ACA

MAa

Ub

i

Addeatctatcamtti(aiTetctttdt(ef

*BPEAtitpsdmNpstyT

Please cite this article in press as: Matos M, et al., Amyloid-beta peptide decreases glutamate uptake in cultured astrocytes: Involve-ment of oxidative stress and mitogen-activated protein kinase cascades, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.022

Neuroscience xx (2008) xxx

0d

ARTICLE IN PRESS

MYLOID-BETA PEPTIDE DECREASES GLUTAMATE UPTAKE INULTURED ASTROCYTES: INVOLVEMENT OF OXIDATIVE STRESS

ND MITOGEN-ACTIVATED PROTEIN KINASE CASCADES

eAsdttME

Ks

Acml2cgopppahtAAafs2cavdiFn(bA

igacH

. MATOS,a E. AUGUSTO,a C. R. OLIVEIRAa,b

ND P. AGOSTINHOa,b*

Center for Neuroscience and Cell Biology, Department of Zoology,niversity of Coimbra, 3004 Coimbra, Portugal

Institute of Biochemistry and Faculty of Medicine, University of Co-mbra, 3004 Coimbra, Portugal

bstract—Alzheimer’s disease (AD) is a progressive neuro-egenerative disorder primarily characterized by excessiveeposition of amyloid-� (A�) peptides in the brain. One of thearliest neuropathological changes in AD is the presence ofhigh number of reactive astrocytes at sites of A� deposi-

ion. Disturbance of glutamatergic neurotransmission andonsequent excitotoxicity is also believed as implicated inhe progression of this dementia. Therefore, the study ofstrocyte responses to A�, the main cellular type involved inhe maintenance of synaptic glutamate concentrations, isrucial for understanding the pathogenesis of AD. This studyims to investigate the effect of A� on the astrocytic gluta-ate transporters, glutamate transporter-1 (GLT-1) and glu-

amate–aspartate transporter (GLAST), and their relative par-icipation to glutamate clearance. In addition we have alsonvestigated the involvement of mitogen-activated proteinMAP) kinases in the modulation of GLT-1 and GLAST levelsnd activity and the putative contribution of oxidative stressnduced by A� to the astrocytic glutamate transport function.herefore, we used primary cultures of rat brain astrocytesxposed to A� synthetic peptides. The data obtained showhat A�1-40 peptide decreased astroglial glutamate uptakeapacity in a non-competitive mode of inhibition, assessed inerms of tritium radiolabeled D-aspartate (D-[3H]aspartate)ransport. The activity of GLT-1 seemed to be more affectedhan that of GLAST, and the levels of both transporters wereecreased in A�1-40-treated astrocytes. We demonstrated

hat MAP kinases, extracellular signal-regulated kinaseERK), p38 and c-Jun N-terminal kinase, were activated in anarly phase of A�1-40 treatment and the whole pathways dif-erentially modulated the glutamate transporters activity/lev-

Correspondence to: P. Agostinho, Center for Neuroscience and Celliology, Department of Zoology, University of Coimbra, 3004 Coimbra,ortugal. Tel: �351-239-820190; fax: �351-239-822776.-mail address: pagost@cnc.cj.uc.pt (P. Agostinho).bbreviations: AD, Alzheimer’s disease; APP, amyloid precursor pro-

ein; A�, amyloid-beta peptide; DHK, dihydrokainate; DTT, dithiothre-tol; D-[3H]aspartate, tritium radiolabeled D-aspartate; EAAT1, excita-ory amino acid transporter 1; EAAT2, excitatory amino acid trans-orter 2; ECF, enhanced chemifluorescence; ERK, extracellularignal-regulated kinase; GAPDH, glyceraldehyde 3-phosphate dehy-rogenase; GLAST, glutamate–aspartate transporter; GLT-1, gluta-ate transporter-1; HBSS, Hanks’ balanced salt solution; JNK, c-Jun-terminal kinase; Ki, inhibition constant; MAPK, mitogen-activatedrotein kinase; NMG, N-methyl-D-glucamine; ROS, reactive oxygenpecies; SDS-PAGE, SDS–polyacrylamide resolving gel; TBARS,hiobarbituric acid reactive substances; TBOA, DL-threo-b-benzoylox-

paspartate; TBS-T, 137 mM NaCl, 20 mM Tris–HCl, pH 7.6 with 0.1%ween; TCA, trichloroacetic acid.

306-4522/08 © 2008 IBRO. Published by Elsevier Ltd. All rights reserved.oi:10.1016/j.neuroscience.2008.08.022

1

ls. Moreover it was shown that oxidative stress induced by�1-40 may lead to the glutamate uptake impairment ob-erved. Taken together, our results suggest that A� peptideownregulates the astrocytic glutamate uptake capacity andhis effect may be in part mediated by oxidative stress andhe differential activity and complex balance between theAP kinase signaling pathways. © 2008 IBRO. Published bylsevier Ltd. All rights reserved.

ey words: glutamate transport, amyloid-beta, oxidativetress, astrocytes, MAP kinases.

lzheimer’s disease (AD) is a neurodegenerative disorderharacterized by a progressive decline of cognitive andemory functions, resulting from the selective synaptic

oss, neuronal dysfunction and death (Blennow et al.,006). Two other neuropathological features of AD in-lude: i) the formation of intra-neuronal neurofibrillary tan-les (NFTs), constituted by paired helical filaments (PHFs)f microtubules and associated hyperphosphorylated tauroteins and, ii) the presence of extracellular amyloidlaques, composed primarily by diffuse or compacted de-osits of amyloid-beta peptides (A�), with 40 or 42 aminocids (LaFerla, 2002). Several groups, including ours,ave demonstrated that A�1-40 and A�1-42 are directly

oxic to cultured neurons, supporting the hypothesis that� is involved in the neurodegeneration associated withD (Agostinho et al., 2003; Celsi et al., 2007; Resende etl., 2007 and in press). The senile plaques in AD brain areormed by a core of A� deposits and dystrophic neuritesurrounded by reactive microglia and astrocytes (Streit,004), the last exhibiting a stellate morphology and in-reased expression of glial fibrillary acidic protein (Peknynd Nilsson, 2005; Maragakis and Rothstein, 2006). Acti-ated astrocytes can bind and degrade A�, suggesting airect role for these cells in A� accumulation and clearance

n AD (Wyss-Coray et al., 2003; Koistinaho et al., 2004).urthermore, A� activates glial cells to produce potentiallyeurotoxic substances, such as reactive oxygen speciesROS) and pro-inflammatory cytokines that can exacer-ate the neurodegenerative process (Akiyama et al., 2000;gostinho and Oliveira, 2007).

One potential mechanism of neurodegeneration in ADs the one where A� enhances the neuronal vulnerability tolutamate excitotoxic death, triggered by the sustainedctivation of glutamate receptors, resulting in an enzymaticascade of events leading to cell death (Francis, 2003;ynd et al., 2004). However, excitotoxicity is normally

revented by the astrocytic glutamate transporters gluta-

mmgt(rsCwtss(scgospataem2a(stpanmi

ltcpqbP2aaewwtc

pctclpbT

(vtkdwmic

M

TwAa(NdwwuRcsa(CbapatpM

C

P4pmawtKRc0mhaa((bcp1rqrdf

M. Matos et al. / Neuroscience xx (2008) xxx2

ARTICLE IN PRESS

ate transporter-1 (GLT-1, corresponding to human ho-ologue excitatory amino acid transporter 2 (EAAT2)) andlutamate–aspartate transporter (GLAST, correspondingo human homologue excitatory amino acid transporter 1EAAT1)), which efficiently remove the excess of this neu-otransmitter from the synaptic cleft (Anderson and Swan-on, 2000). These transporters are present throughout theNS, being GLT-1 highly abundant in astroglial cells,hereas GLAST exists at higher levels in Bergmann glia in

he cerebellum (Kanai et al., 1994; Anderson and Swan-on, 2000). GLT-1 plays a critical role in CNS homeosta-is, accounting for up to 70% of glutamate clearanceGegelashvili and Schousboe, 1997; Anderson and Swan-on, 2000), and it was shown that GLT-1 deficiency isrucially involved with amyotrophic lateral sclerosis patho-enesis (Rothstein et al., 1995, 2005). Although the focusn glutamate transporter regulation continues to intensify,carce data are available about alterations of these trans-orters in aging and disease. Previous studies have shownsignificant reduction of glutamate transporter activity in

he cortex, platelets or fibroblasts of AD patients (Scott etl., 1995; Masliah et al., 1996; Ferrarese et al., 2000; Liangt al., 2002; Zoia et al., 2005), as well as in AD transgenicice (Li et al., 1997; Masliah et al., 2000; Scott et al.,002). It is also known that A� can directly or indirectlyffect the functionality of many related membrane proteinsi.e. ion-motive ATPases, calcium channels) and transportystems (i.e. glucose transporters), raising the possibilityhat glutamate transporters may be equally affected (Par-ura-Gill et al., 1997; Mattson and Chan, 2003; Sultanand Butterfield, 2008). However, neither the functional sig-ificance of such changes nor the factors affecting gluta-ate transporter expression in AD models have been clar-

fied so far.Regardless of the crucial role of astrocytes in extracel-

ular glutamate levels’ maintenance, the regulation of as-roglial transporters is still defectively understood. Astro-ytic glutamate transporters possess putative proteinhosphorylation residues in their primary amino acid se-uences, opening the possibility that their regulation maye indeed mediated by protein kinases, mainly PKC andKA (Gegelashvili et al., 2000; Anderson and Swanson,000; Figiel et al., 2004). However, few data concerning thestrocytic glutamate transport-related signaling pathways arevailable in AD conditions. According to one study the MEK–xtracellular signal-regulated kinase (ERK) signaling path-ay seems to be involved with astroglial glutamate transporthen A� peptide is present (Abe and Misawa, 2003a), but

he whole signaling mechanisms implicated remain to belarified.

In the present study we used cortical astrocytes ex-osed to A� to investigate how the peptide affects astro-ytic glutamate transporters’ activity and levels. The rela-ive participation of GLT-1 and GLAST in the glutamatelearance was also evaluated by measuring the tritium radio-abeled D-aspartate (D-[3H]aspartate) uptake in astrocytesre-treated with glutamate transport inhibitors (DL-threo-b-enzoyloxyaspartate (TBOA) and dihydrokainate (DHK)).

he involvement of mitogen-activated protein kinase s

Please cite this article in press as: Matos M, et al., Amyloid-beta peptidment of oxidative stress and mitogen-activated protein kinase cascades022

MAPKs) on glutamate transporter activity/levels was in-estigated by using specific MEK-ERK (U0126), c-Jun N-erminal kinase (JNK) (SP600125) and p38 (SB203580)inase inhibitors. The contribution of oxidative stress, in-uced by A�, in astrocytic glutamate transport impairmentas also evaluated by using the antioxidant trolox, a vita-in E analog. The knowledge of the mechanisms involved

n the regulation of astrocytic glutamate clearance mayontribute to the development of therapies for AD.

EXPERIMENTAL PROCEDURES

aterials

he synthetic amyloid beta peptides A�1-40, A�40-1 and A�1-42

ere bought from American Peptide (Sunnyvale, CA, USA). D-spartate was acquired from Sigma-Aldrich (St. Louis, MO, USA)nd D-[3H]aspartate was bought from Amersham BiosciencesBuckinghamshire, UK). All cell medium components, trolox and-methyl-D-glucamine (NMG) were purchased from Sigma-Al-rich. DHK and TBOA were acquired from Tocris (Bristol, UK),hereas the MAPK inhibitors, U0126, SP600125 and SB203580ere bought from Calbiochem-Merck (Nottingham, UK). Reagentssed in immunoblotting experiments were purchased from Bio-ad Laboratories; except PVDF membranes and enhancedhemifluorescence (ECF) reagent that were obtained from Amer-ham Biosciences (Piscataway, NJ, USA). Polyclonal primaryntibody anti-glyceraldehyde 3-phosphate dehydrogenaseGAPDH) was bought from Chemicon International (Temecula,A, USA), anti-GLAST/EAAT1 (-C-terminus) from Abcam (Cam-ridge, UK) and anti-GLT-1/EAAT2 (-N-terminus) from Alpha Di-gnostic (San Antonio, TX, USA). The total and phospho-specificolyclonal and monoclonal antibodies anti-ERK, anti-JNK andnti-p38 were all purchased from Cell Signaling Technology (Bos-on, MA, USA). The rabbit and mouse alkaline phosphatase-linkedolyclonal secondary antibodies were purchased from Invitrogenolecular Probes (Barcelona, Spain).

ortical astrocyte primary cultures

rimary astrocyte cultures were prepared from cerebral cortices of–5-day postnatal Wistar rats according to previous describedrocedures (Harris et al., 1996; Saura, 2007), but with someodifications. In brief, the mice were killed by cervical dislocationnd the brain was removed. The left and right cerebral corticesere removed and isolated in ice-cold Hanks’ balanced salt solu-

ion (HBSS) (in mM: 137 NaCl, 5.4 KCl, 0.17 Na2HPO4, 0.22H2PO4, 2.3 NaHCO3, 10 Hepes, 2.7 D-glucose, 0.073 Phenoled, pH 7.4). The meninges were then removed and the corticeshopped up and incubated with a digestive medium containing.25% Trypsin and 0.25 mg/ml DNAse in HBSS at 37 °C for 15–20in. Then, the enzymatic digestion was stopped with 10% ofeat-inactivated FBS and the cell suspension centrifuged for 5 mint 180�g. Afterward, the obtained pellet was resuspended instrocyte culture medium, Dulbecco’s modified Eagle mediumDMEM)—high glucose—supplemented with 10% FBS, penicillin50 U/ml), streptomycin (10 mg/ml), Hepes (6 g/l), and sodiumicarbonate (0.84 g/l) and the number of cells in suspensionounted in a hemocytometer. Then, the cells were plated ontooly-L-lysine-coated 75-cm2 culture flasks, at a density at.14�105 cells/cm2, and maintained at 37 °C in a 5% CO2/95%oom-air humidified incubator. The cell culture medium was fre-uently replaced, every 2–3 days, until the mixed-glial cultureseached confluency, which was normally achieved after 13–15ays of culture in vitro (DIV). In order to separate microglial cellsrom the astrocytes monolayer, the mixed glial-cultures were

haken at 200 rpm in an orbital shaker for 3 h. Then, the medium

e decreases glutamate uptake in cultured astrocytes: Involve-, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.

wacttpmn(css

C

CfAmwbewThSamUwcm

G

TtewNcfact4mcstpS0aco

tmcc(bs

A

a

sim7tab9mlaa

ofiumaitcfiahbae

W

AbMmtrtn1BEovSalptCb0rG11obofw5cowwg

M. Matos et al. / Neuroscience xx (2008) xxx 3

ARTICLE IN PRESS

ith the up-layer detached microglial cells was discarded and thestrocytes that remained in the flasks washed with HBSS bufferontaining EDTA (1 mM) and further detached by a mildrypsinization procedure using HBSS with 0.1% trypsin. Finally,he cells were reseeded with fresh astrocyte culture medium onoly-L-lysine-coated plates, at a density of 3�104cells/cm2, andaintained in culture for 1–2 days before the experiments begin-ing. Our cell preparations had a high percentage of astrocytes�93%), which were confirmed by immunostaining against astro-yte specific protein glial fibrillary acidic protein and microglialpecific antibody CD11b (Matos and Agostinho, unpublished ob-ervations).

ell treatments

ultured astrocytes were treated with A�1-40, A�40-1 or A�1-42

or different periods as indicated in the figure legends. The�1-40 and A�40-1 peptides were reconstituted according to theanufacturer’s instructions, being first prepared in sterile MilliQater into 6 mg/ml and afterward diluted in sterile phosphateuffered saline solution and aged for 5–7 days at 37 °C (Garçãot al., 2006). The synthetic peptide A�1-42 was dissolved in sterileater, at a concentration of 1 mg/ml. (Resende et al., in press).he non-transportable glutamate transporter inhibitors TBOA (in-ibition constant (Ki)�5.7 �M, as reported by Anderson andwanson, 2000) and DHK (Ki�3–23 �M, as reported by Andersonnd Swanson, 2000) were added to the transport saline buffer 30in prior to the initiation of uptake assays. The MAPK inhibitors,0126 (0.4 �M), SP600125 (0.4 �M) and SB203580 (3 �M), asell as the antioxidant trolox (100 �M), were added to the astro-yte culture medium simultaneously to peptide incubation andaintained for 24 h.

lutamate transport: D-[3H]aspartate uptake assays

he uptake analysis for D-[3H]aspartate was evaluated accordingo the methods previously described (Agostinho et al., 1997; Berryt al., 2005), with specific adaptations. The cultured astrocytesere incubated with Krebs buffer (in mM: 132 NaCl, 4 KCl, 1.2a2HPO4, 1.4 MgCl2, 6 glucose, 10 Hepes, 1 mM CaCl2, pH 7.4)ontaining D-[3H]aspartate (0.1 �Ci/ml) and 100 �M D-aspartateor 10 min at 37 °C. After this period, the medium was removednd the cultured cells were placed on ice and washed twice withold Krebs buffer to terminate the uptake assay. Subsequently,he cells were lysed overnight with 500 �l of 0.5 M NaOH and00 �l of the cell sample was transferred to a scintillation vial to beixed with universal scintillation fluid (4 ml). The radioactivity

ontent (disintegrations per minute) was determined using liquidcintillation spectroscopy in a TRI CARB® 2900TR liquid scintilla-ion analyzer. The remaining cell suspension was assayed forrotein content with the Bio-Rad reagent (Garção et al., 2006).ome experiments were performed in the absence of Na� and at°C in order to determine the Na�-dependence of the transport

nd non-specific uptake, respectively. The uptake rates for theells, in each well, were expressed as the uptake per min and mgf protein.

For saturation kinetics assays the total D-aspartate concen-rations ranged from 10 to 800 �M, while the rest of the experi-ents were similar to the previous uptake assays. The kinetic

onstants (i.e. maximum velocity, Vmax, and Michaelis-Mentenonstant, Km) were determined by using nonlinear regressioncurve fit) analysis followed by the application of the one-siteinding (hyperbola) equation to the data, using GraphPad Prismoftware.

ssessment of oxidative stress parameters

Lipid peroxidation. The extent of lipid oxidation was evalu-

ted by measuring the levels of thiobarbituric acid reactive sub- w

Please cite this article in press as: Matos M, et al., Amyloid-beta peptidment of oxidative stress and mitogen-activated protein kinase cascades022

tances (TBARS), as previously described (Agostinho and Ol-veira, 2003). After cell treatment, the culture medium was re-

oved and replaced by an ice-cold solution of 15 mM Tris (pH.4). The cells were scrapped off and diluted three times with 15%richloroacetic acid (TCA), 0.375% thiobarbituric acid, 0.25 M HCl,nd 0.015% 2,6-di-tert-butyl-4-methylphenol (BHT), and incu-ated for 15 min at 100 °C. Then, the samples were centrifuged at5.5�g for 10 min, and the absorbance of supernatants waseasured at 530 nm. The amount of TBARS formed was calcu-

ated using a molar extinction coefficient of 1.56�105 mol�1cm�1

nd normalized per milligram of protein. The data are expresseds arbitrary units.

Protein oxidation. The most general indicator of proteinxidation is carbonyl group content (Harris et al., 1995; Butter-eld et al., 1997). Protein carbonyls levels were measuredsing the Cayman Protein Carbonyl assay kit, according to theanufacturer’s instructions but with some modifications. In brief,strocyte cultures were lysed with 50 mM MES/1 mM EDTA and

ncubated 10 mM DNPH (1:4) for 1 h. Then, 20% TCA was addedo each sample (1:1) and incubated on ice for 10 min. Afterentrifugation for 10 min at 10,000�g, the pellets were washed,rst with 10% TCA and then with ethanol-ethylacetate, to removeny unreacted DNPH. The pellets were solubilized with guanidineydrochloride and centrifuged to remove insoluble material. Car-onyl content was calculated from the absorbance measurementt 595 mm, as described in the instruction of the assay kit, andxpressed as arbitrary units.

estern blotting

fter astrocytes treatment they were lysed with ice-cold isolationuffer (250 mM sucrose, 20 mM Hepes, 10 mM KCl, 1.5 mMgCl2 (pH 7.4), 0.002 mM dithiothreitol (DTT), 0.1 mM phenyl-ethylsulfonyl fluoride (PMSF), 0.001% cocktail protease inhibi-

or, 2 mM orthovanadate and 50 mM NaF). The lysates wereapidly frozen/defrosted three times and assayed for protein con-ent with the Bio-Rad reagent. Afterward, the samples were de-atured by addition of sample buffer (500 mM Tris, 600 mM DTT,0.3% SDS, 30% glycerol, 10% SDS, and 0.012% Bromophenollue, pH 6.8) and loaded into the Western blot gels for separation.qual amounts of protein were then separated by electrophoresisn 10% SDS–polyacrylamide resolving gels (SDS-PAGE) inside aat with running buffer (123 mM Tris, 1000 mM bicine, 0.15%DS, pH 8.3). To make an accurate identification of the proteins,pre-stained precision protein standard (Bio-Rad) was also

oaded in one well gel. After electrophoresis, the proteins on theolyacrylamide resolving gels were transferred electrophoreticallyo a PVDF membrane, during 75 min, in a CAPS buffer (10 mMAPS, pH 11, 10% methanol). Then, these membranes werelocked in TBS (137 mM NaCl, 20 mM Tris–HCl, pH 7.6) with.1% Tween (TBS-T) and 3% fatty acid free BSA, for 45 min atoom temperature. Incubation with the primary antibodies anti-LAST (1:1000), anti-GLT-1 (1:1000), anti-phospho-ERK 1/2 (1:000), anti-phospho-JNK1/2/3 (1:1000), or anti-phospho-p38 (1:000) in TBS-T with 3% fatty acid free BSA was carried outvernight at 4 °C. After extensive washing with TBS-T, the mem-ranes were incubated with an alkaline phosphatase–linked sec-ndary antibody (1:20,000 in TBS-T with 3% fatty acid free BSA)or 2 h at RT. Finally, bands of the desired immunoreactive proteinere visualized, after membrane incubation with ECF reagent formin, on a Versadoc image system. Densities of blot bands were

alculated in the Quantity One program (Bio-Rad). The reprobingf the membranes was performed by removing the ECF reagentith 40% methanol during 30 min, after which TBS-T was used toash the membranes. The antibodies were then removed usinglycine (0.1 M pH 2.3), during 30 min at RT (stripping). After a new

ash in TBS-T, membranes were blocked again with TBS-T with

e decreases glutamate uptake in cultured astrocytes: Involve-, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.

3a1is

S

D(ftam

IptaAa(Aoii2ssfawca2

Aian

Tet2wsaarPfcaumdAd

1imuAws2t(

AD

tfcD

wtsubdD

ttt1tdap

Ga

Tf

Fuwt�Dtdf

M. Matos et al. / Neuroscience xx (2008) xxx4

ARTICLE IN PRESS

% fatty acid free BSA, followed by incubation with new primaryntibodies for the total forms of GAPDH (1:2000), ERK 1/2 (1:000), JNK (1:1000) and p38 (1:1000). Secondary antibodies

ncubation, membrane washing and revelation were done as de-cribed before.

tatistical data analysis

ata were expressed as percentages of values obtained in controluntreated cells) cultures, and were presented as means�S.E.M.or the number of experiences indicated in figure captions. Statis-ical significance was determined using F test for one-variablenalysis of variance (one-way ANOVA) followed by Dunnett’sultiple comparison post hoc tests.

RESULTS

n order to investigate the effect of AD-associated amyloideptides on astrocytic glutamate uptake capacity, we usedhe synthetic peptides A�1-40 (the major component ofmyloid plaques) and A�1-42 (the most toxic species).lthough these peptides tend to form amyloid fibrils andggregates, it is widely accepted that soluble A� oligomersor protofibrils) are the major cause of early synaptic loss inD (Sultana and Butterfield, 2008; Klein et al., 2004). A�ligomers were also shown to activate initial stages of

nflammation, while fibrillar A� is more involved in sustain-ng the chronic inflammation in this disease (White et al.,005). In this study we used “aged” A�1-40 and A�1-42

olutions (see Experimental Procedures), which werehown by our group to have a higher amount of fibrillarorms than of oligomers (Resende et al., in press; Resendend Oliveira, unpublished data for A�1-40). These peptidesere used at a concentration of 5 �M, which was shown toause neuronal injury and microglia activation (Agostinhond Oliveira, 2003; Garção et al., 2006; Ferreiro et al.,006; Resende et al., 2007).

� inhibits glutamate uptake in cultured astrocytesn a non-competitive manner - The astrocytic GLASTnd GLT-1 uptake activity is inhibited in aon-competitive manner”

he effect of A� on astrocytic glutamate transport wasvaluated by measuring the D-[3H]aspartate uptake in cul-ured astrocytes treated with A�1-40, A�40-1 and A�1-42 for4 h (Fig. 1). The common glutamate analog, D-aspartate,as used in these studies because it is an excellent sub-trate for the high-affinity glutamate carriers—36–90 �M offfinity for D-aspartate (Anderson and Swanson, 2000)—nd is not metabolized inside the cells permitting an accu-ate measure of the total transport (Berry et al., 2005).reliminary experiments of D-[3H]aspartate uptake per-

ormed at 0 °C or using a medium without Na� (isosmoti-ally substituted by NMG) showed that the uptake of thismino acid was almost negligible (Matos and Agostinho,npublished observations), suggesting that in our experi-ental conditions the D-aspartate uptake was mainly me-iated by Na�-dependent transporters. Fig. 1 shows that�1-40 (5 �M) and A�1-42 (5 �M) significantly (P�0.05)

ecreased the D-[3H]aspartate uptake by 22.0�7.4% and e

Please cite this article in press as: Matos M, et al., Amyloid-beta peptidment of oxidative stress and mitogen-activated protein kinase cascades022

4.1�4.3%, respectively. Given that both peptides inhib-ted the glutamate uptake and A�1-40 reduced it slightly

ore than A�1-42, we performed the following experimentssing A�1-40. The reverse sequence of this peptide,�40-1, did not affect the uptake of D-[3H]aspartate (Fig. 1),hich demonstrate the amino acid order specificity. Ithould be noted that higher concentrations of A�1-40 (10–0 �M) did not change significantly the proportion of glu-

amate uptake inhibition observed with 5 �M of peptidedata not shown).

For a more detailed study of the inhibition caused by�1-40 in astrocytic glutamate uptake, saturation kinetics of-[3H]aspartate uptake assays was performed using cul-ured astrocytes untreated (control) or treated with A�1-40

or 24 h (Fig. 2). In these experiments the uptake wasarried out in the presence of increasing concentrations of-aspartate (10–800 �M). The saturation binding curvesere then plotted and the kinetic constants determined for

he two conditions (control and A�-treated cells). As can beeen in Fig. 2, a general decrease on the D-[3H]aspartateptake activity was observed in A�1-40-treated cells,eing the maximum rate (Vmax) of D-aspartate uptakeecreased by 38% from the control levels. The Vmax of-[3H]aspartate in control cells (0.40�0.03 nmol/mg pro-ein/min) was higher than those determined for A�-reated cells (0.25�0.02 nmol/mg protein/min), whereashe apparent Km values were similar in control (33.08�1.86 �M) and A�-treated cells (34.44�12.18 �M) (seeable below Fig. 2). Given that Vmax was significantly re-uced in A�-treated astrocytes, with no changes in thepparent Km values, it can be suggested that a non-com-etitive inhibition occurs under these conditions.

LT-1 is a more efficient transporter than GLASTnd seems preferentially inhibited by A�

o investigate whether the A� treatment differentially af-ected GLT-1 and GLAST activity, we performed a set of

ig. 1. Effect of different A� peptide sequences on D-[3H] aspartateptake by astrocytes. Cultured astrocytes were treated or not (control)ith A�1-40, A�40-1 and A�1-42 (5 �M) for 24 h. After peptide treatments

he astrocytes were incubated with 100 �M D-aspartate (and 0.1Ci/ml D-[3H]aspartate) for 10 min to assess the amino acid uptake.ata are expressed as the percentage of control transport obtained in

he absence of A� and are the means�S.E.M. of at least six indepen-ent experiments done in duplicate. * P�0.05, significantly different

rom control (untreated) cells.

xperiments using non-transportable glutamate uptake in-

e decreases glutamate uptake in cultured astrocytes: Involve-, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.

huGD

pbacgftD

Ttupsa

paAatspptcw

rebpGD

acptsTh

Ag

Ibddbcws4lAcsd(Asl

Fuwptwpti

FtwtdD

KuoPl

M. Matos et al. / Neuroscience xx (2008) xxx 5

ARTICLE IN PRESS

ibitors: TBOA and DHK (Fig. 3). These experiments leds to conclude that the collective uptake, mediated byLAST and GLT-1, accounted for 72�3.4% of the total-[3H]aspartate taken up by astrocytes, because in theresence of TBOA alone or with DHK (both transporterslocked) the amount of D-[3H]aspartate uptake was onlybout 28% of control. Considering that in our experimentalonditions GLT-1 and GLAST mediated 72% of the totallutamate transport, the GLT-1 was probably responsibleor about 42% (shown by the transporter selective inhibi-ion with DHK) and GLAST for 30% of the total astrocyte-aspartate uptake in untreated (control) cells (Fig. 3).hese data suggest that GLT-1 mediates glutamate up-

ake more efficiently than GLAST. The D-[3H]aspartateptake that is not blocked by TBOA and DHK (28%) wasrobably mediated by other mechanisms, such as non-pecific diffusion or chloride-dependent glutamate/cystinentiporters (Anderson and Swanson, 2000).

A significant reduction of about 20% in the D-[3H]as-artate uptake was observed in A�1-40-treated astrocytes,s compared with control cells, whereas in the presence of��DHK the uptake was reduced by 42%. In A�-treatedstrocytes the selective inhibition of GLT-1 only reducedhe glutamate transport by 22%, contrasting with that ob-erved in control cells (42% inhibition). Furthermore, in theresence of this GLT-1 inhibitor the amount of D-[3H]as-artate taken up by control astrocytes (58.7�12.4% con-rol) was similar to that determined in A�1-40�DHK-treatedells (58�4.3% control), suggesting that GLT-1 activity

ig. 2. Saturation binding curves of sodium-dependent D-[3H]aspar-ate uptake by A�-treated and control astrocytes. Cultured astrocytesere treated or not (control) with A�1-40 (5 �M) for 24 h. After peptide

reatment the uptake assays were carried out in the presence ofifferent concentrations (0–800 �M) of D-aspartate (and 0.1 �Ci/ml-[3H]aspartate) for 10 min at 37 °C. In the table, represented are the

m and Vmax values determined based on saturation binding curves,sing nonlinear regression (curve fit) analysis followed by applying thene-site binding (hyperbola) equation to the data, using GraphPadrism software. The values are represented as the mean�S.E.M. of at

east five separate experiments, measured in duplicate.

as affected by A� peptide. Given that A�1-40 by itself canes

Please cite this article in press as: Matos M, et al., Amyloid-beta peptidment of oxidative stress and mitogen-activated protein kinase cascades022

educe the uptake activity by about 20%, if its inhibitoryffect was also due to alteration in GLAST activity, it woulde expected a higher decline in the amount of D-[3H]as-artate taken up by astrocytes treated with A�1-40 andLT-1 inhibitor. The inhibitory effect of TBOA and DHK on-[3H]aspartate uptake was not additive, in both controlnd A�1-40-treated astrocytes, which confirms that TBOAan block GLT-1 and GLAST. Moreover, when both trans-orters are blocked (in the presence of DHK and/or TBOA)he amount of D-[3H]aspartate taken up by control cells wasimilar to that determined for A�1-40-treated astrocytes.aken together the results suggest a more prominent in-ibitory effect of A�1-40 in GLT-1 activity than in GLAST.

� peptide decreases the levels of both astrocyticlutamate transporters

n order to study if the (non-competitive) inhibition inducedy A�1-40 in D-aspartate uptake was an outcome of aecrease in astrocytic glutamate transporter levels, weetermined the levels of GLT-1 and GLAST by Westernlot (Fig. 4). The amount of GLT-1 and GLAST in astro-ytes treated with A�1-40 (5 �M) for 24 h was comparedith that determined in untreated (control) astrocytes. Den-itometric analysis of the immunoblot band of GLT-1 (Fig.A) showed a significant (P�0.05) decrease (�18%) in the

evels of this transporter subtype in astrocytes exposed to�1-40 (81.8�1.8% of control) as compared with controlells. The Western blot analysis for GLAST displayed re-ults similar to those observed for alterations in GLT-1ensity (Fig. 4B). GLAST levels were also significantlyP�0.05) decreased (�13%) after treatment with 5 �M�1-40 (87.4�3.4% of control). Taken together, these re-ults show that A�1-40 reduced both GLT-1 and GLAST

evels.

ig. 3. Effect of glutamate transporter inhibitors on D-[3H]aspartateptake by control and A�-treated astrocytes. The cells were incubatedith A�1-40 (5 �M) for 24 h. The nontransportable glutamate trans-orter inhibitors TBOA (nonselective, 70 �M) and DHK (GLT-1 selec-

ive, 25 �M) were incubated 30 min before the uptake assays, whichere carried out using 100 �M D-aspartate (and 0.1 �Ci/ml D-[3H]as-artate) for 10 min. Data are expressed as the percentage of control

ransport obtained in the absence of A�1-40 or glutamate transportnhibitors and are the means�S.E.M. of at least six independent

xperiments done in duplicate. * P�0.05 and ** P�0.01, indicatetatistical significance related to control.

e decreases glutamate uptake in cultured astrocytes: Involve-, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.

Ap

Icanap

WaJtct3i

FAabo essed asi ol cells.

Fcaav

M. Matos et al. / Neuroscience xx (2008) xxx6

ARTICLE IN PRESS

� peptide causes early MAPKshosphorylation/activation

t has been reported that A� peptide activates the MAPKsascades in certain cell types (Daniels et al., 2001; Zhu etl., 2002). Therefore, we investigated whether MAPK sig-al transduction cascades were active in our astrocytesnd whether A� peptide affected the level of activation/hosphorylation of ERK, JNK and p38 kinases. Thus, a

ig. 4. Alterations in GLT-1/EAAT2 (A) and GLAST/EAAT1 (B) levels�1-40 (5 �M) for 24 h. Cell lysates were examined by immunoblotting wnd reprobed with a monoclonal antibody anti-GAPDH to estimate the ty scanning on a Versadoc Image system, and the levels of proteins wf GLT-1 and GLAST, compared with GAPDH levels, and were expr

ndependent experiments. ** P�0.01, significantly different from contr

ig. 5. Representative blots of the time course experiments for ERK1ultured astrocytes were exposed to 5 �M A�1-40 for 3–24 h, and the cnti-phospho-ERK1/2 ((P)ERK1/2), anti-phospho-JNK1/2 ((P)JNK1/2)

nti-ERK1/2, anti-JNK1/2 and anti-p38 antibodies (total levels) to estimate the prisualized by scanning on a Versadoc Image system, and the levels of protein

Please cite this article in press as: Matos M, et al., Amyloid-beta peptidment of oxidative stress and mitogen-activated protein kinase cascades022

estern blot assay was performed using antibodiesgainst the phosphorylated and total forms of ERK1/2,NK1/2 and p38 kinases (Fig. 5). The possible modifica-ions in the phosphorylation state of each MAPK in astro-ytes were determined at different time periods of incuba-ion with 5 �M A�1-40 (3, 6, 12 and 24 h). Immediately after

h incubation with A�1-40 a significant (P�0.01) increasen the levels of phosphorylated ERK1/p44 (118.6�4.8% of

by A� peptide. Cultured astrocytes were treated or not (control) withti-GLT-1 and anti-GLAST monoclonal antibody. The blot was strippednt of protein loaded in the gel. Immunoreactive bands were visualizedtified in the Quantity One program. Bars represent the relative levelspercentage of control. Data are the means�S.E.M. of at least five

/2 and p38 phosphorylation in astrocytes treated or not with A�. Thet was subjected to SDS-PAGE followed by Western blot analysis withhospho-p38 ((P)p38) antibodies. Then, the blots were reprobed with

inducedith an an

otal amouere quan

/2, JNK1ell extracor anti-p

oportion of each kinase phosphorylation. Immunoreactive bands weres were quantified in the Quantity One program.

e decreases glutamate uptake in cultured astrocytes: Involve-, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.

csayirElpmsim3e

Eia

Skdptttesc(id 3

trJc

t(tcigJm0si

Ei

IcpGMUA

Fa2iit�T(dst

Fcw(awabnspGt*a

M. Matos et al. / Neuroscience xx (2008) xxx 7

ARTICLE IN PRESS

ontrol) and ERK2/p42 (122.3�6.8% of control) was ob-erved. However the total levels of these kinases were notffected by A� peptide. The increase in ERK1/2 phosphor-lation induced by A�1-40 lasted only for the first 3 h ofncubation, after which the levels of phospho-ERK wereestored to control values. Similarly with that observed forRK1/2, there was a significant (P�0.01) increase in the

evels of phospho-JNK1/2 (124.5�6.1% of control), andhospho-p38 (165.7�28.7% of control) upon 3 h of treat-ent with A�1-40, after which the levels returned to values

imilar to those obtained for control cells. The observedncrease in phospho-p38 at 3 h was higher than the aug-

ents observed for phospho-ERK and phospho-JNK (ath). Taken together the results suggest that A� caused anarly activation of all MAPKs.

RK, JNK and p38 signaling cascades are involvedn the regulation of astrocytic glutamate transporterctivity

ince in our experimental conditions A� affected MAPinase activity we investigated whether these signal trans-uction cascades modulate the astrocytic glutamate trans-ort and the possible relationship of these pathways withhe A�-induced inhibition of glutamate uptake (Fig. 6). Forhis purpose we have measured the D-[3H]aspartate up-ake in astrocytes treated or not with A�1-40 in the pres-nce of three different MAPK inhibitors (Fig. 6). The expo-ure to MEK1/2/ERK1/2 inhibitor, U0126, led to a signifi-ant (P�0.01) 30% decrease in amino acid uptake70.8�3.9% of control). In addition, incubation with p38nhibitor SB203580 led to similar consequences, causing aecline by about 20% of D-[3H]aspartate uptake (80.5�

ig. 6. Effect of MAPK inhibitors on D-[3H]aspartate uptake by controlnd A�-treated astrocytes. Cultured astrocytes were incubated for4 h in the absence (control) or presence of A�1-40 (5 �M), with MAPK

nhibitors U0126 (MEK1/2 inhibitor, 0.4 �M), SP600125 (JNK1/2/3nhibitor, 0.4 �M) and SB203580 (p38 inhibitor, 3 �M). Subsequentlyhe astrocytes were incubated with 100 �M D-aspartate (and 0.1Ci/ml D-[3H]aspartate) for 10 min to assess the amino acid uptake.he results are expressed as the percentage of control transportuntreated cells) and are the mean�S.E.M. of at least four individualeterminations measured in triplicate. * P�0.05, ** P�0.01, indicate

#

itatistical significance related to control and P�0.05, indicates sta-istical significance to the A� condition.

Please cite this article in press as: Matos M, et al., Amyloid-beta peptidment of oxidative stress and mitogen-activated protein kinase cascades022

.7% of control, P�0.05). These results clearly showedhat ERK and p38 pathways were positively involved in theegulation of astrocytic glutamate uptake. By contrast, theNK1/2/3 inhibitor, SP600125, had no effect on the astro-ytic D-[3H]aspartate uptake (Fig. 6).

The amount of D-[3H]aspartate taken up by astrocytesreated with A�1-40�U0126 (71.4�3.5%) or SB20358073.3�3.9%) was similar to that observed in control cellsreated with these inhibitors (70.8�3.9% and 80.5�3.8%ontrol, respectively), which suggests that A�1-40 abol-

shed the positive effects of ERK and p38 pathways inlutamate uptake. On the other hand, the inhibition of theNK pathway in A�1-40 treated astrocytes caused an aug-ent in the D-[3H]aspartate uptake (�17.5�3.4%, P�.05) as compared with the cells treated with A� alone,uggesting that this pathway contributed to the uptake

nhibition caused by A�.

RK, JNK and p38 signaling cascades are involvedn the regulation of GLT-1 and GLAST levels

n order to determine whether the MAPK signaling cas-ades were involved in the regulation of glutamate trans-orter levels, we performed experiments to determine theLT-1 (Fig. 7) and GLAST (Fig. 8) levels using the sameAPK inhibitors. Fig. 7 shows that in the presence of0126 the GLT-1 levels were not significantly changed by� as compared with untreated (control) cells. By contrast,

ig. 7. Effect of MAPK inhibitors on GLT-1/EAAT2 protein levels inontrol and A�-treated astrocytes. Cultured astrocytes were incubatedith MAPK inhibitors U0126 (MEK1/2 inhibitor, 0.4 �M), SP600125

JNK1/2/3 inhibitor, 0.4 �M) and SB203580 (p38 inhibitor, 3 �M) in thebsence (control) or presence of A�1-40 (5 �M) for 24 h. Cell lysatesere examined by immunoblotting with an anti-GLT-1 monoclonalntibody. The blot was reprobed with anti-GAPDH monoclonal anti-ody to estimate the total amount of protein loaded in the gel. Immu-oreactive bands were visualized by scanning on a Versadoc Imageystem, and the levels of proteins were quantified in the Quantity Onerogram. Bars represent the relative levels of GLT-1, compared withAPDH levels, and were expressed as percentage of control. Data are

he means�S.E.M. of at least five independent experiments.P�0.05, ** P�0.01, indicate statistical significance related to controlnd ## P�0.01, indicates statistical significance to the A� condition.

n A�1-40-treated astrocytes the blockage of ERK by U0126

e decreases glutamate uptake in cultured astrocytes: Involve-, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.

ccJliteletpcslpi

GcgtGpm((pcsuil

hgetg6c8ApPdp

Pg

OtCgekcTdbamesfpvonuc(it

To

C

CATA

(adTumd*t

Fcw(awabnspGm*

M. Matos et al. / Neuroscience xx (2008) xxx8

ARTICLE IN PRESS

aused a significant augment in GLT-1 levels (28%) whenompared with astrocytes treated only with A�1-40. TheNK inhibitor, SP600125, led to a slight increase in theevels of GLT-1 (117.6�1.5% control), which suggests annhibitory role of this pathway in the expression of thisransporter in untreated (control) astrocytes. In the pres-nce of A�1-40 the effect of SP600125 was more evident,

eading to a significant increase (P�0.01) in the GLT-1xpression (increase of 29%), when comparing with cellsreated only with A�1-40. Finally, in astrocytes treated with38 inhibitor (SB203580) a slight, but not significant de-rease in GLT-1 levels (88.4�4.5% of control) was ob-erved, as compared with untreated (control) cells. Thisow decline of about 10% was maintained when A�1-40 wasresent, which suggests that p38 signaling pathway is

mplicated in the downregulation of GLT-1 levels.In Fig. 8 it can be seen that the expression level of

LAST was significantly (P�0.05) increased after astro-yte treatment with U0126 (113.8�4.9% of control), sug-esting an inhibitory effect of ERK pathway in the levels ofhis transporter. Assuming that ERK downregulatesLAST, if this signaling cascade was repressed the trans-orter levels should increase. Indeed, concomitant treat-ent of U0126 with A�1-40 slightly increased GLAST

93.6�4.6% of control) as compared with A�1-40 condition87.4�2.4% of control), which seems to indicate that ERKathway did not mediate the GLAST level reductionaused by A�. On the other hand, both JNK and p38ignaling cascades seem significantly involved in GLASTpregulation. The JNK inhibitor, SP600125, led to a signif-

cant (P�0.05) decrease by approximately 15% in GLAST

ig. 8. Effect of MAPK inhibitors on GLAST/EAAT1 protein levels inontrol and A�-treated astrocytes. Cultured astrocytes were incubatedith MAPK inhibitors U0126 (MEK1/2 inhibitor, 0.4 �M), SP600125

JNK1/2/3 inhibitor, 0.4 �M) and SB203580 (p38 inhibitor, 3 �M) in thebsence (control) or presence of A�1-40 (5 �M) for 24 h. Cell lysatesere examined by immunoblotting with an anti-GLAST monoclonalntibody. The blot was reprobed with anti-GAPDH monoclonal anti-ody to estimate the total amount of protein loaded in the gel. Immu-oreactive bands were visualized by scanning on a Versadoc Imageystem, and the levels of proteins were quantified in the Quantity Onerogram. Bars represent the relative levels of GLAST, compared withAPDH levels, and were expressed as arbitrary units. Data are theeans�S.E.M. of at least five independent experiments. * P�0.05,

* P�0.01, indicate statistical significance related to control.

evels (84.8�5.1% of control) in control astrocytes, and #

Please cite this article in press as: Matos M, et al., Amyloid-beta peptidment of oxidative stress and mitogen-activated protein kinase cascades022

ad no effect when A�1-40 was present. These data sug-est that JNK pathway regulated positively the GLASTxpression in the control condition, but not in A�1-40-reated cells. The p38 inhibitor SB203580 elicited an evenreater GLAST downregulation in control cells (77.1�.0% of control, P�0.01), which means that this signalingascade also promoted the expression of GLAST (see Fig.). This decrease was maintained in the presence of�1-40, suggesting that the peptide may restrain the p38athway, causing a decrease in GLAST levels (76.2�4.5%,�0.01). Thus, it appears that A�1-40 mediated GLASTownregulation through impairment of JNK and p38 signalingathways.

rotein and lipid oxidation induced by A� inhibitlutamate uptake

xidative stress is widely implicated in AD since A� pep-ide has been found to generate ROS (Chauan andhauan, 2006) and the oxidation of lipids and proteins canreatly affect the activity of glutamate transporters (Harrist al., 1995; Lauderback et al., 2001). Moreover, the MAPinase signaling pathways can all be activated by oxidantonditions (Zhu et al., 2002; McCubrey et al., 2006).herefore, we analyzed whether in our experimental con-itions the astrocytic glutamate uptake impairment inducedy A� could be related with oxidative stress. First wessessed if A�1-40 causes oxidative astrocytic damage byeasuring the levels of protein carbonyl groups and thextent of lipid peroxidation. Table 1 shows that A�1-40

ignificantly augmented the levels of carbonyls groups (0.5-old) and of lipid peroxidation products (twofold), as com-ared with control cells. The antioxidant trolox (100 �M), aitamin E analog, completely prevented this increase ofxidative markers induced by A�1-40 in astrocytes. Weext went to study if the reduction in astrocytic glutamateptake caused by A�1-40 could be prevented by trolox. Asan be seen in Fig. 8, this antioxidant significantlyP�0.05) reverted the decrease in D-[3H]aspartate uptakenduced by A�1-40 (about 14%). Trolox per se did not affecthe uptake of this amino acid. These results clearly impli-

able 1. Protein and lipid oxidation in astrocytes treated with A�: effectf vitamin E analogue trolox

onditions/assays Carbonyls (a.u.) TBARS (a.u.)

ontrol 1.05�0.11 1.00�0.21�1-40 1.48�0.10* 2.89�0.75*rolox 0.99�0.07# 1.12�0.47�1-40�trolox 1.11�0.06# 1.05�0.21#

The astrocytes were incubated with A�1-40 (5 �M) and/or trolox100 �M) for 24 h. The levels of protein and lipid oxidation weressessed by measuring the levels of carbonyl groups and TBARS, asescribed in Experimental Procedures. The levels of carbonyls andBARS were normalized for mg of protein and expressed as arbitrarynits (a.u.) relative to control cells. The values are represented as theean�S.E.M. of at least three separate experiments, measured inuplicate.P�0.05, indicates statistical significance related to control (un-

reated) cells.

P�0.05, indicates statistical significance to cells treated only with A�.

e decreases glutamate uptake in cultured astrocytes: Involve-, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.

cian

TrAFapeTGrtlik

pobcatrm3s3Kadstracw

togcoi(stti2itGm

aamocsgsca12Gda

tasdlsmoGsaatdttiidctapbaGi((caswrac(t

hln

M. Matos et al. / Neuroscience xx (2008) xxx 9

ARTICLE IN PRESS

ate oxidative stress as a mediator of glutamate uptakenhibition in A�-treated astrocytes, and in addition provide

possible explanation for the altered activity of MAP ki-ases previously found.

DISCUSSION

he data obtained in this study confirm previous findingseporting that glutamate transporter activity is impaired inD conditions (Scott et al., 1995; Masliah et al., 1996;errarese et al., 2000; Liang et al., 2002; Zoia et al., 2005)nd that the uptake activity is specifically decreased by A�eptide (Harris et al., 1996; Parpura-Gill et al., 1997; Laud-rback et al., 1999; Mattson and Chan, 2003; Fernández-omé et al., 2004). Furthermore, we have shown thatLAST and GLT-1 levels and activity are decreased by A�

esulting in an apparent non-competitive mode of inhibi-ion. This impairment of glutamate transporter function andevels seems also to be due to oxidative stress conditionsnduced by A� that result in the altered activity of MAPinase cascades.

In the present study we used brain cortical astrocytereparations with more than 93% of astrocytes (about 5%f microglia), which in accordance with Saura (2007) cane classified as “astroglial-enriched” cultures. The astro-yte cultures were then treated with A� synthetic peptidesnd assayed for glutamate uptake capacity. In astrocytesreated with A�1-40 and A�1-42, we observed a significanteduction of glutamate uptake activity (Fig. 1). The maxi-um rate (Vmax) of D-aspartate uptake decreased by about8% after A�1-40 treatment, whereas the affinity for theubstrate (expressed by the apparent Km values, �33–4 �M) was not affected by the A� treatment (Fig. 2).nowing the limited information that can be extracted frompparent kinetic constants, the fact that Vmax significantlyeclined with A�, with no change in the apparent Km, mayuggest that the peptide induced a non-competitive inhibi-ion of glutamate transport. That is A� and/or productsesulting from its action (i.e. ROS) may act: i) through anllosteric site in the transporters, changing their structuralonfiguration and activity, and/or ii) through signaling path-ays that lead to the decrease in the transporters levels.

Our data also suggest that A� affected more the up-ake activity of GLT-1 than the GLAST one. In fact, webserved that: i) the GLT-1 inhibitor (DHK) reduced thelutamate uptake to a lower extent in A�-treated astro-ytes (22%) than in control cells (42%), and ii) the amountf D-[3H]aspartate taken up by DHK-treated cells was sim-

lar to that measured in astrocytes treated with A��DHKFig. 3). By contrast the activity of GLAST seems not to beignificantly changed by A�, because if it had affected thisransporter a bigger decline in the uptake activity in cellsreated with A��DHK would be expected. Although TBOAs not a specific GLAST inhibitor (Anderson and Swanson,000), the data obtained when GLT-1 was blocked, show-

ng similar amounts of D-[3H]aspartate uptake in DHK-reated and A��DHK- treated astrocytes, suggest thatLAST uptake activity was unchanged by A�. Further-

ore, as previously suggested by other studies (Kanai et t

Please cite this article in press as: Matos M, et al., Amyloid-beta peptidment of oxidative stress and mitogen-activated protein kinase cascades022

l., 1994; Anderson and Swanson, 2000; Lauderback etl., 2001), our results corroborate the fact that GLT-1 is aore efficient transporter than GLAST, accounting for 42%f the total glutamate transport (Fig. 3). There is a view thatonceives GLAST expression predominating at earlytages of CNS development, while GLT-1 expression pro-ressively augments during its maturation (Abe and Mi-awa, 2003b; Gegelashvili et al., 1997). However, our dataorroborate other observations showing dynamic GLT-1ctivity in cultured postnatal astrocytes (Thorlin et al.,998; Fernández-Tomé et al., 2004; Rothstein et al.,005). The controversial reports concerning the GLT-1 andLAST prevalence in cultured astrocytes might be due toifferent culturing procedures or settings, such as levels ofstrocytic differentiation and confluence.

The putative non-competitive inhibition of glutamateransport uptake induced by A� might also be due tolterations in glutamate transporter levels. In fact, our datahow that the levels of both GLT-1 and GLAST wereecreased in A�-treated astrocytes (Fig. 4). Although the

evels of GLAST were decreased by A�1-40, this reductioneems not to be so significant to the final astrocytic gluta-ate uptake activity inhibition as that of GLT-1. Indeed webserved, using glutamate transporter inhibitors, thatLT-1 had a greater participation in glutamate uptake and

eems preferentially affected by the peptide (Fig. 3). It islso possible that GLAST is normally expressed at lowerbsolute levels than GLT-1 in our cultured astrocytes, andhus a similar reduction in their levels after A� treatmentid not contribute equally to the general decrease in glu-amate uptake. The data obtained suggest that in A�-reated astrocytes the decreased glutamate uptake activitys characterized by a lower Vmax and a more pronouncednhibition in GLT-1 activity than in control cells. In accor-ance with our results, it has been shown that A� de-reases astrocytic glutamate uptake through GLT-1 inhibi-ion in AD brain (Li et al., 1997; Lauderback et al., 2001)nd the reduced GLT-1 expression was associated with arogressive decrease in glutamate transport capacity inrain of transgenic mice developing tau pathology (Dabir etl., 2006). Downregulation or abnormal expression ofLAST/EAAT1 or GLT-1/EAAT2 has also been detected

n amyloid precursor protein (APP) transgenic miceMasliah et al., 2000) and postmortem human brain tissueScott et al., 2002). It is now known that GLT-1 and GLASTontain sites for intracellular phosphorylation (Andersonnd Swanson, 2000) and, alterations in phosphorylationtate of these transporters, induced by intracellular path-ays acting on phosphatases or kinases, can possibly

egulate transporter activities. It is also recognized that A�,cting at the cell membrane or after being endocytosed,an trigger intracellular signaling pathways in astrocytesWyss-Coray et al., 2003), which may affect glutamateransporters levels/activity (Fernández-Tomé et al., 2004).

To clarify the mechanisms involved in A�-induced in-ibition of astrocytic glutamate transporters activity and

evels, we tested the possible involvement of MAPKs sig-aling pathways. Previous studies have reported that as-

roglial MAPK signaling can be involved in the regulation of

e decreases glutamate uptake in cultured astrocytes: Involve-, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.

g2mrbaaoodprgIintiwptbEtrEtstJAcdtlt

ti(Atsia2ahpFtv2aamst1

FcathEiteEntffG

fTtsa(esAts(2sRotAeTAccRolecaotbwrwtitagJo

M. Matos et al. / Neuroscience xx (2008) xxx10

ARTICLE IN PRESS

lutamate transporters expression (Gegelashvili et al.,000; Rodriguez-Kern et al., 2003) and these pathways,ainly MAPK/ERK, are activated under AD conditions (Fer-

er et al., 2001; Webster et al., 2006). However, it remains toe clarified if all MAPK pathways can affect the regulation ofstrocytic glutamate transporters and whether they areffected in AD. Our data show for the first time a clear rolef ERK, JNK and p38 signaling cascades in the regulationf astrocytic glutamate transporter activity (Fig. 6). Theata obtained also imply the existence of an intricate com-ensatory balance between all these MAPK cascades,esulting in diverse consequences at the general astrocyticlutamate uptake activity and levels of GLT-1 and GLAST.ndeed, ERK and p38 signaling cascades may act asmportant stimulatory glutamate uptake signaling mecha-isms in control cells and seem to be abolished in A�1-40-

reated astrocytes. The JNK cascade seems to be onlynvolved in the glutamate uptake inhibition when A�1-40

as present, probably acting as an inhibitor of the trans-orter activity. Therefore, the decrease in glutamate up-ake activity observed in A�-treated astrocytes seems toe due to alterations in signaling pathways operated byRK, JNK and p38 kinases caused by the peptide. Fur-

hermore, we have found that JNK pathway can down-egulate GLT-1 levels, whereas p38 upregulates it andRK has no visible effect in control astrocytes (Fig. 7). In

he presence of A�1-40 the ERK and JNK pathwayseemed to have a high contribution to GLT-1 level reduc-ion. The GLAST expression seems to be promoted by theNK and p38, but not ERK, pathways in control cells.lterations in the pathways operated by ERK and p38,aused by A�, might be responsible for the GLAST levelsecrease (Fig. 8). Likewise, Abe and Misawa (2003a) foundhat the ERK signaling pathway mediates GLAST downregu-ation, whereas Jayakumar and colleagues (2006) reportedhat JNK and p38 may upregulate its levels.

In our cultured astrocytes A�1-40 transiently increasedhe phosphorylation state of ERK, JNK and p38, the max-mum of activation being achieved upon 3 h incubationFig. 5), which indicates an early indiscriminate action of� on the activation of each MAPK. An extensive activa-

ion of ERK in astroglial cells was also observed in earlytages of AD, whereas in advanced AD the phospho-ERK

mmunoreactivity was associated with neuronal cell bodiesnd dystrophic neurites around plaques (Webster et al.,006). A previous report stated that JNK and p38 MAPKsre not active in rat astrocytes (Abe and Misawa, 2003a),owever, we observed that A� induced astrocytic JNK and38 activation in an early phase of peptide treatment.urthermore, our data are in agreement with the observa-

ion that the MEK1/2-ERK1/2 signaling pathway is acti-ated by A� (Mandell and VandenBerg, 1999; Abe et al.,003) and regulates astrocytic glutamate transport (Abend Misawa, 2003a,b). Nevertheless the studies by Abend Misawa (2003b) suggested that A� promotes gluta-ate uptake, which is in contrast with our data and other

tudies that report a downregulation of glutamate uptakeriggered by A� (Harris et al., 1996; Parpura-Gill et al.,

997; Lauderback et al., 1999; Mattson and Chan, 2003; e

Please cite this article in press as: Matos M, et al., Amyloid-beta peptidment of oxidative stress and mitogen-activated protein kinase cascades022

ernández-Tomé et al., 2004). Our data suggested that A�auses the repression of ERK and p38 signaling cascadesnd exacerbation of JNK pathway, resulting in the reduc-ion of astrocytic glutamate transporters activity. Since weave also found that A� increased the phosphorylation ofRK, JNK and p38 in an early time of exposure (Fig. 5), it

s likely that the modulatory effect of peptide on glutamateransporters activity/levels occurs at a downstream step ofach MAPK signaling cascade. Although A�1-40 led toRK, JNK an p38 activation/phosphorylation, it does notecessarily mean that whole pathway is stimulated, sincehe peptide may also interfere with other downstream ef-ectors/kinases, adapter proteins or related transcriptionactors, causing an inhibitory effect on the GLT-1 andLAST transporter activity.

There are several possible mechanisms responsibleor glutamate uptake impairment in A�-treated astrocytes.he Na�,K�-ATPase inhibition by A� could affect the

rans-membrane gradients of Na� and K�, which are es-ential for glutamate transport (Harris et al., 1996). Alter-tions in intracellular Ca2� (LaFerla, 2002), growth factorsRodriguez-Kern et al., 2003) and secreted �APP (Masliaht al., 2000) induced by A� may also control the differentignaling pathways operated by kinases or phosphatases.dditionally, there is now a plentiful of data demonstrating

hat oxidative stress may trigger and modulate the MAPKignaling pathways, mainly JNK and p38 pathwaysDaniels et al., 2001; Zhu et al., 2002; McCubrey et al.,006). Furthermore, there are evidences of oxidativetress in the AD brain, and A� has been found to generateOS (Chauan and Chauan, 2006). High catalase, super-xide dismutase (SOD-1) and glutathione peroxidase ac-ivities confer astrocytes with a larger resistance against� injury than neurons (Abe and Misawa, 2003a). How-ver, astrocytes are also vulnerable to oxidative stress.heir reaction to low concentrations of H2O2 induced by� involves generation of ROS, lipoperoxidation, andhanges in antioxidant defenses that can all affect astro-ytic membrane transporter systems (Brera et al., 2000;ohrdanz et al., 2001; Mattson and Chan, 2003). Thus,xidation of glutamate transporters and/or of membrane

ipids, mainly through the production of 4-hydroxy-2-non-nal (HNE), may be involved in A�-induced glutamatelearance impairment (Harris et al., 1996; Lauderback etl., 2001). In our studies we also found that A� inducesxidative stress in astrocytes through the increase in pro-ein carbonyls groups and lipid peroxidation products (Ta-le 1). Furthermore, these increases in oxidative productsere largely prevented by the addition of trolox, a free

adical scavenger analog of vitamin E. Moreover, troloxas able to revert almost completely the decrease in glu-

amate uptake induced by A� (Fig. 9). These results clearlymplicate the harmful effects of oxidative stress in relationo glutamate uptake impairment, linking to the possibility ofn altered MAPK signaling function that ultimately affectslutamate transporter activity and/or levels. Thus, ERK,NK and p38 pathways may act in concert (possibly withther mechanisms) and depending on the nature of the

xtracellular/intracellular factor (s) (e.g. A�, ROS), the

e decreases glutamate uptake in cultured astrocytes: Involve-, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.

et

IgcdgaJtiiatstts

ABr

A

A

A

A

A

A

A

A

B

B

B

B

C

C

D

D

F

F

F

FaAtD

acfic#

M. Matos et al. / Neuroscience xx (2008) xxx 11

ARTICLE IN PRESS

quilibrium between them may favor inhibition or stimula-ion of the glutamate transporter activity and levels.

CONCLUSION

n conclusion, the data obtained show that A� decreaseslutamate uptake in cultured astrocytes through a non-ompetitive mode of inhibition that can be explained by theecrease in GLAST and GLT-1 levels and activity. Thiseneral inhibition in glutamate transport function seemslso partly mediated through the altered activity of ERK,NK and p38 signaling pathways that can be triggered byhe oxidative stress conditions induced by A�. These find-ngs strongly suggest that neurodegeneration in AD maynvolve an excitotoxic process caused by a decreasedstrocytic glutamate uptake function as a result of oxida-ive stress that can lead to an altered activity of MAPKignaling pathways. Further elucidation of these signalingransducing pathways and the fine details of their cross-alk may contribute to the development of new therapeutictrategies against this dementia.

cknowledgments—The authors would like to thank Fundaçãoissaya Barreto (Coimbra, Portugal) for supporting Marco Matos’s

esearch fellowship.

REFERENCES

be K, Misawa M (2003a) The extracellular signal-regulated kinase

ig. 9. The effect of trolox on the transport of D-[3H] aspartate intostrocytes treated or not with A�. The astrocytes were incubated with�1-40 (5 �M) and/or vitamin E analog trolox (100 �M) for 24 h before

he uptake. Subsequently the astrocytes were incubated with 100 �M-aspartate (and 0.1 �Ci/ml D-[3H]aspartate) for 10 min to assess themino acid uptake. The results are expressed as the percentage ofontrol transport (untreated cells) and are the mean�S.E.M. of at leastve individual determinations measured in triplicate. * P�0.05, indi-ates statistical significance related to control while # P�0.05 and# P�0.01, indicate statistical significance to the A� condition.

cascade suppresses amyloid beta protein-induced promotion of

Please cite this article in press as: Matos M, et al., Amyloid-beta peptidment of oxidative stress and mitogen-activated protein kinase cascades022

glutamate clearance in cultured rat cortical astrocytes. Brain Res979:179–187.

be K, Misawa M (2003b) Amyloid beta protein enhances the clear-ance of extracellular L-glutamate by cultured rat cortical astro-cytes. Neurosci Res 45:25–31.

be K, Hisatomi R, Misawa M (2003) Amyloid beta peptide specificallypromotes phosphorylation and nuclear translocation of the extra-cellular signal-regulated kinase in cultured rat cortical astrocytes.J Pharmacol Sci 93:272–278.

gostinho P, Duarte CB, Oliveira CR (1997) Impairment of excitatoryamino acid transporter activity by oxidative stress conditions inretinal cells: effect of antioxidants. FASEB J 11:154–163.

gostinho P, Oliveira CR (2003) Involvement of calcineurin in theneurotoxic effects induced by amyloid-beta and prion peptides. EurJ Neurosci 17:1189–1196.

gostinho P, Oliveira CR (2007) Neuroinflammation and mitochondrialdysfunction in Alzheimer’s and prion’s diseases. In: Interactionbetween neurons and glia in aging and disease (Malva JO, RegoAC, Cunha RA, Oliveira CR), pp 309–332. New York: SpringerPublishers.

kiyama H, Barger S, Barnum S, Bradt B, Bauer J, Cole GM, CooperNR, Eikelenboom P, Emmerling M, Fiebich BL, Finch CE, Fraut-schy S, Griffin WS, Hampel H, Hull M, Landreth G, Lue L, Mrak R,Mackenzie IR, McGeer PL, O’Banion MK, Pachter J, Pasinetti G,Plata-Salaman C, Rogers J, Rydel R, Shen Y, Streit WJ, Strohm-eyer R, Tooyoma I, Van Muiswinkel FL, Veerhuis R, Walker D,Webster S, Wegrzyniak B, Wenk G, Wyss-Coray T (2000) Inflam-mation and Alzheimer’s disease. Neurobiol Aging 21:383–421.

nderson CM, Swanson RA (2000) Astrocyte glutamate transport:review of properties, regulation, and physiological functions. Glia32:1–14.

erry CB, Hayes D, Murphy A, Wiessner M, Rauen T, McBean GJ(2005) Differential modulation of the glutamate transporters GLT-1,GLAST and EAAC1 by docosahexaenoic acid. Brain Res 1037:123–133.

lennow K, de Leon MJ, Zetterberg H (2006) Alzheimer’s disease.Lancet 368:387–403.

rera B, Serrano A, Ceballos ML (2000) Beta-amyloid peptide arecytotoxic to astrocytes in culture: a role for oxidative stress. Neu-robiol Dis 7:395–405.

utterfield DA, Howard BJ, Yatin S, Allen KL, Carney JM (1997) Freeradical oxidation of brain proteins in accelerated senescenceand its modulation by N-tert-butyl-alpha-phenylnitrone. ProcNatl Acad Sci U S A 94:674 – 678.

elsi F, Svedberg M, Unger C, Cotman CW, Carrì MT, Ottersen OP,Nordberg A, Torp R (2007) Beta-amyloid causes downregulation ofcalcineurin in neurons through induction of oxidative stress. Neu-robiol Dis 26:342–352.

hauan V, Chauan A (2006) Oxidative stress in Alzheimer disease.Pathophysiology 13:195–208.

abir DV, Robinson MB, Swanson E, Zhang B, Trojanowski JQ, LeeVM, Forman MS (2006) Impaired glutamate transport in amouse model of tau pathology in astrocytes. J Neurosci 26:644 – 654.

aniels WM, Hendricks J, Salie R, Taljaard JJ (2001) The role of theMAP-kinase superfamily in beta-amyloid toxicity. Metab Brain Dis16:175–185.

ernández-Tomé P, Brera B, Arévalo M, Ceballos ML (2004) �-Amy-loid 25–35 inhibits glutamate uptake in cultured neurons and as-trocytes: modulation of uptake as a survival mechanism. NeurobiolDis 15:580–589.

errarese C, Begni B, Canevari C, Zoia C, Piolti R, Frigo M, AppollonioI, Frattola L (2000) Glutamate uptake is decreased in platelets fromAlzheimer’s disease patients. Ann Neurol 47:641–643.

erreiro E, Resende R, Costa R, Oliveira CR, Pereira CM (2006) An

endoplasmic-reticulum-specific apoptotic pathway is involved in

e decreases glutamate uptake in cultured astrocytes: Involve-, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.

F

F

F

G

G

G

H

H

H

J

K

K

K

L

L

L

L

L

M

M

M

M

M

M

P

P

R

R

R

R

R

R

S

S

S

S

S

T

W

M. Matos et al. / Neuroscience xx (2008) xxx12

ARTICLE IN PRESS

prion and amyloid-beta peptides neurotoxicity. Neurobiol Dis 23:669–678.

errer I, Blanco R, Carmona M, Puig B (2001) Phosphorylated mito-gen-activated protein kinase (MAPK/ERK-P), protein kinase of 38kDa (p38-P), stress-activated protein kinase (SAPK/JNK-P), andcalcium/calmodulin-dependent kinase II (CaM kinase II) are differ-entially expressed in tau deposits in neurons and glial cells intauopathies. J Neural Transm 108:1397–1415.

igiel M, Maucher T, Rozyczka J, Bayatti N, Engele J (2004) Regula-tion of glial glutamate transporter expression by growth factors.Exp Neurol 183:124–135.

rancis PT (2003) Glutamatergic systems in Alzheimer’s disease. Int JGeriatr Psychiatry 18:S15–S21.

arção P, Oliveira CR, Agostinho P (2006) Comparative study ofmicroglia activation induced by amyloid-beta and prion peptides:role in neurodegeneration. J Neurosci Res 84:182–193.

egelashvili G, Schousboe A (1997) High affinity glutamate trans-porters: regulation of expression and activity. Mol Pharmacol52:6 –15.

egelashvili G, Dehnes Y, Danbolt NC, Schousboe A (2000) Thehigh-affinity glutamate transporters GLT-1, GLAST, and EAAT4are regulated via different signaling mechanisms. Neurochem Int37:163–170.

arris ME, Carney JM, Cole PS, Hensley K, Howard BJ, Martin L,Bummer P, Wang Y, Pedigo NW Jr, Butterfield DA (1995) Beta-amyloid peptide-derived, oxygen-dependent free radicals inhibitglutamate uptake in cultured astrocytes: implications for Alzhei-mer’s disease. Neuroreport 6:1875–1879.

arris ME, Wang Y, Pedigo NW, Hensley K, Butterfield DA, Carney JM(1996) Amyloid � peptide (25–35) inhibits Na�-dependent gluta-mate uptake in rat hippocampal astrocyte cultures. J Neurochem67:277–286.

ynd MR, Scott HL, Dodd PR (2004) Glutamate-mediated excitotox-icity and neurodegeneration in Alzheimer’s disease. NeurochemInt 45:583–595.

ayakumar AR, Panickar KS, Murthy CHRK, Norenberg MD (2006)Oxidative stress and mitogen-activated protein kinase phos-phorylation mediate ammonia-induced cell swelling and gluta-mate uptake inhibition in cultured astrocytes. J Neurosci26:4774 – 4784.

anai Y, Smith CP, Hediger MA (1994) A new family of neurotrans-mitter transporters: the high-affinity glutamate transporters.FASEB J 8:1450–1459.

lein WL, Stine WB Jr, Teplow DB (2004) Small assemblies of un-modified amyloid beta-protein are the proximate neurotoxin in Alz-heimer’s disease. Neurobiol Aging 25:569–580.

oistinaho M, Lin S, Wu X, Esterman M, Koger D, Hanson J, Higgs R,Liu F, Malkani S, Bales KR, Paul SM (2004) Apolipoprotein Epromotes astrocyte colocalization and degradation of depositedamyloid-beta peptides. Nat Med 10:719–726.

aFerla FM (2002) Calcium dyshomeostasis and intracellular signalingin Alzheimer’s disease. Nat Rev Neurosci 11:862–872.

auderback CM, Harris-White ME, Wang Y, Pedigo NW Jr, CarneyJM, Butterfield DA (1999) Amyloid beta-peptide inhibits Na�-de-pendent glutamate uptake. Life Sci 65:1977–1981.

auderback CM, Hackett JM, Huang FF, Keller JN, Szweda LI,Markesbery WR, Butterfield DA (2001) The glial glutamate trans-porter, GLT-1, is oxidatively modified by 4-hydroxy-2-nonenal inthe Alzheimer’s disease brain: the role of Abeta1-42. J Neurochem78:413–416.

i S, Mallory M, Alford M, Tanaka S, Masliah E (1997) Glutamatetransporter alterations in Alzheimer disease are possibly associ-ated with abnormal APP expression. J Neuropathol Exp Neurol56:901–911.

iang Z, Valla J, Sefidvash-Hockley S, Rogers J, Li R (2002) Effects ofestrogen treatment on glutamate uptake in cultured human astro-cytes derived from cortex of Alzheimer’s disease patients. J Neu-

rochem 80:807–814.

Please cite this article in press as: Matos M, et al., Amyloid-beta peptidment of oxidative stress and mitogen-activated protein kinase cascades022

andell JW, VandenBerg SR (1999) ERK/MAP kinase is chronicallyactivated in human reactive astrocytes. NeuroReport 10:3567–3572.

aragakis NJ, Rothstein JD (2006) Mechanisms of disease: astro-cytes in neurodegenerative disease. Nat Clin Pract Neurol2:679 – 689.

asliah E, Alford M, DeTeresa R, Mallory M, Hansen L (1996) Defi-cient glutamate transport is associated with neurodegeneration inAlzheimer’s disease. Ann Neurol 40:759–766.

asliah E, Alford M, Mallory M, Rockenstein E, Moechars D, LeuvenFV (2000) Abnormal glutamate transport function in mutant amy-loid precursor protein transgenic mice. Exp Neurol 163:381–387.

attson MP, Chan SL (2003) Neuronal and glial calcium signaling inAlzheimer’s disease. Cell Calcium 34:385–397.

cCubrey JA, LaHair MM, Franklin RA (2006) Reactive oxygen spe-cies-induced activation of the MAP kinase signaling pathways.Antioxid Redox Signal 8:1775–1789.

arpura-Gill A, Beitz D, Uemura E (1997) The inhibitory effects ofbeta-amyloid on glutamate and glucose uptakes by cultured astro-cytes. Brain Res 754:65–71.

ekny M, Nilsson M (2005) Astrocyte activation and reactive gliosis.Glia 50:427–434.

esende R, Pereira C, Agostinho P, Vieira AP, Malva JO, Oliveira CR(2007) Susceptibility of hippocampal neurons to Abeta peptidetoxicity is associated with perturbation of Ca2� homeostasis. BrainRes 1143:11–21.

esende R, Ferreiro E, Pereira C, Oliveira CR (2008) Neurotoxiceffect of oligomeric and fibrillar species of Abeta1-42 peptide.Involvement of ER calcium release in oligomers-induced celldeath. Neuroscience, in press.

odriguez-Kern A, Gegelashvili M, Schousboe A, Zhang J, Sung L,Gegelashvili G (2003) Beta-amyloid and brain-derived neurotro-phic factor, BDNF, up-regulate the expression of glutamate trans-porter GLT-1/EAAT2 via different signaling pathways utilizing tran-scription factor NF-kappaB. Neurochem Int 43:363–370.

ohrdanz E, Schmuck G, Ohler S, Tran-Thi QH, Kahl R (2001)Changes in antioxidant enzyme expression in response to hy-drogen peroxide in rat astroglial cells. Arch Toxicol 75:150 –158.

othstein JD, Van Kammen M, Levey AI, Martin LJ, Kuncl RW (1995)Selective loss of glial glutamate transporter GLT-1 in amyotrophiclateral sclerosis. Ann Neurol 38:73–84.

othstein JD, Patel S, Regan MR, Haenggeli C, Huang YH, BerglesDE, Jin L, Hoberg MD, Vidensky S, Chung DS, Toan SV, Bruijn LI,Su Z, Gupta P, Fisher PB (2005) B-lactam antibiotics offer neuro-protection by increasing glutamate transporter expression. Nature433:73–77.

aura J (2007) Microglial cells in astroglial cultures: a cautionary note.J Neuroinflammation 15:4–26.

cott HL, Tannenberg AE, Dodd PR (1995) Variant forms of neuronalglutamate transporter sites in Alzheimer’s disease cerebral cortex.J Neurochem 64:2193–2202.

cott HL, Pow DV, Tannenberg AE (2002) Aberrant expression of theglutamate transporter excitatory amino acid transporter 1 (EAAT1)in Alzheimer’s disease. J Neurosci 22:1–5.

treit WJ (2004) Microglia and Alzheimer’s disease pathogenesis.Neurosci Res 77:1–8.

ultana R, Butterfield DA (2008) Alterations of some membrane trans-port proteins in Alzheimer’s disease: role of amyloid �-peptide. MolBiosyst 4:36–41.

horlin T, Roginski RS, Choudhury K, Nilsson M, Rönnbäck L, Hans-son E, Eriksson PS (1998) Regulation of the glial glutamate trans-porter GLT-1 by glutamate and delta-opioid receptor stimulation.FEBS Lett 425:453–459.

ebster B, Hansen L, Adame A, Crews L, Torrance M, Thal L, MasliahE (2006) Astroglial activation of extracellular-regulated kinase inearly stages of Alzheimer disease. J Neuropathol Exp Neurol

65:142–151.

e decreases glutamate uptake in cultured astrocytes: Involve-, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.

W

W

Z

Z

M. Matos et al. / Neuroscience xx (2008) xxx 13

ARTICLE IN PRESS

hite JA, Manelli AM, Holmberg KH, Van Eldik LJ, Ladu MJ (2005)Differential effects of oligomeric and fibrillar amyloid-beta 1-42on astrocyte-mediated inflammation. Neurobiol Dis 18:459 –465.

yss-Coray T, Loike JD, Brionne TC, Lu E, Anankov R, Yan F,Silverstein SC, Husemann J (2003) Adult mouse astrocytes

degrade amyloid-beta in vitro and in situ. Nat Med 9:453– 457.

Please cite this article in press as: Matos M, et al., Amyloid-beta peptidment of oxidative stress and mitogen-activated protein kinase cascades022

hu X, Lee HG, Raina AK, Perry G, Smith MA (2002) The role ofmitogen-activated protein kinase pathways in Alzheimer’s disease.Neurosignals 11:270–281.

oia CP, Tagliabue E, Isella V, Begni B, Fumagalli L, Brighina L,Appollonio I, Racchi M, Ferrarese C (2005) Fibroblast glutamatetransport in aging and in AD: correlations with disease severity.

Neurobiol Aging 26:825–832.

(Accepted 11 August 2008)

e decreases glutamate uptake in cultured astrocytes: Involve-, Neuroscience (2008), doi: 10.1016/j.neuroscience.2008.08.