Production of Ciliary Neurotrophic Factor Is Required for the Β Trauma

Home , Astrogliosis, Growth factor, Neurotrophic factors, Tumor necrosis factor

Central Nervous System-Initiated Inflammation and Neurotrophism in Trauma: IL-1β Is Required for the Production of Ciliary Neurotrophic Factor This information is current as of September 25, 2021. Leonie M. Herx, Serge Rivest and V. Wee Yong J Immunol 2000; 165:2232-2239; ; doi: 10.4049/jimmunol.165.4.2232 http://www.jimmunol.org/content/165/4/2232 Downloaded from

References This article cites 51 articles, 8 of which you can access for free at: http://www.jimmunol.org/content/165/4/2232.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 25, 2021 *average

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

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2000 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Central Nervous System-Initiated Inﬂammation and Neurotrophism in Trauma: IL-1␤ Is Required for the Production of Ciliary Neurotrophic Factor1

Leonie M. Herx,* Serge Rivest,‡ and V. Wee Yong2*†

Injury to the CNS results in the production and accumulation of inflammatory cytokines within this tissue. The origin and role of inflammation within the CNS remains controversial. In this paper we demonstrate that an acute trauma to the mouse brain results in the rapid elevation of IL-1␤. This increase is detectable by 15 min after injury and significantly precedes the influx of leukocytes that occurs hours after. To confirm that IL-1␤ up-regulation is initiated by cells within the CNS, in situ hybridization for cytokine transcript was combined with cell type immunohistochemistry. The results reveal parenchymal microglia to be the sole source of IL-1␤ at 3 h postinjury. A role for CNS-initiated inflammation was addressed by examining the expression of the Downloaded from neurotrophic factor, ciliary neurotrophic factor (CNTF). Analysis of their temporal relationship suggests the up-regulation of CNTF by IL-1␤, which was confirmed through three lines of evidence. First, the application of IL-1 receptor antagonist into the lesion site attenuated the up-regulation of CNTF. Second, the examination of corticectomized animals genetically deficient for IL-1␤ found no CNTF up-regulation. Third, the lack of CNTF elevation in IL-1␤ null mice was rescued through exogenous application of IL-1␤ into the lesion site. These findings provide the first evidence of the requirement for IL-1␤ in the production

of CNTF following CNS trauma, and suggest that inﬂammation can have a beneﬁcial impact on the regenerative capacity of the http://www.jimmunol.org/ CNS. The Journal of Immunology, 2000, 165: 2232–2239.

he reaction of the CNS to trauma is a predeterminant of ies by Nieto-Sampedro et al. (10) showed that extracts collected the ability of the CNS to recover. From the investigations from sites of brain injury promoted the survival of sympathetic, T of numerous groups, it has become evident that an in- parasympathetic, and sensory neurons in culture. More recent crease in the levels of proinflammatory cytokines is a normal and work has identified these factors to include nerve growth factor early feature of the CNS response to trauma (1–3). In particular, (NGF)3 (11), ciliary neurotrophic factor (CNTF) (12, 13), basic we have shown that IL-1 and TNF-␣ become significantly elevated fibroblast growth factor (13), and insulin-like growth factor-1 (14). within the CNS parenchyma by 3 h after a brain stab wound injury, If this transient release of trophic factors could be prolonged, the by guest on September 25, 2021 and are localized to the lesion site (4). ability of the CNS to recover from injury may be improved. The source of inflammatory cytokines in the CNS following An important step to being able to manipulate the duration and trauma remains unresolved. Infiltrating leukocytes are obvious magnitude of neurotrophic factor activity is to identify the molec- candidates, but the CNS is unlike other organs in that the influx of ular mediators involved in their production. In this regard, the leukocytes is delayed in response to an acute insult. Thus, neutro- relationship between inflammatory cytokines and neurotrophic phils do not infiltrate appreciably until about 6 h after injury, factors, given their similar spatial expression following injury, whereas T cells and monocytes appear between 12 to 24 h or later may be critical. In vitro, a number of reports have shown inflam- (5–9). These results suggest that the CNS mounts an early and matory cytokines to influence the induction of neurotrophic factor intrinsic inflammatory response upon injury, before leukocyte in- production, with much emphasis placed on the induction of NGF filtration occurs. The origin of the early increase in inflammatory by IL-1 (reviewed in Ref. 15). This work has also been carried over cytokines within the CNS following trauma remains unresolved, to an in vivo context in which exogenous administration of IL-1 into although microglia have been suggested as possible sources (2, 3). the brain has resulted in the up-regulation of NGF (16, 17). Another hallmark of CNS injury is the posttrauma transient ex- In contrast, very little work has been done to examine whether pression of neurotrophic factors around the lesion site. Early stud- inflammatory cytokines affect CNTF production. CNTF is of particular interest because, in addition to being a survival factor for various neuronal populations, it has potent effects on cells of the Departments of *Clinical Neurosciences and †Oncology, Faculty of Medicine, Uni- versity of Calgary, Calgary, Alberta, Canada; and ‡Laboratory of Molecular Endo- oligodendroglial lineage (reviewed in Ref. 18). CNTF has been crinology and Department of Anatomy and Physiology, Laval University, Quebec, shown to be an important maturation factor for oligodendrocytes, Canada to promote their synthesis of myelin proteins (19, 20), and to pro- Received for publication April 7, 2000. Accepted for publication June 6, 2000. tect oligodendrocytes from apoptotic death induced by several The costs of publication of this article were defrayed in part by the payment of page agents (19, 21). Thus, the regulation of CNTF expression in the charges. This article must therefore be hereby marked advertisement in accordance CNS following injury may be particularly important for attenuat- with 18 U.S.C. Section 1734 solely to indicate this fact. ing neuronal and oligodendrocyte death. 1 This work was supported by an operating grant from the Medical Research Council of Canada (MRC). V.W.Y. is a Senior Scholar of the Alberta Heritage Foundation for Medical Research (AHFMR) and a MRC Scientist. S.R. is also a MRC Scientist. L.M.H. gratefully acknowledges AHFMR and MRC for studentship support. 2 Address correspondence and reprint requests to Dr. V. Wee Yong, Departments of 3 Abbreviations used in this paper: NGF, nerve growth factor; CNTF, ciliary neuro- Clinical Neurosciences and Oncology, Faculty of Medicine, 3330 Hospital Drive NW, trophic factor; IL-1ra, IL-1 receptor antagonist; ISH, in situ hybridization; ICC, com- Calgary, Alberta T2N 4N1, Canada. E-mail address: [email protected] bined immunocytochemistry; GFAP, glial fibrillary acidic protein.

The principal aims of this study were to elucidate the earliest inflicted wound was comparatively large, in accordance with previous de- period posttrauma in which elevation of inflammatory cytokines scriptions (23). Following removal of the dura mater, corticectomy was occurs, to define the cell type responsible for the initial elevation performed by aspiration of the cortex just down to the dorsal aspect of the corpus callosum. Recombinant murine IL-1␤ (50, 200, or 1000 U) (R&D of cytokines, and to establish if a causal relationship exists in vivo Systems, Minneapolis, MN), recombinant human IL-1 receptor antagonist between the elevation of inflammatory cytokines and CNTF fol- (IL-1ra; 50 ␮g/ml) (Amgen, Thousand Oaks, CA), or a saline control were lowing CNS corticectomy trauma. administered locally into the lesion site by means of a piece of absorbent gelatin sponge (Gelfoam, Upjohn, Kalamazoo, MI). Gelfoam was cut to ϳ1mm3 dimensions, soaked in a 10 ␮l volume of the test compound, and Materials and Methods applied directly over the corticectomy site for the duration of the experi- Mice ment, as detailed previously (23). All animals were sacrificed by cervical dislocation. Adult retired female breeders (4–6 mo) of the CD1 strain (Charles River, Montreal, Canada), IL-1␤Ϫ/Ϫ mice (129Sv/C57BL6, back-crossed for three generations with the B10RIII strain as previously described) (22), and RT-PCR the B10RIII genetic controls (The Jackson Laboratory, Bar Harbor, ME) The levels of transcripts encoding inflammatory cytokines IL-1␣, IL-1␤, were housed on a 12-h light/dark cycle with ad libitum access to food and and TNF-␣, as well as CNTF and GAPDH, were determined by semiquan- water. All experimental procedures were approved by the institution’s an- titative RT-PCR. Total RNA was isolated from tissue resected from around imal care committee and were in accordance with the guidelines instituted the lesion site (Fig. 1a) using the TRIzol reagent (Life Technologies, Bur- by the Canadian Council of Animal Care. CD1 mice were used in the lington, Ontario). Because the volume of the inflicted tissue ranged from 12 experiments of Figs. 1–6, whereas the IL-1␤-deficient mice and their wild- to 15 mm3, care was taken to dissect a rim of tissue measuring 1 mm type controls were used in the experiments of Figs. 7 and 8. encircling the lesion cavity. As an indication that comparable amounts of Corticectomy injury tissue were resected for analyses, the total RNA extracted from all samples Downloaded from was similar. Four samples were collected for each set of controls and var- Corticectomy injury, in which a 12–15 mm3 volume of parietal-occipital ious experimental time points. RNA (0.5 ␮g) was reverse-transcribed and cortex is removed by gentle aspiration, was used as a model of CNS trauma amplified in a single-step process, using oligonucleotide primers designed (Fig. 1a). This previously established model is advantageous for the local for murine IL-1␣ (5Ј-AAGTTTGTCATGAATGATTCCCTC-3Ј,3Ј delivery of test compounds to the injury site and mimics similar procedures TGAGTAGTGTCCATCACTCTG-5Ј); IL-1␤ (5Ј-CAGGATGAGGA performed by neurosurgeons for the resection of brain tumors, eplileptic CATGAGCACC-3Ј,3Ј-CACCTCAAACTCAGACGTCTC-5Ј); TNF-␣ foci, etc. (23). Briefly, animals were anesthetized with ketamine (200 (5Ј-ATGAGCACAGAAAGCATGATC-3Ј,3Ј-TTAAGCTCACTGTTCG http://www.jimmunol.org/ mg/kg i.p.) and xylazine (10 mg/kg i.p.) and immobilized in a stereotaxic GACAT-5Ј); and GAPDH (5Ј-CGGAGTCAACGGATTTGGTCGTAT3Ј, frame. Although anesthetics may have protective or detrimental properties 3Ј-CAGAAGTGGTGGTACCTCTTCCGA-5Ј), previously described in on the parameters analyzed in this manuscript, it should be noted that all detail (4), as well as CNTF (5Ј-GGCTAGCAAGGAAGATTCGT-3Ј,3Ј- experimental groups were subjected to the same anesthesia regimen, thus AATGGCATGGAAGGTTCCCT-5Ј) (24). All primers were purchased controlling for unknown factors that may be introduced in this regard. from Life Technologies. The number of cycles was predetermined to be in A midline incision was made, followed by a unilateral circular (2 mm the linear range of amplification for IL-1␣, IL-1␤, TNF-␣, and CNTF (35 diameter) craniectomy over the left hemisphere, 1 mm lateral of the mid- cycles) and GAPDH (25 cycles). cDNA products were electrophoresed on line suture, and exactly midway between the lambda and bregma sutures. 1.5% agarose gels and visualized with ethidium bromide incorporation Precise stereotaxic coordinates were not used given that the size of the under UV light. NIH Image Analysis software was used to quantify the size of each cDNA product, which was then expressed as a ratio to the mean

size of the cDNA products of unoperated controls. GAPDH housekeeping by guest on September 25, 2021 gene was used as an internal control to demonstrate that equivalent amounts of RNA were loaded per sample for all samples. Statistical analysis was performed using ANOVA and Bonferroni posttest.

In situ hybridization (ISH) Three and 24 h after corticectomy, animals were deeply anesthetized and then transcardially perfused with 0.9% cold saline, followed by 4% paraformaldehyde in 0.1 M borax buffer (pH 9.5 at 4°C). Brains were removed, postfixed for 2–8 days, and then placed in a solution containing 10% su- crose diluted in 4% paraformaldehyde-borax buffer overnight at 4°C. The frozen brains were mounted on a microtome (Reichert-Jung, Cambridge Instruments, Deerfield, IL) and cut into 20-␮m coronal sections from the olfactory bulb to the end of the medulla. The sections were collected in a cold cryoprotectant solution (0.05 M sodium phosphate buffer (pH 7.3), 30% ethylene glycol, and 20% glycerol) and stored at Ϫ20°C. Hybridiza- tion histochemical localization of IL-1␣, IL-1␤, and TNF-␣ transcripts was conducted on every sixth section of the whole rostro-caudal extent of each brain as described previously (25). The sections were exposed at 4°C to x-ray films (Kodak, Rochester, NY) for 24 h, dipped into NTB2 nuclear emulsion (Kodak; diluted 1:1 with distilled water), and exposed for 8–15 days before being developed and counterstained with thionin (0.25%). The IL-1␣ and -1␤ cRNA probes were generated from the full-length mouse IL-1␣ and -1␤ cDNAs (Dr. P. Gray, Genentech, South San Fran- cisco, CA) subcloned into PCR II (Invitrogen, Carlsbad, CA) and linearized with XhoI (IL-1␣ antisense, IL-1␤ sense), KpnI (IL-1␤ antisense), or BamHI (IL-1␣ sense). The 1.3-kb mouse TNF cDNA (Dr. Dr. D. Radzioch, ϩ FIGURE 1. A corticectomy injury to the adult mouse brain results in Laval University, Quebec, Canada) was subcloned into pBluescript SKII elevation of inflammatory cytokines. a, Line drawing of a coronal section and linearized with BamHI and PstI for the sense and antisense riboprobes, of the adult mouse brain shows the corticectomy lesion (L). Dotted line respectively. Radioactive cRNA copies were synthesized by incubating 250 ng of linearized plasmid in 6 mM MgCl , 40 mM Tris (pH 7.5), 2 mM delineates the volume of tissue resected from around the injury site for 2 spermidine, 10 mM DTT, 0.2 mM ATP/GTP/CTP, [␣-35S]UTP, and 40 U RNA extraction and RT-PCR analysis. b, Elevation of proinflammatory RNAsin (Promega, Madison, WI), and 20 U of T7 (IL-1␣ sense, IL-1␤ cytokine transcripts at 3 h following corticectomy injury in adult mice. antisense, TNF-␣ sense), SP6 (IL-1␣ antisense, IL-1␤ sense), or T3 Each lane represents target RNA transcript of a single mouse. Note that (TNF-␣ antisense) RNA polymerase for 60 min at 37°C. Unincorporated GAPDH levels remain relatively constant to indicate similarity of RNA nucleotides were removed using the ammonium-acetate precipitation meth- loading in all groups. od; 100 ␮l of DNase solution (1 ␮l DNase, 5 ␮l of 5 mg/ml tRNA, and 94 2234 IL-1␤ REGULATION OF CNTF FOLLOWING CNS TRAUMA

␮ ␣ l of 10 mM Tris/10 mM MgCl2) was added, and 10 min later an extrac- nificantly elevated within 15 min after injury, followed by IL-1 tion was accomplished using a phenol-chloroform solution. The cRNA was and TNF-␣ which increased by 1 h (Fig. 2a). Levels of all three ␮ ␮ precipitated with 80 l of 5 M ammonium acetate and 500 l of 100% proinflammatory cytokines peaked at 3 h following injury. The ethanol for 20 min on dry ice. The pellet was washed with ethanol, dried, ␣ ␤ ␣ and resuspended in 100 ␮l of 10 mM Tris/1 mM EDTA (pH 8.0). A con- magnitude of increase of TNF- is less than that of IL-1 or IL-1 centration of 107 cpm probe was mixed into 1 ml of hybridization solution and is related to the higher basal level of TNF-␣ in normal unin- (500 ␮l formamide, 60 ␮l of 5 M NaCl, 10 ␮l of 1 M Tris (pH 8.0), 2 ␮l jured animals (see Fig. 1b). By ISH, IL-1␣, IL-1␤, and TNF-␣ are of 0.5 M EDTA (pH 8.0), 50 ␮lof20ϫ Denhart’s solution, 200 ␮lof50% localized to cells in the brain parenchyma around the lesion site at dextran sulfate, 50 ␮l of 10 mg/ml tRNA, 10 ␮l of 1 M DTT, and 118 ␮l Depc water minus volume of probe used). This solution was mixed and 3 and 24 h postinjury (Fig. 3). heated for 5 min at 65°C before being spotted on slides. The relationship of the up-regulation of cytokines to that of CNTF mRNA was examined (Fig. 2b), and it was discovered that Combined immunocytochemistry (ICC) and ISH CNTF transcripts become significantly elevated first at 1 h and peaked ICC was combined with ISH to determine the type(s) of cells that express by 12 h postinjury, which follows the up-regulation of IL-1␤. IL-1␤ transcript in the parenchymal brain at the edge of the cortical lesion. Microglia were labeled with an Ab against ionized calcium binding adapter The up-regulation of IL-1␤ following injury is a CNS intrinsic molecule 1 (Iba1), previously described in detail to be specific for micro- response glial cells (26), and astrocytes were labeled with an Ab against glial fibrillary acidic protein (GFAP). Every sixth brain section was processed by the To determine whether the observed up-regulation of IL-1␤ was avidin-biotin method with peroxidase as a substrate. Briefly, slices were due to the infiltration of neutrophils, known to be one of the ear- washed in sterile Depc-treated 50 mM potassium PBS (KPBS) and incubated at room temperature with either Iba1 (rabbit anti-rat Iba1; generously liest cells to respond to injury, tissues were stained with hematox- provided by Dr. Y. Imai, National Institute of Neuroscience, Kodaira, To- ylin and eosin to visualize polymorphic nuclei. Polymorphonuclear Downloaded from kyo, Japan) or GFAP (rabbit anti-cow GFAP; Chemicon International, Te- mecula, CA) Ab diluted in sterile KPBS, 0.4% Triton X-100, 0.25% hep- arin sodium salt USP (ICN Biomedicals, Aurora, OH), and 1% BSA (fraction V; Sigma, St. Louis, MO). Two hours after incubation with the primary Ab, sections were rinsed in sterile KPBS and incubated with bi- otinylated secondary Abs (Vector Laboratories, CA) for 60 min. Sections

were then rinsed with KPBS and incubated at room temperature for 60 min http://www.jimmunol.org/ with an avidin-biotin-peroxidase complex (Vectastain ABC elite kit, Vec- tor Laboratories). After several rinses in sterile KPBS, the brain slices were reacted in 0.05% diaminobenzidine (DAB) and 0.003% hydrogen peroxide

(H2O2). Thereafter, sections were rinsed in sterile KPBS, immediately mounted onto gelatin- and poly-L-lysine-coated slides, dessicated under vacuum for 30 min, fixed in 4% paraformaldehyde for 20 min, and digested by pro- teinase K (10 ␮g/ml in 100 mM tris HCl (pH 8.0) and 50 mM EDTA (pH 8.0)) at 37°C for 25 min. Prehybridization, hybridization, and posthybrid- ization steps were performed according to the above description with the difference of dehydration (alcohol 50, 70, 95, 100%), which was shortened by guest on September 25, 2021 to avoid decoloration of immunoreactive cells (brown staining). After being dried for 2 h under vacuum, sections were exposed overnight at 4°C to x-ray film (Kodak), defatted in xylene, and dipped in NTB2 nuclear emulsion (Kodak; diluted 1:1 with distilled water). Slides were exposed for 14 days, developed in D19 developer (Kodak) for 3.5 min at 15°C, and fixed in rapid fixer (Kodak) for 5 min. Sections were then rinsed in running distilled water for 1–2 h, rapidly dehydrated through graded concentrations of alcohol, cleared in xylene, and coverslipped with DPX. The presence of IL-1␤ transcript was detected by the agglomeration of silver grains in perikarya, whereas Iba1 immunoreactivity within the cell cytoplasm and ramifications was indicated by a brown homogenous coloration. Determi- nation of the double-labeled cells was performed visually for each cell exhibiting clear brown cytoplasm and a number of silver grains within the cell body, delineating convincing hybridized message. Hematoxylin and eosin staining Animals were anesthetized and transcardially perfused with 0.9% saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer. Brains were removed, postfixed in 4% paraformaldehyde, and then embedded in par- affin and cut into 4–6 ␮m sections. Sections were heated for1hat60°C, deparaffinized in xylene, and rehydrated through a graded series of alcohol to water and PBS. Sections were then stained with hematoxylin and eosin, dehydrated through a graded series of alcohol and xylene, and coverslipped with Permount. Results Proinflammatory cytokines are rapidly elevated around the FIGURE 2. The elevation of CNTF follows the elevation of inflamma- lesion site before the increase in CNTF tory cytokines after CNS trauma. a, Changes in levels of transcripts encoding inflammatory cytokines following corticectomy injury. b, Tran- Brain tissues from uninjured adult CD1 mice express low to un- script profile for CNTF elevation following corticectomy. Note that the detectable levels of cytokine transcripts when measured by the elevation of IL-1␤ precedes that of CNTF. The levels of GAPDH remain highly sensitive method of RT-PCR (Fig. 1b). In contrast, inflam- stable across the different groups to indicate relative similarity in RNA matory cytokine transcripts are rapidly up-regulated around the loading across all groups. Each point represents mean Ϯ SEM of four mice. p Ͻ ,ء) lesion site following a corticectomy injury (Fig. 1b). When com- Asterisks indicate significant difference from unoperated controls .(p Ͻ 0.001 ,ءءء ;p Ͻ 0.01 ,ءء ;pared with unoperated controls, mRNA encoding IL-1␤ was sig- 0.05 The Journal of Immunology 2235

FIGURE 4. Polymorphonuclear cells (PMNs) are present around the injury site only at later time points. PMNs, large producers of inflammatory cytokines, are not present in the early hours following corticectomy injury when the peak increase in cytokine transcripts occur. These cells are usu- ally the first infiltrating leukocyte to respond to the injury. Here, at 2 days after injury, PMNs (examples indicated by black arrows) are clearly evident in tissue surrounding the lesion. Downloaded from

positive for IL-1␤ mRNA overlapped entirely with Iba1-positive microglia in the parenchymal tissue surrounding the lesion site.

However, we noted that not all Iba1-positive microglia express http://www.jimmunol.org/ IL-1␤ transcripts (Fig. 5), suggesting heterogeneity in the response of microglia to injury. GFAP-immunoreactive cells were not present in regions depicting hybridization message for the proin- ﬂammatory cytokine, which does not support astrocytes as a potential source of IL-1␤ during the rapid inﬂammatory response that takes place after corticectomy (results not shown).

IL-1ra blocks the up-regulation of CNTF

Given the very early increase and peak of IL-1␤ elevation before by guest on September 25, 2021 the increase of CNTF transcripts, we examined if a relationship of IL-1␤ affecting CNTF could be evidenced. This was first ap- proached by determining the effect of blocking the actions of IL-1␤ around the lesion site of CD1 mice, considering that this cytokine has been shown by ISH to be localized immediately around the area of injury (Fig. 3). Direct application of IL-1ra was made into the corticectomy site through an absorbent gelatin sponge, gelfoam, which overlaid the lesion cavity. We have previously described this technique to apply drugs locally to the injury FIGURE 3. Expression of proinflammatory cytokines in the mouse site and have shown that the application of IL-10, a cytokine syn- brain in response to corticectomy. The dark-field photomicrographs of thesis inhibitor, attenuated the extent of astrogliosis and inflam- dipped NTB-2 emulsion slides (left column) depict IL-1␣, IL-1␤, and matory cytokine production (23). The application of IL-1ra sup- TNF-␣ transcripts 3 and 24 h after insult. The bright-field photomicro- pressed the up-regulation of CNTF at 12 h postcorticectomy (Fig. graphs (right column) are representative examples of cytokine-expressing 6), the time point corresponding to peak CNTF transcript elevation cells in the parenchymal brain surrounding the lesion site. The mRNAs in normal corticectomized mice. encoding proinflammatory cytokines were not detected in the brain of control mice (top panels), whereas a very localized and robust signal was IL-1␤Ϫ/Ϫ mice fail to up-regulate CNTF detected along the lesion site (open arrows). Adjacent sections hybridized with the sense probes did not exhibit positive signal (data not shown). Mice genetically deficient for IL-1␤ were subjected to a corticec- Magnifications: dark-field, ϫ25, scale bar ϭ 100 ␮m; bright-field, ϫ250, tomy insult to address their capacity to generate proinflammatory Ϫ Ϫ scale bar ϭ 10 ␮m. cytokines and CNTF. IL-1␤ / mice did not express IL-1␤ or IL-1␣ cytokine transcripts under basal conditions or following corticectomy injury (data not shown) but did express TNF-␣ which is cells were only observed in the CNS by 1 and 2 days after injury elevated upon injury, confirming the competence of these animals (Fig. 4) and not at 1 and 3 h, the earliest time points that we to mount an inflammatory response to brain injury (Fig. 7a). A examined and which correspond to peak elevation of IL-1␤ tran- subsequent examination of CNTF mRNA after corticectomy in script. Because monocytes, which are also sources of IL-1␤,donot IL-1␤Ϫ/Ϫ animals (Fig. 7a) revealed no elevation of this trophic infiltrate the CNS until much later, the initial production of cyto- factor at time points corresponding to elevated and peak increase kines appears to be a CNS intrinsic response. in wild-type animals (Fig. 7b). Basal levels of CNTF transcripts in Combined ISH and ICC revealed the cellular source of IL-1␤ to uninjured animals were equivalent in the IL-1␤ mutants and their be the CNS resident macrophage, the microglia (Fig. 5). Cells genetic controls (Fig. 7c). 2236 IL-1␤ REGULATION OF CNTF FOLLOWING CNS TRAUMA Downloaded from http://www.jimmunol.org/

FIGURE 5. IL-1␤ is expressed within parenchymal microglial cells 3 h after the lesion. Microglial cells were labeled by the immunoperoxidase technique using iba1 (black arrowheads) (26), and IL-1␤ mRNA was thereafter hybridized on the same sections by means of a radioactive ISH technique by guest on September 25, 2021 (silver grains, white arrowheads). Note the presence of mRNA encoding IL-1␤ within parenchymal microglial cells (agglomeration of silver grains within the cytoplasms, bottom panels). Although all IL-1␤-expressing cells in the brain parenchyma were immunoreactive to iba1, numerous microglia were devoid of positive hybridization signal. Magnifications: dark-field (top left), ϫ23; dark-field and bright-field (top middle and right), ϫ62.5; bottom panels, ϫ230.

Rescue of CNTF mRNA elevation in IL-1␤Ϫ/Ϫ mice for some time that these hematogenous cells enter the CNS only If IL-1␤ is required for CNTF production in CNS trauma, then the after a signiﬁcant delay (28, 29). As shown in this study, the ele- ␤ ␣ ␣ phenotype of lack of CNTF elevation in IL-1␤Ϫ/Ϫ mice should be vation and peak of IL-1 , IL-1 , and TNF- occur during a period restored by the application of exogenous IL-1␤. Indeed, when where no leukocytes are apparent. This suggests a CNS intrinsic IL-1␤ was introduced into the lesion site by gelfoam, an elevation of CNTF was observed at 12 h postcorticectomy, a time point corresponding to peak CNTF elevation in normal, corticectomized mice (Fig. 8). Collectively, the data strongly support the requirement for IL-1␤ in the up-regulation of CNTF following traumatic brain injury.

Discussion The presence of various inflammatory cell types is now recognized to be a common finding in the diseased CNS. This occurs in diseases classically associated with widespread inflammation, such as multiple sclerosis, as well as in neurodegenerative diseases including Alzheimer’s disease. Typically, these conditions involve the migration of hematogenous cells such as monocytes, lymphocytes, and NK cells into the CNS (27). In the context of acute trauma, FIGURE 6. IL-1ra blocks the elevation of CNTF in corticectomized such as the infliction of a stab, aspiration (as in this study), or mice. The graph demonstrates the CNTF mRNA profile at 12 h following contusive injury to the CNS, there has also been recent recognition IL-1ra (50 ␮g/ml) application into the corticectomy injury site. The asterisk that elevation of inflammatory cytokines is attendant to trauma indicates CNTF level of PBS-gelfoam controls, which was significantly (1–4). In these circumstances, while infiltrating leukocytes are ob- different from unoperated controls and 12-h corticectomized animals .(p Ͻ 0.05 ,ء) vious candidates as sources of cytokines, it has been appreciated treated with IL-1ra The Journal of Immunology 2237

FIGURE 7. CNTF is not elevated following corticectomy to IL-1␤Ϫ/Ϫ mice. Cytokine and CNTF transcript proﬁles in mice deﬁcient for IL-1␤ and their genetic controls are displayed. a, Elevation of TNF-␣ and CNTF in IL-1␤Ϫ/Ϫ mice following corticectomy injury. Not shown is the absence of IL-1␣ and IL-1␤ in all

groups. Note that GAPDH levels remain relatively con- Downloaded from stant to indicate similarity of RNA loading in all samples. b, Elevation of CNTF following corticectomy injury in wild-type controls to indicate that this strain of mice can up-regulate CNTF following injury. The apparent difference in basal CNTF expression in unoperated IL-1␤Ϫ/Ϫ or wild-type mice is due to gel exposure variations, and this is conﬁrmed in c where CNTF levels http://www.jimmunol.org/ in unoperated IL-1␤Ϫ/Ϫ mice and their genetic controls show similar basal expression levels when samples were analyzed in the same PCRs and resolved on the same gel. by guest on September 25, 2021

source and is confirmed by the localization of IL-1␤ transcript to of CNTF in vivo following CNS trauma. A temporal profile for microglia alone, at least at 3 h following injury. That microglia are proinflammatory cytokines and CNTF transcript elevation was es- the sole sources of IL-1␤ early on in trauma fulfills the prediction tablished using the corticectomy model of injury. The increase in of Kreutzberg (30) that microglia are the sensors of CNS pathology. As mentioned earlier, several other studies have also reported an early rise in the inflammatory cytokines IL-1␤ and TNF-␣ following CNS trauma (2, 3) or in response to excitotoxic damage (31). Our finding that IL-1␤ transcripts increase by 15 min following corticectomy is the earliest that anyone has observed in CNS trauma. Neurotrophic factors are produced following CNS injury, but the mechanisms that regulate their elevation remain unresolved. In the peripheral nervous system, it has been shown that the lesion- mediated increase in NGF is primarily regulated by the production of IL-1 by activated macrophages (32). In the CNS, studies have similarly implicated a role for IL-1␤ in the induction of NGF expression by astrocytes in vitro and in vivo (16, 33). The role of FIGURE 8. The absence of CNTF up-regulation in IL-1␤Ϫ/Ϫ mice can IL-1 in the production of other trophic factors, such as CNTF, is be corrected by IL-1␤ treatment. CNTF transcript levels following exog- largely unknown, although IFN-␥ (34) as well as IL-1␤ and TNF-␣ enous IL-1␤ application directly into the corticectomy injury site. The as- (35) can up-regulate CNTF production in astroglial cultures. In this terisk indicates that the CNTF level was significantly different from unop- .(p Ͻ 0.05 ,ء) paper we have addressed whether IL-1␤ mediates the production erated and PBS-treated controls 2238 IL-1␤ REGULATION OF CNTF FOLLOWING CNS TRAUMA

CNTF temporally follows the increase in IL-1␤, which is sugges- the application of inflammatory cytokines to lesioned areas was tive of a causative role for IL-1␤ in the up-regulation of CNTF. We neuroprotective or promoted the regeneration of axons (47, 48). confirmed such an inferred relationship by showing that blocking More recently, the transplant of cultured microglial cells (49) or the activity of IL-1␤ with its receptor antagonist attenuated the monocytes (50) into the injured spinal cord led to enhanced neurite production of CNTF, and furthermore, that CNTF elevation was outgrowth. To complicate matters further, the role of inflammation inhibited in IL-1␤Ϫ/Ϫ animals. Importantly, the application of ex- in the CNS may be structure or region specific. A recent study ogenous IL-1␤ into the injury site of IL-1␤Ϫ/Ϫ mice was able to observed that IL-1ra infusion into the striatum reduced excitotox- rescue the phenotype of lack of CNTF elevation in these animals. icity-induced neuronal damage, but failed to protect neurons in the Taken together, these findings show for the first time a requirement cortex (51). Finally, inflammation in the CNS likely leads to the for IL-1␤ in the production of CNTF following CNS trauma. Thus, evolution of astrocyte reactivity (23, 52), and the multiple bene- the CNS-initiated IL-1␤ increase serves to attempt preservation of ficial or detrimental effects of astrogliosis (15) will have to be the integrity of the CNS by producing factors that aid the survival taken into consideration. The results here of IL-1␤ regulating of both neurons and resident glial cells such as the CNTF clearly suggest the potential of inflammation in the injured oligodendrocyte. CNS to have neurotrophic potential. We are currently investigating the mechanisms by which IL-1␤ We conclude that an acute trauma to the CNS initiates the mi- regulates CNTF production. It is possible that this occurs through croglial production of IL-1␤, and this is required to up-regulate the the induction of the transcription factor NF-␬B or AP-1. NF-␬B neurotrophic factor CNTF. This novel observation invites impor- and AP-1 consensus sequences have been reported in the CNTF tant future directions, including the role that IL-1 plays in the promoter (34), and many biological effects of IL-1 are known to regulation of other neurotrophic factors following injury such as Downloaded from occur through NF-␬B and AP-1-mediated mechanisms (36). How- brain-derived neurotrophic factor, glial-derived neurotrophic fac- ever, other possibilities exist, since an elevation of TNF-␣ but not tor, and basic fibroblast growth factor. Also of interest, and cur- CNTF is observed in the IL-1␤Ϫ/Ϫ mice (Fig. 7a), and TNF-␣ can rently a focus of this laboratory, is whether IL-1␤ levels can be also induce the activation of NF-␬B or AP-1 transcription factors. used to manipulate the production of CNTF to enhance oligoden- It is noteworthy that the IL-1␤Ϫ/Ϫ mice did not express IL-1␣ drocyte survival and remyelination. Resolving the neurotrophic

following corticectomy, since IL-1␣ and IL-1␤ are located on sep- potential of CNS inflammation will have impact on the recovery of http://www.jimmunol.org/ arate genes in the mouse. The finding that TNF-␣ is promptly the CNS to injury. elevated in the IL-1␤-deficient mice supports that the failure of IL-1 elevation is not indicative of an inability of these animals to Acknowledgments mediate an inflammatory response within the CNS. A recent study We thank Dr. Barry Rewcastle for his help with interpreting our hematox- examining IL-1␤Ϫ/Ϫ and IL-1␣Ϫ/Ϫ mice has shown that an IL-1␤ ylin and eosin results. deficiency abrogated IL-1␣ expression, as well as IL-1␤, but not vice versa, while TNF-␣ expression level was not affected in either References case (37). These authors suggest that the expression of IL-1␣, but 1. Taupin, V., S. Toulmond, A. Serrano, J. Benavides, and F. Zavala. 1993. Increase in IL-6, IL-1 and TNF levels in rat brain following traumatic lesion. J. Neuro- ␣ ␤ by guest on September 25, 2021 not TNF- , is mostly dependent on the expression of IL-1 ; our immunol. 42:177. current findings support such a hypothesis. Furthermore, our re- 2. Bartholdi, D., and M. E. Schwab. 1997. Expression of inflammatory cytokine and sults indicate that TNF-␣ is not adequate to elevate CNTF in the chemokine mRNA upon experimental spinal cord injury in the mouse: an in situ hybridization study. Eur. J. Neurosci. 9:1422. brain parenchyma following injury, at least within the time points 3. Streit, W. J., S. L. Semple-Rowland, S. D. Hurley, R. C. Miller, P. G. Popovich, studied. and B. T. Stokes. 1998. Cytokine mRNA profiles in contused spinal cord and axotomized facial nucleus suggest a beneficial role for inflammation and gliosis. The cellular source of production of trophic factors postinjury is Exp. Neurol. 152:74. an area of interest. Although it is clear that the microglia are the 4. Rostworowski, M., V. Balasingam, S. Chabot, T. Owens, and V. W. Yong. 1997. early source of IL-1␤, we were unsuccessful in our attempts to Astrogliosis in the neonatal and adult murine brain post-trauma: elevation of inflammatory cytokines and the lack of requirement for endogenous interferon-␥. identify the source of CNTF. Nonetheless, the most likely source J. Neurosci. 17:3664. of CNTF up-regulation following traumatic CNS injury is the as- 5. Giulian, D., J. Chen, J. E. Ingeman, J. K. George, and M. Noponen. 1989. The trocytes, since CNTF has been localized to astrocytes in the normal role of monnuclear phagocytes in wound healing after traumatic injury to adult mammalian brain. J. Neurosci. 9:4416. brain (38, 39). Furthermore, the up-regulation of CNTF following 6. Moreno-Flores, M. T., P. Bovolenta, and M. Nieto-Sampedro. 1993. Polymor- CNS trauma has been noted to occur in reactive astrocytes around phonuclear leukocytes in brain parenchyma after injury and their interaction with purified astrocytes in culture. Glia 7:146. the lesion site (12, 13, 39, 40). 7. Dusart, I., and M. E. Schwab. 1994. Secondary cell death and the inflammatory It remains controversial as to whether inflammation in the in- reaction after dorsal hemisection of the rat spinal cord. Eur. J. Neurosci. 6:712. jured CNS serves a beneficial or detrimental purpose (reviewed in 8. Frank, M., and H. Wolburg. 1996. Cellular reactions at the lesion site after crush- ␣ ing of the rat optic nerve. Glia 16:227. Refs. 15 and 41). A number of studies have observed that TNF- 9. Popovich, P. G., P. Wei, and B. T. Stokes. 1997. Cellular inflammatory response produces apoptotic death to oligodendrocytes and neurons (19, after spinal cord injury in Sprague-Dawley and Lewis rats. J. Comp. Neurol. 21), and some reports have also implicated IL-1␤ in the mediation 377:443. 10. Nieto-Sampedro, M., E. R. Lewis, C. W. Cotman, M. Manthorpe, S. D. Skaper, of neuronal death, particularly following ischemic and excitotoxic G. Barbin, F. M. Longo, and S. Varon. 1982. Brain injury causes a time-depen- brain injury (42–44). Inflammation following CNS trauma would dent increase in neuronotrophic activity at the lesion site. Science 217:858. 11. Arendt, T., M. K. Bruckner, T. Krell, S. Pagliusi, L. Kruska, and R. Heumann. therefore be detrimental in the context of potential injury to neu- 1995. Degeneration of rat cholinergic basal forebrain neurons and reactive rons and oligodendrocytes, and reports of the correspondence of changes in nerve growth factor expression after chronic neurotoxic injury. II. the number of macrophages/microglia with the amount of tissue Reactive expression of NGF gene in astrocytes. Neurosci. 65:647. 12. Ip, N. Y., S. J. Wiegand, J. Morse, and J. S. Rudge. 1993. Injury-induced regu- damage at each level of the spinal cord in contusion injury (e.g., lation of ciliary neurotrophic factor mRNA in the rat brain. Eur. J. Neurosci. Carlson et al. (45)) would be consistent with such a hypothesis. In 5:25. contrast, other reports indicate that inflammation may be beneficial 13. Kirsch, M., T. Schneider, M. Y. Lee, and H. D. Hofmann. 1998. Lesion-induced changes in the expression of ciliary neurotrophic factor and its receptor in rat to recovery. In earlier work, David et al. (46) showed that the optic nerve. Glia 23:239. nonpermissive nature of the rat optic nerve could support neurite 14. Yao, D. L., N. R. West, C. A. Bondy, M. Brenner, L. D. Hudson, J. Zhou, G. H. Collins, and H. D. Webster. 1995. Cryogenic spinal cord injury induces extension from dorsal root ganglia if treated with macrophages astocytic gene expression of insulin-like growth factor 1 and insulin-like growth isolated from the injured brain. Several reports demonstrate that factor binding protein 2 during myelin regneration. J. Neurosci. Res. 40:647. The Journal of Immunology 2239

15. Yong, V. W. 1996. Cytokines, astrogliosis and neurotrophism following CNS 35. Kamiguchi, H., K. Yoshida, M. Sagoh, H. Sasaki, M. Inaba, H. Wakamoto, trauma. In Cytokines and the CNS: Development, Defense and Disease. M. Otani, and S. Toya. 1995. Release of ciliary neurotrophic factor from cultured R. M. Ransohoff, and E. N. Benveniste, eds. CRC Press, Boca Raton, FL, p. 309. astrocytes and its modulation by cytokines. Neurochem. Res. 20:1187. 16. Spranger, M., D. Lindholm, C. Bandtlow, R. Heumann, H. Gnahn, 36. Dinarello, C.A. 1997. Interleukin-1. Cytokine Growth Factor Rev. 8:253. M. Naher-Noe, and H. Thoenen. 1990. Regulation of nerve growth factor (NGF) 37. Horai, R., M. Asano, K. Sudo, H. Kanuka, M. Suzuki, M. Nishihara, synthesis in the rat central nervous system: comparison between the effects of M. Takahashi, and Y. Iwakura. 1998. Production of mice deficient in genes for interleukin-1 and various growth factors in astrocyte cultures in vitro. Eur. interleukin (IL)-1␣, IL-1␤, IL-1␣/␤, and IL-1 receptor antagonist shows that J. Neurosci. 2:69. IL-1␤ is crucial for turpentine-induced fever development and glucocorticoid 17. DeKosky, S. T., J. R. Goss, P. D. Miller, S. D. Styren, P. M. Kochanek, and secretion. J. Exp. Med. 187:1463. D. Marion. 1994. Upregulation of nerve growth factor following cortical trauma. Exp. Neurol. 130:173. 38. Stockli, K. A., F. Lottspeich, M. P. Sendtner, P. Masiakowski, R. Carroll, 18. McLaurin, J., and V. W. Yong. 1995. Oligodendrocytes and myelin. Neurol. Clin. D. Gotz, D. Lindholm, and H. Thoenen. 1989. Molecular cloning expression and North Am. 13:23. regional distribution of rat ciliary neurotrophic factor. Nature 342:920. 19. Louis, J.-C., E. Magal, and S. Takayama. 1993. CNTF protection of oligoden- 39. Guthrie, K. M., A. G. Woods, T. Nguyen, and C. M. Gall. 1997. Astroglial ciliary drocytes against natural and tumor necrosis factor ␣-induced death. Science 259: neurotrophic factor mRNA expression is increased in fields of axonal sprouting 689. in deafferented hippocampus. J. Comp. Neurol. 386:137. 20. Mayer, M., K. Bhakoo, and M. Noble. 1994. Ciliary neurotrophic factor and 40. Lee, M. Y., C. J. Kim, S. L. Shin, S. H. Moon, and M. H. Chun. 1998. Increased leukemia inhibitory factor promote the generation, maturation and survival of ciliary neurotrophic factor expression following spinal cord injury in the rat. oligodendrocytes in vitro. Development 120:143. Neurosci. Lett. 255:79. 21. D’Souza, S. D., K. A. Alinauskas, and J. P. Antel. 1996. Ciliary neurotrophic 41. Schwartz, M., G. Moalem, R. Leibowitz-Amit, and I. R. Cohen. 1999. Innate and factor selectively protects human oligodendrocytes from tumor necrosis factor- adaptive immune responses can be beneficial for CNS repair. Trends Neurosci. mediated injury. J. Neurosci. Res. 43:289. 22:295. 22. Zheng H., D. Fletcher, W. Kozak, M. Jiang, K. J. Hofmann, C. A. Conn, D. Soszynski, C. Grabiec, M. E. Trumbauer, A. Shaw, et al. 1995. Resistance to 42. Relton, J. K., and N. J. Rothwell. 1992. Interleukin-1 receptor antagonist inhibits fever induction and impaired acute-phase response in interleukin-1␤-deficient ischaemic and excitotoxic neuronal damage in the rat. Brain Res. Bull. 29:43. mice. Immunity 3:9. 43. Betz, A. L., G. Y. Yang, and B. L. Davidson. 1995. Attenuation of stroke size in Downloaded from 23. Balasingam, V., and V. W. Yong. 1996. Attenuation of astroglial reactivity by rats using an adenoviral vector to induce overexpression of interleukin-1 receptor interleukin-10. J. Neurosci. 16:2945. antagonist in brain. J. Cereb. Blood Flow Metab. 15:547. 24. Watabe, K., T. Fukuda, J. Tanaka, H. Honda, K. Toyohara, and O. Sakai. 1995. 44. Hagan, P., J. D. Barks, M. Yabut, B. L. Davidson, B. Roessler, and Spontaneously immortalized adult mouse Schwann cells secrete autocrine and F. S. Silverstein. 1996. Adenovirus-mediated over-expression of interleukin-1 paracrine growth-promoting activities. J. Neurosci. Res. 41:279. receptor antagonist reduces susceptibility to excitotoxic brain injury in perinatal 25. Rivest, S., N. Laflamme, and R. E. Nappi. 1995. Immune challenge and immo- rats. Neuroscience 75:1033. bilization stress induce transcription of the gene encoding the CRF receptor in 45. Carlson, S. I., M. E. Parrish, J. E, Springer, K. Doty, and L. Dossett. 1998. Acute

selective nuclei of the rat hypothalamus. J. Neurosci. 15:2680. inflammatory response in spinal cord following impact injury. Exp. Neurol. 151: http://www.jimmunol.org/ 26. Ito, D., Y. Imai, K. Ohsawa, K. Nakajima, Y. Fukuuchi, and S. Kohsaka. 1998. 77. Microglia-specific localization of a novel calcium binding protein, Iba1. Brain Res. Mol. Brain Res. 57:1. 46. David, S., C. Bouchard, O. Tsatas, and N. Giftochristos. 1990. Macrophages can 27. Lotan, M., and M. Schwartz. 1994. Cross talk between the immune system and modify the nonpermissive nature of the adult mammalian central nervous system. the nervous system in response to injury: implications for regeneration. FASEB J. Neuron 5:463. 8:1026. 47. Schwartz, M., A. Solomon, V. Lavie, S. Ben-Bassat, M. Belkin, and A. Cohen. 28. Anderson, P. B., V. H. Perry, and S. Gordon. 1992. The acute inflammatory 1991. Tumor necrosis factor facilitates regeneration of injured central nervous response to lipopolysaccharide in the CNS parenchyma differs from that in other system axons. Brain Res. 545:33. tissues. Neuroscience 48:169. 48. Ikeda, K., Y. Iwasaki, T. Shiojima, and M. Kinoshita. 1996. Neuroprotective 29. Perry, V. H., P. B. Anderson, and S. Gordon. 1993. Macrophages and inflam- effect of various cytokines on developing spinal motoneurons following axotomy. mation in the central nervous system. Trends Neurosci. 16:268. J. Neurol. Sci. 135:109.

30. Kreutzberg, G. W. 1996. Microglia: a sensor for pathological events in the CNS. 49. Rabchevsky, A. G., and W. J. Streit. 1997. Grafting of cultured microglial cells by guest on September 25, 2021 Trends Neurosci. 19:312. into the lesioned spinal cord of adult rats enhances neurite outgrowth. J. Neuro- 31. Pearson, V. L., N. J. Rothwell, and S. Toulmond,. 1999. Excitotoxic brain dam- sci. Res. 47:34. age in the rat induces interleukin-1␤ protein in microglia and astrocytes: corre- 50. Rapalino, O., O. Lazarov-Spiegler, E. Agranov, G. J. Velan, E. Yoles, lation with the progression of cell death. Glia 25:311. M. Fraidakis, A. Solomon, R. Gepstein, A. Katz, M. Belkin, et al. 1998. Implan- 32. Lindholm, D., R. Heumann, M. Meyer, and H. Thoenen. 1987. Interleukin-1 tation of stimulated homologous macrophages results in partial recovery of para- regulates synthesis of nerve growth factor in non-neuronal cells of rat sciatic plegic rats. Nat. Med. 4:814. nerve. Nature 330:658. 33. Dekosky, S. T., S. D. Styren, M. E. O’Malley, J. R. Goss, P. Kochanek, 51. Lawrence, C. B., S. M. Allan, and N. J. Rothwell. 1998. Interleukin-1␤ and the D. Marion, C. H. Evans, and P. D. Robbins. 1996. Interleukin-1 receptor antag- interleukin-1 receptor antagonist act in the stiatum to modify excitotoxic brain onist suppresses neurotrophin response in injured rat brain. Ann. Neurol. 39:123. damage in the rat. Eur. J. Neurosci. 10:1188. 34. Carroll, P., M. Sendtner, M. Meyer, and H. Thoenen. 1993. Rat ciliary neuro- 52. Balasingam, V., T. Tejada-Berges, E. Wright, R. Bouckova, and V. Wee Yong. trophic factor (CNTF): gene structure and regulation of mRNA levels in glial cell 1994. Reactive astrogliosis in the neonatal mouse brain and its modulation by cultures. Glia 9:176. cytokines. J. Neurosci. 14:846.