Brain Research 972 (2003) 110–118www.elsevier.com/ locate/brainres

Research report

N eurotrophin-3 specifically increases mature oligodendrocytepopulation and enhances remyelination after chemical demyelination of

adult rat CNSa , a a b* ´Isabelle Jean , Celine Lavialle , Annick Barthelaix-Pouplard , Catherine Fressinaud

aCell Biology Laboratory, UPRES EA 3143,University Hospital, 4 rue Larrey, F 49033 Angers cedex 01, FrancebNeurology Department, UPRES EA 3143,University Hospital, 4 rue Larrey, F 49033 Angers cedex 01, France

Accepted 20 February 2003

Abstract

In human central nervous system (CNS) demyelinating diseases, spontaneous remyelination is often incomplete. Therefore, we havetested whether neutrotrophin-3 (NT-3) accelerates CNS myelin repair after a chemically-induced demyelination. One group of adult ratswas injected in the corpus callosum (CC) with 1ml of 1% lysophosphatidylcholine (LPC) and 1ml of NT-3 (1 mg/ml), and 15 days afterinjury (D15) remyelination was compared to control rats (receiving 1ml of LPC11 ml of vehicle buffer of NT-3). The demyelinated

1volume decreased by 56% in NT-3-treated rats at D15, and immunohistochemistry showed an increase in mature MBP oligodendrocytes1(OL) (166%) in treated animals (whereas less mature (CNP ) OL were unchanged). Since less than 3% axons degenerate in this model,

and as astrocytic gliosis was not modified, these data suggest that NT-3 acts directly on cells of the OL lineage to enhance remyelinationin vivo. 2003 Elsevier Science B.V. All rights reserved.

Theme: Development and regeneration

Topic: Neurotrophic factors: biological effects

Keywords: Myelin basic protein; Neurotrophin-3; 29,39-Nucleotid cyclic 39-phosphodiesterase; Oligodendrocyte; Remyelination

1 . Introduction posed for the origin of remyelinating cells. First, theycould arise from the proliferation of precursor cells (O-2A

In central nervous system (CNS) demyelinating diseases progenitors, which are labelled by A2B5 antibody (Ab;such as multiple sclerosis, the myelin sheaths and/or [44,49]) and/or anti-PDGFaReceptor Ab [43]); aftermyelinating cells, i.e. oligodendrocytes (OL), are de- several divisions, these cells differentiate in immature OL

1 2 1stroyed. Although some spontaneous remyelination occurs, (CNP /MBP ) and in mature myelinating OL (CNP /1 1this process is not consistent enough for complete repair MBP /MOG ) [21,40]. In vivo, the proliferation of these

(e.g. Ref. [17]). This phenomenon depends on molecules, POL is thought to be driven by several growth factors,such as growth factors, but they are not clearly identified such as PDGF (platelet-derived growth factor) for example[53]. Furthermore, the origin of the cells involved in [22,45]. Or, secondly, remyelination could be ensured byremyelination is debated: they may be either OL precursors mature OL which are present in, or around, the injured(POL) which are present in the adult CNS, are mobile and area, and which have been identified by some authors asproliferate in vitro [32,41], or surviving mature OL, proliferating cells (e.g. Refs. [4,19]). Finally, the pool ofalthough these latter are generally considered as post- new myelinating cells could also arise from these twomitotic cells [20,24]. Thus, several hypotheses are pro- pathways [51]. In addition, neural stem cells could pos-

sibly participate to the remyelination process [3,35,46,47].Indeed, some growth factors enhance the survival,*Corresponding author. Tel.:133-2-4135-5434; fax:133-2-4135-

proliferation and/or migration of POL and even of mature4138.E-mail address: isabelle.jean@med.univ-angers.fr(I. Jean). OL in vitro, such as basic fibroblast growth factor (bFGF)

0006-8993/03/$ – see front matter 2003 Elsevier Science B.V. All rights reserved.doi:10.1016/S0006-8993(03)02510-1

111I. Jean et al. / Brain Research 972 (2003) 110–118

[15,16,19], PDGF [9,40,45], insulin-like growth factor-1 12 controls), four i.p. booster injections of 250 mg/kg(IGF-1) [37], transforming growth factor (TGFb) [10,36], BrdU (Sigma, USA) in NaCl 0.9%/Tris-buffered salineciliary neurotrophic factor (CNTF) [33], glial growth (TBS) pH 7.6/ethanol 10% (29 mg BrdU/ml) werefactor-2 (GGF2) [12] and neurotrophin-3 (NT-3) [6]. In realised every 6 h, beginning 48 h after injections ofvivo, this neurotrophin was reported to enhance OL LPC1NT-3 or LPC1acetic acid. Each animal received 1survival and POL proliferation during development [7], g /kg of BrdU in 24 h [29] and was sacrificed 72 h afterand it improved axonal growth together with remyelination LPC injection (D3), as described below.in a model of spinal cord contusion [38]. So, NT-3 mightdirectly improve remyelination in vivo—independently 2 .3. Sacrifice and tissue processingfrom its neurotrophic effects on axons, which could beinvolved in the model used by McTigue et al. [38]—after Previous results obtained with this protocol [2] revealedchemical or autoimmune demyelination, by promoting the that spontaneous remyelination was complete after 3survival and/or the proliferation of cells of the OL lineage. months and demonstrated an effect of PDGF after 15 days.Increased remyelination by administration of exogenous This time for sacrifice (D15) was therefore chosen in thegrowth factors in vivo has only been proved for IGF-1 [55] present experiments to assess remyelination (except inand PDGF [1,2]. BrdU experiments). Animals were anaesthetised with 5%

Therefore, in the present study, we used a chemical halothane in oxygen and perfused via the left ventriclemodel of ‘pure’ demyelination induced by lysophospha- with 4% paraformaldehyde (PFA; Merck, Germany) intidylcholine (LPC) [4,23,26] in the corpus callosum of phosphate buffer, pH 7.4. The brains were dissected, andadult rats to test the efficiency of NT-3 on remyelination. after overnight post-fixation in PFA at14 8C, they wereThis model is devoid of axonal sections, thus allowing the embedded in paraffin. Blocks were then cut in 5mm serialstudy of NT-3 direct effects on cells of the OL lineage frontal sections throughout the entire CC (Bregma20.2during the process of remyelination. cm to Bregma10.2 cm) with a microtome (Reichert-Jung,

Germany). Procedures were the same at D3 and D15.

2 . Material and methods 2 .4. Histological staining

2 .1. Animals and stereotaxic lesions Brain sections (one in five), recovered after sacrifice atD3 or D15, were stained with Luxol Fast Blue (LFB) for

Adult male Wistar rats (270–330 g) from our breeding 24 h at137 8C, and then with hematein and phloxin.facility (Faculty of Medicine, University Hospital, Angers, These sections were used for the assessment and measure-France) were used for all experiments. Anaesthesia was ment of the area and volume of demyelination at D3 andinduced by i.m. injection of 1.5 ml /kg Vetranquil 0.5% D15 by optical microscopy, using an ocular morphometric

´(Sanofi-Sante nutrition animale, Libourne, France), and grid (see below). In all animals, the needle track and the`maintained by injection of 1.5 ml /kg Imalgene 500 (i.p., injection site were confirmed by this method.

´Rhone Merieux, Lyon, France).Animals were then positioned in the stereotaxic frame 2 .5. Immunohistochemistry

(Stoelting, USA). Group 1 (controls,n527) received 1mlof 1% LPC (Sigma, St Louis, MO, USA) dissolved in Immunolabelling was carried out on D3 or D15 sections,phosphate buffer, pH 7.4 and 1ml of 10 mM acetic acid; at room temperature (RT), after saturation of endogenousgroup 2 (n527) received 1ml of 1% LPC and 1ml of peroxidase with 0.3% H O (10 min) and permeation with2 2

NT-3 (1 mg/ml in 10 mM acetic acid; R&D Systems, cold methanol (220 8C, 10 min). Sections were thenAbingdon, UK). Nine experiments were run in triplicate incubated in 10% normal sheep serum (NSS) (Sigma, St(i.e. for each separate experiment three rats were treated Louis, MO, USA) for 30 min.with NT-3 and three others were not). All substances were We used the following antibodies (Ab): (i) a mouseinjected into the corpus callosum (CC) with a 10-ml monoclonal A2B5 IgM (hybridoma culture supernatant,Hamilton syringe. The coordinates were chosen as char- clone 105, European Collection of Cell Cultures, UK;acterised in Paxinos and Watson’s atlas [42]: Bregma undiluted), as previously described [2]; (ii) a mouse anti-10.12 cm, lateral20.07 cm and meninx20.35 cm for 29,39-nucleotid cyclic 39-phosphodiesterase (CNP) mono-LPC, and20.33 cm for NT-3 or acetic acid. The injections clonal IgG1 (Sigma, diluted 1/20 in NSS); (iii) a rabbitwere realised over 15 min and the needle removal by 5 anti-human myelin basic protein (MBP) polyclonal Abmin to reduce the reflux up the needle track. (Dako, Denmark, diluted 1/100 in NSS); and (iv) a rabbit

anti-cow gliofibrillary acidic protein (GFAP) polyclonal2 .2. BromodeoxyUridine (BrdU) incorporation Ab (Dako, diluted 1/100 in NSS). Sections were incubated

for 1 h in primary Ab solutions, rinsed several times inIn four experiments (i.e. 12 rats treated with NT-3 and phosphate buffer solution (PBS) pH 7.4, and incubated in

I. Jean et al. / Brain Research 972 (2003) 110–118112

biotin-conjugated species specific secondary Abs (Phar- controls), the mean number of positive cells were com-macia-Amersham Biotech, Denmark, diluted 1/100 in pared in the lesion area, using one-way ANOVA (F-test)NSS) for 40 min. and Student’st-test adapted to small samples [48].

After washing in PBS, sections were incubated inAvidin–Biotin Complex (ABC) compound (Vector Lab-oratories, CA, USA) for 40 min and revealed with 3 . Resultsdiaminobenzidin (DAB; 0.66 mg/ml in Tris buffer saline:0.05 M Tris, 0.15 M NaCl, pH 7.6; Sigma). Nuclei were 3 .1. Morphometry of the demyelinated lesioncounter-stained with hematein, and sections were dehy-drated and mounted in Eukitt. At the acute phase (D3), the volume of demyelination in

3Double immunolabelling of sections mounted on NT-3-treated animals (0.06460.05 mm (n510)) was not3gelatined Superfrost slides was used in experiments with significantly different from controls (0.11160.09 mm

BrdU. Bone marrow sections were used for immuno- (n59)), according to Student’st-test and one-way ANOVAhistochemical positive controls. Primary Abs (A2B5, anti- (due to individual variations in both groups) (not shown).CNP, anti-MBP, or anti-GFAP) were used as described As previously reported [2,23], the lesion was characterisedabove, and after DAB revelation, sections were double- by the disappearance of LFB staining in the CC, just belowlabelled with a mouse anti-BrdU monoclonal Ab (BioT- the needle track, in NT-3-treated animals as well as in

¨rend, Koln, Germany; diluted 1/400 in NSS) for 1 h. controls. Nevertheless, after 15 days (D15), although a lossBiotin-conjugated secondary Abs and ABC compound of myelinated fibres was still observed in LPC and inwere incubated for 40 min each. The revelation used LPC1NT-3-treated animals, the demyelinated volume wasDAB–10% NiCl as chromogen, and sections were notably reduced compared to D3. This decrease reached2

mounted in Eukitt. 244% for controls (F.F ) and257% for NT-3-treated0.05

rats (P,0.05, F.F ), demonstrating the presence of0.025

2 .6. Measurements and cell counting protocol spontaneous remyelination in controls.At D15, the area of demyelination varied far less within

Treatment was blinded before analysis. The whole one group and it was obviously reduced in NT-3-treatedlesion was scanned on serial sections stained with LFB. Its rats (Fig. 1A) compared to controls (Fig. 1B). This wasdepth was calculated by counting the number of sections confirmed by the measurements of the LPC lesions inwhere the lesion appeared, and its area was calculated at treated versus control animals (Fig. 2). The demyelinatedthe maximal demyelination site, under the needle track, area spread up from the hemispheric fissure, to the

2using standardised microscopic fields defined by an ocular cingulum (0.1460.03 mm for treated animals compared2morphometric grid (Optiphot-2 Nikon, Japan) as previous to 0.19860.02 mm for controls) and in antero-posterior

described [2]. The areas and volumes for each group length (0.19560.027 mm for treated animals vs.(NT-3-treated or controls) were averaged and compared by 0.31260.038 mm for controls). The volume of demyelina-Student’st-test adapted to small samples [48] and one-way tion was reduced by 56% in the NT-3-treated rats versus

3 3ANOVA (F-test, one factor changed: the presence or controls (0.027360.01 mm vs. 0.06260.01 mm , respec-absence of NT-3). tively; and this difference was significant according to the

Immunohistochemistry was carried out as described t-test: P,0.05 and ANOVAF-test: F.F ).0.001

above on four serial (consecutive) sections at the maximumof the demyelinated area, and the number of immuno- 3 .2. Immunohistochemistrylabelled cells with hematein-stained nucleus were countedat 3400 final magnification using the morphometric grid At day 3, in the centre of the demyelinated area, double-(Optiphot-2 Nikon, Japan). The use of this grid and of labelled cells, expressing OL markers and stained for BrdUmicrometric focusing ensures cells were not counted twice. (Fig. 3A,B), were very few (whereas numerous boneFor double labelling, we determined the number of cells marrow cells, used as positive controls, were stronglywith cytoplasmic brown DAB labelling and nuclear black labelled). Their number was not significantly modified inDAB-NiCl labelling. For each rat, an average number of NT-3-treated animals (Table 1; the mean number of all2

1 29–12 optic fields (depending on the surface of demyelina- BrdU cells from the OL lineage was 148.6 cells /mm for2tion, see below) covering the entire lesion were counted. NT-3-treated rats compared to 169.7 cells /mm for con-

Immunolabelled cells were counted as well in 9–12 trols), 3 days after injection. Double immunohistochemis-corresponding optic fields in the opposite side of the CC try was used to identify the different stages of development(non-injured). Since the area and volume of demyelination of dividing cells. Given the lack of neuronal cell bodies inwere much lower in NT-3-treated animals (see results the CC, we used A2B5 Ab to label POL cells [38,41], asbelow), the number of labelled cells was expressed per previously described [2]. At D3 (Table 1), there were

2 1 2 1 2optic field (0.031 mm ) to allow comparison between 123.9611 A2B5 cells /mm , 63.166.6 CNP cells /mm1 2treated and control rats. For each group (NT-3-treated and and 9466.3 MBP cells /mm in NT-3-treated rats, versus

113I. Jean et al. / Brain Research 972 (2003) 110–118

in NT-3-treated animals compared to controls. The ob-servation of the subventricular zone and of the contralater-al CC at D3 and at D15 did not reveal clusters of cellslabelled for BrdU or for OL markers (not shown).

In the epicentre of the lesion, where the demyelinatedarea was most extensive, immunolabelling with myelinmarkers (CNP and MBP) showed, at day 15, a disappear-ance of labelled sheaths and only a few surviving OL, inthe treated animals (Fig. 3C,E) as well as in controls (Fig.3D). This lack of labelled fibres corresponded to theoutlines of staining observed with LFB.

In the inner rim of the demyelinated area, at D15 (Table1 11), we observed some A2B5 (Fig. 3F) and CNP cells

(Table 1), which were slightly decreased, in the NT-3-1 2treated group (134.6618 A2B5 cells /mm and

1 2113.669.5 CNP cells /mm ), but the differences for these1markers compared to controls (142.3624.6 A2B5 cells /

2 1 2mm and 136.7614.1 CNP cells /mm ) were not signifi-cant (Student’st-test and one-way ANOVA). In contrast,

1the number of MBP cells in the lesion at D15 (Fig. 3E)was significantly higher in the treated animals than incontrols (166% of labelled cells by optic field; i.e.

1 2181.869.6 MBP OL/mm in NT-3-treated rat, vs.Fig. 1. Histological staining (LFB-hematein-phloxin) of the corpus 109.468.3 in controls) (Student’st-test: P,0.05, andcallosum (CC) of a NT-3-treated animal (A) and a control rat (B), 15

ANOVA: F.F ). Thus between D3 and D15, the0.001days after LPC injection. The demyelination is detectable by the1number of mature (MBP ) OL had increased, moredisappearance of the blue LFB staining in the CC (star in (A) and (B)).

significantly, by 93.4% in NT-3-treated rats (Student’sNote the reduced area of demyelination in NT-3-treated animal ((A), the25entire lesion covers only one optic field), versus control ((B), the lesion t-test:P,5310 , and ANOVA:F.F ), versus 47.2%0.001

23covers two optic fields). Bar5100 mm in (A) and (B). in controls (Student’st-test: P,5310 , and ANOVA:1F.F ), whereas A2B5 POL had increased, respective-0.011 2 1 291.767 A2B5 cells /mm , 69.465 CNP cells /mm and ly by 8.6% in NT-3-treated animals (non significant)

1 274.365.7 MBP cells /mm in control rats (n59; non compared to 55.2% in controls (ANOVA:F.F ). In the0.011significant). NT-3-treated animals presented a non sig- lesion area, the total number of POL and MBP OL

11 1nificantly increased number of mature dividing (MBP / (A2B5 and MBP cells) was 26% greater in NT-3-treated

1BrdU ) cells compared to controls; whereas in control rats compared to controls (respectively 316.4627.6 and1 1 2rats, differentiated CNP /BrdU OL increased signifi- 251.7632.9 cells /mm ).

1cantly (Fig. 3A,B; Table 1). The number of A2B5 / Immunolabelling for GFAP, used to identify astrocytes,1BrdU double-labelled cells was not significantly different was restricted to the epicentre of the demyelinated area and

to the needle track. At D15, there was no significantdifference between the two groups (Table 1) either in the

2number (474.4619.5 cells /mm for NT-3-treated rats vs.2479623 cells /mm for controls,n518) or in the dis-1tribution of GFAP cells.

4 . Discussion

Lysolecithins, such as LPC, are well known for injuringspecifically myelin sheaths by rehydration at the in-traperiodic line, and for producing little axonal damage[11,23]. Using a stereotaxic approach, LPC injections can

Fig. 2. Size of the demyelinated lesion, at day 15 after LPC injection. be made in particular structures and the volume of theThe area and the length were measured on serial sections stained with

lesion is controlled. In the injected area, and contrary toLFB. In the presence of NT-3, the length is reduced by 37%, the surfaceother models with axonal sections, ‘naked’ axons, whichby 29% and the volume by 56%. *Significant difference between the two

groups of animals, according to the Student’st-test and ANOVAF-test. have been demyelinated, are still present [2]. The presence

I. Jean et al. / Brain Research 972 (2003) 110–118114

Fig. 3. Immunolabelling of the lesion at day 3 (A, B) and day 15 (C–F) in control (A, B, D, F) and in NT-3-treated animals (C, E) with anti-CNP (A, B),1 1anti-MBP (C, D, E) and A2B5 (F) antibodies. Double CNP /BrdU immunolabelling, at D3, in the corpus callosum (CC) of a control animal (A) and

same zone at higher magnification with a double-stained cell (B, arrowhead) in the border of the demyelinated area. MBP immunolabelling in thedemyelinated area (star) of the CC at day 15 in NT-3-treated animal (C, E), and in control (D). Note the decrease in the area of demyelination in

1NT-3-treated rat (C) compared to control (D). (E) Detail of a MBP OL (arrowhead) in the rim of the demyelinated zone (C), in NT-3-treated rat, athigher magnification. (F) A2B5 immunolabelled cells (arrows, hematein counterstain), in the rim of the demyelinated area at D15, in a control rat. Bar550mm in (A), 2 mm in (B), 100mm in (C) and (D), 10mm in (E) and (F).

Table 1Cell proliferation 3 days after lesion and immunolabelling 3 and 15 days after lesion

2Antibodies Number of cells /mm

Day 3 Day 15

LPC LPC1NT-3 LPC LPC1NT-31BrdU 169.7617.2 148.6617.11 1BrdU /A2B5 49.366 (53.7%) 50.565.6 (40.7%)1A2B5 91.767 123.9611 142.3624.6 134.66181 1BrdU /CNP 69.265.6 (99.7%)* 3865.5 (60.4%)1CNP 69.465 63.166.6 136.7614.1 113.669.51 1BrdU /MBP 51.265.6 (68.9%) 60.166 (63.9%)1MBP 74.365.7 9466.3 109.468.3 181.869.6*1GFAP 479623 474.4619.5

BrdU was injected 48 h after lesion and immunolabelling was quantified on day 3 in the totality of the demyelinated area (mean6S.D. from four separateexperiments run in triplicate). The percentages in parentheses refer to the proportion of dividing cells relative to the total number of cells labelled by thesame marker (A2B5, CNP or MBP).

24*Significant difference between NT-3-treated and control animals (P,5310 , F.F ).0.001

115I. Jean et al. / Brain Research 972 (2003) 110–118

of axons is believed to allow myelin repair by maintaining significant from controls. Since NT-3 was reported toaxonal signals, which are needed for a correct remyelina- enhance OL survival in vitro [6], it could have rescuedtion (for review, see Ref. [34]). This model was therefore cells of the OL lineage at the time of LPC injury.chosen in the present experiment to test whether NT-3 Nevertheless, the differences between the numbers of

1 1 1could directly improve myelin repair, in contrast with that A2B5 , CNP and MBP cells in NT-3-treated rats andused by McTigue et al. [38], which comprised axonal controls at D3 are too small and non significant to approvesections susceptible to modification by the remyelination this hypothesis. Therefore, a protective effect of NT-3process. Previous data revealed that 1ml of LPC is toward LPC-induced demyelination appears unlikely.necessary and enough to induce maximal demyelination of In contrast, the supra-lesional injection of NT-3 allowedthe ipsilateral corpus callosum in the adult rat (lower to decrease significantly the lesion volume by 56%, at dayvolumes induce a small area of demyelination which is 15 after injury, compared to controls. Since, as stateddifficult to detect after several days; higher volumes above, spontaneous remyelination also occurs in ourimplicate major tissue damage) [2]. model, this result indicates a more rapid speed of remyeli-

After the demyelinating injury induced by LPC, a nation. This observation is corroborated by the pronounced1spontaneous repair begins around the 8th post-lesion day in increase in the number of mature MBP OL (166%) in

the rat, and is completed after 1–3 months [2,4,11]. It was the injured area at D15, in NT-3-treated rats versusalso demonstrated in our experiments by a 44% decrease in controls. In contrast, the number of POL was not modifiedthe lesion volume between D3 and D15 in control animals. by NT-3 either at D3 or at day 15. Unexpectedly, BrdUThis spontaneous remyelination can also be observed in incorporation experiments did not show, at D3, an increasehuman demyelinating diseases such as multiple sclerosis, in the total number of dividing cells in the lesion area, butnevertheless, as it is incomplete in that case [51], several a modest modification of their maturation stage in NT-3-approaches, such as growth factor injections [2,55,56], treated animals. So, at the same time (day 3), control

1 1have been explored in order to improve the reconstruction animals had more differentiated (CNP /BrdU ) dividingof myelin sheaths. However, although several growth OL than treated rats, in which there was a tendency for

1 1factors have been identified in vitro for their effects on the more mature (MBP /BrdU ) dividing OL. Consideringproliferation or on the differentiation of the cells of the OL the duration between BrdU injections and animal sacrificecell lineage (for review, see Ref. [52]), such as PDGF (24 h), it is conceivable that this observation results from

1 1[5,8,40], NT-3 [7], IGF-1 [37], and bFGF [15,19], little is CNP or MBP cell proliferation following LPC and/orknown about their role in vivo. Only IGF-1 [55], PDGF [2] LPC1NT-3 injection. Indeed, cell cycle duration for adult

1 1and possibly NT-3 [38] are known to improve remyelina- POL is around 65 h in vitro [50]. Therefore CNP /BrdU1 1tion in vivo. Therefore, in this study, we tested the ability or MBP /BrdU double immunolabelling may arise from

1of NT-3, a molecule of the neurotrophin family, to enhance the division of already differentiated (CNP ) and/or1remyelination specifically. NT-3 is involved in POL prolif- mature (MBP ) OL, rather than from the incorporation of

2 2 1eration and in long-term survival of mature OL in vitro BrdU by POL (CNP /MBP /BrdU ) followed by their1 1 1 1[7,13,30], although its mitogenic effect is debated [7,16]. It differentiation into CNP /BrdU or MBP /BrdU cells.

was reported to enhance remyelination in vivo together Proliferation of such differentiated or mature OL, althoughwith axonal sprouting in a model of spinal cord injury [38]. controversial, has already been observed in vitro (reviewedThus, NT-3 effects on remyelination could be secondary to in Ref. [51]) as well as in vivo (e.g. Ref. [4]). Our result

1axonal regeneration in that case, and we decided to test may also suggest either an accelerated maturation of CNP1 1NT-3 efficacy in a model of ‘pure’ demyelination (i.e. toward MBP cells, or an increased tendency of MBP

devoid of axonal sections), by comparing spontaneous OL to proliferate, in the presence of NT-3. Such arepair to that observed with NT-3. phenomenon was reported in vitro with bFGF and PDGF

The effects that we recorded are NT-3 specific since its [18]. Other origins for remyelinating OL have also beenreplacement by the vehicle buffer alone (control rats), or an checked by careful observation of contralateral corpusirrelevant protein (bovine serum albumin) was ineffective. callosum and of the subventricular zone. At the timesThe dose of 1mg of NT-3 was chosen in comparison with studied we did not notice evidence that cells of the OLin vitro studies and on the basis of our experience with lineage were recruited from these zones, nor did wePDGF [2]. Concentrations ranging from 2 pg/ml to 2 observe differences between control and treated group.ng/ml of NT-3 are generally used in vitro, and the effects The enhanced remyelination that we observe in NT-3-of NT-3 are recorded after a period ranging from 24 h to treated animals, might result from a direct effect of NT-314 days [6]. Previous studies in vivo with PDGF [2] on cells of the OL lineage. Since axonal sections areshowed that lower doses of this growth factor resulted in minimal in this type of toxic demyelination [2,26,52], it isnon significant improvement in myelin repair, and that unlikely that NT-3 has promoted remyelination mainlyhigher doses resulted in too large volumes for injections. through an indirect, neurotrophic effect. Nevertheless,

Although the volume of demyelination was smaller at axons in the demyelinated area could have been targetedD3 in NT-3-treated animals, this was not statistically by NT-3 and present more subtle changes favouring

I. Jean et al. / Brain Research 972 (2003) 110–118116

1remyelination. For example, in vitro NT-3 is necessary for MBP OL in the NT-3-treated animals. These observa-the terminal differentiation of neurons [8]. Since NT-3 did tions extend those of McTigue et al. [38] and demonstratenot modify the number of astrocytes in the injured area that NT-3 belongs to the group of growth factors able tocompared to controls, and we did not observe any spread enhance remyelination in vivo, together with IGF-1 [55,56]of astrocytic gliosis in NT-3-treated rats (compared to and PDGF [1,2]. These results may allow to set NT-3controls), thus the involvement of astrocytes is also rather among the growth factors with putative therapeutic effectsunlikely. Although astrocytes express trk C receptors at for human demyelinating diseases.low level in vitro, their proliferation is not affected byNT-3 [14,28]. However, it cannot be excluded that theymight respond to NT-3 by secretion of other growth factors A cknowledgementsfavouring remyelination. Therefore it can only be hypoth-esized that NT-3 might enhance remyelination directly by This work was supported by the ARSEP (Associationtargeting cells of the OL lineage, and act by subtly driving ´pour la Recherche contre la Sclerose En Plaques). IJ was atheir proliferation and maturation, as it occurs in vitro to recipient from the LFSEP (Ligue Francaise contre lapromote the maturation of OL [54]. Whereas POL number ´Sclerose En Plaques).and POL proliferation were not significantly increased byNT-3, our data tend to demonstrate that differentiated

1 1(CNP ) and mature (MBP ) OL proliferate duringR eferencesspontaneous remyelination, and that NT-3 enhances myelin

repair, either by increasing the speed of maturation of1 1 [1] C. Allamargot, M. Brouillard, A. Pouplard-Barthelaix, C. Fres-CNP toward MBP OL, or by increasing the number of1 sinaud, PDGF improves remyelination in vitro and in vivo. J.MBP proliferating cells. The presence of a limited

1 Neurochem. 66(S2) (1996) Abs S50S.number of dividing CNP cells at D3 in NT-3-treated[2] C. Allamargot, A. Pouplard-Barthelaix, C. Fressinaud, A single

animals, compared to controls, is in favour of the second intracerebral microinjection of platelet-derived growth factorhypothesis. Such an accelerated maturation, or the prolifer- (PDGF) accelerates the rate of remyelination in vivo, Brain Res. 918

(2001) 28–39.ation of mature OL, may represent a very efficient way to[3] Y. Akiyama, O. Honmou, T. Kato, Y. Uede, K. Hashi, J.D. Kocsis,improve remyelination, since these mature cells are far

Transplantation of clonal neural precursor cells derived from adultmore abundant than POL in the adult CNS. The fact thathuman brain established functional peripheral myelin in the rat

the volume of demyelination decreases, and the number of spinal cord, Exp. Neurol. 167 (2001) 27–39.1MBP cells increases significantly only after 15 days in [4] L.S. Arenella, R.M. Herndon, Mature oligodendrocytes. Division

following experimental demyelination in adult animals, Arch.NT-3-treated rats versus controls, also favours the hypoth-Neurol. 41 (1984) 1162–1165.esis that cells of the OL lineage are not rescued at the time

[5] R. Armstrong, V.L. Friedrich Jr., K.V. Holmes, M. Dubois-Dalcq, Inof LPC injection (no significant differences at D3), but thatvitro analysis of the oligodendrocyte lineage in mice during de-

differentiated OL slowly proliferate and remyelinate axons myelination and remyelination, J. Cell Biol. 111 (1990) 1183–1195.with more efficiency in the presence of NT-3. [6] B.A. Barres, R. Schmid, M. Sendtner, M.C. Raff, Multiple extracel-

Despite these properties, the transduction mechanisms of lular signals are required for long-term oligodendrocyte survival,Development 8 (1993) 283–295.NT-3 in the glial cells are still unclear; NT-3 binds a

[7] B.A. Barres, M.C. Raff, F. Gaese, I. Bartke, G. Dechant, Y.A. Barde,receptor with tyrosine kinase activity, belonging to the trkA crucial role for neurotrophin-3 in oligodendrocyte development,

family, which induces activation of the MAPKs pathway Nature 367 (1994) 371–375.[13,27]. These trk receptors are present on astrocytes, [8] L. Bertollini, M.T. Ciotti, E. Cherubini, A. Cattaneo, Neurotrophin-3neurons, and cells of the oligodendrocyte lineage such as promotes the survival of oligodendrocyte precursors in embryonic

hippocampal cultures under chemically defined conditions, BrainPOL and OL [30,31]; the expression of their mRNA is notRes. 746 (1997) 19–24.modified by lysolecithin-induced demyelination in spinal

[9] F. Besnard, F. Perraud, M. Sensenbrenner, G. Labourdette, Platelet-cord [25] whereas it is increased by CNS injury [39]. Our derived growth factor is a mitogen for glial but not for neuronal ratstudy suggests that such receptors are also functional in brain cells in vitro, Int. J. Dev. Neurosci. 7 (1987) 401–409.mature OL of the adult CNS, when they are triggered by [10] F. Besnard, F. Perraud, M. Sensenbrenner, G. Labourdette, Effects

of acidic and basic fibroblast growth factors on proliferation andexogenous NT-3.maturation of cultured rat oligodendrocytes, Int. J. Dev. Neurosci. 7(1989) 401–409.

[11] W.F. Blakemore, R.A. Eames, K.J. Smith, W.I. McDonald, Remyeli-5 . Conclusion nation in the spinal cord of the cat following intraspinal injections of

lysolecithin, J. Neurol. Sci. 33 (1977) 31–43.[12] P.D. Canoll, J.M. Musacchio, R. Hardy, R. Reynolds, M.A. Mar-After a chemical demyelination induced by LPC in the

chionni, J.L. Salzer, GGF/neuregulin is a neuronal signal thatadult rat CNS, our study discloses an accelerated speed ofpromotes the proliferation and survival and inhibits the differentia-

remyelination in the presence of NT-3. This phenomenon tion of oligodendrocyte progenitors, Neuron 17 (1996) 229–243.was demonstrated after 15 days by the decrease in the[13] R.I. Cohen, R. Marmur, W.T. Norton, M.F. Mehler, J.A. Kessler,lesion volume, as well as the increased number of mature Nerve growth factor and neurotrophin-3 differentially regulate the

117I. Jean et al. / Brain Research 972 (2003) 110–118

proliferation and survival of developing rat brain oligodendrocytes, [34] C. Lubetzki, B. Stankoff, The role of axonal signals in theJ. Neurosci. 16 (1996) 6433–6442. myelination of the central nervous system, Pathol. Biol. 48 (2000)

[14] D.F. Condorelli, T. Salin, P. Dell’Albani, G. Mudo, M. Corsato, T. 63–69.Timmusk, M. Metsis, N. Belluardo, Neurotrophins and their trk [35] W. McDonald, M.J. Howard, Repairing the damaged spinal cord: areceptors in cultured cells of the glial lineage and in white matter of summary of our early success with embryonic stem cell transplanta-the central nervous system, J. Mol. Neurosci. 6 (1995) 237–248. tion and remyelination, Prog. Brain Res. 137 (2002) 299–309.

[15] P.A. Eccleston, D.H. Silberberg, Fibroblast growth factor is a [36] R.D. McKinnon, G. Piras, J.A. Ida Jr, M. Dubois-Dalcq, A role formitogen for oligodendrocytes in vitro, Brain Res. 353 (1985) 315– TGF-beta in oligodendrocyte differentiation, J. Cell Biol. 121318. (1993) 1397–1407.

[16] U. Engel, G. Wolswijk, Oligodendrocyte-type-2 astrocyte (O-2A) [37] F.A. McMorris, M. Dubois-Dalcq, Insulin-like growth factor 1progenitor cells derived from adult rat spinal cord: in vitro charac- promotes cell proliferation and oligodendroglial commitment in ratteristics and response to PDGF, bFGF and NT-3, Glia 16 (1996) glial progenitor cells developing in vitro, J. Neurosci. Res. 2116–26. (1988) 199–209.

[17] R.J. Franklin, J.M. Gilson, W.F. Blakemore, Local recruitment of [38] D.M. McTigue, P.J. Horner, B.T. Stockes, F.H. Gage, Neurotrophin-remyelinating cells in the repair of demyelination in the central

3 and Brain Derived neurotrophic factor induce oligodendrocytesnervous system, J. Neurosci. Res. 50 (1997) 337–344.

proliferation and myelination of regenerating axons in the contused`[18] C. Fressinaud, L.L. Sarlieve, G. Labourdette, Regulation of cere-

adult rat spinal cord, J. Neurosci. 18 (1998) 5354–5365.broside sulfotransferase activity in cultured oligodendrocytes: effect

[39] I. Mocchetti, J.R. Wrathall, Neurotrophic factors in central nervousof growth factors and insulin, J. Cell. Physiol. 141 (1989) 667–674.system trauma, J. Neurotrauma 12 (1995) 853–870.[19] C. Fressinaud, P. Laeng, G. Labourdette, J. Durand, J.M. Vallat, The

[40] M. Noble, K. Murray, P. Stroobant, D. Waterfield, P. Riddle, Plateletproliferation of mature oligodendrocytes in vitro is stimulated byderived growth factor promotes division and motility and inhibitsbasic fibroblast growth factor and inhibited by oligodendrocyte-typepremature differentiation of the oligodendrocyte/ type 2 astrocyte2 astrocyte precursors, Dev. Biol. 158 (1993) 317–329.progenitor cell, Nature 333 (1988) 560–562.[20] A.L. Gard, S.E. Pfeiffer, Two proliferative stages of the oligoden-

[41] M. Noble, G. Wolswijk, Development and regeneration in the O-2Adrocyte lineage (A2B51O42 and O41GalC2) under differentlineage: studies in vitro and in vivo, J. Neuroimmunol. 40 (1992)mitogenic control, Neuron 5 (1990) 615–625.287–293.[21] J.A. Gensert, J.E. Goldman, Endogenous progenitors remyelinate

[42] G. Paxinos, C. Watson, The Rat Brain in Stereotaxic Coordinates,demyelinated axons in the adult CNS, Neuron 19 (1997) 197–203.Academic Press, San Diego, CA, 1986, 280 pp.[22] A. Hall, N.A. Giese, W.D. Richardson, Spinal cord oligodendrocytes

[43] N.P. Pringle, H.S. Mudhar, E.J. Collarini, W.D. Richardson, PDGFfrom ventrally derived progenitor cells that express PDGF alpha-receptors in the rat CNS: during late neurogenesis, PDGF alphareceptors, Development 122 (1996) 4085–4094.receptor expression appears to be restricted to glial cells of the[23] S.M. Hall, The effect of injection of LPC into the white matter of

the adult mouse spinal cord, J. Cell Sci. 10 (1972) 535–546. oligodendrocyte lineage, Development 115 (1992) 535–551.[24] R. Hardy, R. Reynolds, Proliferation and differentiation potential of [44] M.C. Raff, R.H. Miller, M. Noble, A glial progenitor cell that

rat forebrain oligodendroglial progenitors both in vitro and in vivo, develops in vitro into an astrocyte or an oligodendrocyte dependingDevelopment 111 (1991) 1061–1080. on culture medium, Nature 302 (1983) 390–396.

[25] G.L. Hinks, R.J. Franklin, Distinctive patterns of PDGF-A, FGF-2, [45] W.D. Richardson, N. Pringle, M.J. Mosley, B. Westermark, M.IGF-1, and TGF-beta gene expression during remyelination of Dubois-Dalcq, A role for platelet-derived growth factor in normalexperimentally-induced spinal cord demyelination, Mol. Cell. gliogenesis in the central nervous system, Cell 53 (1988) 309–319.Neurosci. 14 (1999) 153–168. [46] B. Rogister, S. Belachew, G. Moonen, Oligodendrocytes: from

[26] I. Jean, C. Allamargot, A. Barthelaix-Pouplard, C. Fressinaud, development to demyelinated lesion repair, Acta Neurol. Biol. 99Axonal lesions and PDGF enhanced remyelination in the rat corpus (1999) 32–39.callosum after lysolecithin demyelination, Neuroreport 13 (2002) [47] B. Rogister, T. Ben-Hur, M. Dubois-Dalcq, From neural stem cells627–631. to myelinating oligodendrocytes, Mol. Cell. Neurosci. 14 (1999)

[27] J.R. Johnson, A.K. Chu, C. Sato-Bigbee, Possible role of CREB in 287–300.´ ` ´the stimulation of oligodendrocyte precursor cell proliferation by [48] D. Schwartz, Methodes statistiques a l’usage des medecins et des

´neurotrophin 3, J. Neurochem. 74 (2000) 1409–1417. biologistes, Flammarion Medecine Sciences, Paris, 1989, p. 306.[28] M.A. Kahn, S. Kumar, D. Liebl, R. Chang, L.F. Parada, J. deVellis, [49] N.J. Scolding, P.J. Rayner, D.A.S. Compston, Identification of

Mice lacking NT-3, and its receptor TrkC, exhibit profound de- A2B5-positive putative oligodendrocyte progenitor cells and A2B5-ficiencies in CNS glial cells, Glia 26 (1999) 153–165. positive astrocytes in adult human white matter, Neuroscience 89

[29] H.S. Keirstead, J.M. Levine, W.F. Blakemore, Response of the (1999) 1–4.oligodendrocyte precursor cell population (defined by NG2 label- [50] G. Wolswijk, M. Noble, Identification of an adult specific glialling) to demyelination of the adult spinal cord, Glia 22 (1998) progenitor cell, Development 102 (1989) 387–400.161–170. [51] P.M. Wood, R.P. Bunge, The origin of remyelinating cells in the

[30] S. Kumar, M.A. Kahn, J. de Vellis, NT-3-mediated trk C receptor adult central nervous system: the role of the mature oligodendrocyte,activation promotes proliferation and cell survival of rodent Glia 4 (1991) 225–232.progenitor oligodendrocytes cells in vitro and in vivo, J. Neurosci. [52] R.H. Woodruff, R.J. Franklin, The expression of myelin proteinRes. 54 (1998) 754–765. mRNAs during remyelination of lysolecithin-induced demyelination,

[31] M.B. Lachyankar, P.J. Condon, P.J. Quesenberry, N.S. Litofsky, Neuropathol. Appl. Neurobiol. 25 (3) (1999) 226–235.L.D. Recht, A.H. Ross, Embryonic precursor cells that express trk [53] R.H. Woodruff, R.J. Franklin, Demyelination and remyelination ofreceptors: induction of different cell fates by NGF, BDNF, NT-3, the caudal cerebellar peduncle of adult rats following stereotaxicand CNTF, Exp. Neurol. 144 (1997) 350–360. injections of lysolecithin, ethidium bromide, and complement /anti-

[32] J.H. Levine, R. Reynolds, J.W. Fawcett, The oligodendrocyte galactocerebroside: a comparative study, Glia 25 (1999) 216–228.precursor cell in health and disease, Trends Neurosci. 24 (2001) [54] H. Yan, P.M. Wood, NT-3 weakly stimulates proliferation of adult rat

2 139–47. O1 O4 oligodendrocyte-lineage cells and increases oligoden-[33] L.E. Lillien, M. Sendtner, M.C. Raff, Extracellular matrix-associated drocyte myelination in vitro, J. Neurosci. Res. 62 (2000) 329–335.

molecules collaborate with ciliary neurotrophic factor to induce [55] D.L. Yao, X. Liu, L.D. Hudson, H. Webster, Insulin-like growthtype-2 astrocyte development, J. Cell Biol. 111 (1990) 635–644. factor I treatment reduces demyelination and upregulates gene

I. Jean et al. / Brain Research 972 (2003) 110–118118

expression of myelin-related proteins in experimental autoimmune factor-I given subcutaneously reduces clinical deficits, decreasesencephalomyelitis, Proc. Natl. Acad. Sci. USA 92 (1995) 6190– lesion severity and upregulates synthesis of myelin proteins in6194. experimental autoimmune encephalomyelitis, Life Sci. 58 (1996)

[56] D.L. Yao, X. Liu, L.D. Hudson, H.D. Webster, Insulin-like growth 1301–1306.