Oligodendrocytes Inhibition Mediates Microglial Toxicity to and Glutamate

System xc− and Glutamate Transporter Inhibition Mediates Microglial Toxicity to Oligodendrocytes

This information is current as María Domercq, María Victoria Sánchez-Gómez, Catherine of October 1, 2021. Sherwin, Estibaliz Etxebarria, Robert Fern and Carlos Matute J Immunol 2007; 178:6549-6556; ; doi: 10.4049/jimmunol.178.10.6549 http://www.jimmunol.org/content/178/10/6549 Downloaded from

References This article cites 49 articles, 18 of which you can access for free at: http://www.jimmunol.org/content/178/10/6549.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on October 1, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

؊ System xc and Glutamate Transporter Inhibition Mediates Microglial Toxicity to Oligodendrocytes1

Marıá Domercq,* Marıá Victoria Sańchez-Go´mez,* Catherine Sherwin,† Estibaliz Etxebarria,* Robert Fern,† and Carlos Matute2*

Elevated levels of extracellular glutamate cause excitotoxic oligodendrocyte cell death and contribute to progressive oligodendrocyte loss and demyelination in white matter disorders such as multiple sclerosis and periventricular leukomalacia. However, the mechanism by which glutamate homeostasis is altered in such conditions remains elusive. We show here that microglial cells, in their activated state, compromise glutamate homeostasis in cultured oligodendrocytes. Both activated and resting microglial cells ؊ release glutamate by the cystine-glutamate antiporter system xc . In addition, activated microglial cells act to block glutamate transporters in oligodendrocytes, leading to a net increase in extracellular glutamate and subsequent oligodendrocyte death. ؊ ␣ The blocking of -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/kainate receptors or the system xc anti- Downloaded from porter prevented the oligodendrocyte injury produced by exposure to LPS-activated microglial cells in mixed glial cultures. In a whole-mount rat optic nerve, LPS exposure produced wide-spread oligodendrocyte injury that was prevented by ؊ AMPA/kainate receptor block and greatly reduced by a system xc antiporter block. The cell death was typiﬁed by swelling and disruption of mitochondria, a feature that was not found in closely associated axonal mitochondria. Our results reveal a novel mechanism by which reactive microglia can contribute to altering glutamate homeostasis and to the pathogenesis of white matter disorders. The Journal of Immunology, 2007, 178: 6549–6556. http://www.jimmunol.org/

ctivation of microglia is a natural response to CNS in- In addition, the paralysis of microglia and/or macrophages ar- fection to eliminate cell debris after injury and to support rests the progression of experimental autoimmune encephalitis A tissue repair. Microglia can also act as APCs and can (EAE) (14), an animal model of MS showing oligodendrocyte release soluble factors such as cytokines, heat shock proteins, death, myelin loss, and axonal damage. Thus, understanding and acute phase proteins that propagate and perpetuate the ce- how microglia contributes to oligodendrocyte cell death is cru- rebral immune response. However, an excessive and sustained cial for the development of therapeutic approaches to white activation of microglial cells is detrimental to neurons and oli- matter disorders. godendrocytes (1–3). Microglia activation is present in most A mechanism of cellular injury unique to the CNS is excitotox- by guest on October 1, 2021 pathological conditions in the CNS (4), but whether it plays a icity, which is caused by excessive extracellular glutamate and the beneficial or destructive role is still under debate (5). In auto- subsequent overstimulation of glutamate receptors in neurons and 3 immune diseases such as multiple sclerosis (MS), most data oligodendrocytes (15–18). Glutamate homeostasis is altered in point to a detrimental role of microglia (6, 7). Indeed, LPS- EAE because glutamate-metabolizing enzymes and glutamate activated microglia induce oligodendrocyte death by releasing transporters are down-regulated (19, 20) and treatment with an proinflammatory cytokines and chemokines and mediators of antagonist of the glutamate receptor ameliorates both oligodendro- cell injury such as reactive oxygen species and NO (1, 8–13). cyte cell death and neurological deficits associated with this experimental condition, although it does not prevent inflammation (21, 22). These findings suggest that autoimmunity to myelin and *Departamento de Neurociencias, Universidad del Paı´s Vasco, Leioa, Vizcaya, Spain; and †Department of Cell Physiology and Pharmacology, University of Leicester, the subsequent microglial activation induce an alteration in gluta- United Kingdom mate homeostasis leading to oligodendrocyte excitotoxicity, which Received for publication September 14, 2006. Accepted for publication March contributes to tissue damage in MS (23). Indeed, active MS lesions 5, 2007. are characterized by the presence of activated microglia and/or The costs of publication of this article were defrayed in part by the payment of page macrophage/monocyte infiltrates and by the loss of oligodendro- charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. cytes concomitant with an increase in glutamate levels (24). A 1 This work was supported by the University of Paı´s Vasco, Ministerio de Educacioń possible source of glutamate is from the cystine-glutamate anti- Ϫ y Ciencia and by Ministerio de Salud y Consumo (to C.M.) and National Institutes of porter, called system xc , which exchanges extracellular cystine Health Grant NS44875 (to R.F.). Ϫ for intracellular glutamate (25). The system xc is a heterodimeric 2 Address correspondence and reprint requests to Dr. Carlos Matute, Departamento de protein complex consisting of a catalytic L chain (xCT) and a Neurociencias, Universidad del Paı´s Vasco, Leioa, Vizcaya, Spain. E-mail address: [email protected] regulatory H chain (4F2hc), which is essential for membrane Ϫ 3 Abbreviations used in this paper: MS, multiple sclerosis; AAA, aminoadipic acid; AM, localization of the transporter (26). The system xc appears to acetoxymethyl ester; AMPA, ␣-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; be especially abundant in macrophage/microglia and immune aCSF, artificial cerebrospinal fluid; CNQX, 6-cyano-7-nitroquinoxaline-2,3-dione; EAE, cells (27, 28). experimental autoimmune encephalitis; JC-1, 5,5Ј,6,6Ј-tetrachloro-1,1Ј,3,3Ј-tetraethylbenzimidazolcarbocyanine iodide; NBQX, 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo- In this study we analyze the effects of resting and LPS-acti- (f)quinoxaline; RON, rat optic nerve; ROS, reactive oxygen species; TBOA, vated microglia in glutamate homeostasis and excitotoxicity in DL-threo-␤-benzyloxyaspartate. cocultured oligodendrocytes. We show that microglia release Ϫ Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 glutamate by system xc and alter glutamate homeostasis by www.jimmunol.org 6550 MICROGLIA INDUCED OLIGODENDROGLIAL EXCITOTOXICITY

inhibiting glutamate transporters in oligodendrocytes. In addi- genase in the presence of NADPϩ (1 mM) to ␣-ketoglutarate with the tion, we provide evidence indicating that extracellular gluta- formation of NADPH and fluorescence emission at 430 nM (excitation mate accumulation after microglia activation is toxic to oligo- light 335 nm). Release was quantified in reference to standard curves con- structed with exogenous glutamate. dendrocytes in vitro and that a similar mechanism operates in isolated optic nerves. Western blotting Materials and Methods Western blot analysis of cystine-glutamate antiporter expression was done in microglial cultures by conventional SDS-PAGE polyacrylamide Oligodendrocyte and microglia cultures Ϫ electrophoresis. The system xc was visualized with Abs against the Oligodendrocyte cultures were obtained from optic nerves of 12-day-old cystine-glutamate antiporter xCT (1/250) generously provided by Dr. P. Sprague-Dawley rats using a previously described procedure (29) with Kalivas (Medical University of South Carolina, Charleston, SC; Ref. modifications (16). Cells were seeded onto 24-well plates bearing 10 ␮g/ml 33). Densitometric analysis was performed using the NIH Image pro- poly-D-lysine-coated coverslips at a density of 5,000 or 25,000 cells per gram (n ϭ 3). well for toxicity and uptake experiments, respectively. Cell culture medium (see Refs. 16 and 29) was supplemented with N-acetyl-cysteine to avoid glutathione depletion and oxidative glutamate toxicity (30). Cultures were LPS toxicity assays used after 3–5 days in vitro when almost all cells (Ͼ98%) were O1ϩ/ galactocerebroside Cϩ differentiated oligodendrocytes. Microglial cells LPS (E. coli O11:B4; Sigma-Aldrich) was added to mixed cultures or were isolated from confluent monolayers of cultured astrocytes derived oligodendrocyte cultures 24 h after plating and cell death was determined from the cerebral cortex of newborn rats as previously described (31). after 48 h. To determine oligodendrocyte viability, we counted cells stained Astrocytes flasks were mechanical shacked and free-floating microglia with the oligodendrocyte O1 Ab marker and the vital dye calcein-AM, as

␮ Downloaded from were collected and puriﬁed by preplating on plastic dishes for 1 h. For follows. Cells were loaded with calcein-AM (1 M) for 30 min and ﬁxed oligodendrocyte-microglia coculture, microglial cells were seeded on oli- with 4% p-formaldehyde for 20 min. Cells were then processed with an ␮ godendrocyte cultures at a density 1,000 or 5,000 cells per well for toxicity mouse Ab to the O1 oligodendroglial marker (10 g/ml; Chemicon) and and uptake experiments, respectively. To activate microglia, cocultures labeling was visualized with Alexa 594-conjugated goat anti-mouse IgG (1/200; Molecular Probes). The number of living oligodendrocytes (O1ϩ were treated with LPS (100 ng/ml; Escherichia coli O11:B4 (Sigma- ϩ Aldrich) 1 day after plating and used for experiments after 48 h. Media were and calcein cells) in each well was counted blind to the treatments done conditioned by microglial cultures for 48 h and added immediately to oli- and results were expressed as the percentage of cell death vs nontreated oligodendrocyte pure cultures considered as control.

godendrocytes cultures. Cocultures were characterized by immunocyto- http://www.jimmunol.org/ chemistry to the oligodendrocyte marker galactocerebroside C (5 ␮g/ml; The changes in mitochondrial membrane potential were monitored with Ј Ј Ј Ј Boehringer-Mannheim) and by labeling with Griffonia simplicifolia isolec- 5,5 ,6,6 -tetrachloro-1,1 ,3,3 -tetraethylbenzimidazolcarbocyanine iodide tin B4 (10 ␮g/ml; Vector Laboratories), a selective marker of both resting (JC-1; Molecular Probes). In the cytosol the monomeric form of this dye and activated microglia. fluoresces green (emission read at 527 nm), whereas within the mitochondrial matrix highly concentrated JC-1 forms aggregates that fluoresce red. After drug treatment, cells were loaded with JC-1 (3 ␮M) at 37°C for 30 Electrophysiology min. JC-1 aggregates were detectable in the propidium iodide channel Standard whole-cell recordings were performed at room temperature on an (emission at 590), JC-1 monomers were detectable in the FITC channel inverted microscope (CK-40; Olympus) using the EPC-7 patch-clamp am- (emission at 520), and the changes in mitochondrial potential were calcu- plifier (HEKA) as described (32). Recordings were low-pass filtered at 2 lated as the red/green ratio in each condition. kHz, digitized at 5 kHz, and stored as data files on a computer using the by guest on October 1, 2021 pClamp 8.2 program (Axon Instruments) for later analysis. The extracel- Isolated rat optic nerve (RON) experiments lular bath solution contained 140 mM NaCl, 5.4 mM KCl, 2 mM CaCl2,1 mM MgCl2, and 10 mM HEPES (pH 7.3). Glutamate transporter currents RONs were dissected from 12-day-old lister hooded rats and placed in were recorded in the presence of intracellular thiocyanate to potentiate the artificial cerebrospinal fluid (aCSF) composed of 124 mM NaCl, 3 mM transporter anionic conductance. Patch clamp pipettes (3–5 megaohms) KCl, 2 mM CaCl2, 2 mM MgCl2, 26 mM NaHCO3,2mMNaH2PO4, and were filled with a solution containing 140 mM KSCN, 2 mM CaCl2,2mM 10 mM dextrose. LPS was dissolved in aCSF immediately before the ex- MgCl2, 10 mM HEPES, 11 mM EGTA, and 2 mM sodium ATP (pH 7.3). periment to give a final concentration of 1 ␮g/ml. RONs were incubated at Ϫ Currents were recorded at a holding membrane potential of 70 mV. Local 37°C in 1 ml of aCSF with or without LPS for 3 h, the period during which microperfusion of agonist was performed with a multibarreled apparatus maximal cell death was previously observed (12). Death cells were deter- connected to an electronically driven rotatory motor (RSC-100; mined using ethidium bromide staining, a lipophobic nuclear dye that does Bio-Logic). not stain intact living cells. RONs were incubated for 10 min at room temperature with 0.5 nM ethidium bromide and then washed in aCSF and 3 L-[ H]Glutamate uptake and glutamate release live mounted on a Nikon LABOPHOT-2 epifluorescence microscope equipped with a ϫ20 fluorescence objective and appropriate filter. Whole 3 L-[ H]Glutamate uptake was assessed at 37°C in 300 ␮l of saline solution RON images (excluding the cut ends) were captured with a charge-coupled containing 140 mM NaCl, 5 mM KCl, 2 mM MgCl2, 2 mM CaCl2,10mM device camera averaging 16 frames. Cell counting was performed blind HEPES, and 4.5 g/L glucose (pH 7.4). Cells were equilibrated for 5 min in using NIH ImageJ. The mean number of dead cells found following ex- the absence or presence of uptake inhibitors and then incubated for 10 min posure to any given condition was compared with the number seen in 3 with 100 nM L-[ H]glutamate and unlabeled glutamate (1 ␮M) as previ- control RONs (180 min aCSF only) and is expressed as a percentage ously described (31). Uptake was stopped with two washes in ice-cold change relative to this level of control cell death. PBS, followed immediately by cell lysis in 0.1 N NaOH and 0.1% Triton X-100. Sodium-dependent uptake was calculated to be the difference between the amount of radioactivity in the presence of sodium and the Electron microscopy amount observed in the choline-containing buffer. Sodium-dependent glutamate uptake was normalized to the number of viable cells analyzed by RONs were exposed to 1 ␮g/ml LPS in aCSF for 180 min before washing fluorescence measurement using a Synergy-HT fluorometer (BioTek In- in Sorenson’s buffer and, postfixation, in 3% glutaraldehyde and Soren- struments) after loading cells with the vital dye calcein-acetoxymethyl es- son’s buffer for 90 min at room temperature. The RONs were subsequently ter (AM) (Molecular Probes) as described below. Cells were treated with postfixed with 2% osmium tetroxide and dehydrated in ethanol and pro- TNF-␣ (100 ng/ml; PeproTech), IL-1␤ (10 ng/ml; PeproTech), paraquat pylene oxide. The nerves were infiltrated in epoxy resin and the ends were (10 ␮M; Sigma-Aldrich), and the Abs anti-TNF-␣ (100 ␮g/ml; Genzyme) cut back before ultrathin sections were taken. Sections were counterstained or Trolox (10 ␮M; Sigma-Aldrich) for 48 h before measuring glutamate with uranyl acetate and lead citrate and examined with a JEOL 100CX uptake. electron microscope. To avoid bias in the data, electron micrographs were Glutamate release in pure and mixed cultures of oligodendrocytes and collected blind (by R.F.) to the experimental procedure used to produce microglia was monitored by an enzymatic assay based on the activity of the each sample. Mitochondrial areas were drawn by hand using ImageJ. Ap- enzyme L-glutamate dehydrogenase (36 U/ml; Sigma-Aldrich) as previ- optosis was examined by immunoelectron microscopy of annexin-V (1:50; ously described (3, 32). Glutamate was oxidized by L-glutamate dehydro- Santa Cruz Biotechnology), an early marker. The Journal of Immunology 6551 Downloaded from http://www.jimmunol.org/

FIGURE 1. Activated microglia inhibits Naϩ-dependent glutamate transporter in oligodendrocytes. a, Control (left) and LPS-treated (right) by guest on October 1, 2021 cocultures of oligodendrocytes (galactocerebroside Cϩ cells (GalC) in red) and microglia (IB4ϩ cells (IB4) in green). Note the reactive appearance of microglia in the presence of LPS. Scale bar, 50 ␮m. b,Naϩ-dependent glutamate uptake in pure oligodendrocyte cultures (OL), oligodendrocyte- microglia cocultures (OLϩMG) with or without LPS (100 ng/ml), or OL cultures treated with microglia-conditioned medium with or without LPS (100 ng/ml). Glutamate uptake was normalized to the number of oligodendrocytes in each condition and expressed as a percentage of that measured p Ͻ 0.001. c, Membrane currents ,ءء .in pure OL cultures. Data are mean Ϯ SEM from at least three independent experiments performed in triplicate (top) in oligodendrocytes (OL) clamped at Ϫ70 mV in response to 1 mM D-aspartate applied for the duration indicated by the bars in the same culture conditions described before. The addition of microglia (OLϩMG) did not signiﬁcantly affect glutamate transporter currents in oligodendrocytes, whereas LPS-activated microglia (OLϩMGϩLPS) greatly diminished D-aspartate currents in oligodendrocytes. Bottom, Histogram depicts mean ϩ p Ͻ 0.01. d,Na -dependent glutamate ,ء .amplitudes Ϯ SEM of steady-state currents induced by D-aspartate application to 10–20 oligodendrocytes uptake in pure oligodendrocyte (OL) cultures and in LPS-treated oligodendrocyte plus microglia (OLϩMG) cultures in the absence or presence of TNF-␣ (100 ng/ml), IL-1␤ (10 ng/ml), paraquat (15 ␮M), anti-TNF Ab (100 ␮g/ml) and Trolox (10 ␮M). Data are mean Ϯ SEM from at least three independent experiments performed in triplicate. Symbols denote statistical signiﬁcance vs OL cultures or vs LPS-treated OLϩMG mixed cultures .p Ͻ 0.01; #, p Ͻ 0.05 ,ءء .(and #, respectively ء)

Statistical analysis that LPS-activated microglia, but not microglia in its resting Data are presented as mean Ϯ SEM. The significance between two data state, blocked glutamate uptake in oligodendrocyte-microglia sets was tested with unpaired t test or ANOVA as appropriate. cocultures (n ϭ 6; Fig. 1b). This effect was mainly due to im- paired glutamate transport in oligodendrocytes because control Results and LPS-treated microglia cultured alone lack significant Naϩ- Microglia alters glutamate homeostasis in cultured dependent glutamate uptake (data not shown). The blockade of oligodendrocytes glutamate transport by activated microglia was not secondary to Oligodendrocytes express functional glutamate transporters and an effect on cell viability, because glutamate uptake was nor- contribute to maintaining glutamate homeostasis in white mat- malized to the number of oligodendrocytes in each condition. In ter tracts (18, 31–32, 34). We therefore studied whether the addition, medium conditioned by LPS-activated microglia also function of glutamate transporters in these cells is altered by an induced a partial inhibition of glutamate uptake in oligodendro- activated microglia such as that found in inflammation. To this cytes (n ϭ 3; Fig. 1b). However, to further exclude any effect end, we added microglial cells to oligodendrocytes in vitro and on cell viability, we analyzed the effect of microglia on gluta- cells were subsequently incubated with LPS (100 ng/ml) to ac- mate transporter function in single oligodendrocytes by elec- tivate microglial cells (Fig. 1a). LPS alone had no significant trophysiology. The application of D-aspartate (1 mM) elicited effect in glutamate uptake (n ϭ 4; Fig. 1b). However, we found an inward current in oligodendrocytes clamped at Ϫ70 mV 6552 MICROGLIA INDUCED OLIGODENDROGLIAL EXCITOTOXICITY Downloaded from

FIGURE 2. Activated microglial cells increase extracellular glutamate levels in oligodendrocyte cultures. a, Top, Representative traces showing glutamate levels measured in the medium after the addition of the L-glutamate dehydrogenase (GDH) enzyme to an oligodendrocyte-microglia coculture (OLϩMG) with or without LPS (100 ng/ml) and microglia alone (MG). Bottom, Histogram illustrates extracellular glutamate levels in the culture conditions referred to above and also in oligodendrocyte alone or after addition of LPS (OL and OLϩLPS, respectively), or the cystine- glutamate antiporter inhibitor AAA (2.5 mM). Note that glutamate levels are similar in oligodendrocyte plus microglia (OLϩMG) cultures incubated http://www.jimmunol.org/ with LPS or the glutamate transporter inhibitor TBOA (1 mM) and in microglia alone. b, Extracellular glutamate levels in microglia in the absence (Control) or in the presence of TBOA (1 mM), the cystine-glutamate antiporter inhibitor AAA (2.5 mM) and LPS (100 ng/ml). Data are mean Ϯ p Ͻ 0.001. c, Up-regulation of the cystine-glutamate ,ءء .SEM from three determinations in sister wells from at least three independent experiments antiporter xCT in microglia after LPS (100 ng/ml) treatment for 48 h. Data were normalized to actin and expressed as the percentage of xCT .p Ͻ 0.05 ,ء .(expression vs control (n ϭ 3

(mean amplitude ϭ 26.5 Ϯ 7.3, n ϭ 20; Fig. 1c) that is depen- source of glutamate under these conditions (n ϭ 10; Fig. 2a). In by guest on October 1, 2021 dent on extracellular sodium and inhibited by the glutamate addition, the levels of glutamate in mixed cultures treated with transporter inhibitor DL-threo-␤-benzyloxyaspartate (TBOA) LPS were similar to those after incubation with the glutamate (data not shown) as shown previously (32). Similarly as in up- transporter inhibitor TBOA (1 mM; n ϭ 5; Fig. 2a). Together, take assays, currents elicited by D-aspartate in oligodendrocytes these data suggest that glutamate released by microglial cells is cultured alone were not significantly affected by the presence of normally taken up by oligodendrocytes in control conditions, resting microglia (n ϭ 10, Fig. 1c) but were greatly reduced by whereas in the presence of LPS glutamate accumulates in the LPS-activated microglia (n ϭ 15; Fig. 1c). medium following an inhibition of oligodendrocyte glutamate We tested the effect of different factors released by microglia uptake. upon activation of glutamate transporter activity in oligodendro- We next investigated the mechanisms by which microglia re- cytes. Neither TNF-␣ (100 ng/ml) nor IL-1b (10 ng/ml) have any leases glutamate. We found that glutamate released by microglial effect on glutamate uptake in oligodendrocytes (n ϭ 3; Fig. 1d). cells is not inhibited by TBOA and thus, it is not due to a reversal However, paraquat (15 ␮M), a redox-cycling agent capable of gen- function of transporters (1 mM; n ϭ 3; Fig. 2b). Another possible erating reactive oxygen species (ROS) intracellularly (35), blocked glutamate release mechanism by microglia is the cystine-glutamate glutamate transporters in oligodendrocytes (n ϭ 3; Fig. 1d), and antiporter, which exchanges extracellular cystine for intracellular Trolox (10 ␮M), a ROS scavenger, prevented glutamate trans- glutamate. Indeed, aminoadipic acid (AAA; 2.5 mM), an inhibitor ϭ Ϫ porter inhibition by LPS-activated microglia (n 3; Fig. 1d). of cystine-glutamate antiporter xc (2, 28), induced a significant We next checked whether the inhibition of glutamate transport- reduction in glutamate release by microglia (n ϭ 3; Fig. 2b). In ers by LPS-activated microglia increased glutamate levels in the addition, AAA also inhibited glutamate release in mixed oligoden- culture medium. The concentration of glutamate in control and drocyte-microglia cultures treated with LPS (n ϭ 3; Fig. 2a). Fi- LPS-treated oligodendrocyte cultures was 0.7 ␮M (Fig. 2a, OL; nally, a moderate increase in glutamate levels was detected after n ϭ 9) and 0.65 ␮M, respectively (Fig. 2a,OLϩLPS; n ϭ 4). the exposure of microglial cells to LPS (100 ng/ml, n ϭ 5; Fig. 2b), Ϫ Coculture of microglial cells with oligodendrocytes during a 48-h indicating that glutamate release by the system xc is enhanced in period raised glutamate concentration to 12 ␮M(n ϭ 9; Fig. 2a). activated microglia. In agreement with this finding, we observed Notably, the addition of LPS to activate microglia in these co- that the expression of the cystine-glutamate antiporter xCT is in- cultures further elevated glutamate levels to 34 ␮M(n ϭ 10; creased in microglia after LPS treatment as revealed by Western Fig. 2a). To investigate the cellular source of glutamate, we blot analysis (n ϭ 3; Fig. 2c). analyzed extracellular glutamate levels in pure microglia cultures at the same density. Interestingly, we detected the same Activated microglia induced oligodendroglial excitotoxicity levels of extracellular glutamate in microglia alone as in mixed We next checked whether activated microglia contributes to LPS cultures treated with LPS, suggesting that microglia is a major toxicity in oligodendrocyte-microglia cocultures. The addition of The Journal of Immunology 6553 Downloaded from

FIGURE 3. LPS-activated microglia induces oligodendroglial excitotoxicity. Pure oligodendrocyte cultures or mixed oligodendrocyte-microglia ␮ cultures were treated with LPS (100 ng/ml; 48 h) in the presence or absence of the AMPA/kainate receptor antagonist CNQX (30 M) and the system http://www.jimmunol.org/ Ϫ xc inhibitor AAA (2.5 mM). a, Double fluorescence staining with Abs to oligodendroglial marker O1 (red) and with the vital-dye calcein-AM (green) in untreated oligodendrocyte cultures (OL) and in mixed oligodendrocyte-microglia cultures (OLϩMG) with or without LPS (100 ng/ml; 48 h). Only LPS-activated microglia is toxic to oligodendrocytes. Scale bar, 50 ␮m. b, Histogram shows LPS-activated microglia toxicity to oligodendrocytes, which is blocked in the presence of CNQX (30 ␮M) and Trolox (10 ␮M) (n ϭ 3–4). c, Effect of LPS on mitochondrial membrane potential. Mixed oligodendrocyte-microglia cultures (OLϩMG) or microglia alone cultures (MG) were treated with LPS (100 ng/ml) during a 3-h period and mitochondrial membrane potential was analyzed by fluorometry after loading cells with the fluorescent dye JC-1 (3 ␮M). Data was Ϫ expressed as a JC-1 red/green ratio in percentage vs nontreated control cells. d, Blockade of the system xc by AAA in LPS-activated microglia (MG) prevents oligodendrocyte (OL) excitotoxicity. Data are mean Ϯ SEM of at least three independent experiments performed in triplicate. Symbols .p Ͻ 0.01; #, p Ͻ 0.05; ##, p Ͻ 0.01 ,ءء .(and #, respectively ء) denote statistical significance vs control or LPS-treated cultures by guest on October 1, 2021

resting microglia or LPS did not alter oligodendrocyte viability Oligodendrocytes in white matter tracts express glutamate trans- (n ϭ 3 or 4, respectively; Fig. 3, a and b), in agreement with porters and contribute to glutamate homeostasis (31, 32, 34). In previous data showing no alteration in glutamate homeostasis (see addition, acute exposure to LPS produces oligodendrocyte cell Figs. 1b and 2a). However, LPS-activated microglia induced oli- death in isolated neonatal RONs (12). Oligodendrocytes were eas- godendrocyte cell death, which was prevented by the AMPA and ily identified by the presence of myelinating axons imbedded kainate receptor antagonist 6-cyano-7-nitroquinoxaline-2,3-dione within the cells in addition to features such as the presence of ␮ ϭ (CNQX) (30 M; n 4; Fig. 3, a and b) and by the ROS scav- numerous mitochondria, a light cytoplasm, and a thin endoplasmic ␮ ϭ enger Trolox (10 M; n 3; Fig. 3b). Oligodendroglial excito- reticulum (Fig. 4a). Oligodendrocytes in nerves exposed to LPS (1 toxicity is associated with calcium overload, alteration of the mi- ␮g/ml for 3 h) did not exhibit the kind of frank necrotic changes tochondrial membrane potential and a rise in ROS (36). Because of produced by acute ischemia (37–39) and were typified by the that, we next analyzed whether LPS-activated microglia alters swelling and disruption of mitochondria and, to a lesser degree, mitochondrial membrane potential in oligodendrocytes. After of the endoplasmic reticulum (Fig. 4, b–g). LPS treatment pro- treatment with LPS (100 ng/ml), we detected an early increase duced a significant increase in the mitochondrial area in oligo- in the JC-1 ratio that is indicative of mitochondria membrane dendrocytes, whereas a similar change was not found in the hyperpolarization in mixed oligodendrocyte-microglia cultures mitochondria of neighboring axons (Fig. 4h). To detect early (3h; n ϭ 3; Fig. 3c). The same treatment did not induce any changes related to apoptosis, we analyzed annexin V expression change in mitochondrial membrane potential in pure microglial cell cultures (Fig. 3c). in LPS treated nerves. Annexin V translocation to the cell mem- Because a major source of extracellular glutamate in mixed brane was detected in naturally apoptotic cells in control nerves, Ϫ whereas no annexin V translocation was apparent in damaged cultures is the microglial system xc , we next examined whether the inhibition of the cystine-glutamate antiporter pre- oligodendrocytes in LPS-treated nerves (Fig. 5). Oligodendro- vented LPS toxicity to oligodendrocytes. The inhibition of sys- cyte death induced by LPS was completely abolished by the Ϫ non-N-methyl-D-aspartate (non-NMDA) glutamate receptor an- tem xc with 2.5 mM AAA had no significant effect on cell viability in pure oligodendrocyte cultures (n ϭ 3; Fig. 3d). tagonist 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo(f)quinoxaline ␮ Ϫ However, AAA abolished LPS-activated microglial toxicity to (NBQX) (10 M; Fig. 4i). In addition, the xc antiporter blocker oligodendrocytes (n ϭ 3; Fig. 3d). Taken together, these results AAA (2.5 mM) significantly reduced the number of dead cells suggest that glutamate excitotoxicity contributes to LPS-acti- found in LPS-treated optic nerves, implicating the antiporter in the vated, microglia-induced oligodendrocyte cell death. glutamate release cascade in the whole mount optic nerve. 6554 MICROGLIA INDUCED OLIGODENDROGLIAL EXCITOTOXICITY Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 5. LPS-mediated oligodendrocyte injury in an isolated P10 RON is not apoptotic. a, Natural cell death in the control optic nerve showing a cell with highly condensed chromatin that appears to be in the process of forming apoptotic bodies. Annexin V staining is localized by immunogold electron microscopy to the cell membrane (gold particles are indicated by the arrow) with no staining apparent in the cytoplasm. b and c, Oligodendrocytes in an LPS-treated nerve showing signiﬁcant areas of mitochondrial vacuolization (asterisks) and only cytoplasmic annexin V staining (e.g., arrows). Scale bars ϭ 400 nm (a) and 1 ␮m(b and c).

FIGURE 4. LPS-induced toxicity in optic nerve oligodendrocytes is Discussion glutamate receptor mediated. a, Oligodendrocytes (OL) in control optic In this study we describe how activated microglia can kill oligo- nerve showing numerous mitochondria (arrows) and endoplasmic re- dendrocytes by a dual mechanism leading to excitotoxicity. First, ticulum (arrowheads). Note the healthy appearance of the organelles activated, but not resting, microglia release ROS and inhibit glu- within the cytoplasm and the numerous myelinated and premyelinated tamate uptake in oligodendrocytes, which results in an increase in axons surrounding and embedded within the cells. b–d, Structural al- extracellular glutamate levels. Then we provide evidence that glu- terations of an oligodendrocyte (OC) within an LPS-treated optic nerve (1 ␮g/ml; 3 h). The boxed areas are shown at a higher gain in c and d. tamate release by the cystine-glutamate antiporter expressed in mi- Note the swollen and disrupted mitochondria (arrows) and the swollen croglia contributes substantially to the observed increase in extra- endoplasmic reticulum (arrowheads). Vacuoles (asterisks) are also ap- cellular glutamate. Finally, we show that elevated glutamate levels parent in the processes while axons (AX) embedded within the cells resulting from microglial activation causes oligodendrocyte exci- appear healthy. e–g, Appearance of another oligodendrocyte (OL) in an totoxicity by the activation of AMPA and kainate receptors both in LPS-treated optic nerve. Boxed areas are shown at a higher gain in f and mixed cultures and in isolated optic nerves and that blocking this g for viewing the alterations in the mitochondria and endoplasmic reticulum. h, The mean mitochondrial area measured in oligodendrocytes (OL) and neighboring axons (AX) in control and LPS-treated (ϩLPS) optic nerves. Note the significant increase in the mitochondrial area in increase in the number of dead cells following LPS treatment that is pre- oligodendrocytes but not in axons after LPS treatment. i, Cell death vented by NBQX (10 ␮m). j, Cell death in optic nerve in LPS and LPS with in optic nerve treated with LPS and LPS with NBQX (NBQX ϩ KOS) AAA (AAA ϩ LPS). Note the significantly lower degree of cell death in .p Ͻ 0.001 ,ءءء ;as compared with nontreated nerves (Control). Note the significant the presence of AAA (2.5 mM). Scale bars ϭ 1 ␮m The Journal of Immunology 6555

Ϫ pathway alleviates oligodendrocyte death in culture and in the iso- croglial system xc is increased in LPS-treated microglia. This lated optic nerve. feature is consistent with the up-regulation of cystine-glutamate The immediate or innate immune response is the first line of antiporter xCT transcripts in macrophages and microglia in re- defense against diverse microbial pathogens and requires the ex- sponse to bacterial LPS or ␤-amyloid that others and we have pression of recently discovered TLRs. TLR4 serves as a specific detected (27, 28). Together, these results suggest that the micro- Ϫ receptor for LPS, which in culture binds exclusively to microglia glial system xc is relevant to the maintenance of glutamate ho- (11). Consistent with this idea, we found that LPS toxicity to oli- meostasis in axonal tracts. Indeed, the tonic release of glutamate godendrocytes requires the presence of microglial cells, as previ- that contributes to maintaining extracellular glutamate in the mi- ously demonstrated (11). NO and the cytokines TNF-␣ and IL-1␤ cromolar range outside synapse is mainly dependent on system Ϫ Ϫ are the main mediators of activated microglia-induced oligoden- xc (47). The system xc has been implicated in neurological con- drocyte death (1, 10–12). A recent study (13), moreover, suggests ditions such as ␤-amyloid toxicity (28, 48), virally induced en- that peroxynitrite, produced by the reaction between NO and su- cephalopathy (49), and periventricular leukomalacia (30). The peroxide anion, also mediates oligodendrocyte demise by LPS- present findings bring about the possibility that blocking micro- Ϫ stimulated microglia. We show here that excitotoxicity is an ad- glial system xc during inflammation could contribute to the pre- ditional mechanism contributing to activated microglia-induced vention of oligodendrocyte damage in white matter disorders. oligodendrocyte death. In summary, this study provides evidence that the concomitant Our findings indicate that activated microglia inhibit Naϩ-de- glutamate release from activated microglia by the cystine-gluta- pendent glutamate transporters in cultured oligodendrocytes, re- mate antiporter and the inhibition of Naϩ-dependent glutamate sulting in extracellular glutamate accumulation and the subsequent uptake by activated microglia can induce a local increase in ex- overactivation of ionotropic AMPA/kainate glutamate receptors. tracellular glutamate that leads to excitotoxic oligodendrocyte Downloaded from These results are consistent with those reporting that LPS-acti- death. In addition, these results suggest that counteracting these vated microglia can also cause neuronal excitotoxicity (2). In ad- deleterious mechanisms of microglia could be exploited to treat dition, previous evidence indicates that TNF-␣ or IL-1␤ renders white matter disorders. oligodendrocytes vulnerable to excitotoxicity in mixed glial cultures by inhibition of the function of glutamate transporters in Disclosures astrocytes (40, 41). In contrast, we did not detect any change in The authors have no financial conflict of interest. http://www.jimmunol.org/ glutamate uptake in oligodendrocytes after treatment with either cytokine, TNF-␣ or IL-1␤. The different sensitivity between as- References trocytes and oligodendrocytes could be explained by the expres- 1. Merrill, J. E., L. J. Ignarro, M. P. Sherman, J. Melinek, and T. E. Lane. 1993. sion of different subtypes of transporters in each cellular type. Microglial cell cytotoxicity of oligodendrocytes is mediated through nitric oxide. J. Immunol. 151: 2132–2141. Thus, whereas astrocytes express mainly the GLT-1 subtype, oli- Ϫ 2. Piani, D., and A. Fontana. 1994. Involvement of the cystine transport system xc godendrocytes also express GLAST and EAAC1 (31, 32, 34), the in the macrophage-induced glutamate-dependent cytotoxicity to neurons. J. Im- latter being up-regulated after treatment with TNF-␣ (42). Nota- munol. 152: 3578–3585. 3. Bezzi, P., M. Domercq, L. Brambilla, R. Galli, D. Schols, E. De Clercq, bly, we observed that ROS inhibited glutamate uptake in oligo- A. Vescovi, G. Bagetta, G. Kollias, J. Meldolesi, and A. Volterra. 2001. CXCR4- by guest on October 1, 2021 dendrocytes and that the antioxidant Trolox prevented activated activated astrocyte glutamate release via TNF␣: amplification by microglia triggers neurotoxicity. Nat. Neurosci. 4: 702–710. microglia-induced glutamate transporter inhibition and toxicity to 4. Block, M. L., and J. S. Hong. 2005. Microglia and inflammation-mediated neu- oligodendrocytes. In addition to ROS, peroxynitrite, generated by rodegeneration: multiple triggers with a common mechanism. Prog. Neurobiol. the reaction between the superoxide anion and NO, has been iden- 76: 77–98. 5. Schwartz, M., O. Butovsky, W. Bruck, and U. K. Hanisch. 2006. Microglial tified as a major oxidant released by LPS-activated microglia (13). phenotype: is the commitment reversible? Trends Neurosci. 29: 68–74. Indeed, peroxynitrite may indirectly induce excitotoxicity because 6. Hartung, H. P., S. Jung, G. Stoll, J. Zielasek, B. Schmidt, J. J. Archelos, and it is a potent and fast inhibitor of the glutamate transporters K. V. Toyka. 1992. Inflammatory mediators in demyelinating disorders of the CNS and PNS. J. Neuroimmunol. 40: 197–210. GLAST, GLT-1, and EAAC1 (43). 7. Carson, M. J. 2002. Microglia as liaisons between the immune and central ner- A link between inflammation and excitotoxic neurodegeneration vous systems: functional implications for multiple sclerosis. Glia 40: 218–231. 8. Robbins, D. S., Y. Shirazi, B. E. Drysdale, A. Lieberman, H. S. Shin, and due to glutamate transporter inhibition has been proposed as a M. L. Shin. 1987. Production of cytotoxic factor for oligodendrocytes by stim- component in the etiology of MS. In MS white matter active le- ulated astrocytes. J. Immunol. 139: 2593–2597. sions, which are characterized by the presence of activated micro- 9. Pang, Y., Z. Cai, and P. G. Rhodes. 2000. Effects of lipopolysaccharide on oligodendrocyte progenitor cells are mediated by astrocytes and microglia. J. Neu- glia and/or macrophage/monocyte infiltrates and the loss of oligo- rosci. Res. 62: 510–520. dendrocytes, the expression of glutamate transporters is severely 10. Molina-Holgado, E., J. M. Vela, A. Arevalo-Martin, and C. Guaza. 2001. LPS/ down-regulated (44) and glutamate levels are increased (24). No- IFN-␥ cytotoxicity in oligodendroglial cells: role of nitric oxide and protection by the anti-inflammatory cytokine IL-10. Eur. J. Neurosci. 13: 493–502. tably, glutamate levels are also altered in serum, suggesting that 11. Lehnardt, S., C. Lachance, S. Patrizi, S. Lefebvre, P. L. Follett, F. E. Jensen, monocyte/macrophage lineage contributes to glutamate homeosta- P. A. Rosenberg, J. J. Volpe, and T. Vartanian. 2002. The toll-like receptor TLR4 is necessary for lipopolysaccharide-induced oligodendrocyte injury in the CNS. sis alteration in this disease (45). In this study we demonstrate for J. Neurosci. 22: 2478–2486. the first time a causative link between inflammation and the loss of 12. Sherwin, C., and R. Fern. 2005. Acute lipopolysaccharide-mediated injury in glutamate transporter function in oligodendrocytes and white neonatal white matter glia: role of TNF-␣, IL-1␤, and calcium. J. Immunol. 175: 155–161. matter. In addition to microglia, a recent study has demon- 13. Li, J., O. Baud, T. Vartanian, J. J. Volpe, and P. A. Rosenberg. 2005. Peroxyni- strated that autoantigen-activated myelin basic protein-specific trite generated by inducible nitric oxide synthase and NADPH oxidase mediates T cells are also able to inhibit glutamate transporter expression microglial toxicity to oligodendrocytes. Proc. Natl. Acad. Sci. USA 102: 9936–9941. in glial cells (46). Altogether, these data could explain gluta- 14. Heppner, F. L., M. Greter, D. Marino, J. Falsig, G. Raivich, N. Hovelmeyer, mate homeostasis alterations in EAE as well as in MS plasma A. Waisman, T. Rulicke, M. Prinz, J. Priller, et al. 2005. Experimental autoim- and cerebrospinal fluid. mune encephalomyelitis repressed by microglial paralysis. Nat. Med. 11: ϩ 146–152. Under conditions in which Na -dependent glutamate transport- 15. Yoshioka, A., B. Bacskai, and D. Pleasure. 1996. Pathophysiology of oligoden- ers are not functional, tonic glutamate release by the microglial droglial excitotoxicity. J. Neurosci. Res. 46: 427–437. Ϫ 16. Matute, C., M. V. Sańchez-Go´mez, L. Martıńez-Millań, and R. Miledi. 1997. system xc accumulates and contributes to oligodendrocyte death. Glutamate receptor-mediated toxicity in optic nerve oligodendrocytes. Proc. In addition, we observed that the release of glutamate by the mi- Natl. Acad. Sci. USA 94: 8830–8835. 6556 MICROGLIA INDUCED OLIGODENDROGLIAL EXCITOTOXICITY

17. McDonald, J. W., S. P. Althomsons, K. L. Hyrc, D. W. Choi, and M. P. Goldberg. 34. Li, S. G., A. R. Mealing, P. Morley, and P. K. Stys. 1999. Novel injury mech- 1998. Oligodendrocytes from forebrain are highly vulnerable to AMPA/kainate anism in anoxia and trauma of spinal cord white matter: glutamate release via receptor-mediated excitotoxicity. Nat. Med. 4: 291–297. reverse Naϩ-dependent glutamate transport. J. Neurosci. 19: RC16. 18. Fern, R., and T. Moller. 2000. Rapid ischemic cell death in immature oligoden- 35. Patel, M., B. J. Day, J. D. Crapo, I. Fridovich, and J. O. McNamara. 1996. drocytes: a fatal glutamate release feedback loop. J. Neurosci. 20: 34–42. Requirement for superoxide in excitotoxic cell death. Neuron 16: 345–355. 19. Hardin-Pouzet, H., M. Krakowski, L. Bourbonniere, M. Didier-Bazes, E. Tran, 36. Sańchez-Go´mez, M. V., E. Alberdi, G. Ibarretxe, I. Torre, and C. Matute. 2003. and T. Owens. 1997. Glutamate metabolism is down-regulated in astrocytes dur- Caspase-dependent and caspase-independent oligodendrocyte death mediated by ing experimental allergic encephalomyelitis. Glia 20: 79–85. AMPA and kainate receptors. J. Neurosci. 23: 9519–9528. 20. Ohgoh, M., T. Hanada, T. Smith, T. Hashimoto, M. Ueno, Y. Yamanishi, 37. Thomas, R., M. G. Salter, S. Wilke, A. Husen, N. Allcock, M. Nivison, M. Watanabe, and Y. Nishizawa. 2002. Altered expression of glutamate trans- A. N. Nnoli, and R. Fern. 2004. Acute ischemic injury of astrocytes is mediated porters in experimental autoimmune encephalomyelitis. J. Neuroimmunol. 125: by Na-K-Cl cotransport and not Ca2ϩ influx at a key point in white matter de- 170–178. velopment. J. Neuropathol. Exp. Neurol. 63: 856–871. 21. Pitt, D., P. Werner, and C. S. Raine. 2000. Glutamate excitotoxicity in a model 38. Wilke, S., M. Salter, M. R. Thomas, N. Allcock, and R. Fern. 2004. Mechanism of multiple sclerosis. Nat. Med. 6: 67–70. of acute ischemic injury of oligodendroglia in early myelinating white matter: the 22. Smith, T., A. Groom, B. Zhu, and L. Turski. 2000. Autoimmune encephalomy- importance of astrocyte injury and glutamate release. J. Neuropathol. Exp. Neu- elitis ameliorated by AMPA antagonists. Nat. Med. 6: 62–66. rol. 63: 872–881. 23. Matute, C., E. Alberdi, M. Domercq, F. Pe´rez-Cerda´, A. Pe´rez-Samartıń, and 39. Salter, M., and R. Fern. 2005. NMDA receptors are expressed in developing M. V. Sańchez-Go´mez. 2001. The link between excitotoxic oligodendroglial oligodendrocyte processes and mediate injury. Nature 438: 1167–1171. death and demyelinating diseases. Trends Neurosci. 24: 224–230. 40. Pitt, D., I. E. Nagelmeier, H. C. Wilson, and C. S. Raine. 2003. Glutamate uptake 24. Srinivasan, R., N. Sailasuta, R. Hurd, S. Nelson, and D. Pelletier. 2005. Evidence by oligodendrocytes: implications for excitotoxicity in multiple sclerosis. Neu- of elevated glutamate in multiple sclerosis using magnetic resonance spectros- rology 61: 1113–1120. copy at 3 T. Brain 128: 1016–1025. 41. Takahashi, J. L., F. Giuliani, C. Power, Y. Imai, and V. W. Yong. 2003. Inter- 25. McBean, G. J. 2002. Cerebral cystine uptake: a tale of two transporters. Trends leukin-1␤ promotes oligodendrocyte death through glutamate excitotoxicity. Ann. Pharmacol. Sci. 23: 299–302. Neurol. 53: 588–595. 26. Sato, H., M. Tamba, T. Ishii, and S. Bannai. 1999. Cloning and expression of a 42. Romera, C., O. Hurtado, S. H. Botella, I. Lizasoain, A. Cardenas, P. Fernandez- plasma membrane cystine/glutamate exchange transporter composed of two dis- Tome, J. C. Leza, P. Lorenzo, and M. A. Moro. 2004. In vitro ischemic tolerance tinct proteins. J. Biol. Chem. 274: 11455–11458. involves upregulation of glutamate transport partly mediated by the TACE/ Downloaded from 27. Sato, H., K. Kuriyama-Matsumura, T. Hashimoto, H. Sasaki, H. Wang, T. Ishii, ADAM17-tumor necrosis factor-␣ pathway. J. Neurosci. 24: 1350–1357. G. E. Mann, and S. Bannai. 2001. Effect of oxygen on induction of the cystine 43. Trotti, D., D. Rossi, O. Gjesdal, L. M. Levy, G. Racagni, N. C. Danbolt, and transporter by bacterial lipopolysaccharide in mouse peritoneal macrophages. A. Volterra. 1996. Peroxynitrite inhibits glutamate transporter subtypes. J. Biol. J. Biol. Chem. 276: 10407–10412. Chem. 271: 5976–5979. 28. Qin, S., C. Colin, I. Hinners, A. Gervais, C. Cheret, and M. Mallat. 2006. System 44. Werner, P., D. Pitt, and C. S. Raine. 2001. Multiple sclerosis: altered glutamate Ϫ xc and apolipoprotein E expressed by microglia have opposite effects on the homeostasis in lesions correlates with oligodendrocyte and axonal damage. Ann. neurotoxicity of amyloid-␤ peptide 1–40. J. Neurosci. 26: 3345–3356. Neurol. 50: 169–180.

29. Barres, B. A., I. K. Hart, H. S. Coles, J. F. Burne, J. T. Voyvodic, 45. Westall, F. C., A. Hawkins, G. W. Ellison, and L. W. Myers. 1980. Abnormal http://www.jimmunol.org/ W. D. Richardson, and M. C. Raff. 1992. Cell death and control of cell survival glutamic acid metabolism in multiple sclerosis. J. Neurol. Sci. 47: 353–364. in the oligodendrocyte lineage. Cell 70: 31–46. 46. Korn, T., T. Magnus, and S. Jung. 2005. Autoantigen specific T cells inhibit 30. Oka, A., M. J. Belliveau, P. A. Rosenberg, and J. J. Volpe. 1993. Vulnerability glutamate uptake in astrocytes by decreasing expression of astrocytic glutamate of oligodendroglia to glutamate: pharmacology, mechanisms, and prevention. transporter GLAST: a mechanism mediated by tumor necrosis factor-␣. FASEB J. Neurosci. 13: 1441–1453. J. 19: 1878–1880. 31. Domercq, M., M. V. Sańchez-Go´mez, P. Areso, and C. Matute. 1999. Expression 47. Baker, D. A., Z. X. Xi, H. Shen, C. J. Swanson, and P. W. Kalivas. 2002. The of glutamate transporters in rat optic nerve oligodendrocytes. Eur. J. Neurosci. origin and neuronal function of in vivo nonsynaptic glutamate. J. Neurosci. 22: 11: 2226–2236. 9134–9141. 32. Domercq, M., E. Etxebarria, A. Pe´rez-Samartıń, and C. Matute. 2005. Excitotoxic 48. Barger, S. W., and A. S. Basile. 2001. Activation of microglia by secreted amy- oligodendrocyte death and axonal damage induced by glutamate transporter in- loid precursor protein evokes release of glutamate by cystine exchange and at- hibition. Glia 52: 36–46. tenuates synaptic function. J. Neurochem. 76: 846–854.

33. Szumlinski, K. K., M. H. Dehoff, S. H. Kang, K. A. Frys, K. D. Lominac, 49. Espey, M. G., Y. Kustova, Y. Sei, and A. S. Basile. 1998. Extracellular glutamate by guest on October 1, 2021 M. Klugmann, J. Rohrer, W. Grifﬁn, 3rd, S. Toda, N. P. Champtiaux, et al. 2004. levels are chronically elevated in the brains of LP-BM5-infected mice: a mech- Homer proteins regulate sensitivity to cocaine. Neuron 43: 401–413. anism of retrovirus-induced encephalopathy. J. Neurochem. 71: 2079–2087.