Basic fibroblast growth factor expression and tenascin C immunoreactivity after partial unilateral hemitransection of the rat brain

BRAIN RESEARCH ELSEVIER

Brain Research 730 (I 996) I - 16

Research report

Basic fibroblast growth factor expression and tenascin C immunoreactivity after partial unilateral hemitransection of the rat brain A n d r e a L i p p o l d t ~,.b,*, B e t h A n d b j e r b, H e l l m u t G e r s t ~', D e t l e v O a n t e n ~', Kiell F u x e h "' <'ffa~ l)<'lhrih'k-('e#lter lbr Molecular Medicine (MDC). t{vp,,rten,~ion Research, Rohert-RiJs.vle-Str. 10, 13125 Berlin Bm'h, ( , e n m m v i, l)~lutrtmeHr of Ne,rosciem'e. Dil ision of Cellular and Molecular Ncurochemi,~t*3", Karoli#t,~l,'a lnstitule 5'lo<'/,h~ds*l. 5tv,'deH

Accepted 13 February 1996

~bstract Basic fibrobhist growth factor (bFGF) gene expression as well as its inmmnoreactivity were studied after partial unilaleral tlemitransection of the rat brain during a time course of 24 h, 72 h, 7 and 14 days. The mechanical injury restllted m a global increase of bFGF gene expression at the 24-h time interval. This global increase was seen at the ipsilateral site at the level of the lesion as well as rostral to the lesion in the ipsilateral hemisphere. The upregulation in bFGF gene expression was in most of the areas investigated due to an upregulation in glial cells as seen by means of nonradioactive in situ hybridization compared with immunocytochemistry lot glial fibrillary acidic protein (GFAP). Basic F G F immunoreactivity (IR) was increased around the lesion in glial cell nuclei 7 days after the injury. This increase was also detected in GFAP positive glial cells surrounding small vessels in the lesioned area. Moreover, in the present paper we demonstrate increased tenascin immunoreactivity in the lesioned area 7 days after injury. The tenascin IR was increased at the edges of the lesion as well as in vessel like structures. The tenascin IR was partially codistributed ~aith GFAP 1R in the lesioned area. The lesion was also characterized by an increase in vimentin IR as well as in laminin IR. It is suggested thai lhe observed changes in lhe expression of bFGF, matrix proteins (laminin, tenascin) and intermediate filaments (vimentin) are inw~h,ed m hi) lissue repair. (b) protection of neuronal cells from excitotoxic influences and (c) lkmnation of new vessels in the lesioned area. KevworcLw bFGF: Tenascin; I,aminin: Hemitransection; Rat brain: Hybridization, m situ: mRNA

I. Introduction

Basic fibroblast growth factor (bFGF) is one of the main components influencing brain de- and regeneration [16,17,20,23,36,41]. Its mRNA is expressed in high amounts in the CA2 region of the hippocampal formation, the fasciola cinereum and the induseum griseum in the rat brain [14]. The bFGF protein is seen within large populanons of astroglial cell nuclei and of nerve cell bodies of the brain [21,64]. Basic FGF is acting as a mitogenic and angiogenic agens, described in many tissue systems, and has a role in wound repair [12,16,22,25,44,55,58]. It belongs to the family of heparin binding proteins [6,18,28,65] and enhances survival and growth of neurons and glial cells in several brain regions in vivo and in vitro [45,60,61]. In lesions of the nigrostriatal dopaminergic system induced by 6-OH-dopamine in the rat brain bFGF mRNA and protein have been demonstrated to be upregulated in the

(2orresponding author at address "a'. Fax: +49 (30) 949-4161.

nigral astroglial cells 4 - 2 4 h after injection. This upregulation was also seen in the neostriatum on the ipsihlteral side [7]. In other types of lesions bFGF upreguhition has been observed shortly after the injury and to be prominent from several days until weeks in thc affected tissue [17,20,23,30,56]. It may act in the lesioned tissue to p,'otect neurons by stabilizing intracellular Ca: '-levels. by, reduction of Ca 2- influx, enhanced Ca 2 extrusion or buffering and by stabilizing or improving mitochondrial function 4~. Furthermore, bFGF is able to protect neurons against excitotoxicity, e.g., by suppressing the expression of the NMDA receptor protein in the hippocampal neurons [9,41]. Another function of bFGF is its ability to induce expression of extracellular matrix components. It has been shown to influence the expression of tenascin in primitive neuroectodermal tumor cells [48] and in cultured rat cerebral cortical astrocytes [43J. The effect of lenascin induction in astrocytes was transient and enhanced by heparin. Tenascin is an extracellular matrix glycoprotein and has a limited spatiotemporal distribution in embryonal tissue tindergoing morphogenesis [2,49,63]. In the adul! tenascin is

01)06-8993/96/SI5.1)0 Copyright i 1996 Elsevier Science B.V. All rights leserved. PSI S 0 0 0 6 - ~ o o ~ i 9 ~ i t ) 0 2 4 2 - 9

2

A. Lippoldt el al. / Brain Research 730 (1996) 1 - 1 6

found in small amounts, e.g., in the central nervous system and in regions of the basement membrane [2,15]. Tenascin is involved in neural and nonneuronal cellular interactions. It is expressed in astrocytes [10,52] but also in the vasculature in subendothelial and medial regions [38]. In the adult organism tenascin may be reinduced during wound healing and regeneration [39,40]. In the present paper we report on the induction of bFGF after unilateral partial hemitransection at the di- and mesencephalic junction [1] of the rat brain on the basis of in situ hybridization and immunocytochemistry. Furthermore, events influenced by bFGF were investigated such as angiogenesis and the expression of the extracellular matrix glycoprotein tenascin. The aim of the present experiments was to study bFGF gene expression in head trauma and its possible involvement in angiogenesis and expression of tenascin.

2. Materials and methods 2.1. Animals Twenty-eight male pathogen-free adult Sprague-Dawley rats (200 g average body weight) were used for the experiment. They were kept under regular lighting conditions (lights on 06.00 h and off 20.00 h) at constant temperature (23°C) and had free access to food pellets and water. 2.2. Partial meso-diencephalic hemitransection The anaesthetized (chloralhydrate, 0.35 g / k g body weight) rats were mounted in a stereotactic instrument (David Kopf, Tujunga, CA, USA). After drilling out a circular-shaped piece of scull bone, the dura was penetrated with a rectangular knife (4 mm wide, 1 mm thick at the edge), which was tilted 20 ° to the frontal plane and

l Fig. 1. Schematical illustration of the partial hemitransection. The hemitransection was performed as indicated in Section 2. The hemitransection causes a large lesion within the t h a l a m u s / h y p o t h a l a m u s . Bregma: 4.5 mill.

inserted 1 mm caudal to Bregma and l mm lateral to the midline (Bregma: - 1 . 0 0 mm, L + 1.00 ram). The knife was lowered 10 mm ventrocaudally into the right hemisphere, withdrawn 2 mm, moved laterally until touching the bone and then removed. The lesion cuts ascending and descending pathways including the nigrostriatal dopaminergic system with the exception of its most medial components [1] (Fig. 1). As controls sham-operated (four rats) rats were used. The rats were studied at several timepoints after hemitransection (four in each group): 24 h, 72 h, 7 days and 14 days. For immunocytochemistry two time points were studied (three rats per group): 72 h and 7 days. 2.3. Dissection and fixation For in situ hybridization studies the animals were anaesthetized with pentobarbital (60 m g / k g body weight) and transcardially perfused through the left cardiac ventricle with icecold 0.9% NaC1. The brains were removed and snap frozen in isopentan ( - 3 5 ° C ) . All material was sectioned at a 12 ~m thickness in a cryostat (Leica, Frigocut, Germany). For immunocytochemistry the rats were anaesthetized with pentobarbital and perfused through the left cardiac ventricle with ice cold 0.9% saline and subsequentely with 300 ml icecold 2% paraformaldehyde/0.1% picric acid. The brains were taken out and postfixed for 90 min in the same fixative at 4°C. Subsequently, the brains were transferred to 10% sucrose in phosphate buffer and immersed for 48 h at 4°C and snap frozen with COy. The brains were cut at 12 Ixm thickness in a cryostat (Leica, Frigocut, Germany). 2.4. RNA probe synthesis The RNA probes were synthesized by in vitro transcription from a bFGF fragment, amplified by polymerase chain reaction (PCR) according to Bean et al. [3] containing the promotor regions for SP6 RNA polymerase at the 3' end and for T7 RNA polymerase at the 5' end. The fragment represents the bFGF-cDNA sequence between nucleotide 442 and nucleotide 660 [33]. Transcription with T7 RNA polymerase (Boehringer, Mannbeim, Germany) resulted ill a 218 nucleotide long sense RNA, whereas a 218 nucleotide long antisense RNA was obtained after in vitro transcription using SP6 RNA polymerase (Boehringer, Mannheim, Germany). The labelling was performed with either >S-c~-UTP (Dupont NEN, Boston, MA. USA) o1 D i g o x i g e n i n - l l - U T P ( D I G - I 1 - U T P ) (Boehringer, Mannheim, Germany). The radioactive labelled transcripts were purified using Nensorb columns (Dupont NEN, Boston, MA, USA) and the DIG-1 l-UTP-labelled probes were ethanol precipitated. Concentrations were measured in a UV-spectrophotometer at 260 nm. The labelling quality was checked by separating the probes via a denaturat-

A. Lippoh# el al./Brain Re,~earch 730 (1996) / 16

ing formaldehyde gel electrophoresis. The specific activity of the probes was in the order of 5 • l0 s dpm/tj, g RNA. For detection of the digoxygenin by an alkaline phos-

phatase conjugated antibody (Boehringer, Mannheim, Germany), the RNA was transferred onto a Nylon membrane (Nytran N, Schleicher and Schtill, Germany) [5,36]. The

7days

e

B:-3.3

72 h

b

B:-3.2 sham

B'-3.2 Fig. 2. Distribution of bFGF mRNA in coronal sections of the hemitransected rat brain at different time points after tile lesion e~ent al lhe level of the lesion. (a) b F G I mRNA expression 24 h after the lesion event. Note the large increase in bFGF mRNA in the ipsilateral cortex, a, ,,ell as around the lesioned area. (b) bFGF mRNA expression as seen at 72 h after the mechanical lesion. The bFGF mRNA is still upregulated, hm~evcr, is reduced in the cortical parts and is high abundant surrounding the lesioned area and in the caudate putamen. (c) At day 7 after the surgery the bFGF mRNA is almost normalized in the ipsilateral cortex, but still increased around the lesioned area including the caudate putamen. (d) The bFGF mRNA expression at the 14 day time interxal is almost back to control values. (e) bFGF mRNA expression in the sham operated animal, bFGF mRNA is highh, abundant in the CA2 region of the hippocampal formation, in the induseum gfiseum and in the lasciola cinereum. (f) represents a control section at the lesion level hybridized with the [~x-~Ss[I :TP-labelled bFGF sense RNA. (* = lesion: B = bregma level in mm).

4

A. Lippoldt et al. / Brain Research 730 (1996) 1 - 1 6

2.5. In situ hybridization

tions were acetylated to avoid unspecific binding. Alter dehydration in graded ethanols the sections were prehybridized in a humidified chamber with 150 ~1 prehybridization buffer (50% deionized formamide, 50 mM

The procedure is described earlier in detail [5]. Briefly, the sections were brought to room temperature, fixed with 4% buffered paraformaldehyde, washed in PBS and deproteinized in 0.1 M HC1. After additional washing the sec-

Tris-HC1, pH 7.6, 25 mM EDTA, pH 8.0, 20 mM NaCI, 0.25 m g / m l yeast tRNA, 2 . 5 × Denhardt's solution (0.05% Ficoll, 0.05% polyvinylpyrrolidone, 0.05% bovine serum albumin)) for 2 - 4 h. After draining the prehybridization buffer off" the slides, the sections were hy-

specifity of the probes was checked by Northern blot analysis (Y. Cao, unpublished data).

bFGF mRNA cingulate cortex, rostral 0,25

0,20

0,15 cgx sham

•,~

0,10

0,05

0,00 0

i

1

i

i

i

i

i

48

96

144

192

240

288

336

time (hours)

bFGF mRNA frontoparietal cortex, rostral 0,30 -

0,25 '

0,20

'

0,15

'

.4

frpctx sham e~ 0,10

iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii

0.05

0,00

b

i

i

i

t

48

96

144

192

i

240

i

288

i

336

time (hours)

Fig. 3, Semiquantitative evaluation of the b F G F m R N A expression during the time course rostral to the lesion and at the lesion level. The b F G F m R N A levels of the sham operated animals (::::) and at the ipsilateral side of the lesion were expressed as optical densities + SD. The levels of the sham operated animals are the measured optical densities at the ipsilateral side +_ standard deviation (S.D.). (a), (b) and (c) demonstrate the levels rostral to the lesioned area, whereas (d) are the optical densities surrounding the lesioned area. The statistical analysis was done using the multiple t-test ( . . . . . P < 0.001, ~ P < 0.01, ~ P < 0.05).

A. I.ipl~oldt et al. / Brain Research 730 (I 996) 1 - 1 6

bFGF mRNA caudate putamen, rostral 0,30 -

0,25 '

9,20 '

cpu 0,15 ' sham e~ Q

0,10 '

0,05 '

3, 00

i

L

I

I

48

9G

144

192

time

I

i

240

i

288

336

(hours)

bFGF m R N A caudate putamen, lesion 0,3 -

0,2

cpu sham e,

0,1

0,0

I

I

f

I

i

48

96

144

192

240

I

288

i

336

time (hours) Fig. 3 (continued). bridized according to the standard procedure using 0.15 ng labelled sense and antisense RNA, respectively. The hybridization was done at 37°C in a buffer containing 50% formamide, 1 X Denhardt's solution, 10% dextransulfate, 0.5 m g / m l yeast tRNA, 0.1 m g / m l poly-A and 0.2 M DTT (for details see 5) for 16-18 h. Thereafter, the slides were washed in 0.5 X S S C / 5 0 % formamide at 48°C for 2 h, 20 rain in 1 X SSC at 48°C and subsequently treated with RNase A (10 p,g/ml) for 30 rain at 37°C and washed several times in 1 X SSC, 0 . 5 x SSC and 0 . 2 x SSC at 48°C, dehydrated in graded ethanols, dried and exposed on Hyperfilm-SH X-ray films (Amersham. UK) at 20°C for several weeks.

The specificity of the method used was determined by hybridization with 35S-o~-UTP labelled sense RNA at the same specific activity, length and concentration as the antisense RNA. 2.6. Computer-assisted microdensitomet O, Semiquantitative evaluations of the bFGF mRNA levels were obtained by measuring the gray values of the film autoradiograms using the microdensitometrical program of the IBAS 2.5 (Kontron, Germany) according to a previously published microdensitometric method [69]. The changes in bFGF mRNA were measured directly in the

6

,4 Lippoldt el ctl./Brain Research 730 (1996) 1-16

Fig.4 b F G F IR

•

L L

L

c| b F G F - I R / G F A P IR

bFGF I R / V i m e n t i n - l R

L

b

Fig.6

50urn e

25un

A. Lippohlt et al. / Brain Research 730 (1996) / /6

center of the lesion and in the ipsilateral nonlesioned part about 500 g m rostral to the lesion site. Three measurements were performed for each region: the total value, i.e., measurements of the areas in question in the sections hybridized with ~SS-o~-UTP-labelled antisense RNA, the unspecific value, i.e., measurements of the corresponding areas in the control sections hybridized with )5S-c~-UTPlabelled sense RNA and the background value, i.e., measurements of the fihn background outside the sections. The autoradiograms were digitized directly via a T.V. camera (CCD 72/MTi-camera) on a screen, allowing the measurement of an entire coronal section at once. The areas of interest were selected by means of a light pen and the transmissions were measured. The specific values were defined as those obtained by subtracting the nonspecific values (sections incubated with the labelled sense RNAprobe) from the total values (sections incubated with labelled antisense RNA-probe) after correcting for the background. The sense values were in all cases background values. Photographs of the in situ autoradiograms were token from the screen. Transmissions as well as optical densities were obtained utilizing post-processing programs. For statistical analysis multiple t-test was used. As the basis for the tests served the transmission mean values and transmission standard deviations (SD).

7

protein (GFAP) a monoclonal anti-GFAP antibody (Boehringer Mannheim, Germany) (1:10) was used. The astroglial bFGF was detected using a polyclonal anti-rat bFGF antibody [24]. The incubations with the primary antibodies were made in PBS containing 1(/~ bovine serum albumin (BSA) and 0.3% Triton X-100 in a humidified chamber over night at 4°C and processed with the immunofluorescence technique. The following protocol was used: after incubation with the primary antibodies and washing in PBS the sections were incubated with lluoresceine isothiocyanate (FITC) lot rabbit antibodies or Texas Red conjugated secondary antibodies (Amcrsham, UK) for mouse antibodies in the dilutions 1:100. For double-labelling, the slides were incubated with the primary antibodies as described above. Subsequently, the slides were rinsed in PBS and incubated with goat anti-mouse Texas Redlabelled antibody (Dianova) and goat anti-rabbit FITClabelled antibody (Dianova) each diluted 1:100 (60 mm at RT). After rinsing in PBS the slides were coverslipped with an antifading medium (glycerol/PBS 3:1, 0.1c~ pphenylendiamine) and examined in a Zeiss microscope (Axioplan, Zeiss Oberkochen, Germany) for fluorescence microscopy using the appropriate excitation and barrier filter combinations. Control sections were performed by omitting the primary antibodies.

2. Z hnmunocvtochemistrv 3. Results

hnmunocytochemistry was performed using the immunofluorescence technique. For the detection of extracellular matrix components the sections were incubated with anti-tenascin [26] (1:200), anti-laminin antibodies (Sigma Chemicals, USA) (1:300) and for intermediate filament proteins with anti-vimentin (Boehringer Mannheim, Germany) (1:3). The tenascin antiserum was characterized by Western Blot analysis and detected only two bands of ca. 240 and 200 kDa, confirming the described subunit molecular weight. The antiserum did not react with plasma fibronectin [39]. The anti-laminin antibody specifically labells basement membranes and does not react with fibronectin, vitronectin, collagen IV or chondroitin sulfate types A, B and C. An anti-vimentin antibody was used to stain reactive astroglial cells within the lesion in a dilution of 1:4 (Boehringer Mannheim, Germany) and does specifically recognize vimentin. To detect glial fibrillary acidic

3.1. hi situ hybridization: b F G F mRNA

The bFGF mRNA levels are strongly upregulated at 24 h around the lesion site and also in the surrounding cortical hemispheres (retrosplenial and frontoparietal cortices; amygdaloid cortex) of the ipsilateral side. This significant increase can still be observed 72 h after surgery. At the 7 day time interval the bFGF gene expression returned to sham related levels in all areas measured. The control value was also prominent at the 14 day time interval (Fig. 2 and Fig. 3d). Rostral to the lesion site bFGF mRNA expression appeared to be significantly increased in the cingulate and parietal cortices at the 24 h and 72 h time intervals (Fig. 3a,b). This upregulation returned to control levels at the 7 day time interval. Additionally, in the caudate putamen roslral to the lesion site bFGF mRNA

Fig. 4. The cellular expression of the upregulated bFGF mRNA is demonstrated: 24 h after the mechanical injury the bFGF mRNA (dark-red hybridized material, thick arrow) is localized in GFAP positive cells (brownish material, small arrm~) around the lesion site.The cell nuclei are counterstamed wilh netltra] red. F~g. 6. Localization of the bFGF IR. (a) demonstrates the bFGF IR in the lesioned area in some cases surrounding small ~esse[s (arrow). (b) Double mununocytochemislr,, lor bFGF- and GFAP-IR. Note that the bFGF IR (green fluorescence) and the IR lk)r GFAP (red fluorescence) are codislribuled as visualized by the yellowish color of the cell nuclei (arrow). (c) and (d) demonstrate a higher magnification of a vessel in the lesioned area. (c) demonstrales hl;GF IR surrotmding a vessel in the lesioned area (arrows). Note in (d) the codistribution of bFGF- and GFAP-IR ( a r r o ~ ) . [, hlmen. :+ lexion. (e) Double immunocytochenfistry for bFGF and vimentin-lR. Note that the bFGF IR (green fluorescence) and the '~inlentin IR (red fluorescence) arc o,distribatcd m some cells (yellowish fluorescence).

8

A. Lippoldt et al. / Brain Research 730 (1996) 1 - 16

expression was significantly upregulated after 24 h and returned to control values until the 7th day after the surgery (Fig. 3c). The upregulation of bFGF mRNA in the lesioned area and rostral to the lesion was related to an upregulation in GFAP-positive cells as demonstrated by non-radioactive in situ hybridization coupled with GFAPimmunocytochemistry (Fig. 4). 3.2. lmmunocytochemistr 3, bFGF immunoreactit,i~, Basic FGF immunoreactivity (IR) in the sham operated rats was seen in large numbers of astroglial cell nuclei in the rat thalamus, the hypothalamus and the cerebral cortex and in the neurons of the CA2 region of the hippocampal

formation (Fig. 5a). 72 h after the surgery bFGF-IR is increased around the lesioned area. It surrounds the lesioned area and is exclusively located in astroglial cell nuclei, showing a pronounced increase in density in bFGF IR reflected as an increased intensity of bFGF ir glial profiles (Fig. 5b). After 7 days an increased density and intensity of bFGF ir glial cell nuclei was found surrounding the lesion (Fig. 5c). The bFGF IR was localized in astrocytes in the lesioned area as well as in astrocytes surrounding the vessels in the lesioned area (Fig. 6a-d) and was partly codistributed with the vimentin positive reactive astroglial cells (Fig. 6e). Moreover, the nerve cells of the CA2 region of the hippocampal formation showed increased bFGF IR as well (Fig. 5c).

Fig. 5. b F G F i m m u n o r e a c t i v i t y (IR) in the hemitransected area at different time points. (a) demonstrates b F G F IR in the ipsilateral hemisphere of the lesioned level. Note the increased density of b F G F IR profiles surrounding the lesioned area (arrows). (b) At the 7 day time interval the b F G F ir profiles are still increased s u r r o u n d i n g the lesioned area and in the C A 2 of the h i p p o c a m p a l formation (arrows). (c) demonstrates the distribution of b F G F IR in a coronal section of the s h a m operated animal at the lesion level. * - lesion: HiF = h i p p o c a m p a l formation; C A 2 = C A 2 area of the h i p p o c a m p a l formation: D G = dentate gyrus, cc corpus callosum; 3V ~ third ventricle.

A. Lippoldt et al. / Brain Research "LeO(1796) I /6 3.3. T e n a s c i n C - -

immunoreacticitv

a n d r e l a t i o n s h i p to

~)

4. Discussion

astroglia and zessels

Tenascin was almost not detectable in the sham operated rat brain at the lesion level. It appeared within and around the lesioned area 72 h after the injury and was still visible after 7 days (Fig. 7). Tenascin IR outlined the cavity of the lesion and showed partial codistribution with the vimentin positive glial cells (Fig. 6 and Fig. 7). Tenascin was related in some cases to GFAP positive astrocytes with long processes (Fig. 8a,b). However, in some cases the observed tenascin immunoreactivities were probably related to the basal membrane of the vessel wall (Fig. 8c). The edges of the lesion cavity stained for tenascin IR were also enriched in bFGF positive glial cells (Fig. 5c).

Basic FGF is a potent mitogen for a variety of cell types. In the brain bFGF functions as survival and differentiation factor for neurons and may also exert its neurotrophic effects via its expression by glial cells. Basic' FGF produced by different cell types can be bound to components of the extracellular matrix (ECM) probably via heparan sulfate. The ECM may serve as a store for the growth factor. Basic FGF is a powerful growth factor for endothelial cells and vascular smooth muscle cells and stimulates as a potent angiogenic factor neovascularization. In head trauma, a model of which was used in the present paper, angiogenesis is a pathophysiological response in order to maintain the blood flow. in the damaged

Fig. 5 (continued).

b FGF

IR ~ sham

CA2

HiF DG

".

6~ e g m a

: t . i:,¢

-2~Om~s

A. l.q,p~,/dr ~,l (/I./Brcmz R~'se¢lrch 730 (1996) 1 - 1 6

]I

.8 .=_

E..,-¢ o

© ~J

e-

~d

0

Z

%.2

r--: rr

-7

12

A. lJl~poldt et al. / Brain Research Z~O (1996) I - 16

Fig. 8. The cellular distribution of the tenascin IR is shown here. (a) demonstrates the tenascin IR in the lesioned area. (b) demonstrates the GFAP-IR of lhe same area as in (a). In (c) double fluorescence ¢~f tenascin IR and GFAP IR is demonstrated. Note the partial overlap of both immunoreactivities. The arrow demonstrates the GFAP positive cells, which partially express mnascin IR. The arrowhead is pointing to an tenascin ir part what is not costained for GFAP.

A. Lippol& el al./Br~l& Rexearch 730 ( 1996J I- /6 area. Moreover, mechanical injury leads to glial cell activation and the production of a glial scar. The damaged nerve cells in the injured area need to be supported by growth factors to be able to survive. The present paper demonstrates a global increase in bFGF mRNA expression after mechanical injury. This increase is last (within 24 h) and strong. It spreads over the ipsilateral hemisphere and takes place predominantly in astrocytes. 4.1. Characterization o/" the le.ffo/z The intermediate filament protein vimentin (Fig. 9) was highly upregulated near the edges of the lesion and surrounded the cavity of the injured area. Vimentin expression is observed in immature glial cells and is a marker for radial glia. After several types of brain injury an increase in expression of vimentin is observed around the injured area [29,46,54]. This increase in intermediate filament IR demonstrates a change in cell metabolism of these astrocytes, i.e., the cells become hyperplastic and tend to proliferate and could become mobile as well [54]. Moreover, there is evidence that astrocytic cells in close contact with wide extracellular fluid spaces become vimentin positive [46]. Thus, in the present model the activation of astrocytes with the feature of vimentin IR may be the consequence of (a) a destroyed blood brain barrier by the hemitransection and their exposure to large fluid spaces,

13

(b) the action of growth factors released by dying cells after the insult and (c) the release of several ions like Ca~' ' and K ~ in the affected area [29,31]. The activated astrocytes may function to protect neurons from homeostatic changes, such as ionic imbalances and to stabilize the blood brain barrier in the affected tissue [29]. Indeed, the present results demonstrate an increased laminin IR in the injured area as a marker for basement membrane production and capillary formation fl)llowing the lesion event [32,59,68]. Capillary formation leads to neovascularization in wounds of the central nervous system [4,32] to compensate for the impaired circulation in the damaged area [47]. These events are accompanied and made possible with the fast and strong induction of astroglial bFGF by the partial unilateral hemitransection. 4.2. Di,~'tribution of b F G F mRNA am/ immunoreactirilv q/'ter mechanical injuo' Partial unilateral hemitransection induced a transient and fast increase in bFGF mRNA. These data are in agreement with previous reported results. After mechanical lesions in the cerebrum a rapid increase in bFGF genc expression was found around the lesioned site within 24 48 h and had disappeared within 14 days [35]. In cortical brain injury bFGF mRNA was increased within 4 h after lhe lesion and remained increased for at least 2

Fig. 9 Vimentin [R in the lesioned area. Note the vimentin positive cells surrounding the lesion,

14

A. Lippoldt et al. / Brain Research 730 (1996) 1-16

weeks [20]. This longlasting effect is in contrast to our present results and may be dependent on the kind of lesion used. The increase in bFGF mRNA was paralleled by increased bFGF IR 72 h and 7 days alter the head trauma. The lesion was surrounded by a dense population of small bFGF ir nuclei, representing presumably the nuclei of the reactive vimentin positive astroglia at the edges of the lesion. The bFGF produced by the glial cells may support the astrogliosis in the injured area and glial scar formation [13]. Moreover, since bFGF is suggested to regulate nerve growth factor synthesis in astrocytes via an autocrine and paracrine stimulation [61,66,67], to stabilize intracellular free calcium levels and to protect neurons against excitotoxic injury by suppressing the expression of NMDA receptors in hippocampal neurons [41] it may support neuron survival in the lesioned area. Furthermore, bFGF may have a function in new vessel formation in the injured area because of its known angiogenic properties [8,19,42,50].

4.3. Tenascin immunoreactivio' after mechanical injuo' The extracellular matrix protein tenascin was upregulated in the present type of mechanical injury at the 7 day time point. This upregulation was seen in GFAP positive glial cells and in other cell types as well. Tenascin is a six-armed glycoprotein existing in at least three isoforms generated by differential splicing of its mRNA [27,51]. It is only expressed in special populations of glial cells [2,53] in the adult brain and its exact function is still a matter of debate. In the brain a few reports are available demonstrating the upregulation of tenascin around a mechanical lesion in agreement with the present findings [34]. In in vitro investigations tenascin production in astrocytes has been shown to be regulated by bFGF [43]. Basic FGF specifically upregulated the tenascin expression in cultures of cerebral cortical astrocytes, whereas nerve growth factor or epidermal growth factor had no effect [43]. Contradictory functions have been described in vitro. Thus, it has been found to be an inhibitory substrate for several types of neurons and to promote neurite outgrowth when it is bound to other extracellular matrix proteins [37,57]. Tenascin has been described to be an inhibitor for glial cell migration [62]. In the case of increased shear stress induced by hypertension it has been seen in the vascular smooth muscle cells of the vessel wall [26]. In healing skin wounds in rats a marked increase in tenascin expression has been reported at the wound edges [39]. The lesion was characterized by an increase in vimentin IR around the lesion site, increased laminin IR found in an increased number of laminin immunoreactive basement membranes associated with an upregulation of bFGF and tenascin expression in the lesioned area and the edges of the wound. After mechanical injury active bFGF heparansulfate complexes may be released by disruption of cell mem-

branes, proteolytic activities and by platelets and activated macrophages appearing in the injured area due to their attachment to the subendothelium. These soluble bFGF complexes may in turn activate astroglial cells, as seen by the presence of vimentin IR, to induce the production of bFGF, nerve growth factor and other neurotrophic factors to prevent damage of the nerve cell populations. The high amount of bFGF as well as other growth factors produced in the injured area may also lead to the production of extracellular matrix proteins like tenascin [43] and laminin [11 ] and to angiogenesis [8,19]. These events in turn may lead to: (a) counteraction of the Ca =` rise in this area: (b) downregulation of NMDA receptors to prevent further excitotoxic nerve cell damage: and (c) new vessel formation and wound repair. Since microvessels in the brain are characterized by a linkage of astroglial cells to the vessel wall, the bFGF seen in astrocytes surrounding the vessels may support the production of extracellular matrix proteins and thus the formation of new capillaries along these substances also via its action on endothelial cells. In conclusion, the present paper demonstrates the upregulation of bFGF, intermediate filaments (vimentin) and matrix proteins (laminin, tenascin) in a model of head trauma. It is suggested that these molecules may be involved in (a) tissue repair, (b) protection of neuronal cells and (c) formation of new vessels in the lesioned area.

Acknowledgements This study was supported by grants of the Swedish Medical Research Council (04X715), a grant from the Verum Foundation, Munich, Germany, grants from the Deutsche Forschungsgemeinschaft (SFB 317, Li 604/1 - 1). We are very grateful for the kind gift of the bFGF antibody by A. Baird (Scripps Research Institute. La Jolla) and the kind gift of the tenascin C antibody by A.W.A. Hahn (University of Basel).

References [I] Agnati, L.F.. Fuxe. K., Calza, L., Zini, I., Benfenati, F., Farabegoli. C. and Goldstein, M.. Chronic treatment with L-DOPA plus carbidopa in hemitransected rats: preferential effects at intact dopamine synapses leading to behavioural signs of dopamine receptor super sensitivity, Acre PMsiol. Stand., 199 (1983) 27 34, [2] Bartsch. S., Bartsch. U., D/Srries, U., Faissner, A.. Weller, A.. Ekblom, P. and Schachner, M., Expression of tenascm in the developing and adult cerebellar cortex, ,/. Neurosei.. 12 (1992) 736-749. [3] Bean, A.J.. Simons. J.F. and H~ikfelt. T., Differential expression od acidic and basic FGF in the rat substantia nigra during development. NeuroReport, 3 (1992) 993 996. [4] Beggs, J.l~. and Waggener. J.D., Microvascular regeneration ft)llowing spinal cord injury: the growth sequence and permeability properties of new vessels, Adu. Neurol., 22 (1979) 191-206.

A. Lippoldt et al./' Brain Resear{'h Z{O (1996) / 16

[5] Bunnemann, B.. Fuxe, K.. Metzger. R., Bjelke, B. and Ganten, D.. The semi-quantitative distribution and cellular localization of an giolensinogen-mRNA in the rat brain, .I. Chum Neuroanat., 5 (1992) 245-262 [6] Burgess, WH. and Maciag, T., The heparin-binding (fibrobhlsl) growth factor family of proteins. Annu. Rer. Biochem., 58 (1989) 575 6(){~ [7] Chadi. G . Cao. Y.. Pettersson, R.F., Fuxe, K., Temporal and spatial iucrease nl astroglial basic fibroblast growth factor synthesis after 6-hydrox_~doparnine-induced degeneration of the nigrostriatal dopanlinc neunms, Neuroseienee. 61 (1994) 891-910. [8] Chen. HII.. Chien, C.H. and Liu, H.M., Correlation between anglogenesis and basic fibroblast growth factor expression in experimental brain infarct, Stroke, 25 (1994) 1651 - 1657. [9] Choi, D.~,V., Excilotoxic cell death, .I. Neurohiol., 23 (1992) 12t, I 1276. [10] Crossin, K.L, Hoflman. S.. Tan, S.S., Edelman, G.M., Cytotaciin and its prote~glycan ligand mark structural and functional boundaries m somatosensory cortex of the early postnatal nlOllSe, l)et'. Biol.. 136(It)89)381 392. [1 I] l)rago, J.. Nurcombe. V., Pearse, M.J., Murphy, M. and Bartlett. P.F., Basic fihn~blast growth factor npregulates steady-state levels of lanfinin B I and B2 chain mRNA in cultured neuroepithelial cells, Etp. ('ell Re,~., 196 (1991) 246-254. [I 2] Ecclcslonc, PA. and Silberberg, D.H., Fihroblast growth factor is a mitogcn for oligodendrocyles in vitro, Brain Rev.. 353 (1985) 315-.318. [13] Eclancher. F.. Perraud, F.. Faltin, ,1., Labourdette. G.. Sensenbrenner, M.. Reactive astrogliosis after basic fibmblasi growth factor (bFGF) inleclion in injured neonatal rat brain, Glia. 3 (1990) 5O2 5O9 [14] Emotu. N. (ionzalcz, A.-M., Walicke, P.A., Wada, E., Simmons, I).M., Shimasaki, S.. Baird. A.. Basic fibroblasi growth factor (FGF) in the ceniral nervous system: identificatkm of specific loci of basic FGF exprcssi
15

[24] Gonzalez, A.M., Berry, M., Maher, P.A.. l,ogan. A.. Baird, A., A comprehensive analysis of FGF-2 and FGFRI m the rat brain, Brain Res., 7(11 (1995) 201-226. [25] Gospodarowicz. D., Neufeld, G. and Schwcigerer. L., Fibrobhlsi growth factor. Mol. (Fell. Endocrine/., 46 (1986) 187-206. [26] Hahn. A.W.A.. Kern. F., Jonas. 1;.. John, M.. Bfihler, F.R., Resinck, T.J., Functional aspects of vascular tcnascin C expresqon, ,/ Va,~c. Re,~., :Q (1995) 162-174. [27] Jones, F.S,, Hoffman, S., Cunnmgham, B.A., Edclman, G.M., A detailed xtructnre model of cylolaclin; protein homologies, ahernafive RNA splicing, and binding regions. Pro< Nail. AccuL Sci. USA, 86 (1980) 1905 1909. [28] Johnson, I).E.. Williams, L,T., Structural and hmctional diversity irl the F(IF receptor multigene family..Ida. ('am'or Rev.. 60 (19931 I 42. [29] Kin(lx, M.S., Bhat, A.N., Bhat, NR., Transient ischemia stimulates glial fibrillary acidic protein and vimeniin gene expression m the gerbil neocortex, striatum and hippocampus, 1,Io/, Brain Res., 13 (1992) 199 206. [30] Kiyota, Y., Takami, K,, lwane, M., Shine, A., Miyarnoto, M., Tsukuda, R., Nagaoka, A.. Increasc in basic fibroblast growth factor likc immunoreaciivity in rat brain after lin'cbrain ischemia, Bra#1 Re~., 545 (1991) 322-328. [31] Kraig. R.P., Jaeger. C.B., Ionic concnmilants of asm)glial tnmsformarion i{~ reactive species, Stroke. 21 (Suppl. III)(1091)) 111-184 111-187. [32] Krum. J.M., More. N.S. and Rosenstcin. JM.. Brain angiogenesis: variatinns ill vascular basement inellflhranu glycoprolein inlmunoreactixit 3, E.W~. NeuroL, I l l {1991) 152 1~5. [33] Kurokawa. T.. Seno. M. and Igarashi. K.. Nucluotide sequence of rat basic fibroblast grov, th fi.lclor eDNA, \"la-h,ic ,~'id~ Re~,. 16 (1988) 5201 [34] Laywell, E.D., l)6rries, U., Barlsch. 1.. Faissner. /\.. Schachner, M. and Sleindler, D.. Enhanced cxpressi~m ~l" tl~c developmentally regulated extracelhllar matrix molecule tenscin following adult brain injur }. Prec. ,'Vat/. Acad. Nci. I,'X4, 89 (1992} 2634 2,538. [35] l~cdoux, I)., Mereau, A.. Pieri, 1., BarrJtault. D and Courty, J,. High affinit 3 rcceptoI> to acidic and basic fibroblast growth t'actor t FG]:) arc detected in aduh brain membrane preparations bul not in liver, kidne',, intestine, lung or slon/~lch. (;#outll [-a,'toes, 5 (1991) 221 231. [36] I,ippoldl, A.. Andbjer, B., Rosen, 1,., Richter, I.i., Ganten, D,. Cau. Y., Peuersson. R.F. and Fuxe, K., Pholochemicall,< reduced local cerebral ischemia in nit: time dependent and global increase in expression of basic fibroblasl gro'~th factor mRNA, I'IraiJl Rc~., 625 (1993) 45 46. [37] Lochlcr, A., Vaughau, 1,., Kaphmy, A., Prochiantz. A.,. Schachner. M. and Faissller, A., J l/tenascJn in subslrate-I'Jound and sohihle form displays contrary ef|)cis nn n(uriie ~mtgrn,.,,th..I ('eli Biol., 113(It5'91) 1159 1171. [38] Mackie. E..J.. Tenascm in conncctixe ti',,StlC de;elopment and palhogenesis, t'epWwct. Def. Neurnhiol.. 2 (!c)94) 125 132. [39] Mackic, E.J., Halfter. W. and Livemni, I). Induction of tcnascm iu healing wounds. ,I. Cell Biol., 107 (I~,188) 2757- 27e.7. [40] Martini. R., Schachner, M. and Faissner, .\,, l:,nhanccd expression of lhc extracelluhir matrix molecule J I/lenascin in the regenerating adnlt mouse sciatic nerve, ,/. Neurocvt
16

[45]

[46]

[47] [48]

[49]

[50] [51]

[52]

[53]

[54]

[55]

[56]

[57]

A. Lipl~ohtt et al. / Brain Researeh 730 (1996) 1 /6

L.F,, Hormones and growth factors induce the synthesis of glial fibrillary acidic protein in rat brain astrocytes, J. Neurosci. Res., 14 (1985) 167-176. Morrison, R.S.. Sharma, A., De Vellis, J. and Bradshaw, R., Basic fibroblast growth factor supports the survival of cerebral cortical neurons in primary culture, Ptwe. Natl. Aead. Sci. USA, 83 (1986) 7537-7541. Pixley, S.K., De Vellis, J., Transition between immature radial gila and mature astrocytes studied with a monoclonal antibody to vimentin, Der. Brain Res.. 15 (1984)201-209. Poulsml, O.B.. Strandgaard> S.. Edvinsson, L,, Cerebral autoregulation. Cerebtweasc. Brain Metab. Ret'., 21 (990) 161 192. Rettig, W.J., Erickson, H.P., Albino, A.P. and Garin-Chesa. P.. Induction of human tenascin (neuronectin) by growth factors and cytokines: cell type-specific signals and signalling pathways..l. Cell Set., 107 (1994) 487-497. Riou, J.F.. Umbhauer, M., Shi, D.L., Boucaut, J.C., Tenascin: u potential modulator of cell-extracellular matrix interactions during vertebrate embryogenesis, Biol. Cell.. 75 (1992) 1-9. Risau, W.. Angiogenic growth factors, Pro~4. Growth Factor Re~.. 2 (1990) 71-79. Spring, J., Beck, K., Chiquet-Ehrismann, R., Two contrary functions of tenascin: dissection of the active site by recombinant tenascin fragments, Cell, 69 (1989) 325-334. Steindler, D.A.. Cooper. N.G.. Faissner, A.. Schachner, M., Bound aries defned by adhesion molecules during development of the cerebral cortex: the JI/tenascin glycoprotein in the mouse somatosensory cortical barrel field, Dev. Biol., 131 (1989) 243-260. Steindler, D.A., O'Brien, T.F., Laywell. E., Harrington. K., Faissner, A., Schachner, M., Boundaries during normal and abnormal brain development: in vivu and in vitro studies of glia and glycoconjugates, Exp. Neurol., 109 (1990)35-56. Schiffer, D., Giordana, M.T., Migheli, A.. Giaccone. G,, Pezzotta, S., Mauro, A.. Glial fibrillary acidic protein and vimentin in the experimental glial reaction of the rat brain. Brain Res., 374 (1986) I10-118. Schweigerer, L., Neuteld, G., Friedman, J., Abraham, J.A., Fiddes, J.C. and Gospodamwicz, D., Capillary endothelial cells express basic fibroblast growth factor, a mitogen that promotes their own growth, Noture, 325 (1987) 257-259. Takami, K., lwane, M., Kiyota, Y., Miyamoto, M., Tsukuda, R., Shiosaka, S., lncrease in basic fibroblast growth factor immunoreactivity and its mRNA level in rat brain following transient lorebrain ischemia, Exp. Brain Res., 90 (1992) 1-10. Taylor, Pesheva, P. and Schachner, M., Influence of janusin and tenascin on growth cone behaviour in vitro. J. Neurosci., 35 (1993) 347-362.

[58] Thomas, K.A., Rios-Candelore, M., Gimenez-Gallego, G.. l)iSalw~, J.. Bennett. C.. Rodkey, J. and Fitzpatrick, S., Pure brain-derived acidic fibroblast growth factor is a potent angiogenic vascular endothelial cell mitogen with sequence homology to interleukm I. Proe. N~ttl. Aead. Set. USA, 82 (1985) 6409-6413. [59] Tryggvason. K.. The laminin family, Curr. O[fin. Cell Biol., 5 (1993) 877-882. [60] Unsicker, K., Reichert-Preibsch, H.. Schmidt, R.. Pettmun, B., Labourdettc. G. and Sensenbrenner, M., Astroglial and fibrobla,,l growth factors have neurotrophic functions for cultured peripheral and central nerwms system neurons, Proe. Natl. Aead. 5'ci. {,',S'A,84 (1987) 5459-5463. [61] Walicke. P.. Cowman, W.M., Ueno, N., Baird, A. and Guillemin. R., Fibroblast growth factor promotes survival of dissociuted hip pocampal neurons and enhances neurite extension, l'ro~. Natl. Acad. Set. USA. 83 (1986) 3012 3016. [62] Wehrleq4aller, B. and Chiquet, M., Dual function ~1 tenascin: simuhaneous promotion of neurite growth and inhibition of glial migration, .I. (_'ell Set., 1116(1993) 597-610. [63] Weller, A.. Beck, DS. and Eckblom. P., Amino acid sequence of mouse lenascin and differential expression of two tcnuscm isotkwms during embryogenesis, ,1. (;'ell Biol., I 12 ( 1991 ) 355-362. [64] Woodward, W.R.. Nishi. R., Meshul, C.K., Williams, T.[L, Conlombe, M. und Eckenstein, F.P., Nuclear and cytoplasmic localization of basic t'ibroblast growth factor in astrocytes and CA2 hippocampal neurons, J. Neuro,sei., 12 (1992) 142 152. [65] Yuyon, A.. Klagsbrun, M., Esko, J.D.. Leder, P. and Ornitz, I).M.. Cell surface, heparin-like molecules are required for binding of basic fibriblasl growth factor tu its high affinity receptor, Cell, 64 (1991) 841-848. [66] Yoshidu, K. and Gage. F.H.. Fibroblast growth factors stimulate nerve growth factor synthesis and secretion by astrocytcs, Braitt Res.. 538 (1991) 118 126. [67] Yoshida, K. and Gage, F.H., Cooperative regulation of nerve growth factor synthesis and secretion in fibroblasts and astrocytes by fibroblast growth factor and other cytokines. Btzfin Re+.. 569 (1992) 14 25. [68] Yurchcnco, P.D.. Cheng. Y.-S. and Colognato, H.. Luminin forms an independent network in basement membranes, J. Cell Biol.. 117 (1992) 1119-1133. [69] Zoli. M.. Zini, 1.. Agnati, L.F.. Guidolin, D., Ferraguti, F. and Fuxe K.. Aspects of neural plasticity in the central nervous system. I. Computer-assisted image analysis methods, Neuroc]wm. I;tl., 16 (1990) 383 418.