of April 14, 2018.This information is current as

BrainOxide-Induced Neurotoxicity in the Mouse

on Nitricγ and IFN-αEffects of TNF-

Véronique Blais and Serge Rivest

http://www.jimmunol.org/content/172/11/7043doi: 10.4049/jimmunol.172.11.7043

2004; 172:7043-7052; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/172/11/7043.full#ref-list-1

, 15 of which you can access for free at: cites 39 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2004 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on April 14, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on April 14, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Effects of TNF-� and IFN-� on Nitric Oxide-InducedNeurotoxicity in the Mouse Brain1

Veronique Blais and Serge Rivest2

The present study investigated the interaction between highly reactive gaseous-free radical NO and cytokines that are producedby activated Th-1 cells on the cerebral immune response and neuronal integrity. CD-1 mice received an intrastriatal infusion ofdifferent solutions containing the NO synthase inhibitor N(G)-nitro- L-arginine methylester, NO-releasing substance sodium ni-troprusside (SNP), IFN-�, and/or TNF-�. The solution containing both cytokines caused a profound and transient transcriptionalactivation of numerous genes encoding proinflammatory proteins in microglial/monocytic cells ipsilateral to infusion site. Thisincrease in gene expression peaked 1 day after the cerebral bolus of cytokines and returned to basal levels from 3 to 7 days postadministration. N(G)-nitro- L-arginine methylester further stimulated this immune reaction to IFN-� and TNF-�, but the brain ofthese mice failed to exhibit signs of neurodegeneration and demyelination. In contrast, a single bolus of SNP in the striatal regioncaused neuronal death and demyelination as early as 1 to 3 days following the infusion with the NO donor. This phenomenon wasgreatly exacerbated by the coadministration of both cytokines, although TNF-� remained the most critical cytokine to enhance thedamage of cerebral elements. These data provide evidence that NO has the ability to modulate the immune response, which is notby itself detrimental for the brain. However, SNP-induced NO production together with TNF-� in the cerebral environment arecritical events leading to intense neurodegeneration and demyelination in vivo.The Journal of Immunology, 2004, 172: 7043–7052.

T he proinflammatory cytokine TNF-� is essential to or-chestrate the immune response in the brain and to protectneuronal cells against pathogens. However, an exagger-

ated response and over production of this molecule may not besuitable and there is accumulating evidence that chronic inflam-mation has a leading role in damaging neurons and other cells ofthe CNS (reviewed in Ref. 1). The duration of the response andlevels of TNF-� in the cerebral environment may be the criticalfactors for the fate of the immune system to protect or cause celldeath.

The detrimental effects of TNF-� in the CNS may also dependon the presence of other cytokines produced by either residentpopulations of cells or infiltrating leukocytes, such as T cells. Ag-stimulated T cells have the ability to produce TNF-� along withIFN-� that acts in both innate and specific cell-mediated immuni-ties (2). Activated CD4� Th1 cells and their secreted cytokines arebelieved to be crucial to the pathogenesis of multiple sclerosis,especially during the demyelinating episodes (3). The mixture ofboth TNF-� and IFN-� may therefore be harmful to neuronal el-ements. In this regard, transgenic mice expressing IFN-� in hip-pocampus exhibit a profound enhanced microglial reactivity to le-sion-induced neuronal injury indicating that IFN-� acts as an

amplifier of the response (4). Four days of treatment with IFN-�failed to alter cell survival or myelin basic protein gene expressionin cultured human oligodendrocytes, but these cells are more sus-ceptible to Fas-mediated apoptosis and this effect is augmented inpresence of TNF-� (5).

NO is another important and versatile player in the immunesystem and accumulating evidence supports the hypothesis that itplays a critical role in chronic degenerative diseases (6). Activationof the inducible form of NO synthase (iNOS)3 and NO productiontake place during many neuropathological conditions (7). iNOS isresponsible for the cytotoxic action of macrophages and neutro-phils and NO has been implicated in neurodegeneration andchronic inflammation (8, 9). NO has also been found to participatein Fas-mediated apoptosis in Jurkat cells (10) and NO donor caninduce apoptosis of murine thymocytes (11). Although inductionof cell death by exogenous administration of NO donor has not yetbeen studied in the mouse brain, implication of this molecule incellular toxicity has been suggested by the use of NO synthaseinhibitors or knockout mice in different models of demyelinatingand neurodegenerative diseases (12–15).

It is therefore possible that cytokines produced by the adaptiveimmune system together with a sharp increase in NO levels pro-mote demyelination and neurological disorders. In the presentstudy, we investigated the effects of TNF-� and IFN-� in the brainof mice challenged with the NO donor sodium nitroprusside (SNP)or NO synthase inhibitor N(G)-nitro-L-arginine methylester (L-NAME). We evaluated the expression of genes encoding severalproinflammatory molecules and investigated the consequences ofsuch a response on the neuronal integrity. Although mice chal-lenged with both cytokines exhibited a profound and transient ce-rebral immune reaction that was enhanced by L-NAME, their

Laboratory of Molecular Endocrinology, Centre Hospitalier de l’Universite LavalResearch Center, and Department of Anatomy and Physiology, Laval University,Quebec City, Quebec, Canada

Received for publication August 7, 2003. Accepted for publication March 18, 2004.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work is supported by The Canadian Institutes in Health Research. S.R. is aCanadian Institutes in Health Research scientist and holds a Canadian Research Chairin Neuroimmunology. V.B. is supported by a PhD studentship from the K.M. Hunter/Canadian Institutes in Health Research foundation.2 Address correspondence and reprint requests to Dr. Serge Rivest, Laboratory ofMolecular Endocrinology, Centre Hospitalier de l’Universite Laval Research Center,2705 boulevard Laurier, Quebec City, Quebec, Canada, G1V 4G2. E-mail address:Serge.Rivest@crchul.ulaval.ca

3 Abbreviations used in this paper: iNOS, inducible NO synthase; FJB, Fluoro-Jade B;NeuN, neuronal nuclei; LFB, Luxol fast blue; L-NAME, N(G)-nitro-L-arginine meth-ylester; MCP-1, monocyte chemoattractant protein 1; SNP, sodium nitroprusside;TLR, Toll-like receptor; TRAF, TNFR1-associated factor.

The Journal of Immunology

Copyright © 2004 by The American Association of Immunologists, Inc. 0022-1767/04/$02.00

by guest on April 14, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

brains did not show any signs of neurodegeneration. However, theneurotoxic effects of SNP were clearly exacerbated by the immuneligands.

Materials and MethodsAnimals

Adult male CD-1 mice (30–35 g of body weight) were acclimated to stan-dard laboratory conditions (14-h light and 10-h dark cycle, lights on at t �0600 h and off at t � 2000 h) with free access to mouse chow and water.Animal breeding and experiments were conducted according to CanadianCouncil on Animal Care guidelines, as administered by the Laval Univer-sity Animal Care Committee. A total of 130 mice were assigned to differentprotocols as described in Table I.

Experimental protocols

Mice were anesthetized with an i.p. injection (10 ml/kg of body weight) ofa mixture of ketamine hydrochloride (15 mg/ml) and xylazine (1 mg/ml)and placed in a stereotaxic apparatus (Kopf Instruments, Tujunga, CA).The striatum was then reached (�2.0 mm lateral and �3.0 mm dorsoven-tral to the bregma) with a 33-gauge stainless steel cannula (Plastics One,Roanoke, VA) that was connected to a 50-�l Hamilton syringe with anintramedic polyethylene tubing (PE-50; Caly Adams, Parsipanny, NJ). Avolume of 1 �l was infused over 2 min by means of a microinjection pump(model A-99; Razel Scientific Instruments, Stanford, CT). These coordi-nates were selected on the basis of preliminary data showing a robusthybridization signal and reliable pattern of cytokine gene expression overthe ipsilateral cerebral cortex, hippocampus, corpus callosum, and basalganglia following a single bolus of immune ligands. Besides, the corpuscallosum was found to exhibit demyelination in response to toxins (16) andthis phenomenon was investigated in the present study. We did not how-ever select these coordinates to mimic a particular neurodegenerative dis-ease, but they were used to represent a more general mechanism of immuneresponse in the cerebral tissue. At different times after the injection, micewere deeply anesthetized via an i.p. injection (100 �l) of a mixture ofketamine hydrochloride (91 mg/ml) and xylazine (9.1 mg/ml) and thenrapidly perfused transcardially with 0.9% saline, followed by 4% parafor-maldehyde in 0.1 M sodium phosphate, pH 7.6, at 4°C.

In situ hybridization histochemistry

After transcardiac perfusions, brains were rapidly removed from the skull,postfixed for 1–3 days, and then placed in a solution containing 20% su-crose diluted in 4% paraformaldehyde-sodium phosphate buffer (pH 7.6)overnight at 4°C. The frozen brains were mounted on a microtome(Reichert-Jung; Cambridge Instruments, Deerfield, IL) and cut into 20-�mcoronal sections from the olfactory bulb to the caudal medulla. The sliceswere collected in a cold cryoprotectant solution and stored at �20°C. Hy-bridization histochemical localization of Toll-like receptor (TLR)2, indexof innate immune response; I�B�, index of NF-�B activation; monocytechemoattractant protein 1 (MCP-1), index of chemokine induction; TNF-�,a general index of inflammation; iNOS, index of phagocyte activation; andIL-12p40, a key cytokine involved in the transfer from the innate to adap-tive immunity, mRNAs were performed on every 12th section of the entirerostrocaudal span of each brain as previously described (17–19). Plasmidswere linearized and sense and antisense riboprobes synthesized as de-scribed in Table II.

Immunocytochemistry was combined with the in situ hybridization pro-tocol to determine the type(s) of cells that express the different transcriptsin the mouse brain. Double-labeling procedures were similar to those de-scribed in other studies published by our group (19–21).

Detection of demyelination, neuronal death, and apoptosis

Demyelination was determined via the Luxol fast blue (LFB) staining.Every sixth section of the whole rostrocaudal extent of each brain wasmounted onto poly-L-lysine-coated slides, dried overnight under vacuum,dehydrated through graded concentrations of alcohol (50, 70, and 95%; 1min each), and incubated at 60°C for 6 h in the LFB solution (solvent blue38 at 1%; Sigma-Aldrich, St. Louis, MO) in ethanol 95% and 0.5% aceticacid. The sections were then rinsed in alcohol 95% (1 min), lithium car-bonate 0.05% (Sigma-Aldrich, 1–5 min), and alcohol 70% (2 dips). There-after, the slides were incubated in eosine Y solution 1% (EM DiagnosticSystems, Gibbstown, NJ) for 40 s, rinsed in distillated water, incubated incresyl violet 0.25% (Sigma-Aldrich) for 40 s, rinsed in distillated water,dehydrated through graded concentrations of alcohol (50, 70, 95, and100%; 1 min each), cleared two times in xylene for 1 min, and coverslippedwith distrene plasticizer xylene (Electron Microscopy Sciences, FortWashington, PA).

Table I. Treatment, dose, and number of mice used in each condition

Treatmenta

No. of Mice in Day

1 d 3 d 7 d 15 d 21 d

Saline solution (1 �l) 4 3 2rmIFN-� (100 ng) 3 4 2rmTNF-� (100 ng) 3 3 2rmIFN-� (100 ng) � rmTNF-� (100 ng) 7 5 6 2 2SNP (1 �g) 4 5 3 3 5SNP (1 �g) � rmIFN-� (100 ng) 4 5 4 3 3SNP (1 �g) � rmTNF-� (100 ng) 2 3 3 3 2SNP (1 �g) � rmIFN-� (100 ng) � rmTNF-� (100 ng) 6 5 3 4 3L-NAME (1 �g) 3 2L-NAME (1 �g) � rmIFN-� (100 ng) � rmTNF-� (100 ng) 2 2

a Treatment using recombinant murine (rm) IFN-�: catalog no. 485-MI, lot no. CFP031051 (R&D Systems): rmTNF-�:catalog no. 410-MT, lot no. CS081031 and CS081071 (R&D Systems): SNP: catalog no. S-0501, lot no. 80K3698 (Sigma-Aldrich); L-NAME: catalog no. N-5751, lot no. 40K0577 (Sigma-Aldrich).

Table II. Plasmids and enzymes used for the synthesis of cRNA probes

Plasmid (mouse) Vector Length (bp)

Enzymes

Sense Probe Antisense Probe

TLR2 PCR-blunt II-topo 2278 Spe I/T7 EcoR V/Sp6I�B� Bluescript II� 1114 Hind III/T3 BamH I/T7MCP-1 pGEMEX-1 578 Sac I/Sp6 BamH I/T7TNF-� Bluescript K II� 1300 BamH I/T7 Pst I/T3iNOS Bluescript SK II� 817 EcoR I/T7 Kpn I/T3IL-12p40 pCL-Neo 1050 Not I/T7 Xho I/T3

7044 ROLE OF CYTOKINES IN NO-INDUCED NEURODEGENERATION

by guest on April 14, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Apoptotic cells were labeled by immunohistochemistry using a cleavedcaspase-3 mAb. Brain sections were washed in sterile diethyl pyrocarbon-ate-treated 50 mM potassium PBS and incubated 48 h at 4°C with thecleaved caspase-3 Ab Asp175 (Cell Signaling Technology, Mississauga,Ontario, Canada), which was diluted in sterile potassium PBS (1/800) plus0.4% Triton X-100 plus 1% BSA (fraction V; Sigma-Aldrich) plus 1%normal goat serum. After incubation with the primary Ab, brain slices wererinsed in sterile potassium PBS and incubated with a mixture of potassiumPBS plus 0.02% Triton X-100 plus 1% BSA plus Cy2-conjugated anti-rabbit IgG Ab (1/500; Jackson ImmunoResearch Laboratory, West Grove,PA) for 2 h in a dark room at 20°C. Tissues were thereafter rinsed in sterilepotassium PBS, mounted onto poly-L-lysine slides, and coverslipped withVectaShield mounting medium with propidium iodide (Vector Laborato-ries, Burlingame, CA). The phenotype of cleaved caspase-3 immunoreac-tive cells was determined via double immunofluorescence. Brain sectionswere incubated overnight at 4°C with an Ab directed against neuronalnuclei (NeuN; Chemicon International, Temecula, CA), glial fibrillaryacidic protein (Chemicon International), or Mac-1 �-chain (CD11b; BDBiosciences, San Jose, CA), which was diluted in sterile potassium PBS(1/1000, 1/1000, or 1/500, respectively) plus 0.4% Triton X-100 plus 1%BSA plus 1% normal goat serum. After incubation with the primary Ab,brain slices were rinsed in potassium PBS and incubated with a mixture ofpotassium PBS plus 0.02% Triton X-100 plus 1% BSA plus Cy3-conju-gated anti-mouse IgG Ab (1/500; Jackson ImmunoResearch Laboratory;for NeuN and glial fibrillary acidic protein) or Cy3-conjugated anti-rat IgGAb (1/500; Jackson ImmunoResearch Laboratory; for mac-1) for 2 h in adark room at 20°C. Tissues were thereafter rinsed in sterile potassium PBSand incubated overnight at 4°C with the cleaved caspase-3 Ab Asp175,which was diluted in sterile potassium PBS (1/800) plus 0.4% Triton X-100plus 1% BSA plus 1% normal goat serum. After incubation with the pri-

mary Ab, brain slices were rinsed in sterile potassium PBS and incubatedwith a mixture of potassium PBS plus 0.02% Triton 100-X plus 1% BSAplus Alexa Fluor 488-conjugated anti-rabbit IgG Ab (1/500; MolecularProbes, Eugene, OR) for 2 h in a dark room at 20°C. Tissues were there-after rinsed in sterile potassium PBS, mounted onto poly-L-lysine slides,and coverslipped with a polyvinyl alcohol (Sigma-Aldrich) mounting me-dium containing 2.5% 1,4-diazabicyclo(2,2,2)-octane (Sigma-Aldrich) inbuffered glycerol.

Cell death by apoptosis was also detected using a TdT-FragEL DNAFragmentation Detection kit (Oncogene Research Products, San Diego,CA). Positive controls were generated from brain sections of animals thatreceived only sham treatments. These sections were mounted on the slidesand covered with 1 �g/�l DNase I in 1� TBS/1 mM MgSO4 for 20 minat room temperature.

Confirmation of neuronal death was determined via the Fluoro-Jade B(FJB) method (22). Briefly, every sixth section of the whole rostrocaudalextent of each brain was mounted onto poly-L-lysine-coated slides, driedunder vacuum 2 h, dehydrated through graded concentrations of alcohol(50, 70, and 100%; 1 min), rehydrated through graded concentrations ofalcohol (100, 70, and 50%; 1 min each) and 1 min in distillated water. Theywere then dipped and shaken into potassium permanganate (0.06%) for 10min, rinsed 1 min in distillated water, and dipped and shaken in a solutioncontaining FJB 0.004% (Histochem, Jefferson, AR) plus acetic acid 0.1%(Sigma-Aldrich) plus DAPI 0.0002% (Molecular Probes) for 20 min. Theslides were thereafter rinsed three times in distillated water (1 min each),dried, dipped in xylene three times (2 min each), and coverslipped withdistrene plasticizer xylene.

Nissl stain was also used as a general index of cellular morphology thatmay be altered in response to the different treatments.

FIGURE 1. Expression pattern ofgenes encoding molecules of the innateimmune system in response to an intrastri-atal infusion of cytokines and the NOdonor SNP. These darkfield photomicro-graphs were taken from nuclear emulsion-dipped coronal sections (20 �m) of micekilled 1 or 3 days after an intraparenchy-mal injection of saline solution (1 �l),IFN-� (100 ng), TNF-� (100 ng), IFN-��TNF-�, SNP (1 �g), or SNP�IFN-��TNF-�. All photomicrographs weretaken at a similar rostrocaudal level near-est the injection site. Please note the ro-bust and widespread hybridization signalfor TLR2 and MCP-1 transcripts all overthe ipsilateral side. Cells expressing IL-12p40 were, conversely, less numerousand scattered across ipsilateral side ofmice challenged with both cytokines ei-ther alone or combined with SNP andkilled 1 day afterward. Magnification,�3.25; scale bar, 1 mm.

7045The Journal of Immunology

by guest on April 14, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Data analysis

The relative intensity of mRNA signals throughout the brain of each animalwas assessed on X-ray film images and graded according to the scale ofundetectable (�), low (�), moderate (��), strong (���), or very strong(����). Dipped emulsion slides were examined under microscopic eval-uation to ascertain the subcellular localization of each transcript. Semi-quantitative analysis of the hybridization signal was conducted on X-rayfilms (Biomax; Kodak, Rochester, NY) over numerous brain sections ip-silaterally to injection site. Transmittance values (referred to in this studyas OD) and extent of positive hybridization signal were measured under aNorthern Light Desktop Illuminator (Imaging Research, St. Catherine’s,Ontario, Canada) using a Sony Camera Video System attached to aMicroNikkor 55-mm Vivitar extension tube set for a Nikon lens and cou-pled to a Macintosh computer (Power PC 7100/66) and Image software(version 1.61, non-FPU; a kind gift from W. Rasband, National Institutesof Health, Bethesda, MD). OD for each pixel was calculated using a knownstandard of intensity and distance measurements from a logarithmic specteradapted from BioImage Visage 110s (Millipore, Ann Arbor, MI). Sectionsfrom experimental and control animal were digitized and subjected to den-sitometric analysis, yielding measurements of integrated OD (area digi-tized � average OD). The OD for each section was corrected for the av-erage background signal, determined on the contralateral section. Data arereported as mean � SEM values for experimental and control mice. Thenumber of FJB-positive neurons was quantified on brain sections that weredigitized at �100 of magnification with a RT-SPOT camera (DiagnosticInstruments, Sterling Heights, MI) mounted directly onto a BX-60 Olym-pus microscope and connected to a PowerMac G4 computer. All the FJB-positive cells were counted manually and the results were reported as themean number of positive cells per area � SEM values. The digitized areaswere taken at the same rostrocaudal level ipsilateral to the injection site.Three different areas were delimited and quantified: surrounding the infu-sion site in the caudate putamen (area a), dorsal to the infusion site in thecortex (area b), and lateral to the infusion site in the cortical region (areac). Statistical analysis was performed by a two-way ANOVA (StatView4.5, Macintosh computer) followed by a Bonferroni-Dunn test procedure aspost hoc comparisons for each time postinjection (1 and 3 days). Quanti-

fication of the degenerated area was conducted on LFB-stained sectionswith an Olympus Optical System (BX-50, Bmax) coupled to a Macintoshcomputer (Power PC 7100/66) and Image software (version 1.61, non-FPU). The damaged area was digitized, and pixel values were measuredunder brightfield illumination at a magnification �3.125. The percentageof the damaged area was calculated from the entire coronal section of eachanimal. Data are reported as mean � SEM percentage values for experi-mental and control mice. Statistical analysis was performed by a two-wayANOVA (StatView 4.5, Macintosh computer) followed by a Bonferroni-Dunn test procedure as post hoc comparisons for each time postinjection.

ResultsEffects of IFN-� and TNF-� on expression of proinflammatorygenes

Transcriptional activation of the gene encoding TLR2 is a verysensitive index of microglial activation in response to a large va-riety of immune stimuli and insults (1, 17, 23–25). However, onlyfew cells were positive for TLR2 mRNA in the brain of mice thatreceived a single intrastriatal bolus of IFN-�, and the hybridizationsignal remained localized near the injection site (Fig. 1, top). Thiswas also the case for the mice that were infused with the vehiclesolution into the striatum; the message was found quite specificallyalong the cannula track. In contrast to such localized induction,transcriptional activation took place in the infused areas of animalschallenged with the proinflammatory cytokine TNF-�. The signaldiffused across the ipsilateral site 24 h after the single TNF-� in-fusion, but intensity and extent of hybridization signal were high-est in mice challenged with both cytokines simultaneously (Fig. 1).Indeed, numerous small-scattered cells exhibited an intense hy-bridization signal across the cerebral tissue of mice killed 24 hafter being infused with both cytokines (Fig. 2). Dual-labeling pro-

FIGURE 2. Cellular distribution of TLR2, MCP-1, and IL-12p40 mRNA in cerebral tissue of mice that received IFN-�, TNF-�, and SNP in the dorsalbasal ganglia. Animals were killed 1 day after the intraparenchymal infusion of a solution (1 �l) containing IFN-� (100 ng), TNF-� (100 ng), and SNP(1 �g). These darkfield and brightfield photomicrographs of dipped coronal sections (20 �m) into nuclear emulsion NTB2 depict positive signal withinsmall-scattered microglial cells (TLR2 and MCP-1 mRNAs) and few larger monocytic cells (IL-12p40 mRNA) across the cerebral tissue. The dual-labelingprocedure provided the anatomic evidence that although TLR2 and MCP-1 mRNAs were largely colocalized within parenchymal microglia (iba1�),IL-12p40 transcript seemed associated with infiltrating monocytes/macrophages. Few microglial cells were nevertheless positive for the mRNA encodingthe latter cytokine. TLR2, MCP-1, or IL-12p40 mRNA was hybridized on the same sections by means of a radioactive in situ hybridization technique usingsilver grains (right column). Arrowheads indicate dual-labeled cells (TLR2 mRNA/microglial cell (top row); MCP-1 mRNA/microglial cell (middle row);IL-12p40 mRNA/monocytic cell (bottom row).

7046 ROLE OF CYTOKINES IN NO-INDUCED NEURODEGENERATION

by guest on April 14, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

cedure provided the anatomical evidence that the gene encodingTLR2 was essentially expressed within microglial cells (see Fig. 2,which shows agglomerations of silver grains within iba1 immu-noreactive cells). The signal for TLR2 transcript decreased 3 daysfollowing the cerebral challenge with IFN-� and TNF-� and re-turned to background levels at time 7 days postinfusion (data notshown).

The mRNAs encoding numerous other genes were assessed inadjacent coronal sections. In this study again, saline-treated miceexhibited an increase in the expression levels of MCP-1 transcriptin cells adjacent to the cannula track at time 24 h postinfusion. Thesignal essentially vanished 72 h after the single bolus of saline inthe basal ganglia. A low but positive hybridization signal forMCP-1 was detected in the region ipsilateral to injection site ofmice challenged with either IFN-� or TNF-�, although the lattercaused a more pronounced increase in MCP-1 gene expression(Fig. 1). However, the signal intensity was clearly highest in micethat received a single bolus with a mixture of both cytokines andkilled 1 day afterward (Figs. 1 and 2). The pattern of MCP-1 geneexpression was very similar to that of TLR2 mRNA, but the cellsexhibited a lower number of silver grains (Fig. 2). Treatment withboth cytokines also triggered transcriptional activation of genesencoding I�B� (index of NF-�B activity) and TNF-�, even if thesignals for both I�B� and TNF transcripts were much lower thanthose of TLR2 and MCP-1 mRNAs (data not shown).

In contrast, a single intracerebral infusion of IFN-�, TNF-�, orboth cytokines together only provoked a very modest increase iniNOS gene expression within few isolated cells that were adjacent

to the endothelium of the brain capillaries (Fig. 3). This was alsothe case for the gene encoding IL-12p40 that was induced only infew small-scattered cells 24 h after the intrastriatal infusion of bothIFN-� and TNF-�. The pattern of TLR2- and MCP-1-expressingcells was highly different from that of IL-12p40-expressing cells(Figs. 1 and 2). This suggests a different type of cells, although alltranscripts were found within iba1-immunoreactive cells (Fig. 2).It is nevertheless possible that the technique used in this study isnot sensitive enough to detect IL-12p40 transcripts in all activatedmicroglial cells or this cytokine is induced only in a selective sub-set of monocytic cells. As depicted by Fig. 3, the innate immuneresponse triggered by the intraparenchymal cytokine infusion es-sentially disappeared at time 3 days posttreatment.

Role of NO in mediating the inflammatory reaction in the CNS

Inhibition of NO pathway enhanced the effects of IFN-� andTNF-� on the innate immune response in the CNS (Fig. 3). Indeed,the nonselective inhibitor of NO synthase L-NAME caused a pro-found transcriptional activation of genes encoding TLR2, MCP-1,I�B�, TNF-�, iNOS, and IL-12p40 24 h after the intracerebralinfusion of IFN-� and TNF-�. The hybridization signal for TNF-�,iNOS, and IL-12p40 transcripts was significantly stronger in thebrain of mice that received a systemic injection of L-NAME intothe intrastriatal area along with the mixture of both cytokines. Allthese transcripts returned to low or undetectable levels at time 3days postinjection, and administration of L-NAME alone remainedwithout notable effect on expression of the genes investigated inthe present study.

FIGURE 3. Semiquantitative analysisof hybridization signals across the ipsilat-eral side of infusion with cytokines, N(G)-nitro-L-arginine methylester (L-NAME),and the NO donor SNP. Measurement ofthe average OD of the different transcriptswas performed at the indicated postinjec-tion day (d) as described in Materials andMethods. The values are means � SEM oftwo to seven mice per group and statisticalanalysis was performed using a two-wayANOVA followed by a Bonferroni-Dunnpost hoc test (StatView 4.5). �, Signifi-cantly different (p � 0.05) from their ap-propriate control groups (saline, SNP, orL-NAME) within the same postinjectiontime. †, Significantly different (p � 0.05)from their corresponding groups of saline-injected mice (saline, IFN-�, TNF-�, orIFN-��TNF-�) within the same postin-jection time. ‡, Significantly different(p � 0.05) from both groups of IFN-� andTNF-�-treated mice and their correspond-ing groups of mice that received either sa-line or SNP alone. IL-12p40, IL-12 p40subunit.

7047The Journal of Immunology

by guest on April 14, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

To further assess the role of NO as mediator of the inflammatoryresponse in the CNS, another group of mice was infused with theNO donor SNP either alone or in combination with IFN-� and/orTNF-�. A low and localized hybridization signal was found for themRNA encoding TLR2 and MCP-1 in the brain of mice chal-lenged with SNP into the basal ganglia and killed 24 h afterward(Figs. 1 and 2). The message for TLR2, MCP-1, and I�B� tran-scripts increased 2 days later and remained detectable up to 21days posttreatment with SNP. The hybridization signal was mod-erate to low and remained mostly in the caudate putamen of ipsi-lateral side, although few positive cells were also found in thehippocampus and cortex. Administration of SNP along with IFN-�or TNF-� increased gene expression as soon as 24 h postadmin-istration, but the signal intensity and distribution patterns weresimilar to those already described for mice challenged only withboth cytokines. The marked differences in gene expression clearlytook place at 3 and 7 days; whereas infusion of both cytokinescaused a transient innate immune response, the mixture of SNP,IFN-�, and/or TNF-� led to a prolonged increase in TLR2,MCP-1, I�B�, TNF-�, iNOS, and IL-12p40 gene expression inregions ipsilateral to the injection site (Fig. 3). The gene encodingTLR2 remained quite strong 21 days following injection of SNPplus both cytokines, and I�B� and TNF-� transcripts were stilldetectable at this time after all SNP treatments (Fig. 3). Expressionof all the genes reported in this study was restricted to damagedareas near the injection site. This localized inflammatory responseseems therefore associated with neuronal cell death and cellular

infiltration (see below). It is of interest to note that both NO in-hibitor and donor exacerbated the immune reaction to IFN-� andTNF-�, but only SNP provoked neurodegeneration. The hybrid-ization signal was not due to unspecific labeling because damagedtissues never depicted positive messages when hybridized withsense probes (data not shown). Both cytokines also increased ex-pression of proinflammatory genes in response to a lower dose ofSNP (500 ng, data not shown).

Consequences of inflammation and NO on the neuronal integrity

LFB staining was used for the histologic analysis of myelinatedfibers (blue) and neurons (purple) in the CNS. This staining prep-aration also allowed identification of infiltrating cells, such as neu-trophils and monocytes/macrophages. Except for the lesion causedby implantation of the infusion cannula, the brain of saline-injectedmice never exhibited signs of neurodegeneration at any of thetimes evaluated in this study. Fibers of the corpus callosum andmyelinated fibers in the caudate putamen in proximity of injectionsite were unchanged compared with contralateral side, and the in-jection did not affect the shape and number of neurons (Fig. 4).Clusters of blood cells were detected along the cannula track in themeninges, cortex, and caudate putamen, but not within the brainparenchyma. Similarly, a single intrastriatal bolus of IFN-�,TNF-�, or both cytokines together failed to cause neurodegenera-tion and demyelination at all the times evaluated. This contrastswith the brain of animals that received the NO donor SNP, which

FIGURE 4. Effect of cytokines and NO donor SNP on the neuronal integrity in the mouse brain. LFB staining was used to label the myelinated fibers,neurons, and immigrating cells in the mouse brain. Examples of LFB-stained coronal sections of mice (top row) that were killed 3 days after a singleintrastriatal infusion with solutions containing saline, IFN-�, TNF-�, and SNP. The degenerated areas are depicted by the dashed lines. Confirmation ofneuronal death was performed via the FJB method, whereas cell death by apoptosis was detected by means of a TdT-FragEL DNA Fragmentation Detectionkit (bottom row). Photomicrographs (second row) result from the merge of six different pictures taken on the same brain section to cover the entire ipsilateralside at the level of the injection site. The images were merged using Canon PhotoStitch 3.1. Magnification: LFB, �3.25; FJB, �10 (second row) and �100(third row); and FragEL, �250 (bottom row). Scale bar: FJB, 400 �m (second row) and 25 �m (third row); and FragEL, 10 �m (bottom row). Depictionsof FragEL cells are laser scanning confocal microscopic images (Fluoview SV500; Olympus America, Melville, NY).

7048 ROLE OF CYTOKINES IN NO-INDUCED NEURODEGENERATION

by guest on April 14, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

causes neuronal death, demyelination, and infiltration of circulat-ing leukocytes in the infused areas (Fig. 4 and Fig. 5B). The ce-rebral tissue in the cortex and caudate putamen was partially de-structed and filled with cellular debris and damaged neurons,which have a different shape than healthy neurons (Fig. 5, I and J).

The degeneration process spread out to more caudal regions, in-cluding the hippocampus and striatum.

IFN-� did not modify notably the effects of SNP, especiallyat time 3 days. (Figs. 4 and 5). In contrast, the ability of NOdonor to cause neurotoxicity was clearly exacerbated by the

Table III. Number of FJB neurons in the brain of cytokine-infused and NO donor-infused micea

Treatmentc Days

Areab

a b c

IFN-� � TNF-� 1 1.00 � 0.66 31.86 � 3.00 0.00 � 0.00SNP 1 23.50† � 5.50 25.75 � 3.47 0.75 � 0.75SNP � IFN-� 1 30.50† � 9.73 79.00†* � 12.32 18.00 � 10.84SNP � TNF-� 1 70.00†* � 4.16 102.33†* � 14.25 60.00†* � 21.70SNP � IFN-� � TNF-� 1 45.00† � 16.09 61.33†* � 18.35 32.00† � 16.65IFN-� � TNF-� 3 0.00 � 0.00 5.00 � 5.00 0.00 � 0.00SNP 3 32.80† � 6.84 32.00† � 7.23 7.00 � 3.82SNP � IFN-� 3 42.50† � 11.85 48.00†* � 8.99 6.25 � 4.05SNP � TNF-� 3 50.33† � 6.12 59.33†* � 3.18 50.33†* � 9.62SNP � IFN-� � TNF-� 3 84.00†* � 9.98 85.60†* � 9.50 49.20†* � 16.74

a All the FJB-positive cells were counted manually and the results were reported as the mean number of positive cells per area � SEM values. Thedigltized areas were taken at the same rostrocaudal level ipsilateral to the injection site.

b Three different areas were delimited and quantified; surrounding the infusion site in the caudate putamen (area a), dorsal to the infusion site in thecortex (area b), and lateral to the infusion site in the cortical region (area c). *, Significantly different ( p � 0.05) from the group of SNP mice. †,Significantly different ( p � 0.05) from the IFN-� � TNF-� group of mice. See Materials and Methods for details.

c IFN-� � TNF-�, intrastriatal (IS) infusion of IFN-� (100 ng) and TNF-� (100 ng); SNP, IS infusion of SNP (1 �g); SNP�IFN-�, IS infusion ofSNP (1 �g) and IFN-� (100 ng); SNP�TNF-�, IS infusion of SNP (1 �g) and TNF-� (100 ng); SNP�IFN-��TNF-�, IS infusion of SNP (1 �g), IFN-�(100 ng) and TNF-� (100 ng).

FIGURE 5. Role of IFN-� and TNF-� on the cerebral damage caused by NO donor SNP. LFB staining was used to label myelinated fibers, neurons,and immigrating cells in the mouse brain. These photomicrographs depict representative examples of LFB-stained sections from mice that were perfusedtranscardially with paraformaldehyde 3 days (d) after being infused with a solution containing IFN-��TNF-� (A), SNP (B), SNP�IFN-� (C), SNP�TNF-�(D), or SNP�IFN-��TNF-� (E). F–H, Hippocampal region of mice treated with IFN-��TNF-� (3 days), SNP�IFN-��TNF-� (3 days) or SNP�TNF-�(7 days), respectively. The shape of neurons in the contralateral (I) and ipsilateral (J) cerebral cortex of mice at time 3 days and infiltration of neutrophilsat 3 days (K), 7 days (L), and 15 days (M) following the cerebral administration of SNP�TNF-�. Note the damage 7 days (N), 15 days (O), and 21 days(P) after the intraparenchymal injection of SNP�IFN-��TNF-�. Q, The average percentage of the degenerated area as detailed in Materials and Methods.�, Significantly different (p � 0.05) from the group of SNP mice. †, Significantly different (p � 0.05) from the IFN-��TNF-� group of mice. Black arrowsindicate polymorphonuclear cells. Magnification: A–H, �25; I–M, �250; and N–P, �10. Scale bar: A–H, 100 �m; I–M, 10 �m; and N–P, 250 �m.

7049The Journal of Immunology

by guest on April 14, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

proinflammatory cytokine TNF-�. In these brains, myelinated fi-bers of the corpus callosum and neurons of the cerebral cortex,hippocampus, and dorsal striatum were injured by the solutioncontaining SNP and TNF-�. However, neurodegeneration tookplace essentially within the ipsilateral side and the contralateralside remained intact, at least as revealed by this staining procedure(Figs. 4 and 5D). Several clusters of infiltrating neutrophils weredetected in the injured areas. These cells were associated withblood vessels at time 3 days postinjection and progressively im-migrated across the brain parenchyma (Fig. 5, K–M). The neuro-toxic effects of the solution containing SNP and TNF-� were notexacerbated, but slightly attenuated by IFN-�. Indeed, the extent ofthe degenerated structures was more restricted in the brains ofSNP/IFN-�/TNF-�-administered mice than in those of mice chal-lenged only with SNP and TNF (Figs. 4 and 5). Although L-NAMEenhanced expression of several inflammatory genes in response tointrastriatal IFN-� and TNF-� infusion, these brains did not showany anatomical signs of neuronal cell death (data not shown).

As for the LFB histologic preparations, the brain of mice thatreceived intraparenchymal administration of saline solution,IFN-�, TNF-�, or both cytokines exhibited only background FJBstaining, except along the cannula track (Fig. 4 and Table III). Thislocalized staining decreased at 3 days and essentially vanished 7days after the infusion. One day after the single bolus of SNP in thebasal ganglia, few positive FJB neurons were found along the can-nula track and surrounding areas. The signal greatly increased andspread across the cortex and dorsal striatum 3 days after the infu-sion with the NO donor (Fig. 4). Dying neurons were also foundmore rostrally in the lateral septum and more caudally in the hip-pocampus, dentate gyrus, and striatum. Once again, TNF-� eitheralone or combined with IFN-� greatly exacerbated the neurode-generative processes caused by SNP (Fig. 4 and Table III). Posi-tive FJB neurons were detected almost across the entire ipsilateralside of the brain as well as in the lateral septum and hippocampusof the contralateral side. This was particularly the case in the CNS

of mice that were killed 3 days after being injected with SNP andTNF-�. FJB-positive cells were quantified in different areas ipsi-lateral to the infusion site (Table III). The number of degeneratingneurons was highest in all the areas of mice that were killed 1 dayafter the intracerebral challenge with both SNP and TNF-�. Attime 3 days however, it is the mixture of SNP combined with bothcytokines that led to a more severe neuronal death in all the areasquantified (Table III). The FJB signal was no longer detectable inthe brain of all mice at times 7–21 days postinjection. Adminis-tration of L-NAME either alone or together with cytokines did notprovoke neuronal cell death as revealed by both LFB and FJBmethods (data not shown).

To determine the potential contribution of apoptosis in mediat-ing the effects of different treatments on cell death, we used aTdT-FragEL DNA fragmentation detection kit. As for the otherhistologic procedures, this method failed to detect apoptotic cellsin the brain of mice injected with L-NAME alone or together witheither IFN-� or TNF-� or both cytokines at all postinjection times.SNP alone did not yet induce apoptosis at time 24 h, but apoptoticcells were already numerous in the cortex just above the corpuscallosum of mice challenged with the solution containing SNP andcytokines. As seen in Fig. 4 (bottom row), the number of positivenuclei greatly increased 3 and 7 days after the infusion of SNPalone, but the NO-releasing substance caused more cell death viaapoptosis when combined with TNF-�. It is interesting to note thatalthough LFB and FJB revealed a massive neurodegeneration at 3days, the number and spread of FragEL-positive cells peaked 7days after the different treatments using SNP. The brains of theseanimals exhibited a similar pattern of cleaved caspase-3 immuno-reactive cells across the ipsilateral side (Fig. 6A). These cells werealso immunoreactive to the neuronal marker NeuN (Fig. 6B). Al-though these data show that neuronal death by apoptosis clearlytook place in these brains, cellular necrosis may also contribute tothe neuronal and glial cell loss in mice treated with SNP alone ortogether with TNF-� and/or IFN-�. Astrocytes and microglia did

FIGURE 6. Laser scanning confocal microscopicimage of cleaved caspase-3 immunoreactive (ir) cells inthe brain of mice that were challenged with NO donorSNP either alone or combined with cytokines. Cleavedcaspase-3 (green) was labeled by immunohistochemis-try using a mAb, whereas Cy2 (A) or Alexa Fluor 488(B) secondary Ab was used to bind the primary Ab di-rected against the protein. Nuclei (red) were labeled viapropidium iodide (A). Neurons (red) were labeled byimmunohistochemistry using a NeuN mAb that was re-vealed with a Cy3 secondary Ab (B). Laser scanningconfocal microscopy studies were performed with a�100 Plan-Apo oil immersion objective, numericalapeture 1.35, �2 numerical zoom with a BX-61 micro-scope (Olympus America, Melville, NY). Emissionswere recorded by photomultipliers preset respectivelyfor EGFP (A) or FITC (B) (green pseudo color) and Cy3(red pseudo color) fluorescent dyes in Fluoview SV500imaging software. Thirteen 0.5 �m confocal z-serieswere acquired for each area and corrected by 2 Kahl-man low-speed scans. Acquired z-series images werethen flattened into one image and exported in 24 bitsTIFF format. Please note the strong signal over numer-ous neurons at time 3 days postinjection with SNP aloneor combined with cytokines, but cytokines alone failedto trigger cleaved caspase-3 in infused regions. B, Dou-ble immunofluorescence in SNP�IFN-��TNF-�-in-fused brains at 24 h due to the stronger NeuN signal atthis time; NeuN rapidly vanishes in degenerating neu-rons. Scale bar, 20 �m.

7050 ROLE OF CYTOKINES IN NO-INDUCED NEURODEGENERATION

by guest on April 14, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

not exhibit positive cleaved caspase-3 immunoreactive signal (datanot shown).

DiscussionThe present study provides solid evidence that parenchymal ad-ministration of TNF-� alone or combined with IFN-� provoked arobust inflammatory reaction in the mouse brain. These proinflam-matory signaling events were transient and did not lead to demy-elination or neurodegeneration. This suggests that these immunemolecules are not by themselves capable of causing cell death inthe CNS, at least when administered acutely as performed in thepresent study. However, these cytokines contributed greatly to ex-acerbate the cerebral damage caused by the NO donor SNP. Thetreatment combining TNF-� and SNP was particularly harmful tocerebral elements. In other words, TNF-� clearly accelerated andaccentuated the devastating effects of NO donor in the mouseCNS. The brains of these animals exhibited demyelination (LFB),neurodegeneration (positive FJB neurons), and cellular apoptosis(positive TUNEL and cleaved caspase-3 neurons) across the ipsi-lateral side and sometime in regions of the contralateral side. Thiscerebral damage was also associated with a robust increase in ex-pression of genes encoding proinflammatory molecules in a patternsimilar to that of degenerating groups of cells. Some transcriptswere still expressed 21 days after the treatment combining SNPand cytokines, which suggest that SNP-induced neurotoxicityleads to a subsequent inflammatory response and microglial reac-tivity. The latter is not directly responsible for the neurodegenera-tion, but it contributed to expand the damage in presence of ex-ogenous TNF-�. In this regard, inhibition of NO pathway byL-NAME increased transcriptional activation of TNF-�, iNOS, andIL-12p40 in IFN-� plus TNF-�-injected mice, but these brains didnot exhibit any signs of neurodegeneration. Therefore, inhibitingand stimulating NO enhanced the ability of both cytokines to trig-ger the inflammatory response in the brain, although neurotoxicitytook place only in the groups of SNP-treated mice. In this case,inflammation is believed to be a consequence and not a directcause of cell death, but once engaged this process becomes moresusceptible to TNF-�.

Inflammatory molecules have been studied in the context of sev-eral neurodegenerative diseases. Lee et al. (26) have demonstratedthat neutralizing TNF-� Ab was able to reduce the number ofdying neurons and oligodendrocytes in response to spinal cordinjury. They have also observed a better cell survival in animalstreated with a NO synthesis inhibitor. Besides, constitutive pro-duction of soluble TNF-� type I receptor in transgenic mice sig-nificantly inhibited the neurodegenerative events following mo-toneuron axotomy (27). Moreover, this cytokine plays a criticalrole in the profound neurodegeneration taking place in animalstreated with the glucocorticoid receptor inhibitor RU486 before theintrastriatal bolus of LPS (28). Inhibition of TNF-� is indeed ableto totally abolish the neurotoxic effects of the endotoxin and TNF-�-induced neuronal damage is dependent on both NO and caspasepathways (28).

Taken together these data weigh in favor of TNF-� in playing akey role in neurodegenerative events. Despite this finding, con-flicting results exist in regard to the role of TNF-� in either exac-erbating brain damages or ameliorating brain recovery during dif-ferent pathologic situations. In a very elegant study, Arnett et al.(16) have used mice lacking TNF-� and its associated receptors tostudy a model of demyelination and remyelination. Surprisingly,the lack of TNF-� led to a significant delay in remyelination,which was associated with a reduction in the pool of proliferatingoligodendrocyte progenitors followed by a reduction in the numberof mature oligodendrocytes. This reparative role for TNF-� in the

CNS is mediated via the TNFR2, not TNFR1 (16), indicating thatthe dual role of TNF-� in demyelination and remyelination maydepend on the receptor type and/or the cellular source of the cy-tokine. It is also important to mention that a single bolus of TNF-�either alone or combined with IFN-� did not provoke neurodegen-eration despite the profound inflammatory response.

The binding of TNF-� to its cognate receptors leads to the for-mation of the TNFR1-associated death domain/TNFR1-associatedfactor (TRAF)2 complex, which activates the NF-�B signalingevents. TNF-� is actually one of the most potent effectors ofNF-�B activity through the 55 kDa TNF type I receptor in mosttypes of cells in the systemic immune system as well as in the CNS(29). FADD/MORT1, TRAF2, and the death domain kinase re-ceptor-interacting protein are recruited and may also interact di-rectly with TNFR1-associated death domain (30). Whereas FADD/MORT1 is essential for TNF-induced apoptosis, receptor-interacting proteins and TRAF2 seem to be the key molecules foractivating both NF-�B and mitogen-activated protein kinases (31).The latter is a major survival pathway (32); it leads to the synthesisof several antiapoptotic proteins, and inhibiting NF-�B provokescell death in presence of TNF-� (33–35). However, the increase inthe activity of this pathway by TNF-� clearly failed to prevent theneurotoxic effects of SNP, it actually further increased demyeli-nation and neuronal cell death. These data, together with the robustimmunoreactive signal for the cleaved caspase-3 and positiveTUNEL nuclei, support the concept that apoptotic signaling eventscircumvented the protecting activity of NF-�B in animals chal-lenged with SNP in the CNS.

IFN-� was also found to play a critical role in mediating celldeath, because transgenic mice expressing IFN-� in the hippocam-pus showed increased susceptibility to neuronal damage (4) andIFN regulatory factor-1 (a transcription factor activated by IFN-�)knockout mice are more resistant to experimental autoimmune en-cephalomyelitis (36). On the contrary, Furlan et al. (37) have sug-gested a protective role of IFN-� in this autoimmune demyelinat-ing disease commonly used to model the pathogenetic mechanismsinvolved in multiple sclerosis. They showed that administration ofa vector expressing IFN-� in the fluid of the spinal chord attenu-ated the symptoms. In this particular case, IFN-� was responsibleof the induction of TNFR inhibitor (p55) on infiltrating T cells,helping TNF-� to induce apoptosis of these cells. In the same line,the single injection of IFN-� either alone or combined with TNF-�failed to cause demyelination and neuronal loss, but the mixture ofboth cytokines causes a strong and transient transcriptional acti-vation of proinflammatory genes. This response is therefore notdetrimental for the cerebral elements, although it may depend onthe dose and duration of the treatment. Indeed, we have recentlyfound that chronic TNF-� infusion in the basal ganglia led to aprofound neurodegeneration via both NO and caspase pathways(28). We have yet to determine whether chronic IFN-� infusionmay also be associated with such damage in the brain or the cy-tokine has to be combined with TNF-� to induce cell death inthe CNS.

The exact mechanisms involved the ability of TNF-� and IFN-�to enhance the neurotoxic effects of SNP remain unclear at thispoint. NO is able to reduce ATP generation and induce apoptosisof PC12 cells (38). Reduction of mitochondrial respiration andneurotoxicity also take place in response to antifungal drug am-photericin B, especially when combined with IFN-� (39). Theseauthors have shown that amphotericin B causes a dose-dependentincrease of NO generation in IFN-�-stimulated rat and mouse as-trocytes, as well as in IFN-� plus TNF-�-activated rat astrocytomacell line C6. These data indicate that both cytokines may furtherincrease NO production and seriously compromise the oxidative

7051The Journal of Immunology

by guest on April 14, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

phosphorylation and mitochondrial respiratory chain in neuronsand other cells of the brain. It remains however important to keepin mind that the NO concentration reached with the dose of SNPused in these experiments may not represent the endogenous levelsnormally achieved during pathologic conditions. It is thereforepossible that IFN-� and TNF-� have the ability to amplify cerebraldamages in other models of neurotoxicity.

In conclusion, intrastriatal infusion of TNF-� provoked a robustand transient innate immune response in the brain, especially in thepresence of IFN-�. Surprisingly, such an inflammatory responseand microglial reactivity was not associated with neurodegenera-tion and demyelination. These molecules (in particular TNF-�)were however able to enhance the neurotoxic effects of SNP thatwere associated with a sustained transcriptional activation ofTLR2 gene up to 21 days postinjection. Cooperation betweenTNF-� and NO may therefore play a critical role in neurologicaldisorders that have an immune etiology.

AcknowledgmentsWe thank Dr. Alain Israel (Institut Pasteur, Paris, France) for the gift of theplasmid containing the mouse I�B� cDNA, Dr. S. C. Williams (TexasTech University, Lubbock, TX) for the cDNA containing MCP-1, Dr. D.Radzioch (McGill University, Montreal, Quebec, Canada) for the cDNAcontaining the mouse TNF-�, Dr. M. Olivier (McGill University, Montreal,Quebec, Canada) for the mouse iNOS cDNA, Dr. K. Pahan (University ofNebraska Medical Center, Lincoln, NE) for the mouse IL-12p40 cDNA,and Dr. Y. Imai (National Institute of Neuroscience, Kodaira, Japan) forthe anti-ionized calcium-binding adapter molecule 1 (iba1) Ab.

References1. Nguyen, M. D., J. P. Julien, and S. Rivest. 2002. Innate immunity: the missing

link in neuroprotection and neurodegeneration? Nat. Rev. Neurosci. 3:216.2. Moser, M., and K. M. Murphy. 2000. Dendritic cell regulation of TH1-TH2

development. Nat Immunol. 1:199.3. Hermans, G., P. Stinissen, L. Hauben, E. Van den Berg-Loonen, J. Raus, and

J. Zhang. 1997. Cytokine profile of myelin basic protein-reactive T cells in mul-tiple sclerosis and healthy individuals. Ann. Neurol. 42:18.

4. Jensen, M. B., I. V. Hegelund, N. D. Lomholt, B. Finsen, and T. Owens. 2000.IFN� enhances microglial reactions to hippocampal axonal degeneration. J. Neu-rosci. 20:3612.

5. Pouly, S., B. Becher, M. Blain, and J. P. Antel. 2000. Interferon-� modulateshuman oligodendrocyte susceptibility to Fas-mediated apoptosis. J. Neuropathol.Exp. Neurol. 59:280.

6. Bogdan, C. 2001. Nitric oxide and the immune response. Nat. Immunol. 2:907.7. Griffith, O. W., and D. J. Stuehr. 1995. Nitric oxide synthases: properties and

catalytic mechanism. Annu. Rev. Physiol. 57:707.8. MacMicking, J., Q. W. Xie, and C. Nathan. 1997. Nitric oxide and macrophage

function. Annu. Rev. Immunol. 15:323.9. Gross, S. S., and M. S. Wolin. 1995. Nitric oxide: pathophysiological mecha-

nisms. Annu. Rev. Physiol. 57:737.10. Beltran, B., M. Quintero, E. Garcia-Zaragoza, E. O’Connor, J. V. Esplugues, and

S. Moncada. 2002. Inhibition of mitochondrial respiration by endogenous nitricoxide: a critical step in Fas signaling. Proc. Natl. Acad. Sci. USA 99:8892.

11. Gordon, S. A., W. Abou-Jaoude, R. A. Hoffman, S. A. McCarthy, Y. M. Kim,X. Zhou, X. R. Zhang, R. L. Simmons, Y. Chen, L. Schall, and H. R. Ford. 2001.Nitric oxide induces murine thymocyte apoptosis by oxidative injury and a p53-dependent mechanism. J. Leukocyte Biol. 70:87.

12. Gold, D. P., K. Schroder, H. C. Powell, and C. J. Kelly. 1997. Nitric oxide andthe immunomodulation of experimental allergic encephalomyelitis. Eur. J. Im-munol. 27:2863.

13. Fenyk-Melody, J. E., A. E. Garrison, S. R. Brunnert, J. R. Weidner, F. Shen,B. A. Shelton, and J. S. Mudgett. 1998. Experimental autoimmune encephalo-myelitis is exacerbated in mice lacking the NOS2 gene. J. Immunol. 160:2940.

14. Sahrbacher, U. C., F. Lechner, H. P. Eugster, K. Frei, H. Lassmann, andA. Fontana. 1998. Mice with an inactivation of the inducible nitric oxide synthasegene are susceptible to experimental autoimmune encephalomyelitis. Eur. J. Im-munol. 28:1332.

15. Iravani, M. M., K. Kashefi, P. Mander, S. Rose, and P. Jenner. 2002. Involvementof inducible nitric oxide synthase in inflammation-induced dopaminergic neuro-degeneration. Neuroscience 110:49.

16. Arnett, H. A., J. Mason, M. Marino, K. Suzuki, G. K. Matsushima, and J. P. Ting.2001. TNF� promotes proliferation of oligodendrocyte progenitors and remyeli-nation. Nat. Neurosci. 4:1116.

17. Laflamme, N., G. Soucy, and S. Rivest. 2001. Circulating cell wall componentsderived from gram-negative and not gram-positive bacteria cause a profoundtranscriptional activation of the gene encoding Toll-like receptor 2 in the CNS.J. Neurochem. 70:648.

18. Laflamme, N., and S. Rivest. 2001. Toll-like receptor 4: the missing link of thecerebral innate immune response triggered by circulating Gram-negative bacterialcell wall components. FASEB J. 15:155.

19. Laflamme, N., S. Lacroix, and S. Rivest. 1999. An essential role of interleukin-1�in mediating NF-�B activity and COX-2 transcription in cells of the blood-brainbarrier in response to systemic and localized inflammation, but not during endo-toxemia. J. Neurosci. 19:10923.

20. Nadeau, S., and S. Rivest. 2000. Role of microglial-derived tumor necrosis factorin mediating CD14 transcription and NF-�B activity in the brain during endo-toxemia. J. Neurosci. 20:3456.

21. Nadeau, S., and S. Rivest. 2002. Endotoxemia prevents the cerebral inflammatorywave induced by intraparenchymal lipopolysaccharide injection: role of glu-cocorticoids and CD14. J. Immunol. 169:3370.

22. Soulet, D., and S. Rivest. 2003. Polyamines play a critical role in the control ofthe innate immune response in the mouse CNS. J. Cell Biol. 162:257.

23. Boivin, G., Z. Coulombe, and S. Rivest. 2002. Intranasal herpes simplex virustype 2 inoculation causes a profound thymidine kinase dependent cerebral in-flammatory response in the mouse hindbrain. Eur J. Neurosci. 16:29.

24. Zekki, H., D. L. Feinstein, and S. Rivest. 2002. The clinical course of experi-mental autoimmune encephalomyelitis is associated with a profound and sus-tained transcriptional activation of the genes encoding Toll-like receptor 2 andCD14 in the mouse CNS. Brain Pathol. 12:308.

25. Laflamme, N., H. Echchannaoui, R. Landmann, and S. Rivest. 2003. Cooperationbetween Toll-like receptor 2 and 4 in the brain of mice challenged with cell wallcomponents derived from Gram-negative and Gram-positive bacteria. Eur. J. Im-munol. 33:1127.

26. Lee, Y. B., T. Y. Yune, S. Y. Baik, Y. H. Shin, S. Du, H. Rhim, E. B. Lee,Y. C. Kim, M. L. Shin, G. J. Markelonis, and T. H. Oh. 2000. Role of tumornecrosis factor-� in neuronal and glial apoptosis after spinal cord injury. Exp.Neurol. 166:190.

27. Terrado, J., D. Monnier, D. Perrelet, D. Vesin, S. Jemelin, W. A. Buurman,L. Mattenberger, B. King, A. C. Kato, and I. Garcia. 2000. Soluble TNF receptorspartially protect injured motoneurons in the postnatal CNS. Eur. J. Neurosci.12:3443.

28. Nadeau, S., and S. Rivest. 2003. Glucocorticoids play a fundamental role inprotecting the brain during innate immune response. J. Neurosci. 23:5536.

29. Laflamme, N., and S. Rivest. 1999. Effects of systemic immunogenic insults andcirculating proinflammatory cytokines on the transcription of the inhibitory factor�B� within specific cellular populations of the rat brain. J. Neurochem. 73:309.

30. Siegel, R. M., and M. J. Lenardo. 2001. To B or not to B: TNF family signalingin lymphocytes. Nat. Immunol. 2:577.

31. Chan, F. K., R. M. Siegel, and M. J. Lenardo. 2000. Signaling by the TNFreceptor superfamily and T cell homeostasis. Immunity 13:419.

32. Karin, M., and A. Lin. 2002. NF-�B at the crossroads of life and death. Nat.Immunol. 3:221.

33. Beg, A. A., and D. Baltimore. 1996. An essential role for NF-�B in preventingTNF-�-induced cell death. Science 274:782.

34. Van Antwerp, D. J., S. J. Martin, T. Kafri, D. R. Green, and I. M. Verma. 1996.Suppression of TNF-�-induced apoptosis by NF-�B. Science 274:787.

35. Wang, C. Y., M. W. Mayo, and A. S. Baldwin, Jr. 1996. TNF- and cancertherapy-induced apoptosis: potentiation by inhibition of NF-�B. Science 274:784.

36. Tada, Y., A. Ho, T. Matsuyama, and T. W. Mak. 1997. Reduced incidence andseverity of antigen-induced autoimmune diseases in mice lacking interferon reg-ulatory factor-1. J. Exp. Med. 185:231.

37. Furlan, R., E. Brambilla, F. Ruffini, P. L. Poliani, A. Bergami, P. C. Marconi,D. M. Franciotta, G. Penna, G. Comi, L. Adorini, and G. Martino. 2001. Intra-thecal delivery of IFN-� protects C57BL/6 mice from chronic-progressive ex-perimental autoimmune encephalomyelitis by increasing apoptosis of central ner-vous system-infiltrating lymphocytes. J. Immunol. 167:1821.

38. Yasuda, M., H. Fujimori, and H. Panhou. 1998. NO depletes cellular ATP con-tents via inactivation of glyceraldehyde-3-phosphate dehydrogenase in PC12cells. J. Toxicol. Sci. 23:389.

39. Trajkovic, V., M. Markovic, T. Samardzic, D. J. Miljkovic, D. Popadic, andM. M. Stojkovic. 2001. Amphotericin B potentiates the activation of induciblenitric oxide synthase and causes nitric oxide-dependent mitochondrial dysfunc-tion in cytokine-treated rodent astrocytes. Glia 35:180.

7052 ROLE OF CYTOKINES IN NO-INDUCED NEURODEGENERATION

by guest on April 14, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from