B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

ava i l ab l e a t www.sc i enced i rec t . com

www.e l sev i e r. com/ loca te /b ra in res

Research Report

Transforming growth factor-β1 upregulates keratan sulfateand chondroitin sulfate biosynthesis in microglias afterbrain injury

Jiarong Yina, Kazuma Sakamotoa, Haoqian Zhanga, Zenya Itoa,b, Shiro Imagamaa,b,Satoshi Kishidaa, Takamitsu Natoria, Makoto Sawadac,Yukihiro Matsuyamab, Kenji Kadomatsua,⁎aDepartment of Biochemistry, Nagoya University Graduate School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, JapanbDepartment of Orthopedics, Nagoya University Graduate School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, JapancResearch Institute of Environmental Medicine, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan

A R T I C L E I N F O

⁎ Corresponding author. Fax: +81 52 744 2065.E-mail address: kkadoma@med.nagoya-uAbbreviations: ACM, astrocyte condition

conditioned medium; CNS, central nervousgrowth factor; Gal, galactose; GalNAc, N-acacetylglucosamine 6-O-sulfotransferase-1; Kpolymerase chain reaction; TGF-β1, transform

0006-8993/$ – see front matter. Published bydoi:10.1016/j.brainres.2009.01.042

A B S T R A C T

Article history:Accepted 22 January 2009Available online 3 February 2009

After injury to the adult central nervous system, levels of extracellular matrix moleculesincrease at the injury site and may inhibit the repair of injured axons. Among thesemolecules, the importance of proteoglycans, particularly their chondroitin sulfate chains,has been highlighted. We have recently reported that keratan sulfate-deficient mice showbetter axonal regeneration after injury. Here, we investigated the regulation of keratansulfate and chondroitin sulfate biosynthesis after neuronal injuries. Several key enzymesrequired for glycosaminoglycan biosynthesis (β3GlcNAcT-7 and GlcNAc6ST-1 for keratansulfate; CS synthase-1 and C6ST-1 for chondroitin sulfate) were expressed at significantlyhigher levels in the lesion 7 days after a knife-cut injury was made to the cerebral cortex inadult mice. These increases were accompanied by increased expression of TGF-β1 and bFGF.Since microglias at the injury sites expressed both keratan sulfate and chondroitin sulfate,the effects of these cytokines were examined in microglias. TGF-β1 induced the expressionof the above-named enzymes in microglias, and consequently induced keratan sulfate andchondroitin sulfate biosynthesis as well as the expression of the chondroitin/keratan sulfateproteoglycan aggrecan in these cells. TGF-β1 also induced bFGF expression in microglias.bFGF in turn induced TGF-β1 expression in astrocytes. Astrocyte-conditioned mediumfollowing bFGF stimulation indeed induced keratan sulfate and chondroitin sulfateproduction in microglias. This production was blocked by TGF-β1-neutralizing antibody.Taken together, our data indicate that the biosyntheses of keratan sulfate and chondroitinsulfate are upregulated in common by TGF-β1 in microglias after neuronal injuries.

Published by Elsevier B.V.

Keywords:Chondroitin sulfateCytokineGliaKeratan sulfate

.ac.jp (K. Kadomatsu).ed medium; bFGF, basic fibroblast growth factor; bFGF-ACM, bFGF-activated astrocytesystem; CS, chondroitin sulfate; CSPG, chondroitin sulfate proteoglycan; EGF, epidermaletylgalactosamine; GlcA, glucuronic acid; GlcNAc, N-acetylglucosamine; GlcNAc6ST-1, N-S, keratan sulfate; KSPG, keratan sulfate proteoglycan; RT-PCR, reverse-transcriptioning growth factor-β1

Elsevier B.V.

11B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

1. Introduction

Injuries in the adultmammalian central nervous system (CNS)often leave serious functional defects, since neuronal axonsdo not spontaneously regenerate after injuries. Many factorsare thought to be involved in this inability of axons toregenerate, and can be summarized mainly into two cate-gories: (1) lack of intrinsic regenerative capacity (Neumannand Woolf, 1999; Widenfalk et al., 2001) and (2) production ofinhibitory factors (De Winter et al., 2002; Filbin, 2003; Silverand Miller, 2004). Among the inhibitory factors, myelin-associated molecules, such as Nogo, MAG and Omgp, havebeen studied extensively, although studies using their knock-out mice or their receptor-deficient mice show that thesefactors are not sufficient for the in vivo inhibition of axonalregeneration (Filbin, 2003; Silver and Miller, 2004). Other

Fig. 1 – Expression of KS and CS by microglia after brain injury.immunohistochemistry and RT-PCR. (B–E) 5D4-reactive KSPGwasreactive to KS, and Iba1 specifically recognizesmicroglias. (F–I) Mirecognizes CS. The specimens were treated with ChABC before ishow magnified photos corresponding to the squares in (D and H

chemorepulsive molecules, such as Sema3A and RGMa, alsoplay important roles in inhibiting axonal regeneration (DeWinter et al., 2002; Hata et al., 2006; Kaneko et al., 2006).

Besides these molecules, the importance of chondroitinsulfate proteoglycans (CSPGs) has been recently highlighted(Moon et al., 2001; Bradbury et al., 2002; Grimpe and Silver,2004). Proteoglycans are a group of proteins that link acidicpolysaccharides, i.e., sulfated glycosaminoglycans, of whichthere are threemain forms: chondroitin sulfate (CS)/dermatansulfate, keratan sulfate (KS) and heparan sulfate/heparin(Scott et al., 1990; Johnson-Green et al., 1991). The inhibitoryfunction of CSPGs on axonal outgrowth is largely ascribed totheir covalently attached CS-glycosaminoglycans, since theablation of CS by the use of chondroitinase ABC or a DNAenzyme as to xylosyltransferase enhances neuronal axongrowth at the site of CNS injury (Moon et al., 2001; Bradbury

(A) Schematic localization of samples forexpressed bymicroglia at the site of injury. 5D4 is specificallycroglias expressed CSPG at the site of injury. CS56 specificallymmunohistochemical staining with 5D4 antibody. (E and I)) respectively. Bar, 20 μm.

Fig. 2 – Relationships between astrocytes and expression of KS and CS after brain injury. (A–D) 5D4-reactive KSPG only partlymerged with astrocytes at the site of injury. (E–H) CS56-reactive CSPG expression merged with astrocytes at the site ofinjury. The specimens were treated with ChABC before immunohistochemical staining with 5D4 antibody. (D and H) showmagnified photos corresponding to squares in (C and G) respectively. Bar, 20 μm.

12 B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

et al., 2002; Grimpe and Silver, 2004). For example, the axongrowth of dopamine neurons is enhanced by chondroitinaseABC treatment after nigrostriatal tract transection (Moonet al., 2001). Chondroitinase ABC treatment has been shownto enhance functional recovery after spinal cord injury in a ratmodel (Bradbury et al., 2002).

We have recently demonstrated that KS is important ininhibiting axonal regeneration. Thus, after a stab wound tothe cerebral cortex, mice deficient in KS of the CNS [N-acetylglucosamine 6-O-sulfotransferase-1 (GlcNAc6ST-1)-deficient mice] show less glial scar formation andenhanced neuronal axon regeneration than do wild-typemice, even though the induction of CS expression iscomparable to that in wild-type mice (Zhang et al., 2006).KS ablation ameliorates functional disturbance after spinalcord injury (Imagama et al., unpublished data). These datasuggest that KS plays a critical role in inhibiting axonalregeneration.

This background indicates the importance of verifying themechanisms underlying the induction of KS and CS expres-sion after neuronal injury. It has been recently reported thatthe inhibition of CS chain polymerization in astrocytes via

Fig. 3 – Expression of synthases of KS, CS and aggrecan after brN-acetyl-glucosamine 6-O-sulfotransferase; β3GlcNAcT, β1,3 N-agalactose 6-O-sulfotransferase; β4GalT, β1,4-galactosyltransferaGalT, galactose transferase; GlcAT, glucuronic acid transferase; Cchondroitin sulfate N-acetyl-galactosaminyltransferase; CSGlcATchondroitin 4-O-sulfotransferase; GalNAc4,6ST, N-acetyl-galacto2-O-sulfotransferase. (C) Expression of KS and CS synthases aftecontralateral region (C), tissueswere excised 7 days after brain injalso subjected to RNA extraction. RNA expression was estimatedexpression. The results represent means±SD (n=3). *P<0.05 (ver

RNA interference decreases the inhibitory activity of CSPGagainst neurite outgrowth (Laabs et al., 2007). It was alsoreported that among various CS units the C unit is upregulatedafter neuronal injury (Properzi et al., 2005). Some other studiesalso investigated glycosaminoglycan levels in the centralnervous system with epilepsy or after injury (Perosa et al.,2002; Dobbertin et al., 2003). However, detailed analysesinvolving expression of synthetic enzymes for KS biosynthesisafter neuronal injury are still needed. Here, we demonstratethat the biosyntheses of KS and CS share a regulationmechanism in common which is mediated by TGF-β1 inmicroglias after neuronal injury.

2. Results

2.1. Glial cells expressing KS and CS

We made a knife-cut injury to the right hemisphere of thecerebral cortex in 8-week-old male mice, and localized KS andCS expression in the brain 7 days after injury. 5D4 is anantibody specific to KS, while the antibody CS56 specifically

ain injury. (A) The biosynthesis process of KS. GlcNAc6ST,cetylglucosaminyltransferase; KSGal6ST, keratan sulfatese. (B) The CS synthesis process. XylT, xylose transferase;6ST, chondroitin 6-O-sulfotransferase; CSGalNAcT,, chondroitin sulfate glucuronic acid transferase; C4ST,samine-4-sulfate 6-O-sulfotransferase; 2OST, uronosylr brain injury. For RNA extraction of injured lesion (I) andury. Postnatal day 1 brain (P1) and normal adult brain (N) wereby RT-PCR. (D) Real-time PCR for the quantification of RNA

sus contralateral region).

13B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

recognizes CS. Regions beside the lesion core were subjectedfor the immunofluorescence analysis (Fig. 1A). A significantarea positive for KS and CS expressionmergedwithmicrogliasthat were reactive to antibody Iba1 (Figs. 1B–I). The experi-ments were performed on two individual mice, and showedsimilar results.

In contrast to the results ofmicroglias, areas positive for KSexpression only partly merged with reactive astrocytes,

whereas most astrocytes appeared to express CS (Figs. 2A–H).These results were consistent with those reported for KSexpression in rat spinal cord after injury (Jones and Tuszynski,2002). Although we did not exclude the possibility that aportion of KS is produced by reactive astrocytes, our datasuggested that KS and CS production by microglias was asuitable subject for the study of glycosaminoglycan biosyn-thesis regulation.

14 B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

2.2. Expression profiles of enzymes for KS and CSbiosynthesis after brain injury

KS and CS are biosynthesized by many enzymes (Figs. 3A, B).KS is composed of repeating disaccharide units of galactose(Gal) and N-acetylglucosamine (GlcNAc), where the C6 posi-tion of GlcNAc is always sulfated. The reaction sequence forthe biosynthesis of KS consists of N-acetylglucosaminylation,6-sulfation of a GlcNAc residue exposed at the nonreducingend, and galactosylation (Fig. 3A) (Habuchi et al., 2006;Kitayama et al., 2007). CS is composed of glucuronic acid(GlcA) andN-acetylgalactosamine (GalNAc). The C2 position ofGlcA and the C4 and C6 positions of GalNAc can be sulfated.Xylose is first linked to serine residues, followed by thesequential addition of Gal, Gal and GlcA. The first GalNAc ofthe CS polymer is then added by a CSGalNAc transferase, andthen polymerization (composed of GlcA and GalNAc) isaccomplished by CS synthases (Fig. 3B) (Silbert and Sugu-maran, 2002; Sakai et al., 2007).

Among enzymes required for KS or CS biosynthesis, wepicked up several representative ones as listed in Fig. 3C, sothat we could determine at least some of the enzymesupregulated during KS and CS biosynthesis in vivo. To obtainexpression profiles of these enzymes, tissues from the injurylesion as well as from the contralateral regions (‘I’ and ‘C’,respectively, in Fig. 3C) were excised 7 days after a knife-cutinjury to the right hemisphere of the cerebral cortex in 8-week-old male mice was made (Fig. 1A). The excised tissueswere subjected to RNA extraction and further RT-PCR. RT-PCRwas also performed for uninjured 8-week-old male mousebrain (‘N’) and postnatal day 1 mouse brain (‘P1’). In this case,P1 sampleswere used as a positive control, since KS and CS areknown to be highly expressed in P1 mouse brain (Snow et al.,1990; Emerling and Lander, 1996; Miller et al., 1997; Oohiraet al., 2000; Zhang et al., 2006). We found that GlcNAc6ST-1, -3,-4 and β3GlcNAcT-7 expressionwasmarkedly increased in theinjured lesion compared with that in the contralateral regionand in uninjured brain (Fig. 3C). The expression of two otherenzymes related to KS biosynthesis, e.g., GlcNAc6ST-2 andKSGal6ST, was unchanged (Fig. 3C). As to CS biosynthesis, theexpression of CS synthase-1 and C6ST-1 was at least found tobe increased in the injured lesion compared with that in thecontralateral region and in uninjured brain (Fig. 3C). Theseresults suggest that upregulation of GlcNAc6ST-1, -3, -4 andβ3GlcNAcT-7 expression might be involved in the elevation ofKS biosynthesis, while upregulated CS synthase-1 and C6ST-1expression was related to the increase in CS biosynthesis. We

Fig. 4 – Induction of KS, CS and aggrecan expression in primary cuwere stimulated with indicated cytokines for 24 h. RT-PCR was tenzymes that increased after brain injury, as demonstrated in Fiexpression. The results represent means±SD (n=3). *P<0.05 (versindicated cytokines for 24 h. (C) Time course of enzyme expressiofor the quantification of RNA expression. The results represent mKSPG and CSPG was examined by Western blot analysis. β-actinsince the sources of the samples are the same. (E) Dose-dependecontrols are common for the 5D4 and CS56Western blots, since thinduction of TGF-β1 and bFGF expression after brain injury. Tiss(‘Control’) were collected 7 days after brain injury. RNA extracted

further examined the expression of aggrecan as an example ofa proteoglycan that carries both KS and CS (Fig. 3C). As shownin Fig. 3C, aggrecan expression also was enhanced by braininjury. We performed RT-PCR for Fig. 3C using 2 sets ofsamples, and obtained similar results. Data shown in Fig. 3C isa representative one. The expression upregulation observed inFig. 3C was confirmed by quantitative RT-PCR (Fig. 3D).

2.3. KS and CS expression in microglias under the controlof cytokines

Many cytokines have been implicated in the function ofastrocytes and/or microglias (Faber-Elman et al., 1996; Ridetet al., 1997; Schilling et al., 2001). Therefore, we next examinedthe response of primary cultured microglias to cytokines.Primary cultured microglias were exposed to various cyto-kines for 24 h, after which the expression of enzymes for KSand CS biosynthesis was examined. The examined enzymeshad been upregulated in vivo after brain injury (Figs. 3C, D),and are known to be involved in KS and CS biosynthesis(GlcNAc6ST-1 and β3GlcNAcT-7 for KS; C6ST-1 and CSsynthase-1 for CS: Figs. 3A, B). In this case, we choseGlcNAc6ST-1 instead of GlcNAc6ST-3 and -4, since ourprevious study has demonstrated that GlcNAc6ST-1 is essen-tial for inducing KS biosynthesis after neuronal injury (Zhanget al., 2006). Among the cytokines tested, TGF-β1 and EGFupregulated the expression of GlcNAc6ST-1, β3GlcNAcT-7,C6ST-1 and CS synthase-1 (Fig. 4A). Aggrecan expression wasalso upregulated only by TGF-β1 and EGF (Fig. 4A). QuantitativeRT-PCR verified that these expressions were indeed upregu-lated by TGF-β1 in a dose-dependent manner (Fig. 4B). TGF-β1

at 70 pg/ml could increase expressions of the enzymes(Fig. 4B). Data in Figs. 4A and B were obtained from microgliasexposed to TGF-β1 for 24 h. We then investigated time courseof the expressions. The upregulation of GlcNAc6ST-1 occurredas early as 12 h after TGF-β1 stimulation, whereas expressionsof other molecules showed a tendency of increase at this timepoint (Fig. 4C).

Consistent with these data, TGF-β1 increased production ofKS- and CS-bearing proteoglycans, which appeared as smearbands on Western blot analysis (Fig. 4D). EGF increased theexpression of these proteoglycans, but to a lesser extent thanTGF-β1 did (Fig. 4D). Consistent with data shown in Fig. 4B, theKS- and CS-reactivity on Western blot increased in a dose-dependent manner (Fig. 4E). The increase became apparent12 h after TGF-β1 stimulation (Fig. 4E). Taking the data in Fig.4C into account, a post-transcriptional regulation of enzyme

lturedmicroglia by cytokines. (A) Primary culturedmicrogliashen performed to examine expression of aggrecan andg. 3C. (B) Real-time PCR for the quantification of RNAus control). Primary culturedmicroglias were stimulated withns for KS and CS biosynthesis. Real-time PCR was performedeans±SD (n=3). *P<0.05 (versus 0 hour). (D) The expression ofcontrols are common for the 5D4 and CS56 Western blots,ncy and time course of KSPG and CSPG expressions. β-actine sources of the samples are the same. (F) RT-PCR showed theues from injured lesion (‘Injury’) and contralateral regionsfrom the tissues was then subjected to RT-PCR.

15B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

16 B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

productions may contribute in part to the increase in KS- andCS-reactivity on Western blot.

We then examined the expression of the cytokines listedin Fig. 4A in vivo after brain injury. TGF-β1 expression wasfound to be induced in vivo (Fig. 4F). bFGF and EGF are well-known cytokines that activate astrocytes (Faber-Elman et al.,1996; Smith and Strunz, 2005). bFGF, but not EGF, showed astriking increase of expression in our injury model (Fig. 4F).The induction of TGF-β1 and bFGF after brain injury isconsistent with previous reports (Wiessner et al., 1993;Endoh et al., 1994). Taking into account that TGF-β1 expres-sion, but not EGF expression, was induced after brain injuryalthough both molecules induced KS and CS biosynthesis inmicroglias (Figs. 4A–F), our data suggested that TGF-β1 was anintrinsic factor that could upregulate KS and CS biosynthesisin vivo.

2.4. Cross talk between astrocytes and microglias

TGF-β1 might not only upregulate KS and CS biosynthesis, butalso play a role in intercellular communication. In thiscontext, it is noteworthy that expression of TGF-β1 and bFGFwas increased after brain injury (Fig. 4F). bFGF activatesastrocytes (Meiners et al., 1993; Faber-Elman et al., 1996;Riboni et al., 2001; Brambilla et al., 2003), and primary culturedmicroglias can express bFGF (Araujo and Cotman, 1992). Basedon these backgrounds, we considered a possible link betweenTGF-β1 and bFGF. Thus, we first examined whether or notbFGF expression was induced in microglias by TGF-β1. TGF-β1

exposure for 24 h induced bFGF expression in primary culturedmicroglias in a dose-dependent manner (Fig. 5). We nextaddressed a complementary question: whether or not bFGF-activated astrocytes produced TGF-β1. Astrocytes were pri-mary cultured either with or without bFGF for 24 h, and theconditioned medium was collected (Fig. 6A). We detected theapparent induction of TGF-β1 production by astrocytes afterstimulation with bFGF (Fig. 6B). The conditioned medium wasthen applied to primary cultured microglias. Western blotanalysis confirmed that medium conditioned from bFGF-

Fig. 5 – bFGF expression in primary cultured microglias byTGF-b1. (A and B) Medium was collected from primarycultured microglias 24 h after TGF-β1 stimulation, andsubjected to Western blot analysis for bFGF expression. Tomake it sure that the samples were from cells with similarnumbers, cell lysates were examined for β-actin expressionon Western blot analysis. bFGF expression in primarycultured microglias was induced by exogenous TGF-β1 in adose-dependent manner. Experiments were performed 3times and similar results were obtained.

activated astrocytes (bFGF-ACM), but not astrocytes withoutbFGF treatment (ACM), induced the expression of proteogly-cans bearing KS and CS (Figs. 6C, D). Although bFGF alone alsoinduced KS and CS, the extent of induction was much lowerthan that by bFGF-ACM. KS and CS expression induction wasconfirmed by immunocytochemistry (Fig. 6E). The antibodythat neutralizes TGF-β1 functions effectively blocked theseinductions (Fig. 6F), indicating that the conditioned mediumfrom bFGF-activated astrocytes indeed contained functionalTGF-β1.

3. Discussion

In this study, we identified microglias as a suitable target bywhich to investigate KS and CS expression. We found thatTGF-β1 induces both KS and CS biosyntheses in microglias.Thus, this is the first report showing that the biosyntheses ofKS and CS induced after neuronal injury share a regulationmechanism in common, which is mediated by TGF-β1 (Fig. 7).As aggrecan mRNA expression is also induced by TGF-β1

(Fig. 4), the expression of core proteins of proteoglycans mightutilize this mechanism. These findings suggest that acommon regulatory mechanism contributes to biosynthesi-zing all components of KS/CSPGs (i.e., KS, CS and coreproteins), which function as inhibitory molecules as to axonalregeneration after injury.

In interpreting our data, we suggest that upregulation ofexpression is responsible for the increase in levels of theenzymes. But we cannot rule out the possibility that thedistribution of cell types is altered following injury and thelevels of enzymes simply reflect this change. However, in vitrostudies verified that primary cultured microglias barelyexpress KSPG or CSPG without the stimulation of TGF-β1,whereas expressions of enzymes for KS and CS biosynthesisare indeed upregulated by TGF-β1. Immunohistochemistrydata indicated that the main source of KSPG is microglias.Therefore, our study may suggest that the increased expres-sion through TGF-β1 may at least in part contribute to theincrease in levels of the enzymes.

The extracellular matrix of the adult CNS has a uniquecomposition. Instead of collagens, laminin-1 and fibronectin,this matrix is rich in hyaluronic acid and CSPGs (Ruoslahti,1996). KSPGs and proteoglycans bearing both KS and CSchains are also important components of the CNS extra-cellular matrix. Disorganized re-induction of proteoglycanproduction may be triggered upon injury and thereby inhibitneuronal axon regrowth. From this point of view, it isimportant to note that a common mechanism works in theinduction of KS and CS expression after neuronal injuries.This finding is interesting, as KS ablation and CS ablationindependently promote functional recovery after spinal cordinjury (Bradbury et al., 2002; Imagama et al., unpublisheddata). These collectively suggest a close functional relation-ship between KS and CS. For example, proteoglycans bearingboth KS and CS chains, other than CS-alone or KS-aloneproteoglycans, might play an important role in inhibitingaxonal regeneration.

TGF-β1 also induced bFGF expression in microglias. bFGF,in turn, induced TGF-β1 expression in astrocytes. Indeed,

Fig. 6 – Cross talk between astrocytes and microglias through TGF-b1 and bFGF. (A) Preparation of astrocyte-conditionedmedium (ACM). Primary cultured astrocytes were incubated either with or without bFGF for 24 h. (B) TGF-β1 expression inprimary cultured astrocytes was induced by exposure to bFGF-treated ACM (bFGF-ACM). TGF-β1 amounts were determinedwith ELISA specific for TGF-β1. The results represent means±SD (n=3). *P<N0.05 (versus ACM). (C, D) KSPG and CSPGexpression was induced in primary cultured microglias incubated with bFGF-ACM. Western blotting data are shown.β-actin controls are common for the 5D4 and CS56 Western blots, since the sources of the samples are the same. (E)Immunocytochemical staining for KS and CS confirmed the induction of KSPG and CSPG expression in primary culturedmicroglias incubated with bFGF-ACM. Bar, 150 μm. (F) TGF-β1-neutralizing antibody diminished KS and CS expression inmicroglias treated with bFGF-ACM.

17B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

Fig. 7 – Schematic diagram showing the cross talk between astrocytes andmicroglias. The cross talk is accomplished by TGF-β1

and bFGF, and results in the production of KS and CS, which leads to the inhibition of neuronal axon regeneration.

18 B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

conditioned medium from bFGF-treated astrocytes (bFGF-ACM) induced KS and CS expression in primary culturedmicroglias, and TGF-β1-neutralizing antibody suppressed theexpression. Our data therefore suggest that TGF-β1 and bFGFestablish cross talk between astrocytes and microglias afterneuronal injuries (Fig. 7). bFGF and TGF-β1 induce CS expres-sion in astrocytes and astrocyte migration, which together aresometimes regarded as astrocyte activation, and thus couldpromote glial scar formation (Faber-Elman et al., 1996). Wehave demonstrated that bFGF alone can induce CS expressionin microglias (Fig. 6). However, the extent of KC and CSinduction by bFGF alone in microglias is much smaller thanthat by bFGF-ACM (Fig. 6). Furthermore, immunohistoche-mistry demonstrates that a large portion of KS- and CS-positive signals merge with microglias after neuronal injury(Figs. 1, 2). Based on these findings, it is likely that the crosstalk between astrocytes and microglias mediated by TGF-β1

and bFGF plays an indispensable role in inducing KS and CSbiosynthesis after injury (Fig. 7).

It has been recently reported that ligands for the EGFreceptor (ErbB1), including EGF, TGF-α and HB-EGF, are potentinducers of CSPG production in astrocytes (Smith and Strunz,2005). A specific inhibitor for ErbB1 suppresses CSPG produc-tion by astrocytes after treatment with these ligands.Although TGF-α is consistently expressed, ErbB1 expressionis induced by brain injury. Thus, signaling through ErbB1 inastrocytes plays an important role in CSPG production afterinjury. Smith and Strunz also reported that TGF-β1 inducesCSPG production to some extent (Smith and Strunz, 2005).CSPG production by astrocytes through growth factors orcytokines has also been established in other studies (Asher etal., 2000; Properzi et al., 2005). In this context, our presentstudy has provided an additional notion that KS- and CS-biosynthesis share a regulatory mechanism in common inmicroglias. Taken together, the common mechanism of KS-and CS-biosynthesis may include activation of both astrocytesand microglias by growth factors, such as TGF-β1, bFGF andTGF-α.

TGF-β's actions on inflammatory cells after neuronal injuryare complex and contextual. TGF-β1 induces ramification inmicroglias (Schilling et al., 2001) as well as antioxidant geneexpression in microglia in vitro, and thusmodulates microglialredox status and activation state (Min et al., 2006). Astrocytesappear to suppress microglial production of nitric oxidethrough a mechanism involving activation of the latent formof TGF-β (Vincent et al., 1997, 1998). Overexpression orinjection of TGF-β1 in the CNS suppresses microglial activa-tion and consequently the induction of the pro-inflammatory

chemokines after hypoxic/ischemic injury (Gross et al., 1993;McNeill et al., 1994; Henrich-Noack et al., 1996). In contrast, theoverproduction of TGF-β1 by astrocytes aggravates CNSinflammation (Wyss-Coray et al., 1995), and injection of TGF-β1-neutralizing antiserum suppresses CNS inflammation afterinjury (Logan et al., 1999; King et al., 2004). Our present studyhas shown that bFGF-stimulated astrocytes produce sufficientlevels of TGF-β1 to induce KS and CS expression in microglias.It is noteworthy that raised levels of TGF-β correlate with thedeposition of scar materials after traumatic injury to the CNS(Logan et al., 1992). A combination of antibodies to TGF-β1 andTGF-β2 reduces scar formation after the nigrostriatal tract iscut (Moon and Fawcett, 2001). Furthermore, mice deficient insmad3, a TGF-β-specific intracellular signaling molecule,display reduced scar formation after a stab wound to thecerebral cortex (Wang et al., 2007). These results suggest thatKS and CS production in microglias via TGF-β1 contributes tothe inhibition of axonal regeneration after neuronal injury.

Microglias show morphological changes, including rami-fied and amaeboid, the latter being regarded as an activatedform (Giulian et al., 1986). KS is reportedly expressed inramified resident microglias, and is suppressed in experi-mental autoimmune encephalomyelitis, a T-helper type 1cell-mediated CNS inflammation model (Jander and Stoll,1996). After cytokines stimulation, expression of MHC class II,amarker ofmicroglial activation, shows amirror image profileto that of KS (Jander et al., 2000). Thus, interferon-γ inducesMHC class II but not KS, while TNF-α induces KS but not MHCclass II. In contrast, we and others have shown KS expressioninduction in microglias close to the injury center afterneuronal injury (Jones and Tuszynski, 2002). Furthermore,KS-deficient mice exhibit more axonal regeneration after astab wound in the cerebral cortex (Zhang et al., 2006). Thepresent study has also demonstrated that microglias expressnot only KS but also CS upon stimulation with TGF-β1. Thesedata indicate that KS expression in microglias is contextual,but that it indicates a hazardous status in some types of CNSinjuries.

4. Experimental procedures

4.1. Controlled cortical knife-cut injury

C57BL/6J male mice, 8 weeks of age, were maintained intemperature-controlled rooms on a 12 h light/dark cycle. Themice were maintained in the animal facilities of NagoyaUniversity. All experiments were performed in accordance

19B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

with protocols approved by the institutional animal commit-tee. C57BL/6J mice were anesthetized and placed in a stereo-taxic frame. Cortical injury was induced in the right parietalcortex. For the knife-cut model, a knife cut (2.5 mm depth and6.0 mm length) 1.0 mm lateral to the bregma was made, andthe lesions, the contralateral regions and the correspondingregions from normal mice were isolated at the indicated daysafter injury.

4.2. RT-PCR and real-time PCR

Total RNA was extracted from cells or from brain tissues andwas analyzed by RT-PCR and real-time PCR. All the PCRproducts were taken from the same tissue. GAPDH was usedas an internal control to make it sure that the same inputamounts were applied for RT-PCR. RT-PCR primers used aresummarized in Table 1, under the following PCR conditions:94 °C for 30 s, 57 °C for 30 s, 72 °C for 1min, 30 cycles. Real-timePCRwas performed using the TaqManGene Expression Assaysand 7500 Real-time PCR system (Applied Biosystems, Foster,CA). Real-time PCR primers and probes were from AppliedBiosystems (GlcNAc6st-1: Assay ID, Mm00490018-gl;β3GlcNACt-7: Assay ID, Mm00507533-ml; C6ST-1: Assay ID,Mm00489736-ml; CS synthase: Assay ID, Mm01319178-ml;aggrecan: Assay ID, Mm00573424-gl; GAPH: Assay ID,Mm03302249-gl).

4.3. Immunohistochemistry

Tissues were cut into 5-μm sections with a cryostat andmounted on glass slides. The sections were fixed with coldacetone for 5 min, dried, and then blocked in phosphate-buffered saline (PBS) containing 10% goat serum albumin.Sections were then incubated with monoclonal Cy3-conjuga-tedovernight at 4 °C or 1 h at room temperature. After rinsing,

Table 1 – Nucleotide sequence of RT-PCR primers for target gen

Gene name Forward primer (5′–

GlcNAc6ST-1 AAGCCTACAGGTGGTGCGlcNAc6ST-2 TTCTCAGCCTGCAGGCCCGlcNAc6ST-3 AGACAGCCAAGGCGCTGGlcNAc6ST-4 TGGTGGTCATCAAAGACβ3GlcNAcT-7 CTATGCTGACATCCTACAβ4GalT-4 AGAACTGGGACTGCTTCKSGal6ST-1 ATGATTGTGATCTCTACTC4ST-1 ACTCATCTACTGCTATGTC6ST-1 ACCACTTGACTCAGTTCGalNAcT-1 AGGCGGCTTGCAGAAGTGalNAc4,6ST GAAGATTACCTGGACCT2OST CATGTCCACTTCCTCAACGlcAT-1 AGGTACAGGATGGCCGCCS synthase-1 TGACATGCAGGTCCTGCCXylT-1 GCATCCACACCCTCAGCGGalT-1 CGACCGTAATGCCTACAGalT-2 CTTTGCCATCGCCATGGAAggrecan CAGGGTCACTGTTACCGCEGF TCCTAGAGAAACACCAAbFGF ACACGTCAAACTACAACTGF-β1 CACCATCCATGACATGAGAPDH GGTGGAGGTCGGAGTCA

sections stained with anti-Iba1 antibody were incubated withCy3-conjugatedanti-rabbit IgG (500×dilution; Invitrogen,Carls-bad, CA) for 30 min at room temperature, then rinsed.Monoclonal biotin-conjugated anti-CS CS-56 antibody (Seika-gaku, Tokyo, Japan) or monoclonal biotin-conjugated anti-KS5D4 antibody (Seikagaku) at 100×dilution was then incubatedin a blocking solution overnight at 4 °C or 1 h at room tempe-rature. After rinsing, Cy2-conjugated streptavidin (JacksonImmunoResearch, West Grove, PA) was incubated at 500×dilution for 30 min at room temperature, rinsed, and thenmounted with FluorSave (Calbiochem, San Diego, CA) andexamined by confocal microscopy (MRC 1024; Bio-Rad Labora-tories, Tokyo, Japan).

4.4. Cell culture

Primary cultures of cerebral cortical astrocytes were preparedfromnewborn C57BL/6Jmice as previously described (Smith etal., 1990; Allamanet al., 2004). Briefly, forebrainswere removedaseptically from the skulls, the meninges were excised care-fully under a dissecting microscope, and the cortices wereisolated. The small tissues obtained by mincing the corticeswere cultured in flasks in DMEM containing 10% FCS, thenincubated at 37 °C in a humidified atmosphere containing 5%CO2. The culture medium was renewed every 3–5 days.Experiments were performed on confluent 30-day-old cul-tures. Over 95%of cells obtainedwereGFAP-positive. Astrocytecells were used in 5×105 cells in a 3.5-cm dish.

Microglia-enriched cultures were obtained using themethod of Giulian et al. (1986). Briefly, small pieces of tissueswere obtained by mincing the cortices as described for theastrocyte primary culture and then were cultured in flasks inMi-medium (DMEM, 10% FCS, 0.2% glucose and insulin 5 μg/ml). Themixed glial culture grown for 21 dayswas subjected toshaking at 200 rpm on a gyratory shaker for 30 min. The

es

3′) Reverse primer(5′–3′)

GAA CAGGACTGTTAACCCGCTCATCT GTTCTTGTTGAGATCTGACCTGCA AACTGTCCATGCCTCTTGGCGTG GTCATATTGAATGCGAAGGCGT ACTTGTGTACCACAAGCATG

GTA TGGCATATAAAGGGTTGTGCTT TTCTGTCTTCTTCATTGGATGC CCTCCAGTGTCTCATACTTG

TGCGTGTTCTTCTGGATCATG CCTCTGTGGCTTTATCCAGGCTT GTGCATGGTATACTTGACATTT CGTAGTGGTAGAACTCGTACGTT TGTAGCTGCTCCTCCTGCTTTG AGAGGTCCACATCCTCCAGCAT GCTGCAATGACGTTGACAGG

GGT CCGGTCAAACCTCTGAGGATCA AATGCCGTCCCACTTCATGCCA GTGCAGGTGATTCGAGGCTCGAC AGCAGTGATTAGCCGTGGAATCC CATTGGAAGAAACAGTATGGACC GCACAATCATGTTGGACAAACG CAAAGTTGTCATGGATGACC

20 B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

detached cells (mainly microglia) were reseeded in freshculture flasks, and after 2 h any contaminating oligodendro-cyte progenitors were detached with Tris-buffered salinecontaining 1 mM EDTA. This procedure routinely provides afirmly attached homogeneous population of microglia. Micro-glia were cultured in DMEM containing 10% FBS. More than95% of cells obtainedwere found to be Iba1-positive. Microglialcells were used in 1×105 cells in a 3.5-cm dish.

In all experiments, microglias and astrocytes were grown foran additional 24 h after cytokines were added: IL-1β 10 ng/ml(Sigma), TNF-α 20ng/ml (PeproTechHouse, London,UK), TGF-β1

35 ng/ml (R&D, Minneapolis, MN), EGF 20 ng/ml (BiomedicalTechnologies Inc., Stoughton, MA) or bFGF 20 ng/ml (PeproTechHouse, London, UK) or conditioned medium.

4.5. Western blot analysis

Samples of the supernatant fraction and cells were collectedafter centrifuging at 10,000 g for 15 min and were separated byelectrophoresis on 5% SDS–PAGE. Proteins were then blottedonto nitrocellulose membranes. Blots were blocked with 5%fat-free dry milk in PBS for 60 min and incubated overnight at4 °C with the primary antibody [anti-KS 5D4 (1000×dilution;Seikagaku), anti-CS CS56 (1000×dilution; Seikagaku) or anti-aggrecan (1000×dilution)] in PBS containing 0.3% Triton X-100.They were washed and then were incubated with a secondantibody [horseradish peroxidase-conjugated goat anti-mouseIgG (5000× dilution), anti-mouse IgM (5000×dilution) or anti-rabbit IgG (5000×dilution; Jackson ImmunoResearch)] in PBScontaining 0.3% Triton X-100 at room temperature for 60 min.Anti-bFGF antibody (1000×dilution; Biomedical Technologies,Stoughton, MA) and anti-β-actin antibody (100,000×dilution;Sigma) were also used as indicated. Bound antibodies werevisualized with an ECL Western blotting detection kit (GEHealthcare, Buckinghamshire, UK).

4.6. Enzymatic treatment

Before immunohistochemistry and Western blotting by theuse of 5D4 (anti-KS) antibody were performed, tissue speci-mens or cell media were treated with chondroitinase ABC(1 mU/μg protein, 0.1 M Tris–acetate, pH 7.3; Seikagaku)overnight at 37 °C.

4.7. Quantification of TGF-β1 andTGF-β1-neutralizing antibody

The amounts of TGF-β1 were estimated by ELISA (R&D,Minneapolis, MN). TGF-β1-neutralizing monoclonal antibody(HB 9849, ATCC, Manassas, VA) was obtained from theAmerican Type Culture Collection (Manassas, VA); normalmouse IgGwas from Invitrogen. Anti-TGF-β1 or normalmouseIgG at 5 μg/ml was added to the microglia medium stimulatedby bFGF-ACM for 24 h. After 24 h, KSPG and CSPG expression inmicroglias was examined by Western blotting.

4.8. Statistical analysis

Statistical analysis was performed using the Student's t-test.P<0.05 was considered statistically significant.

Acknowledgments

We thank Hideto Watanabe, Nobuo Sugiura and Kenji Uchi-mura for their helpful comments on this manuscript, andTakashi Muramatsu for his continuous support of this study.This work was supported by the 21st COE program and theGlobal COE program, MEXT, Japan; Grants-in-Aid, MEXT(18390099 and 20390092 to K.K.) and by theUehara Foundation.

R E F E R E N C E S

Allaman, I., Pellerin, L., Magistretti, P.J., 2004. Glucocorticoidsmodulate neurotransmitter-induced glycogen metabolism incultured cortical astrocytes. J. Neurochem. 88, 900–908.

Araujo, D.M., Cotman, C.W., 1992. Basic FGF in astroglial,microglial, and neuronal cultures: characterization of bindingsites and modulation of release by lymphokines and trophicfactors. J. Neurosci. 12, 1668–1678.

Asher, R.A., Morgenstern, D.A., Fidler, P.S., Adcock, K.H., Oohira, A.,Braistead, J.E., Levine, J.M., Margolis, R.U., Rogers, J.H., Fawcett,J.W., 2000. Neurocan is upregulated in injured brain and incytokine-treated astrocytes. J. Neurosci. 20, 2427–2438.

Bradbury, E.J., Moon, L.D., Popat, R.J., King, V.R., Bennett, G.S., Patel,P.N., Fawcett, J.W., McMahon, S.B., 2002. Chondroitinase ABCpromotes functional recovery after spinal cord injury. Nature416, 636–640.

Brambilla, R., Cottini, L., Fumagalli, M., Ceruti, S., Abbracchio, M.P.,2003. Blockade of A2A adenosine receptors prevents basicfibroblast growth factor-induced reactive astrogliosis in ratstriatal primary astrocytes. Glia 243, 190–194.

De Winter, F., Holtmaat, A.J., Verhaagen, J., 2002. Neuropilin andclass 3 semaphorins in nervous system regeneration. Adv. Exp.Med. Biol. 515, 115–139.

Dobbertin, A., Rhodes, K.E., Garwood, J., Properzi, F., Heck, N.,Rogers, J.H., Fawcett, J.W., Faissner, A., 2003. Regulation ofRPTPbeta/phosphacan expression and glycosaminoglycanepitopes in injured brain and cytokine-treated glia. Mol. Cell.Neurosci. 24, 951–971.

Emerling, D.E., Lander, A.D., 1996. Inhibitors and promoters ofthalamic neuron adhesion and outgrowth in embryonicneocortex: functional association with chondroitin sulfate.Neuron 17, 1089–1100.

Endoh, M., Pulsinelli, W.A., Wagner, J.A., 1994. Transient globalischemia induces dynamic changes in the expression of bFGFand the FGF receptor. Brain Res. Mol. Brain Res. 22, 76–88.

Faber-Elman, A., Solomon, A., Abraham, J.A., Marikovsky, M.,Schwartz, M., 1996. Involvement of wound-associated factorsin rat brain astrocyte migratory response to axonal injury: invitro simulation. J. Clin. Invest. 97, 162–171.

Filbin, M.T., 2003. Myelin-associated inhibitors of axonalregeneration in the adult mammalian CNS. Nat. Rev. Neurosci.4, 703–713.

Giulian, D., Baker, T.J., Shih, L.C., Lachman, L.B., 1986. Interleukin 1of the central nervous system is produced by ameboidmicroglia. J. Exp. Med. 164, 594–604.

Grimpe, B., Silver, J., 2004. A novel DNA enzyme reducesglycosaminoglycan chains in the glial scar and allowsmicrotransplanted dorsal root ganglia axons to regeneratebeyond lesions in the spinal cord. J. Neurosci. 24, 1393–1397.

Gross, C.E., Bednar, M.M., Howard, D.B., Sporn, M.B., 1993.Transforming growth factor-beta 1 reduces infarct size afterexperimental cerebral ischemia in a rabbit model. Stroke 24,558–562.

Habuchi, H., Habuchi, O., Uchimura, K., Kimata, K., Muramatsu, T.,2006. Determination of substrate specificity of

21B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

sulfotransferases and glycosyltransferases (proteoglycans).Methods Enzymol. 416, 225–243.

Hata, K., Fujitani, M., Yasuda, Y., Doya, H., Saito, T., Yamagishi, S.,Mueller, B.K., Yamashita, T., 2006. RGMa inhibition promotesaxonal growth and recovery after spinal cord injury. J. Cell Biol.173, 47–58.

Henrich-Noack, P., Prehn, J.H., Krieglstein, J., 1996. TGF-beta 1protects hippocampal neurons against degeneration caused bytransient global ischemia. Dose–response relationship andpotential neuroprotective mechanisms. Stroke 27, 1609–1614discussion 1615.

Jander, S., Stoll, G., 1996. Downregulation of microglial keratansulfate proteoglycans coincident with lymphomonocyticinfiltration of the rat central nervous system. Am. J. Pathol. 148,71–78.

Jander, S., Schroeter, M., Fischer, J., Stoll, G., 2000. Differentialregulation of microglial keratan sulfate immunoreactivity byproinflammatory cytokines and colony-stimulating factors.Glia. 30, 401–410.

Johnson-Green, P.C., Dow, K.E., Riopelle, R.J., 1991. Characteriza-tion of glycosaminoglycans produced by primary astrocytes invitro. Glia 4, 314–321.

Jones, L.L., Tuszynski, M.H., 2002. Spinal cord injury elicitsexpression of keratan sulfate proteoglycans by macrophages,reactive microglia, and oligodendrocyte progenitors.J. Neurosci. 22, 4611–4624.

Kaneko, S., Iwanami, A., Nakamura, M., Kishino, A., Kikuchi, K.,Shibata, S., Okano, H.J., Ikegami, T., Moriya, A., Konishi, O.,Nakayama, C., Kumagai, K., Kimura, T., Sato, Y., Goshima, Y.,Taniguchi, M., Ito, M., He, Z., Toyama, Y., Okano, H., 2006. Aselective Sema3A inhibitor enhances regenerative responsesand functional recovery of the injured spinal cord. Nat. Med. 12,1380–1389.

King, V.R., Phillips, J.B., Brown, R.A., Priestley, J.V., 2004. The effectsof treatment with antibodies to transforming growth factorbeta1 and beta2 following spinal cord damage in the adult rat.Neuroscience 126, 173–183.

Kitayama, K., Hayashida, Y., Nishida, K., Akama, T.O., 2007.Enzymes responsible for synthesis of corneal keratan sulfateglycosaminoglycans. J. Biol. Chem. 282, 30085–30096.

Laabs, T.L., Wang, H., Katagiri, Y., McCann, T., Fawcett, J.W., Geller,H.M., 2007. Inhibiting glycosaminoglycan chainpolymerization decreases the inhibitory activity ofastrocyte-derived chondroitin sulfate proteoglycans.J. Neurosci. 27, 14494–14501.

Logan, A., Frautschy, S.A., Gonzalez, A.M., Sporn, M.B., Baird, A.,1992. Enhanced expression of transforming growth factor beta1 in the rat brain after a localized cerebral injury. Brain Res. 587,216–225.

Logan, A., Baird, A., Berry, M., 1999. Decorin attenuates gliotic scarformation in the rat cerebral hemisphere. Exp. Neurol. 159,504–510.

McNeill, H., Williams, C., Guan, J., Dragunow, M., Lawlor, P.,Sirimanne, E., Nikolics, K., Gluckman, P., 1994. Neuronal rescuewith transforming growth factor-beta 1 afterhypoxic–ischaemic brain injury. NeuroReport 5, 901–904.

Meiners, S., Marone, M., Rittenhouse, J.L., Geller, H.M., 1993.Regulation of astrocytic tenascin by basic fibroblast growthfactor. Dev. Biol. 60, 480–493.

Miller, B., Sheppard, A.M., Pearlman, A.L., 1997. Developmentalexpression of keratan sulfate-like immunoreactivitydistinguishes thalamic nuclei and cortical domains. J. Comp.Neurol. 380, 533–552.

Min, K.J., Yang, M.S., Kim, S.U., Jou, I., Joe, E.H., 2006. Astrocytesinduce hemeoxygenase-1 expression in microglia: a feasiblemechanism for preventing excessive brain inflammation.J. Neurosci. 26, 1880–1887.

Moon, L.D., Fawcett, J.W., 2001. Reduction in CNS scar formationwithout concomitant increase in axon regeneration following

treatment of adult rat brain with a combination of antibodiesto TGFbeta1 and beta2. Eur. J. Neurosci. 14, 1667–1677.

Moon, L.D., Asher, R.A., Rhodes, K.E., Fawcett, J.W., 2001.Regeneration of CNS axons back to their target followingtreatment of adult rat brain with chondroitinase ABC. Nat.Neurosci. 4, 465–466.

Neumann, S., Woolf, C.J., 1999. Regeneration of dorsal columnfibers into and beyond the lesion site following adult spinalcord injury. Neuron 23, 83–91.

Oohira, A., Matsui, F., Tokita, Y., Yamauchi, S., Aono, S., 2000.Molecular interactions of neural chondroitin sulfateproteoglycans in the brain development. Arch. Biochem.Biophys. 374, 24–34.

Perosa, S.R., Porcionatto, M.A., Cukiert, A., Martins, J.R., Passeroti,C.C., Amado, D., Matas, S.L., Nader, H.B., Cavalheiro, E.A., Leite,J.P., Naffah-Mazzacoratti, M.G., 2002. Glycosaminoglycanlevels and proteoglycan expression are altered in thehippocampus of patients with mesial temporal lobe epilepsy.Brain Res. Bull. 58, 509–516.

Properzi, F., Carulli, D., Asher, R.A., Muir, E., Camargo, L.M., vanKuppevelt, T.H., ten Dam, G.B., Furukawa, Y., Mikami, T.,Sugahara, K., Toida, T., Geller, H.M., Fawcett, J.W., 2005.Chondroitin 6-sulphate synthesis is up-regulated ininjured CNS, induced by injury-related cytokines andenhanced in axon-growth inhibitory glia. Eur. J. Neurosci. 21,378–390.

Ridet, J.L., Malhotra, S.K., Privat, A., Gage, F.H., 1997. Reactiveastrocytes: cellular and molecular cues to biological function.Trends Neurosci. 20, 570–577.

Riboni, L., Viani, P., Bassi, R., Giussani, P., Tettamanti, G., 2001.Basic fibroblast growth factor-induced proliferation of primaryastrocytes. Evidence for the involvement of sphingomyelinbiosynthesis. J. Biol. Chem. 276, 12797–12804.

Ruoslahti, E., 1996. Brain extracellular matrix. Glycobiology 6,489–492.

Sakai, K., Kimata, K., Sato, T., Gotoh, M., Narimatsu, H., Shinomiya,K., Watanabe, H., 2007. Chondroitin sulfateN-acetylgalactosaminyltransferase-1 plays a critical role inchondroitin sulfate synthesis in cartilage. J. Biol. Chem. 282,4152–4161.

Schilling, T., Nitsch, R., Heinemann, U., Haas, D., Eder, C., 2001.Astrocyte-released cytokines induce ramification and outwardK+ channel expression in microglia via distinct signallingpathways. Eur. J. Neurosci. 14, 463–473.

Scott, J.E., Cummings, C., Greiling, H., Stuhlsatz, H.W., Gregory, J.D., Damle, S.P., 1990. Examination of corneal proteoglycans andglycosaminoglycans by rotary shadowing and electronmicroscopy. Int. J. Biol. Macromol. 12, 180–184.

Silbert, J.E., Sugumaran, G., 2002. Biosynthesis ofchondroitin/dermatan sulfate. IUBMB Life 54, 177–186.

Silver, J., Miller, J.H., 2004. Regeneration beyond the glial scar.Nat. Rev. Neurosci. 5, 146–156.

Smith, G.M., Strunz, C., 2005. Growth factor and cytokineregulation of chondroitin sulfate proteoglycans by astrocytes.Glia 52, 209–218.

Smith, G.M., Rutishauser, U., Silver, J., Miller, R.H., 1990.Maturation of astrocytes in vitro alters the extent andmolecular basis of neurite outgrowth. Dev. Biol. 138, 377–390.

Snow, D.M., Steindler, D.A., Silver, J., 1990. Molecular and cellularcharacterization of the glial roof plate of the spinal cord andoptic tectum: a possible role for a proteoglycan in thedevelopment of an axon barrier. Dev. Biol. 138, 359–376.

Vincent, V.A., Tilders, F.J., Van Dam, A.M., 1997. Inhibition ofendotoxin-induced nitric oxide synthase production inmicroglial cells by the presence of astroglial cells: a role fortransforming growth factor beta. Glia 19, 190–198.

Vincent, V.A., Lowik, C.W., Verheijen, J.H., de Bart, A.C., Tilders,F.J., Van Dam, A.M., 1998. Role of astrocyte-derived tissue-typeplasminogen activator in the regulation of

22 B R A I N R E S E A R C H 1 2 6 3 ( 2 0 0 9 ) 1 0 – 2 2

endotoxin-stimulated nitric oxide production by microglialcells. Glia 22, 130–137.

Wang, Y., Moges, H., Bharucha, Y., Symes, A., 2007. Smad3 nullmice display more rapid wound closure and reduced scarformation after a stab wound to the cerebral cortex. Exp.Neurol. 203, 168–184.

Widenfalk, J., Lundstromer, K., Jubran, M., Brene, S., Olson, L., 2001.Neurotrophic factors and receptors in the immature and adultspinal cord after mechanical injury or kainic acid. J. Neurosci.21, 3457–3475.

Wiessner, C., Gehrmann, J., Lindholm, D., Töpper, R., Kreutzberg,G.W., Hossmann, K.A., 1993. Expression of transforminggrowth factor-beta 1 and interleukin-1 beta mRNA in rat brain

following transient forebrain ischemia. Acta Neuropathol. 86,439–446.

Wyss-Coray, T., Feng, L., Masliah, E., Ruppe, M.D., Lee, H.S., Toggas,S.M., Rockenstein, E.M., Mucke, L., 1995. Increased centralnervous system production of extracellular matrixcomponents and development of hydrocephalus in transgenicmice overexpressing transforming growth factor-beta 1. Am. J.Pathol. 147, 53–67.

Zhang, H., Muramatsu, T., Murase, A., Yuasa, S., Uchimura, K.,Kadomatsu, K., 2006. N-Acetylglucosamine6-O-sulfotransferase-1 is required for brain keratan sulfatebiosynthesis and glial scar formation after brain injury.Glycobiology 16, 702–710.