CXCL10/CXCR3 signaling in glia cells differentially affects NMDA-induced cell death in CA and DG...

CXCL10/CXCR3 Signaling in Glia Cells Differentially AffectsNMDA-Induced Cell Death in CA and DG Neurons of

the Mouse Hippocampus

Hilmar R.J. van Weering,1 Hendrikus W.G.M. Boddeke,1 Jonathan Vinet,1 Nieske Brouwer,1

Alexander H. de Haas,1 Nico van Rooijen,2 Allan R. Thomsen,3 and Knut P.H. Biber1,4*

ABSTRACT: The chemokine CXCL10 and its receptor CXCR3 are impli-cated in various CNS pathologies since interference with CXCL10/CXCR3signaling alters the onset and progression in various CNS disease models.However, the mechanism and cell-types involved in CXCL10/CXCR3 signal-ing under pathological conditions are far from understood. Here, we investi-gated the potential role for CXCL10/CXCR3 signaling in neuronal cell deathand glia activation in response to N-methyl-D-aspartic acid (NMDA)-inducedexcitotoxicity in mouse organotypic hippocampal slice cultures (OHSCs).Our findings demonstrate that astrocytes express CXCL10 in response toexcitotoxicity. Experiments in OHSCs derived from CXCL10-deficient(CXCL102/2) and CXCR3-deficient (CXCR32/2) revealed that in the absenceof CXCL10 or CXCR3, neuronal cell death in the CA1 and CA3 regions wasdiminished after NMDA-treatment when compared to wild type OHSCs. Incontrast, neuronal cell death in the DG region was enhanced in bothCXCL102/2 and CXCR32/2 OHSCs in response to a high (50 lM) NMDA-concentration. Moreover, we show that in the absence of microglia the dif-ferential changes in neuronal vulnerability between CXCR32/2 and wildtype OHSCs are fully abrogated and therefore a prominent role for microgliain this process is suggested. Taken together, our results identify a region-spe-cific role for CXCL10/CXCR3 signaling in neuron-glia and glia-glia interac-tions under pathological conditions. VVC 2010 Wiley-Liss, Inc.

KEY WORDS: chemokines; excitotoxicity; astrocyte; microglia;liposomic clodronate

INTRODUCTION

Excitotoxicity is a pathological process in the CNS by which persis-tent activation of glutamate receptors leads to the progressive loss of

neurons (Choi, 1988; Choi and Rothman, 1990;Arundine and Tymianski, 2003). For a long time,excitotoxicity-induced neuronal cell death has beenviewed as a result of solely endogenous processes butrecent data suggest a major role for glial cells in exci-totoxic insults. Microglia, the immunocompetent cellsof the CNS, have been shown to rapidly respond toexcitotoxicity by proliferation and migration towardssites of neuronal injury (Acarin et al., 1996; Heppneret al., 1998). Interestingly, inhibition of microglia ac-tivity has been shown to considerably affect the extentof excitotoxicity-induced neurodegeneration (Tikkaand Koistinaho, 2001; Dehghani et al., 2003; Kohlet al., 2003; Lalancette-Hebert et al., 2007), suggest-ing a prominent role for these cells in this paradigm.However, the source and type of signals for excitotox-icity-induced microglial activation are largelyunknown.

In the last years, chemokines have emerged aspotential modulators of glia activity under pathologi-cal conditions (Cartier et al., 2005; De Haas et al.,2007). Of interest here is the chemokine CXCL10that mediates its actions through chemokine receptorCXCR3 (Loetscher et al., 1998). CXCL10 is a well-known chemo-attractant for peripheral immune cellsincluding specific subsets of NK cells and macro-phages, dendritic cells and activated CD41 T-cellsand is therefore implicated in the modulation of bothinnate- and adaptive immune responses under patho-logical conditions (Neville et al., 1997; Liu et al.,2005). Interestingly, increased expression of CXCL10in the CNS has been demonstrated in ischemia(Wang et al., 1998), Alzheimer’s disease (Xia et al.,2000; Galimberti et al., 2006), Multiple sclerosis(MS) (Balashov et al., 1999; Sorensen et al., 1999;Simpson et al., 2000), and human immunodeficiencyvirus (HIV)-encephalitis (Sanders et al., 1998; Cinqueet al., 2005). Depending on the insult, CXCL10expression was demonstrated in neurons (Wang et al.,1998; Rappert et al., 2004) and in astrocytes sur-rounding primary lesions (Wang et al., 1998; Balashovet al., 1999; Simpson et al., 2000; Xia et al., 2000;Omari et al., 2005; Tanuma et al., 2006), whileCXCR3 expression was found in microglia (Biberet al., 2001, 2002; Dijkstra et al., 2004; Rappertet al., 2004; Li et al., 2006; Tanuma et al., 2006

1Department of Neuroscience, Section Medical Physiology, UniversityMedical Center Groningen (UMCG), Rijksuniversiteit Groningen (RUG),Groningen, The Netherlands; 2Department of Molecular Cell biology,Free University Medical Center (VUMC), Amsterdam, The Netherlands;3 Institute of International Health, Immunology and Microbiology, Uni-versity of Copenhagen, Copenhagen, Denmark; 4Molecular Psychiatry,Psychiatry and Psychotherapy, University of Freiburg, Freiburg,Germany

Grant sponsors: BCN (HRJvW); NWO-Vidi grant (KPHB).*Correspondence to: Prof. Knut Biber, Department of Psychiatry, SectionMolecular Psychiatry, University of Freiburg, Hauptstr. 5, 79104 Freiburg,Germany. E-mail: [email protected] for publication 14 October 2009DOI 10.1002/hipo.20742Published online 15 January 2010 in Wiley Online Library(wileyonlinelibrary.com).

Abbreviations used: CA; cornu ammonis, CNS; central nervous system,CXCL102/2; CXCL10-deficient, CXCR32/2; CXCR3-deficient, DG; dentategyrus, ISH; in situ hybridization, Lip-CL; liposome-encapsulated clodro-nate, NMDA; N-methyl-D-aspartic, OHSC; organotypic hippocampal sliceculture, PI; propidium iodide, QPCR; quantitative polymerase chainreaction.

HIPPOCAMPUS 21:220–232 (2011)

VVC 2010 WILEY-LISS, INC.

De Haas et al., 2008) and in reactive astrocytes (Simpsonet al., 2000; Goldberg et al., 2001; Omari et al., 2005;Tanuma et al., 2006). In vitro, CXCL10 has been shown toinduce directed migration of microglia (Biber et al., 2002; Rap-pert et al., 2002) and proliferation in astrocytes (Flynn et al.,2003), suggesting that CXCL10 might be involved in the acti-vation and/or recruitment of local glia cells under pathologicalconditions.

Interestingly, interfering with CXCL10/CXCR3 signaling invarious CNS disease models alters the progression and severityof disease symptoms (reviewed in Liu et al., 2005). Blocking ofCXCL10 with antibodies reduced neuronal apoptosis in spinalcord injury (Glaser et al., 2006), suggesting a proinflammatoryrole for CXCL10. Administration of anti-CXCL10 alsodecreased the clinical and histological disease incidence in ex-perimental autoimmune encephalomyelitis (EAE), a model forMS (Fife et al., 2001). In contrast, induction of EAE inCXCL102/2 or CXCR32/2 mice provoked more severe clini-cal and histological symptoms and earlier onset compared towild type controls (Klein et al., 2004; Liu et al., 2006; Mulleret al., 2007). These conflicting observations suggest thatCXCL10 and its receptor CXCR3 have multiple functions inthe development of CNS pathologies.

In summary, expression of CXCL10 and CXCR3 in bothglia and neurons has been associated with the clinical signs ofseveral excitotoxicity-related CNS disorders. Furthermore, invitro findings suggest a role for this chemokine-receptor couplein the communication between damaged neurons and glia cells.However, the role of CXCL10 in neuron-glia signaling underpathological conditions is far from understood. Therefore, wehave studied the implication of CXCL10/CXCR3 signaling inexcitotoxicity-induced neurodegeneration using N-methyl-D-as-partic (NMDA)-treated mouse organotypic hippocampal slicecultures (OHSCs) as a model.

MATERIALS AND METHODS

Animals

All experiments have been approved by the Dutch animalexperimental committee. The C57BL/6J (Harlan) andCXCR32/2 mice were housed and handled in accordance withthe guidelines of the central animal laboratory facility of Gro-ningen. The CXCR32/2 mice were kindly provided by B. Lu,MD, and Prof. C. Gerard, MD/PhD (Children’s Hospital,Harvard Medical School, Boston (MA)). The CXCL102/2

mice were housed and handled at the animal facilities of thePanum institute, University of Copenhagen, Denmark.

Chemicals

Culture media and supplements were all obtained fromGIBCO1 (Invitrogen), unless mentioned otherwise. Recombi-nant murine CXCL10 was obtained from PeproTech1(London, United Kingdom).

Preparation of OHSCs

OHSCs were prepared from 2 to 3 day old C57BL/6J,CXCL102/2 and CXCR32/2 mouse pups as described withminor modifications (Stoppini et al., 1991). In brief, hippo-campi were acutely isolated in ice cold serum-free Hank’s Bal-anced Salt Solution (HBSS), supplemented with 0.5% glucose(Sigma) and 15 mM HEPES. Isolated hippocampi were cutinto 375 lM thick slices using a tissue chopper (McIlwain)and transferred to 0.4 lM culture plate inserts (Millipore),placed in six-well plates containing 1.2 ml of culture mediumper well. Culture medium (pH 7.2) consisted of 0.53 mini-mum essential medium containing 25% heat-inactivated horseserum, 25% BME basal medium without L-glutamine, 2 mMglutamax and 0.65% glucose. OHSCs were kept at 358C in ahumidified atmosphere (5% CO2) and the culture medium wasrefreshed the first day after preparation and every consecutivetwo days. After six days in culture, OHSCs were placed in cul-ture medium containing 10–50 lM NMDA (Sigma) for 4 hand subsequently the medium was replaced with standard cul-ture medium. OHSCs were kept in culture for another 24 hunless stated otherwise.

Depletion of Microglia From OHSCs UsingClodronate-Liposomes

Multi-lamellar CL2MDP (clodronate)-liposomes (Lip-CL)were prepared as described elsewhere (Van Rooijen and Sand-ers, 1994). Clodronate was a gift of Roche Diagnostics (Mann-heim, Germany), phosphatidylcholine (Lipoid EPC) wasobtained from Lipoid (Ludwigshafen, Germany) and choles-terol was purchased from Sigma (USA). Depletion of microgliawas achieved by incubation of freshly prepared OHSCs withapproximately 0.5 mg/ml Lip-CL solution (1:10 liposome dilu-tion in standard slice culture medium) for 24 h at 358C. Sub-sequently, OHSCs were rinsed in PBS (358C) and placed onfresh culture medium. After depletion the medium wasrefreshed every two days. Liposomes containing PBS (Lip-PBS)served as negative controls.

Immunohistochemistry

OHSCs were shortly rinsed in PBS (358C) and fixated with4% paraformaldehyde (PFA) overnight at 48C. After fixation,OHSCs were pre-incubated with 5% normal goat serum(NGS; Vector) in PBS containing 0.3% Triton X-100 (PBS1)for 1 h prior to incubation with primary antibodies in 1%NGS/PBS1 overnight at 48C. The following primary antibod-ies were used: rabbit-anti-Iba1 (1:1,000; Wako), mouse-anti-GFAP (1:600; Chemicon) and mouse-anti-NeuN (1:1,000;Chemicon) for detection of microglia, astrocytes and neuronalnuclei, respectively. Secondary antibodies used were donkey-antimouse-Alexa488 and donkey-antirabbit-Alexa633 (Molecu-lar Probes) and goat-antimouse-Cy3 (Jackson IR Laboratories)for NeuN, Iba1 and GFAP, respectively.

CXCL10 IN MOUSE HIPPOCAMPAL EXCITOTOXICITY 221

Hippocampus

Quantification of Neuronal Cell Death

OHSCs were incubated with 5 lg/ml propidium iodide (PI)during and up to 24 h after the NMDA-challenge (Vornovet al., 1991; Pozzo Miller et al., 1994). Confocal images (LeicaSP2 AOBS) of the neuronal layers were taken mid-section at403 magnification and both NeuN1 and NeuN1/PI1 cellswere quantified manually using ImageJ software. The percent-age of neuronal cell death was determined by the number ofNeuN1/PI1 cells divided by the total number of NeuN1 cellsper neuronal layer.

Quantitative Polymerase Chain Reaction

OHSCs (6 per condition) were lysed in guanidinium iso-thiocyanate/ monopthiolglycerol buffer (Sigma) and total RNAwas extracted and precipitated with a one-chloroform-phenolstep (Chomczynski and Sacchi, 1987). A total of 1 lg RNAwas transcribed into cDNA as described previously (Biberet al., 1997). Amplification reactions were performed usingTaqMan1 primer-probes (Applied biosystems: b-actin,Mm00607939_s1 and CXCL10, Mm00445235_m1). Reac-tions were run in triplets on the ABI Prism1 7,900 HT realtime PCR instrument. Threshold cycles were determined man-ually by setting thresholds for fluorescence intensity. Relativegene expression levels were analyzed by the 2-DDCT method,using b-actin as a reference gene (Livak and Schmittgen,2001). The amplification efficiency was verified by linearregression with the fluorescence per cycle number (Ramakerset al., 2003).

In situ Hybridization (ISH) Followed byImmunohistochemistry

The CXCL10 PCR product (147 bp) was cloned into a dualPCRII vector (Invitrogen) and probes were synthesized by run-off transcription with SP6 or T7 RNA polymerase and digoxi-genin (DIG)-conjugated UTP (Roche). OHSCs were fixatedwith 4% PFA in 0.1 borate buffer (pH 9.5) overnight at 48C.After rinsing in ice-cold KPBS, OHSCs were dehydrated with20% sucrose in KPBS for 10 h and cut in 10 lM sections(Leica CM3050 S, Leica Microsystems). Subsequently, thesections were mounted on APES (Sigma)-coated glass slides,air-dried and postfixated with 4% PFA for 30 min at 378C.Nonradioactive ISH experiments were performed as describedpreviously (Copray and Brouwer, 1994). For ISH combinedwith immunohistochemistry, sections were subsequentlyincubated overnight in PBS/1%NGS at 48C with primary anti-bodies against GFAP, Iba1 and NeuN as mentioned earlier. Sec-tions were incubated with the appropriate biotinylated second-ary antibodies (1:400, Vector) for one hour (RT). Antibody-antigen reactions were detected using the biotin-streptavidinmethod (Vectastain1 ABC-kit, Vector Labs) and the complexwas visualized with diaminobenzidin/H2O2. Bright-field imageswere taken with an Olympus BX50 system, equiped with anOlympus DP70 camera.

Western Blot Analysis

OHSCs (12 per condition) were lysed in 200 ll ice-coldRIPA buffer (PBS, containing 0.1% Nonidet P40, 0.5% so-dium deoxylate, 0.1% SDS and proteinase inhibitors (Com-plete Mini; Roche)). Subsequently, one volume of 23 samplebuffer (175 mM Tris-HCl (pH 6.8) with 4% SDS, 4M urea,4.5% glycerol, 0.05 M DTT and 0.01% bromophenol) wasadded and samples were heated for 5 min at 958C. 30 lg oftotal protein was loaded on a 10% SDS polyacrylamide geland transferred to a Hybond-ECL nitrocellulose membrane(Amersham Biosciences). Next, the membranes were blockedwith Odyssey1 Blocking Buffer (OBB) in PBS (1:1) for onehour at RT and subsequently incubated overnight at 48C withrabbit-antimouse CXCL10 (1:1,000, Peprotech) and mouse-anti-ß-actin (1:8,000, Abcam) in 1:1 OBB/PBS, containing0.1% Tween-20. Membranes were rinsed with PBS/0.1%Tween-20 (PBS-T) and incubated with donkey-antirabbitIRDeyTM 800CW and donkey-antimouse IRDey

TM

680 (both1:8,000; LI-COR1) in PBS-T for one hour at RT. Labeledproteins were visualized by paired exposure at 700 nm and 800nm using the Odyssey1 Infrared Imaging System (LI-COR1

Biosciences GmbH).

Rapid Isolation of Microglia From OHSCs andFlow Cytometry

OHSCs were triturated in HBSS, containing 0.1% glucoseand 15 mM HEPES. Single-cell suspensions were filtered (70lm cell strainer, BD Biosciences) and rinsed in PBS. Then,cells were pelleted at 1000g for 10 min at 48C, resuspended inice-cold 75% Percoll (GE Healthcare) in PBS, over-layeredwith 5 ml ice-cold 25% Percoll and subsequently with 3 mlPBS. Density gradients were centrifuged at 800g for 25 min at48C. After centrifugation, microglia at the 25–75% interfacewere collected, diluted 3-fold with PBS and centrifuged at1,000g for 10 min at 48C. Cell pellets were resuspended inPBS, containing 10% normal rat serum (Invitrogen) and anti-CD16/32 (eBioscience, 1:100). Next, the microglia werestained for phycoerythyrin-coupled monoclonal antimouseCXCR3 (R&D) or concentration-matched isotype IgG2A con-trol (eBioscience) for 20 min at 48C in PBS containing 1%FCS. Cell-size, granularity and fluorescence intensity were ana-lyzed by flow cytometry (488-nm laser FACSCalibur, BectonDickinson), supported by CellQuest software (Becton Dickin-son) and the acquired data was analyzed using WinMDI ver-sion 2.9 (Joseph Trotter).

Statistical Analysis

Data are represented as the mean 6 standard error of themean (SEM). Statistical comparison between groups was per-formed using one-way ANOVA with Gabriel’s post hoc testand p-values smaller than 0.05 were considered significant. Allstatistical tests were performed in SPSS version 14.0.2 (SPSSInc., Chicago).

222 VAN WEERING ET AL.

Hippocampus

RESULTS

NMDA-Induced Excitotoxicity in Mouse OHSCs

To determine the role of CXCR3 and its ligand CXCL10 inexcitotoxicity we devised a model for NMDA-induced excito-toxicity using mouse OHSCs. After six days in culture, OHSCsshowed preserved hippocampal structure with intact neuronallayers CA1, CA3 and DG (Fig. 1A). While control OHSCsshowed almost no cell death in the three neuronal regions

(CA1; 0.7 6 0.2%, CA3; 0.8 6 0.3%, DG; 1.6 6 0.3%, seealso Fig. 1B), treatment with NMDA for 4 h and subsequentculturing for 24 h in standard culture medium clearly inducedcell death, as determined by PI-uptake. PI1-cells colocalizedwith the neuronal marker NeuN, showing that NMDA specifi-cally induced cell death in neurons (Fig. 1C, arrow). The per-centage of neuronal cell death in response to 0, 10, 15, 25,and 50 lM NMDA was quantified as the number of NeuN1/PI1 cells divided by the total number of NeuN1 cells perregion (Fig. 1D). Note that neurons from the different regions

FIGURE 1. NMDA-induced excitotoxicity in mouse OHSCs.After 6 DIV, control OHSCs showed preserved hippocampal struc-ture with intact neuronal layers CA1, CA3, and DG as determinedby neuronal nuclear marker NeuN (A, in green). Cell death (PI, inred) was hardly detected in untreated (CNTR) slice cultures (>1%,B). Treatment with NMDA for 4 h and subsequent culturing for24 h clearly induced cell death (C), which was restricted to theneuronal layers CA1, CA3, and DG and colocalized with NeuN(C, white arrow). The percentages of neuronal cell death per sub-

region (CA1/CA3/DG) were quantified as the number of NeuN1/PI1 cells divided by the total number of NeuN1 cells (D). Notethat the hippocampal regions display a selective vulnerabilitytowards NMDA-treatment with CA1>CA3>DG. Data are a sum-mary of three separate experiments with at least 5 OHSCs per con-dition. Bars indicate mean 6 SEM. Scale bars indicate 150 lm (A)and 25 lm (B,C). [Color figure can be viewed in the online issue,which is available at wileyonlinelibrary.com.]


Hippocampus

showed selective vulnerability towards NMDA-induced excito-toxicity (CA1>CA3>DG).

NMDA-Excitotoxicity Induces CXCL10Expression in Reactive Astrocytes

Next, we determined the expression of CXCL10 in NMDA-exposed OHSCs. Quantitative polymerase chain reaction(QPCR) analysis of total RNA lysates showed clear upregulationof CXCL10 mRNA in response to NMDA-treatment (Fig. 2A)

with a significant 12-fold increase 8 h after induction (P 50.002). CXCL10 mRNA levels were back at control levels 24 hafter the NMDA-treatment (P 5 1.000). Western Blot analysisof whole OHSC protein lysates (Fig. 2B) showed clear upregula-tion of CXCL10 protein as soon as 4 h up to 24 h after excito-toxicity-induction with a clear peak at 10 h. CXCL10 proteinwas almost nondetectable 48 h after the NMDA-treatment.

To localize the cellular source for CXCL10 we combinednonradioactive ISH for CXCL10 mRNA with immunohisto-chemistry for NeuN (neurons), Iba1 (microglia) and GFAP

FIGURE 2. Astrocytes express CXCL10 in response to excito-toxicity. (A) QPCR analysis revealed upregulation of CXCL10mRNA in NMDA-treated OHSCs as soon as 4 h after inductionwith a twelve fold increase at 8 h (P 5 0.002). Twenty-four hoursafter the challenge expression levels were similar to control levels(P 5 1.000). Data are a summary of four separate experiments.Bars indicate mean 6 SEM, **P < 0.01, Student’s t-test. (B) West-ern blot analysis showed clear upregulation of CXCL10 protein(in green) in OHSCs from 4 h up to 24 h after NMDA-treatment,with a clear reduction after 48 h. b-actin (in red) served as a posi-

tive control. Western blot results are a representation of three indi-vidual sets of experiments. (C–F) ISH experiments revealed multi-ple CXCL101 cells in NMDA-treated OHSCs (C, white arrows).OHSCs hybridized with a sense probe were devoid of signals (D).ISH for CXCL10 (E, black arrows) in combination with GFAP-immunohostochemistry (F) identified the CXCL101 cells to beastrocytes (F, open arrows). Scale bars indicate 100 lm (C,D) and25 lm (E,F). [Color figure can be viewed in the online issue,which is available at wileyonlinelibrary.com.]


Hippocampus

(astrocytes). As QPCR-analysis revealed a significant increase inCXCL10 mRNA expression 8 h after excitotoxicity-induction(Fig. 2A), we used OHSCs that were fixated at this time-point.ISH for CXCL10 mRNA alone revealed multiple CXCL101-cells in NMDA-treated OHSCs (Fig. 2C, white arrowheads).These CXCL101-cells were predominantly located in the stra-tum radiatum, stratum lacunosum-moleculare and the hilararea (Fig. 2E, closed arrows), although some CXCL101-cellswere also observed inside the suprapyramidal and granular celllayers. Slice cultures hybridized with a sense probe were devoidof signal (Fig. 2D). Interestingly, ISH for CXCL10 mRNA(Fig. 2E, closed arrows) in combination with GFAP-immuno-histochemistry showed 100% overlap (Fig. 2F, open arrows),demonstrating that the CXCL101-cells are astrocytes.CXCL10-positivity was pre-dominantly found in ‘‘reactive’’astrocytes, as determined by morphology and upregulation ofGFAP expression (Fig. 2F). In contrast, colocalization for

CXCL10 mRNA with NeuN- or Iba1-immunohistochemistrywas never observed (data not shown).

Regional Changes in NMDA-Induced Cell Deathin CXCL102/2OHSCs

To determine the role of astrocytic CXCL10 on NMDA-induced neuronal cell death, we prepared OHSCs fromCXCL10-deficient (CXCL102/2) mice and exposed these to10–50 lM NMDA to induce excitotoxicity. After six days inculture, CXCL102/2 OHSCs showed preserved hippocampalstructure, and the basal rate of neuronal cell death in the threeneuronal regions was equal to control wild type conditions(CA1; 1.2 6 0.3%, CA3; 1.6 6 0.3%, DG; 1.9 6 0.5%).Interestingly, in response to NMDA-treatment, regional differ-ences in neuronal cell death were observed when compared towild type conditions (Fig. 3A). In the CA1 and CA3 region of

FIGURE 3. Regional differences in NMDA-induced neuronalcell death in CXCL102/2 and CXCR32/2 OHSCs. Graphs repre-sent the percentages of neuronal cell death per hippocampal region(CA1/CA3/DG) in response to 10–50 lM NMDA in wild type vs.CXCL102/2 (A) and CXCR32/2 (B) OHSCs. (A) CXCL102/2

OHSCs (open bars) showed significantly lower percentages in cell-death in response to 10 lM NMDA in the CA1 region, and inresponse to 15 lM and 25 lM NMDA in the CA3 region com-pared to wild types (closed bars). Treatment with 50 lM NMDAresulted in significantly higher cell death in the DG of CXCL102/2

OHSCs compared to wild types. Administration of 1028 M recombi-

nant CXCL10 (light gray bars) directly after NMDA-treatment‘‘rescued’’ the effects observed in CXCL102/2 OHSCs. ThisCXCL10-mediated increase in cell-death in CXCL102/2 OHSCsequalled wild type conditions. (B) Similar results were obtained inCXCR32/2 OHSCs (open bars) when compared with wild type con-ditions. However, administration of 1028 M recombinant CXCL10(dark gray bars) directly after NMDA-treatment had no effect inCXCR32/2 OHSCs at any NMDA-concentration. Data are a sum-mary of three separate experiments with at least 5 OHSCs per condi-tion. Bars indicate mean 6 SEM. *P < 0.05, **P < 0.01, ***P <0.001, ANOVA.


Hippocampus

CXCL102/2 OHSCs (Fig. 3A, white bars) a significantdecrease in neuronal vulnerability to NMDA was observedcompared to wild types (Fig. 3A, black bars), which was de-pendent on the NMDA-concentration. Neuronal cell death inresponse to 10 lM NMDA was significantly lower in the CA1(38.8 6 8.3%) compared to wild type conditions (67.1 616.3%, P 5 0.035). In the CA3 region a decrease in neuronalcell death was observed in response to 15 lM NMDA (11.2 61.9% vs. 41.7 6 11.3% in wild types, P 5 0.001) and inresponse to 25 lM NMDA (28.6 6 5.6% vs. 79.9 6 8.6% inwild types, P < 0.001).

Surprisingly, the detrimental effect of CXCL10 observed inthe CA1 and CA3 regions was not detected in the DG inresponse 10–25 lM NMDA (Fig. 3A; DG). In contrast, weobserved a significant increase in neuronal cell death in theDG in response to 50 lM NMDA (53.3 6 7.3% vs. 30.7 64.1% in wild types, P 5 0.041), suggesting that the presenceof CXCL10 in wild type OHSCs has a protective effect onDG neurons at a high (50 lM) NMDA concentration.

Application of CXCL10 to CXCL102/2 OHSCsMimics Wild Type Conditions

Application of 1028 M recombinant CXCL10 to CXCL102/2

OHSCs directly after the NMDA-challenge (in the aim to‘‘mimic’’ wild type conditions in response to NMDA) partly‘‘rescued’’ the differences observed in CXCL102/2 OHSCscompared to wild type conditions (Fig. 3A, gray bars). In theCA1 region, application of CXCL10 directly after treatmentwith 10 lM NMDA resulted in a significant increase (P 50.015) in neuronal cell death in the CXCL102/2 OHSCs,which was equivalent to wild type conditions (64.2 6 12.3%vs. 67.1 6 16.3% in wild types, P 5 1.000). In the CA3region, the effect of CXCL10 application was less pronouncedafter treatment with 15 lM NMDA in CXCL102/2 OHSCs(18.5 6 4.5% vs. 41.7 6 11.3% in wild types, P 5 0.132).However, application of CXCL10 after treatment with 25 lMresulted in a strong increase (P < 0.001) in neuronal cell deathin the CA3 region of CXCL102/2 OHSCs, which again wasequivalent to wild type conditions (58.2 6 9.7% vs. 79.9 68.6% in wild types, P 5 0.277).

In the DG region, application of CXCL10 to CXCL102/2

OHSCs directly after treatment with 10–25 lM NMDA hadno significant effect on neuronal cell death in this region (Fig.3A). However, neuronal cell death in the DG induced by 50lM NMDA in CXCL102/2 OHSCs (46.5 6 5.4%) wasslightly decreased by application of CXCL10 (44.7 6 9.8%)and did not differ significantly anymore from wild type condi-tions (30.7 6 4.1%, P 5 0.265).

Regional Changes in NMDA-Induced Cell Deathin CXCR32/2 Slice Cultures

To verify that the effects of CXCL10 on NMDA-inducedexcitotoxicity observed in the CXCL102/2 slice cultures aremediated through CXCR3 we performed experiments in

CXCR32/2 slice cultures (Fig. 3B). Also here, a decrease inneuronal sensitivity towards NMDA was observed in the CA1and CA3 region, which was comparable to the effects seen inthe CXCL102/2 slice cultures. In response to 10 lM NMDA,neuronal cell death in the CA1 of CXCR32/2 slice cultures(open bars, 24.5 6 5.5%) was significantly lower comparedwith wild types (black bars, 67.2 6 16.3%, P < 0.001). Inaddition, neuronal cell death in the CA3 was significantly lowerupon treatment with 15 lM NMDA (3.3 6 1.4% vs. 41.7 611.3% in wild types, P < 0.001) and with 25 lM NMDA(55.7 6 6.6% vs. 79.9 6 8.6% in wild types, P 5 0.038).Similar to CXCL102/2 OHSCs, an increase in cell death wasobserved in the DG of CXCR32/2 OHSCs in response to50 lM NMDA (Fig. 4; 47.9 6 3.7% vs. 30.7 6 4.1% inwild types, P < 0.001).

Application of 1028 M recombinant CXCL10 directly afterthe NMDA-insult (gray bars) for 24 h had no effect on theamount of neuronal cell death in CXCR32/2 OHSCs at anycondition, suggesting that the effects observed in CXCL102/2

OHSCs in response to CXCL10 application are mediatedthrough CXCR3.

Activation of Microglia Coincides With RegionalDifferences in Cell Death in CXCR32/2 OHSCs

Acutely isolated microglia from wild type OHSCs (6 DIV)clearly expressed CXCR3 surface protein as determined by flowcytometry (Fig. 4A). Previous research has shown that microgliain vitro respond to CXCL10 with intracellular calcium transi-ents and migration, processes both mediated through CXCR3(Biber et al., 2001, 2002; Dijkstra et al., 2004; Rappert et al.,2004). Therefore, we hypothesized that abrogation of CXCR3-signaling could result in impaired microglia activation inCXCR32/2 OHSCs, partly explaining the differences in neuro-nal cell death observed in CXCR32/2 and wild type OHSCs.Treatment with 10 lM NMDA resulted in prominent celldeath and local accumulation of morphologically activatedmicroglia in the CA1 region of wild type OHSCs (Fig. 4B). Asdiscussed above, neurons in the CA1 region of CXCR32/2

OHSCs were less affected by treatment with 10 lM NMDAwhen compared with wild type conditions. The lower level ofneuronal cell death in CXCR32/2 OHSCs coincided withlower numbers of activated microglia in this region (Fig. 4C).At concentrations of 15–50 lM NMDA no distinction couldbe made in the number and activation of microglia in the CA1region between wild type and CXCR32/2 OHSCs (data notshown).

Similar effects were observed in the CA3 region in responseto 15 lM NMDA, where a decrease in neuronal cell death inCXCR32/2 OHSCs (Fig. 4E) coincided with a decreased accu-mulation of microglia compared with wild type conditions(Fig. 4D). Also here, no differences in microglia activationwere observed in the CA3 region in response to 25 and 50 lMNMDA when compared to wild type conditions (data notshown). In the DG region, no differences in microglia activa-tion were observed between wild type (Fig. 4F) and CXCR32/2


Hippocampus

(Fig. 4G) OHSCs in response to 50 lM NMDA. These resultsindicate that morphological activation and accumulation ofmicroglia at sites of neuronal cell death were not impaired inCXCR32/2 OHSCs, but rather coincided with the delayedneuronal cell death observed in these OHSCs.

CXCL10 Per Se Does Not Induce MicrogliaActivation or Cell Death in Wild Type OHSCs

Next, we determined whether application of CXCL10 alonewould induce activation of microglia and/or neuronal cell death

FIGURE 4. Differences in neuronal cell death in CXCR32/2

OHSCs coincide with differences in microglia activation. Rapidlyisolated microglia from OHSCs express CXCR3 surface protein asdetermined by flow cytometry (A). Gray dotted line indicates iso-type control, black line indicates CXCR3-PE. Confocal images B-G show the correlation between neuronal cell death (NeuN, green/PI, red) and microglia activation (Iba1, black) in wild type(B,D,F) and CXCR32/2 OHSCs (C/E/G) at 10 lM (CA1), 15 lM(CA3) and 50 lM (DG) NMDA. Note that a decrease in neuronal

cell death in CXCR32/2 OHSCs (C,E) coincides with a decreasein microglia activation compared with wild type conditions (B,D).In the DG region, no differences in microglia activation could bedetected between wild type (F) and CXCR32/2 OHSCs (G) inresponse to 50 lM NMDA, although neuronal cell death washigher in CXCR32/2 slice cultures. Scale bar indicates 75 lm.Confocal images for Iba1 were gray-scaled and inversed. [Colorfigure can be viewed in the online issue, which is available atwileyonlinelibrary.com.]


Hippocampus

in wild type OHSCs. Treatment of wild type OHSCs with 1028

M CXCL10 for 24 h did not induce morphological activation ofmicroglia (Figs. 5B,D) when compared to untreated controls(Fig. 5A/C). Microglia in CXCL10-treated OHSCs maintained

their ramified morphology with small somata and secondary/tertiary branching, suggesting that CXCL10 per se is not suffi-cient to induce microglia activation (Fig. 5D). In addition,treatment with 1028 M CXCL10 for 24 h did not induce

FIGURE 5. CXCL10 does not directly induce microglia activa-tion or neuronal cell death, but rather affects neuronal cell deaththrough modulation of microglia activity. After 6 DIV, wild typeOHSCs were treated with or without 1028 M CXCL10 for 24 htogether with 5 lg/ml PI to asses the effects of direct exposure toCXCL10 on microglia activation and neuronal cell death. In bothcontrol (A) and CXCL10-treated slice cultures (B) microglia (Iba1)remained in their ramified state (C, D). No effect of CXCL10-treatment on neuronal (NeuN) cell death (PI) was observed (F)when compared with control conditions (E). However, in the

absence of microglia, differences in NMDA-induced neuronal celldeath between wild type and CXCR32/2 OHSCs are fully abro-gated. Graphs in (G) depict the percentages of neuronal cell death(mean 6 SEM) in the CA1, CA3, and DG regions in response tovarious NMDA-concentrations in microglia-depleted wild type(WT Lip-CL) and CXCR32/2 (CXCR32/2 Lip-CL) OHSCs. Datadepicted in (G) are a summary of three separate experiments withat least 6 OHSCs per condition. Scale bars indicate 300 lm (A–B,E–F) and 30 lm (C, D). [Color figure can be viewed in the onlineissue, which is available at wileyonlinelibrary.com.]


Hippocampus

neuronal cell death in wild type OHSCs (Fig. 5F), demonstrat-ing that the presence of CXCL10 per se, in the absence of anexcitotoxic stimulus, does not induce neurodegeneration in thismodel.

Differences in Neuronal Cell Death BetweenWild Type and CXCR32/2 OHSCs Are FullyAbrogated by Ablation of Microglia

To further investigate the role of microglia in NMDA-induced excitotoxicity we depleted wild type- and CXCR32/2

OHSCs from their endogenous microglia population by treat-ment with liposomic clodronate (Lip-CL), without affectingother cell-types (Marin-Teva et al., 2004; Markovic et al.,2005), and subsequently exposed these OHSCs to various con-centrations of NMDA. The basal rate of neuronal cell death inmicroglia-depleted CXCR32/2 OHSCs (CA1; 1.0 6 0.7%,CA3; 1.2 6 0.4%; DG; 4.0 6 2.0%, P 5 1.000) and wildtype OHSCs (CA1; 2.5 6 0.6%, CA3; 3.1 6 0.8%; DG; 4.16 1.0%, P 5 1.000) did not differ from untreated CXCR32/2

and wild type slice conditions, respectively (Fig. 5G). Inresponse to NMDA-treatment, neuronal cell death wasenhanced in the absence of microglia in both CXCR32/2 andwild type OHSCs. These results are in accordance with previ-ous findings in our lab and have been discussed elsewhere (VanWeering, manuscript submitted). However, although we previ-ously observed significant differences in neuronal cell deathbetween wild type- and CXCR32/2 OHSCs (Fig. 3A), no suchdifferences were observed between microglia-depleted wildtype- and CXCR32/2 OHSCs in response to NMDA-treat-ment at any concentration in any hippocampal region (Fig.5G).

DISCUSSION

CXCL10 Differently Affects CA and DG NeuronsUpon Excitotoxicity

Our primary finding here is that CXCL10/CXCR3 signalingdifferentially affects neuronal survival in the three hippocampalregions upon NMDA-induced excitotoxicity. Accordingly, wedemonstrated that neurons in the CA regions displayed adecreased vulnerability to NMDA-treatment in the absence ofCXCL10 or CXCR3, suggesting that the presence of CXCL10has a negative effect on the survival of these neurons upon exci-totoxicity. In contrast, abrogation of CXCL10/CXCR3 signal-ing did not desensitize neurons of the DG region in responseto excitotoxicity, but resulted in an increased vulnerability ofthese cells to high (50 lM) NMDA-concentrations. Further-more, we observed that exposure to exogenous CXCL10 onlyaffected neuronal survival in the presence of NMDA and didnot lead to neuronal cell death or glia activation in controlOHSCs, suggesting that CXCL10 per se does not induce neu-rodegeneration nor does it directly trigger glia activation under

healthy conditions. Indeed, it has been shown previously thatchronic expression of CXCL10 in the CNS in vivo did notlead to neuropathological signs under healthy conditions(Boztug et al., 2002). However, since abrogation of CXCL10/CXCR3 signaling resulted in changes in the degree of NMDA-induced neuronal cell death, it is suggested that CXCL10 ispart of a signaling system that (in)directly affects the process ofexcitotoxicity-induced neurodegeneration rather than that it isthe direct cause of neuropathological signs in our paradigm.These CXCL10-mediated effects on neuronal survival underexcitotoxic conditions might be through modulation of neuro-nal properties or through modulation of microglia and/or astro-cyte activation (see below).

Effects of CXCL10 on Neurons

Although CXCL10 per se did not induce neuronal celldeath, it is not unlikely that this chemokine has direct effectson neurons as several studies have shown CXCR3-expression inthese cells (Coughlan et al., 2000; Xia et al., 2000; Nelson andGruol, 2004). Recently, it has been demonstrated that acute ex-posure of CXCL10 altered neuronal signaling properties in rathippocampal neuron-glia cultures (Nelson and Gruol, 2004)and reduced long-term potentiation via CXCR3 in acute adulthippocampal slices (Vlkolinsky et al., 2004). Although theinvolvement of glia cells could not be excluded in both studies,the findings suggest that CXCL10 can have an (in)direct mod-ulatory effect on neuronal properties in the hippocampus. Inline with this, it has been shown that exposure of rat hippo-campal neuron-glia cultures to relatively high levels ofCXCL10 resulted in (in)direct downregulation of synaptic pro-teins GAD65/67, GABABR1, and GABAARa2 and upregula-tion of NMDAR1 and mGluR2/3 in the neuronal population(Cho et al., 2009). Whether such changes may account for thedifferences in neuronal cell death in our paradigm is not yetclear. However, an CXCL10-induced downregulation of inhibi-tory synaptic proteins and upregulation of glutamate-subtypereceptors in neurons could in part explain the relative insensi-tivity of CA neurons in CXCL102/2 and CXCR32/2 OHSCstowards NMDA and may also account for the increased vulner-ability of these neurons in CXCL102/2 slice cultures whenexposed to CXCL10.

CXCL10 is Expressed in Astrocytes in Responseto Excitotoxicity

We demonstrated that upon NMDA-induced neuronal celldeath CXCL10 is expressed in astrocytes, but not in neuronsor microglia. Expression of CXCL10 in astrocytes in vivo hasbeen observed in various excitotoxicity-related neurodegenera-tive pathologies such as in stroke (Wang et al., 1998), encepha-litis (Bhowmick et al., 2007), Alzheimer’s disease (Xia et al.,2000), MS (Balashov et al., 1999; Simpson et al., 2000; Omariet al., 2005; Tanuma et al., 2006) and its EAE animal model(Ransohoff et al., 1993; Carter et al., 2007). In these studies,local upregulation of CXCL10 expression in reactive astrocyteswas highly correlated with sites of neuronal injury, suggesting


Hippocampus

that the process of neurodegeneration itself may provide thesignals which trigger CXCL10 expression in these cells. Possiblesignaling candidates are (pro-)inflammatory factors such asIFNb1/2, IFNg, IL-1a/b and TNFa (Oh et al., 1999; Huaand Lee, 2000; Salmaggi et al., 2002; Croitoru-Lamoury et al.,2003; Omari et al., 2005), which all have been shown toinduce CXCL10 expression in astrocytes in vitro. However,direct evidence that these factors also stimulate astrocyticCXCL10 expression in response to excitotoxicity in vivoremains to be established.

Effects of CXCL10 on Neuronal Survival AreMost Likely Mediated Through Microglia

Microglia are considered to play a prominent role in excito-toxic insults as inhibition of microglia activity has been shownto differentially affect the extent of neuronal cell death (Yrjan-heikki et al., 1998; Tikka and Koistinaho, 2001; Kitamuraet al., 2004; Hailer et al., 2005; Turrin and Rivest, 2006; Imaiet al., 2007; Lalancette-Hebert et al., 2007; Xapelli et al.,2008; Hailer, 2008;). Thus, microglia are not inevitably neuro-protective as certain conditions might trigger a (pro-)inflamma-tory phenotype in these cells that is detrimental for neuronalsurvival under excitotoxic conditions. Indeed, it has been dem-onstrated that upon excitotoxicity, activated microglia expresspro-inflammatory factors, such as TNFa, IL-1, NO, iNOS andCOX-2, which are all toxic to neurons when expressed in largequantities (Barone and Feuerstein, 1999; Aloisi, 2001; Campu-zano et al., 2008).

The data presented here demonstrate that CXCL10/CXCR3signaling might be one of the factors that affect microglia activ-ity under excitotoxic conditions. Previously, it has been demon-strated that microglia functionally express CXCR3 in vitro(Biber et al., 2001; Biber et al., 2002; Rappert et al., 2002;Dijkstra et al., 2004). Moreover, we observed pronouncedCXCR3-expession in acutely isolated microglia from wild typeOHSCs. However, we demonstrated that accumulation ofmicroglia at sites of neuronal cell death was not impaired inCXCR32/2 OHSCs compared with wild type conditions, sug-gesting that CXCL10/CXCR3 signaling is not a prerequisitefor microglia recruitment in response to excitotoxicity. In linewith this evidence, it has been described recently that themigration of microglia towards injured neurons in the hippo-campus depends on purinergic signaling (Kurpius et al., 2007).However, since differences in neuronal cell death observedbetween wild type and CXCR32/2 OHSCs were fully abro-gated in the absence of microglia, it suggests that neuronal sur-vival under excitotoxic conditions is affected through microglialCXCR3, which is in agreement with earlier studies from ourgroup (Rappert et al., 2004).

Why CXCL10/CXCR3 signaling is detrimental in the CAregions of the hippocampus and protective in the DG region isnot understood. Regional differences in the response ofmicroglia to CXCR3-stimulation might be instrumental here.Alternatively, regional differences in coexpression of local fac-tors, together with CXCL10, might direct microglia to either a

proinflammatory or a neuroprotective phenotype. More experi-ments are necessary to address these concepts.

In summary, our results identify a region-specific role forCXCL10/CXCR3 signaling in neuron-glia and glia-glia interactionsunder pathological conditions. Since ablation of microglia abro-gated the differences in neuronal cell death between wild type andCXCR32/2 OHSCs, an important role for microglial CXCR3 inneuronal excitotoxicity in the hippocampus is suggested.

REFERENCES

Acarin L, Gonzalez B, Castellano B, Castro AJ. 1996. Microglialresponse to N-methyl-D-aspartate-mediated excitotoxicity in theimmature rat brain. J Comp Neurol 367:361–374.

Aloisi F. 2001. Immune function of microglia. Glia 36:165–179.Arundine M, Tymianski M. 2003. Molecular mechanisms of calcium-

dependent neurodegeneration in excitotoxicity. Cell Calcium34:325–337.

Balashov KE, Rottman JB, Weiner HL, Hancock WW. 1999.CCR5(1) and CXCR3(1) T cells are increased in multiple sclero-sis and their ligands MIP-1alpha and IP-10 are expressed indemyelinating brain lesions. Proc Natl Acad Sci USA 96:6873–6878.

Barone FC, Feuerstein GZ. 1999. Inflammatory mediators and stroke:New opportunities for novel therapeutics. J Cereb Blood FlowMetab 19:819–834.

Bhowmick S, Duseja R, Das S, Appaiahgiri MB, Vrati S, Basu A.2007. Induction of IP-10 (CXCL10) in astrocytes following Japa-nese encephalitis. Neurosci Lett 414:45–50.

Biber K, Dijkstra I, Trebst C, De Groot CJ, Ransohoff RM, BoddekeHW. 2002. Functional expression of CXCR3 in cultured mouseand human astrocytes and microglia. Neuroscience 112:487–497.

Biber K, Klotz KN, Berger M, Gebicke-Harter PJ, van CD. 1997.Adenosine A1 receptor-mediated activation of phospholipase C incultured astrocytes depends on the level of receptor expression.J Neurosci 17:4956–4964.

Biber K, Sauter A, Brouwer N, Copray SC, Boddeke HW. 2001. Is-chemia-induced neuronal expression of the microglia attractingchemokine Secondary Lymphoid-tissue Chemokine (SLC). Glia34:121–133.

Boztug K, Carson MJ, Pham-Mitchell N, Asensio VC, Demartino J,Campbell IL. 2002. Leukocyte infiltration, but not neurodegenera-tion, in the CNS of transgenic mice with astrocyte production ofthe CXC chemokine ligand 10. J Immunol 169:1505–1515.

Campuzano O, Castillo-Ruiz MM, Acarin L, Castellano B, GonzalezB. 2008. Distinct pattern of microglial response, cyclooxygenase-2,and inducible nitric oxide synthase expression in the aged rat brainafter excitotoxic damage. J Neurosci Res 86:3170–3183.

Carter SL, Muller M, Manders PM, Campbell IL. 2007. Induction ofthe genes for Cxcl9 and Cxcl10 is dependent on IFN-gamma butshows differential cellular expression in experimental autoimmuneencephalomyelitis and by astrocytes and microglia in vitro. Glia55:1728–1739.

Cartier L, Hartley O, Dubois-Dauphin M, Krause KH. 2005. Chemo-kine receptors in the central nervous system: Role in brain inflam-mation and neurodegenerative diseases. Brain Res Brain Res Rev48:16–42.

Cho J, Nelson TE, Bajova H, Gruol DL. 2009. Chronic CXCL10alters neuronal properties in rat hippocampal culture. J Neuroim-munol 207:92–100.

Choi DW. 1988. Glutamate neurotoxicity and diseases of the nervoussystem. Neuron 1:623–634.


Hippocampus

Choi DW, Rothman SM. 1990. The role of glutamate neurotoxicityin hypoxic-ischemic neuronal death. Annu Rev Neurosci 13:171–182.

Chomczynski P, Sacchi N. 1987. Single-step method of RNA isolationby acid guanidinium thiocyanate-phenol-chloroform extraction.Anal Biochem 162:156–159.

Cinque P, Bestetti A, Marenzi R, Sala S, Gisslen M, Hagberg L, PriceRW. 2005. Cerebrospinal fluid interferon-gamma-inducible protein10 (IP-10, CXCL10) in HIV-1 infection J Neuroimmunol168:154–163.

Copray JC, Brouwer N. 1994. Selective expression of neurotrophin-3messenger RNA in muscle spindles of the rat. Neuroscience63:1125–1135.

Coughlan CM, McManus CM, Sharron M, Gao Z, Murphy D, JafferS, Choe W, Chen W, Hesselgesser J, Gaylord H, Kalyuzhny A, LeeVM, Wolf B, Doms RW, Kolson DL. 2000. Expression of multiplefunctional chemokine receptors and monocyte chemoattractantprotein-1 in human neurons. Neuroscience 97:591–600.

Croitoru-Lamoury J, Guillemin GJ, Boussin FD, Mognetti B, GigoutLI, Cheret A, Vaslin B, Le GR, Brew BJ, Dormont D. 2003.Expression of chemokines and their receptors in human and simianastrocytes: Evidence for a central role of TNF alpha and IFNgamma in CXCR4 and CCR5 modulation. Glia 41:354–370.

De Haas AH, Boddeke HW, Biber K. 2008. Region-specific expressionof immunoregulatory proteins on microglia in the healthy CNS.Glia 56:888–894.

De Haas AH, Van Weering HR, De Jong EK, Boddeke HW, BiberKP. 2007. Neuronal chemokines: Versatile messengers in centralnervous system cell interaction. Mol Neurobiol 36:137–151.

Dehghani F, Hischebeth GT, Wirjatijasa F, Kohl A, Korf HW, HailerNP. 2003. The immunosuppressant mycophenolate mofetil attenu-ates neuronal damage after excitotoxic injury in hippocampal slicecultures. Eur J Neurosci 18:1061–1072.

Dijkstra IM, Hulshof S, van der Valk P, Boddeke HW, Biber K. 2004.Cutting edge: Activity of human adult microglia in response to CCchemokine ligand 21. J Immunol 172:2744–2747.

Fife BT, Kennedy KJ, Paniagua MC, Lukacs NW, Kunkel SL, LusterAD, Karpus WJ. 2001. CXCL10 (IFN-gamma-inducible protein-10) control of encephalitogenic CD41 T cell accumulation in thecentral nervous system during experimental autoimmune encepha-lomyelitis. J Immunol 166:7617–7624.

Flynn G, Maru S, Loughlin J, Romero IA, Male D. 2003. Regulationof chemokine receptor expression in human microglia and astro-cytes. J Neuroimmunol 136:84–93.

Galimberti D, Schoonenboom N, Scheltens P, Fenoglio C, BouwmanF, Venturelli E, Guidi I, Blankenstein MA, Bresolin N, Scarpini E.2006. Intrathecal chemokine synthesis in mild cognitive impair-ment and Alzheimer disease. Arch Neurol 63:538–543.

Glaser J, Gonzalez R, Sadr E, Keirstead HS. 2006. Neutralization ofthe chemokine CXCL10 reduces apoptosis and increases axonsprouting after spinal cord injury. J Neurosci Res 84:724–734.

Goldberg SH, Van der Meer P, Hesselgesser J, Jaffer S, Kolson DL,Albright AV, Gonzalez-Scarano F, Lavi E. 2001. CXCR3 expressionin human central nervous system diseases. Neuropathol Appl Neu-robiol 27:127–138.

Hailer NP. 2008. Immunosuppression after traumatic or ischemicCNS damage: it is neuroprotective and illuminates the role ofmicroglial cells. Prog Neurobiol 84:211–233.

Hailer NP, Vogt C, Korf HW, Dehghani F. 2005. Interleukin-1betaexacerbates and interleukin-1 receptor antagonist attenuates neuro-nal injury and microglial activation after excitotoxic damage inorganotypic hippocampal slice cultures. Eur J Neurosci 21:2347–2360.

Heppner FL, Skutella T, Hailer NP, Haas D, Nitsch R. 1998. Acti-vated microglial cells migrate towards sites of excitotoxic neuronalinjury inside organotypic hippocampal slice cultures. Eur J Neuro-sci 10:3284–3290.

Hua LL, Lee SC. 2000. Distinct patterns of stimulus-inducible chemo-kine mRNA accumulation in human fetal astrocytes and microglia.Glia 30:74–81.

Imai F, Suzuki H, Oda J, Ninomiya T, Ono K, Sano H, Sawada M.2007. Neuroprotective effect of exogenous microglia in global brainischemia. J Cereb Blood Flow Metab 27:488–500.

Kitamura Y, Takata K, Inden M, Tsuchiya D, Yanagisawa D, NakataJ, Taniguchi T. 2004. Intracerebroventricular injection of microgliaprotects against focal brain ischemia. J Pharmacol Sci 94:203–206.

Klein RS, Izikson L, Means T, Gibson HD, Lin E, Sobel RA, WeinerHL, Luster AD. 2004. IFN-inducible protein 10/CXC chemokineligand 10-independent induction of experimental autoimmune en-cephalomyelitis. J Immunol 172:550–559.

Kohl A, Dehghani F, Korf HW, Hailer NP. 2003. The bisphosphonateclodronate depletes microglial cells in excitotoxically injured orga-notypic hippocampal slice cultures. Exp Neurol 181:1–11.

Kurpius D, Nolley EP, Dailey ME. 2007. Purines induce directedmigration and rapid homing of microglia to injured pyramidalneurons in developing hippocampus. Glia 55:873–884.

Lalancette-Hebert M, Gowing G, Simard A, Weng YC, Kriz J. 2007.Selective ablation of proliferating microglial cells exacerbates ische-mic injury in the brain. J Neurosci 27:2596–2605.

Liu L, Callahan MK, Huang D, Ransohoff RM. 2005. Chemokine re-ceptor CXCR3: an unexpected enigma. Curr Top Dev Biol68:149–181.

Li H, Gang Z, Yuling H, Luokun X, Jie X, Hao L, Li W, ChunsongH, Junyan L, Mingshen J, Youxin J, Feili G, Boquan J, Jinquan T.2006. Different neurotropic pathogens elicit neurotoxic CCR9- orneurosupportive CXCR3-expressing microglia. J Immunol 177:3644–3656.

Liu L, Huang D, Matsui M, He TT, Hu T, Demartino J, Lu B, Ge-rard C, Ransohoff RM. 2006. Severe disease, unaltered leukocytemigration, and reduced IFN-gamma production in CXCR32/2

mice with experimental autoimmune encephalomyelitis. J Immunol176:4399–4409.

Livak KJ, Schmittgen TD. 2001. Analysis of relative gene expressiondata using real-time quantitative PCR, the 2(-Delta Delta C(T))method. Methods 25:402–408.

Loetscher M, Loetscher P, Brass N, Meese E, Moser B. 1998. Lympho-cyte-specific chemokine receptor CXCR3: Regulation, chemokinebinding and gene localization. Eur J Immunol 28:3696–3705.

Marin-Teva JL, Dusart I, Colin C, Gervais A, van Rooijen N, MallatM. 2004. Microglia promote the death of developing Purkinjecells. Neuron 41:535–547.

Markovic DS, Glass R, Synowitz M, Van Rooijen N, Kettenmann H.2005. Microglia stimulate the invasiveness of glioma cells byincreasing the activity of metalloprotease-2. J Neuropathol ExpNeurol 64:754–762.

Muller M, Carter SL, Hofer MJ, Manders P, Getts DR, Getts MT,Dreykluft A, Lu B, Gerard C, King NJ, Campbell IL. 2007.CXCR3 signaling reduces the severity of experimental autoimmuneencephalomyelitis by controlling the parenchymal distribution ofeffector and regulatory T cells in the central nervous system.J Immunol 179:2774–2786.

Nelson TE, Gruol DL. 2004. The chemokine CXCL10 modulatesexcitatory activity and intracellular calcium signaling in culturedhippocampal neurons. J Neuroimmunol 156:74–87.

Neville LF, Mathiak G, Bagasra O. 1997. The immunobiology ofinterferon-gamma inducible protein 10 kD (IP-10): A novel, pleio-tropic member of the C-X-C chemokine superfamily. CytokineGrowth Factor Rev 8:207–219.

Oh JW, Schwiebert LM, Benveniste EN. 1999. Cytokine regulation ofCC, CXC chemokine expression by human astrocytes. J Neurovirol5:82–94.

Omari KM, John GR, Sealfon SC, Raine CS. 2005. CXC chemokinereceptors on human oligodendrocytes: Implications for multiplesclerosis. Brain 128:1003–1015.


Hippocampus

Pozzo Miller LD, Mahanty NK, Connor JA, Landis DM. 1994. Spon-taneous pyramidal cell death in organotypic slice cultures from rathippocampus is prevented by glutamate receptor antagonists. Neu-roscience 63:471–487.

Ramakers C, Ruijter JM, Deprez RH, Moorman AF. 2003. Assump-tion-free analysis of quantitative real-time polymerase chain reac-tion (PCR) data. Neurosci Lett 339:62–66.

Ransohoff RM, Hamilton TA, Tani M, Stoler MH, Shick HE, MajorJA, Estes ML, Thomas DM, Tuohy VK. 1993. Astrocyte expres-sion of mRNA encoding cytokines IP-10 and JE/MCP-1 in experi-mental autoimmune encephalomyelitis. FASEB J 7:592–600.

Rappert A, Biber K, Nolte C, Lipp M, Schubel A, Lu B, Gerard NP,Gerard C, Boddeke HW, Kettenmann H. 2002. Secondary lymph-oid tissue chemokine (CCL21) activates CXCR3 to trigger a Cl-current and chemotaxis in murine microglia. J Immunol168:3221–3226.

Rappert A, Bechmann I, Pivneva T, Mahlo J, Biber K, Nolte C, KovacAD, Gerard C, Boddeke HW, Nitsch R, Kettenmann H. 2004.CXCR3-dependent microglial recruitment is essential for dendriteloss after brain lesion. J Neurosci 24:8500–8509.

Salmaggi A, Gelati M, Dufour A, Corsini E, Pagano S, Baccalini R,Ferrero E, Scabini S, Silei V, Ciusani E, De RM. 2002. Expressionand modulation of IFN-gamma-inducible chemokines (IP-10, Mig,and I-TAC) in human brain endothelium and astrocytes: Possiblerelevance for the immune invasion of the central nervous systemand the pathogenesis of multiple sclerosis. J Interferon CytokineRes 22:631–640.

Sanders VJ, Pittman CA, White MG, Wang G, Wiley CA, AchimCL. 1998. Chemokines and receptors in HIV encephalitis. AIDS12:1021–1026.

Simpson JE, Newcombe J, Cuzner ML, Woodroofe MN. 2000.Expression of the interferon-gamma-inducible chemokines IP-10and Mig and their receptor, CXCR3, in multiple sclerosis lesions.Neuropathol Appl Neurobiol 26:133–142.

Sorensen TL, Tani M, Jensen J, Pierce V, Lucchinetti C, Folcik VA,Qin S, Rottman J, Sellebjerg F, Strieter RM, Frederiksen JL, Ran-sohoff RM. 1999. Expression of specific chemokines and chemo-kine receptors in the central nervous system of multiple sclerosispatients. J Clin Invest 103:807–815.

Stoppini L, Buchs PA, Muller D. 1991. A simple method for organo-typic cultures of nervous tissue. J Neurosci Methods 37:173–182.

Tanuma N, Sakuma H, Sasaki A, Matsumoto Y. 2006. Chemokineexpression by astrocytes plays a role in microglia/macrophage acti-vation and subsequent neurodegeneration in secondary progressivemultiple sclerosis. Acta Neuropathol 112:195–204.

Tikka TM, Koistinaho JE. 2001. Minocycline provides neuroprotec-tion against N-methyl-D-aspartate neurotoxicity by inhibitingmicroglia. J Immunol 166:7527–7533.

Turrin NP, Rivest S. 2006. Tumor necrosis factor alpha but not inter-leukin 1 beta mediates neuroprotection in response to acute nitricoxide excitotoxicity. J Neurosci 26:143–151.

Van Rooijen N, Sanders A. 1994. Liposome mediated depletion ofmacrophages: Mechanism of action, preparation of liposomes andapplications. J Immunol Methods 174:83–93.

Vlkolinsky R, Siggins GR, Campbell IL, Krucker T. 2004. Acute ex-posure to CXC chemokine ligand 10, but not its chronic astroglialproduction, alters synaptic plasticity in mouse hippocampal slices.J Neuroimmunol 150:37–47.

Vornov JJ, Tasker RC, Coyle JT. 1991. Direct observation of the ago-nist-specific regional vulnerability to glutamate, NMDA, and kai-nate neurotoxicity in organotypic hippocampal cultures Exp Neurol114:11–22.

Wang X, Ellison JA, Siren AL, Lysko PG, Yue TL, Barone FC, Shatz-man A, Feuerstein GZ. 1998. Prolonged expression of interferon-inducible protein-10 in ischemic cortex after permanent occlusionof the middle cerebral artery in rat. J Neurochem 71:1194–1204.

Xapelli S, Bernardino L, Ferreira R, Grade S, Silva AP, Salgado JR,Cavadas C, Grouzmann E, Poulsen FR, Jakobsen B, Oliveira CR,Zimmer J, Malva JO. 2008. Interaction between neuropeptide Y(NPY) and brain-derived neurotrophic factor in NPY-mediatedneuroprotection against excitotoxicity: A role for microglia. EurJ Neurosci 27:2089–2102.

Xia MQ, Bacskai BJ, Knowles RB, Qin SX, Hyman BT. 2000.Expression of the chemokine receptor CXCR3 on neurons and theelevated expression of its ligand IP-10 in reactive astrocytes: Invitro ERK1/2 activation and role in Alzheimer’s disease. J Neuro-immunol 108:227–235.

Yrjanheikki J, Keinanen R, Pellikka M, Hokfelt T, Koistinaho J. 1998.Tetracyclines inhibit microglial activation and are neuroprotectivein global brain ischemia. Proc Natl Acad Sci USA 95:15769–15774.


Hippocampus

CXCL10/CXCR3 signaling in glia cells differentially affects NMDA-induced cell death in CA and DG...

Documents

Transcript of CXCL10/CXCR3 signaling in glia cells differentially affects NMDA-induced cell death in CA and DG...