptotic-like

[2,36]. Unexpectedly, we did not observed any effect of

death in glutamate-induced neurotoxicity, we further

de-L

-NNA, a NOS inhibitor, on ERK1 / ERK2 activation in

termined the role of protein phosphatases in

glutamate-glutamate-induced cortical neurotoxicity, although we

induced apoptotic-like death. Our results showed that

found that

-NNA provided significant neuroprotection

inhibition of either PP 1 or PTP could not only prevent the

against the glutamate-induced excitotoxicity. Our finding is

reversion of ERK1 / ERK2 activation but also promote

contrary to that by Yun et al. [50]. The basis for this

glutamate-induced apoptosis. This would suggest that

discrepancy is unclear, but presumably reflects differences

either PP 1 or PTP be involved in the restrain of

gluta-in culture conditions. We used a serum-free medium

mate-induced neurotoxicity. Therefore, they might have an

specific to neuron culture, while they used a serum-con-

opposite effect to PKC or PTK on glutamate-induced

taining medium. In fact, the time course of ERK1 / ERK2

apoptotic-like death through regulating ERK1 / ERK2

acti-activation in our study is also different from theirs. This

vation. However, it is still unclear whether these

phosphat-may also reflects the different mechanism in ERK1 / ERK2

ases act just the reverse action(s) of those kinases.

activation. They found that ERK1 / ERK2 is not fully

In conclusion, we have clearly shown that in

glutamate-activated until 10 min after either NO, or NMDA, or

induced cortical neurotoxicity ERK1 / ERK2 activation was

glutamate treatment, while we found that ERK1 / ERK2

mainly mediated by NMDA-R-induced influx of

extracel-21

activation peaked during glutamate treatment and was

lular Ca

. Subsequently, a pathway dependent on both

rapidly recovered after glutamate treatment. The time

PKC and PTK was mainly involved, which was also

course of ERK1 / ERK2 activation in our study is con-

mainly responsible for ERK1 / ERK2-mediated

apoptotic-sistent with that reported by Fukunaga and Miyamoto [18].

like death, and a CaM kinase II-dependent pathway was

We have previously observed that ERK1 / ERK2 activa-

relatively mildly involved. Reversion of ERK1 / ERK2

tion was involved in the glutamate-induced cortical neuro-

activation was mainly mediated by a pathway dependent

toxicity [26]. Some other studies have also shown in-

on both PP1 and PTP, which might be involved in the

volvements of PKC, PTK and CaM kinase II in excitotox-

restrain of glutamate-induced apoptotic-like death. Further

icity [13,14,21,37]. Therefore, we determined whether

study is ongoing to clarify the exact relationship between

protein kinases, which has been shown involved in ERK1 /

PKC and PTK and between PP1 and PTP in regulating

ERK2 activation in excitotoxicity, are also responsible for

ERK1 / ERK2-mediated apoptosis, the exact PTK

mem-ERK1 / ERK2-mediated apoptotic-like death. Inhibition of

ber(s) involved, and target molecules of those protein

ERK1 / ERK2, PKC, PTK and CaM kinase II, respectively,

kinases and phosphatases.

(5)

mechanisms for neuronal survival, differentiation, and plasticity?,

Acknowledgements

Neuron 16 (1996) 233–236.

[17] D.A. Fruman, C.B. Klee, B.E. Bierer, S.J. Burakoff, Calcineurin

The authors thank Professor Zhida Xu for excellent

_{phosphatase activity in T lymphocytes is inhibited by FK 506 and}

assistance in assessment of the apoptotic-like death.

cyclosporin A, Proc. Natl. Acad. Sci. USA 89 (1992) 3686–3690. [18] K. Fukunaga, E. Miyamoto, Role of MAP kinase in neurons, Mol.