Prolonged Corticosterone Treatmentof Adult Rats Inhibits the

Proliferation of OligodendrocyteProgenitors Present ThroughoutWhite and Gray Matter Regions

of the BrainGERARD ALONSO*

CNRS-UMR5101, CCIPE, Montpellier Cedex 05, France

KEY WORDS glucocorticoids; astrocytes; microglia; germinative zones; remyelina-tion

ABSTRACT It is well established that glucocorticoids inhibit the proliferation ofprogenitor cells that occurs in the hippocampal dentate gyrus of adult mammals. Activecell proliferation also occurs in the subventricular zone (SVZ) of the lateral ventricle and,to a lesser extent, throughout white and gray matter regions of the adult brain. The aimof the present study was to determine whether extrahippocampal cell proliferation isalso affected by glucocorticoids. The cell proliferation marker bromodeoxyuridine (BrdU)was administered to control rats, to adrenalectomized rats, and to rats treated with adaily injection of corticosterone (10 mg/kg) for a period of 15 days. In control andadrenalectomized rats, high to low numerical densities of BrdU-labeled nuclei weredetected within the different forebrain regions examined. In rats treated with cortico-sterone, a dramatic decrease of cell proliferation was detected in the dentate gyrus, butalso throughout all white and gray matter regions examined, except for the SVZ of thelateral ventricle. Double-labeling experiments indicated that throughout the differentwhite and gray forebrain regions examined, except for the SVZ, BrdU-labeled nucleiwere essentially associated with cells immunostained for the marker of oligodendrocyteprogenitors NG2. These data indicate that glucocorticoids inhibit the proliferation ofoligodendrocyte precursors located throughout the white and gray matter regions of theadult rat brain. Since the proliferation of oligodendrocyte precursors plays a major rolein the processes of remyelination, these data raise the question of possible detrimentaleffects of therapeutic treatments of CNS trauma based on the administration of glu-cocorticoids. GLIA 31:219–231, 2000. © 2000 Wiley-Liss, Inc.

INTRODUCTION

In the developing CNS of rodents, neurons and glialcells arise from progenitor cells that actively prolifer-ate during the late embryonic and early postnatal pe-riods (Cowan, 1979). The rate of proliferation of neuralcells then dramatically decreases after the two firstpostnatal weeks (Allen, 1912; Bryans, 1959; Hommesand Leblond, 1967). Nevertheless, it has been knownfor some time that dividing cells still persist within the

CNS of adult and aged mammals. In the adult CNS,active cell proliferation of neuronal progenitors hasbeen reported to occur in two germinative zones: (1) thesubgranular zone of the dentate gyrus (DG), whichgenerates hippocampal interneurons (Altman and Das,

*Correspondence to: Gerard Alonso, CNRS-UMR5101. CCIPE, 141 rue de laCardonille, 34094 Montpellier Cedex 05, France.E-mail: alonso@bacchus.montp.inserm.fr

Received 10 March 2000; Accepted 18 May 2000

GLIA 31:219–231 (2000)

© 2000 Wiley-Liss, Inc.

1965; Kuhn et al., 1996); and (2) the subventricularzone (SVZ) of the lateral ventricle, which generatesinterneurons that migrate into the olfactory bulb (Loisand Alvarez-Buylla, 1993; Morshead et al., 1994; Gold-man, 1995; Hauke et al.; 1995). Although less activethan in the DG and the SVZ, cell proliferation has alsobeen detected in various white and gray matter regionsof the adult CNS (Hommes and Leblond, 1967; Korr etal., 1973; Mares and Loddin, 1974; McCarthy and Leb-lond, 1988). In these regions, most dividing cells werereported to be essentially glial precursors, althoughtheir nature has not been clearly identified (Hommesand Leblond, 1967; McCarthy and Leblond, 1988).

Steroid hormones have long been known to affect awide spectrum of brain functions. Since they easilycross through the blood-brain barrier, these moleculescan convey molecular information to neurons and glialcells. During the last few years, the attention has fo-cused on the effects of glucocorticoids on the neurogen-esis in the adult CNS. Indeed, the levels of circulatingglucocorticoids were found to highly influence the pro-liferation of progenitor cells that occurs in the dentategyrus of the hippocampus (Gould et al., 1992; Cameronand Gould, 1996; Gould and Cameron, 1996). More-over, glucorticoids have been suggested to be, at leastin part, responsible for the decreased neurogenesis ob-served in the dentate gyrus of aged animals (Sapolskyet al., 1985; Landfield and Eldridge, 1994; Lupien etal., 1998; Porter and Landfield, 1998; Cameron andMcKay, 1999).

Surprisingly, little is known on the effects of glu-cocorticoids on the proliferation of neural cells occur-ring in adult brain regions outside from the hippocam-pus. The present study was undertaken in order todetermine whether glucocorticoids could influence theproliferation of neural cells throughout various brainregions. The cell proliferation marker bromodeoxyuri-dine (BrdU) was used to study the effects of withdrawalor overloading of circulating corticosterone (by adrenal-ectomy or of prolonged corticosterone treatment, re-spectively) on the cell proliferation throughout variouswhite and gray regions of the adult rat forebrain.

MATERIALS AND METHODSAnimals

Male adult (2–3-month-old, 250–300-g) Sprague-Dawley rats (Iffa-Credo, l’Arbresle, France) were used.They were kept in light- (12-h light/12-h dark) andtemperature- (24 6 1°C) controlled rooms and had freeaccess to standard dry food and tap water.

Hormonal Treatment

Three groups of animals were considered: (1) controlrats (n 5 5), (2) rats bilaterally adrenalectomized for 2weeks (n 5 6), and (3) animals receiving a daily sub-cutaneous injection for 2 weeks of either corticosterone

(n 5 12), testosterone (n 5 4), or progesterone (n 5 4)(10 mg/kg in 0.2 ml sesame oil). Control and adrena-lectomized animals were treated with a daily injectionof 0.2 ml sesame oil for 2 weeks. Animals were sacri-ficed the day of the last injection of sesame oil or steroidhormone.

BrdU Administration

BrdU (100 mg/kg, Sigma) was administered intra-peritoneally twice a day during the 3 days precedingtheir perfusion (i.e., the 12th, 13th, and 14th days oftreatment). At 12 h after the last BrdU administration,animals were perfused as described in the followingparagraph.

Tissue Preparation

After deep anesthesia with equithesin, animals wereperfused through the ascending aorta with phosphate-buffered saline (PBS), pH 7.4, followed by 500 ml offixative composed of 4% paraformaldehyde in 0.1 Mphosphate buffer, pH 7.4. The forebrain was dissectedand fixed by immersion in the same fixative for 12 h at4°C. It was then cut frontally with a vibratome into40–50-mm-thick sections. The sections were carefullyrinsed in PBS and subsequently treated for single per-oxidase or double fluorescence immunostaining.

Peroxidase Immunostaining

Vibratome sections were incubated with 2 N HCL for30 min at room temperature, carefully rinsed, and in-cubated as follows: (1) for 48 h at 4°C with an anti-BrdU antibody (monoclonal mouse IgG; NovocastraLaboratories, Newcastle, UK; diluted 1:100); (2) for12 h at 4°C with a peroxidase-labeled Fab fragment ofgoat IgG anti-mouse IgG (Biosys, Compiegne, France,diluted 1:1,000; and (3) with 0.1% 3,39-diaminobenzi-dine diluted in 0.05 M Tris buffer, pH 7.3, in the pres-ence of 0.2% H2O2. The primary and secondary anti-bodies were diluted in PBS containing 0.1% TritonX-100, 1% bovine serum albumin (BSA), and 1% nor-mal goat serum. Immunostained sections weremounted in Permount and observed under a light mi-croscope. For each of the different groups of rats, seriesof sections passing through the rostral parts of thestriatum or of the hippocampus were selected andtreated for the quantitative analysis of BrdU-immuno-stained nuclei within specific forebrain regions. In or-der to avoid any bias, the nature of the sections wasmasked during the quantitative analysis. Using the310 objective of an Axioskop Zeiss microscope con-nected to an image analyzer System (Samba, Unilog,France), the number of BrdU-labeled nuclei wascounted within square fields of the 150-mm side (finalmagnification 3450) centered on specific structures,

220 ALONSO

including the corpus callosum, fimbria, optic chiasma,ventromedial hypothalamus, lateral septum, piriformcortex, hippocampal dentate gyrus, and SVZ of thelateral ventricle and of the alveus hippocampus (seeFig. 1). For each structure, 2 to 6 fields per sectionswere analyzed on at least four sections per animal.Within each field analyzed, the numerical density wasexpressed as number of BrdU-labeled nuclei/mm2.

Double Fluorescence Immunostaining

Sections were first treated for the immunofluores-cence detection of various neuronal or glial markers, orfor the lectin labeling of microglial cells. For this theywere incubated for 48 h at 4°C with a specific primaryantibody or with a biotin-labeled isolectin B4 (isoB4,Sigma, diluted 1:500). The primary antibodies usedincluded: (1) mouse IgM monoclonal antibodies againstpolysialylated neural cell adhesion molecule (PSA-NCAM, kindly provided by Dr. Seki, diluted 1:1,000),and myelin/oligodendrocyte specific protein (OligoM,Chemicon, diluted 1:500); and (2) rabbit IgG polyclonalantibodies against glial fibrillary acidic protein (GFAP;Dako, diluted 1:1,000), S-100 (Sigma, diluted 1:500),glucocorticoid receptor (GR; Santa Cruz, diluted 1:200),and NG2 chondroitin sulfate proteoglycan (NG2;Chemicon, diluted 1:500). After careful rinsing withPBS, sections were incubated for 2 h with either (1)fluorescein (FITC)-conjugated secondary antibodiesagainst mouse IgM or rabbit IgG (Sigma, diluted 1:200)for the sections treated with a primary antibody, orextravidin-FITC (Sigma, diluted 1:200) for the sectionstreated with the biotin-labeled lectin. The labeled sec-tions were then treated as described above for immu-nodetection of BrdU. The anti-BrdU antibody was thenrevealed with an anti-mouse IgG secondary antibodyconjugated with Cy3 (Jackson Laboratories, WestGrove, PA). All primary and secondary antibodies werediluted in PBS, pH 7.4 containing 0.1% Triton X-100and 1% normal goat serum.

After careful rinsing, the double-labeled sectionswere mounted in Mowiol (Calbiochem, La Jolla, CA)and examined under a Leica TCS 4D confocal laserscanning microscope equipped with a krypton/argonmixed gas laser. Two laser lines emitting at 488 nmand 568 nm were used for exciting the fluorescein- andCy3-conjugated secondary antibodies, respectively.The background noise of each confocal image was re-duced by averaging eight image inputs. Unaltered dig-italized images were transferred to a PC computer andPowerpoint (Microsoft) was used to prepare and printfinal figures.

Some sections were treated for the quantitative eval-uation of the frequency of the association of BrdU-labeled nuclei with the different types of labeled cells.The x 25 objective was used to collect confocal images ofdouble immunostained sections centered on the differ-ent forebrain regions analyzed.

Controls for both peroxidase and fluorescence immu-nostaining consisted of (1) omitting the primary anti-bodies and applying the secondary antibodies alone, (2)applying each primary antibody sequentially and thenreacting them with an inappropriate secondary anti-body, and (3) exciting each fluorochrome by the inap-propriate laser line. With this approach, we were ableto confirm that the secondary antibodies used did notinduce artifactual labeling.

RESULTSPeroxidase Immunostaining for BrdU

Control rats

In all the sections examined, BrdU immunostainingwas typically associated with the nucleus of a numberof cells distributed throughout the different forebrainregions. The typical distribution of such BrdU-labelednuclei in control rats is schematically illustrated inFigure 1, in two frontal sections cut through two ros-trocaudal forebrain levels.

In the different forebrain regions examined theBrdU-labeled nuclei exhibited various immunostainingintensities, indicating that some cells have dividedsince administration of BrdU was stopped. Whateverthe immunostaining intensity, three main forebrainstructures could be distinguished in terms of their nu-merical density in BrdU-labeled nuclei:

1. Structures exhibiting high numerical densities ofBrdU-labeled nuclei, including the SVZ of the lat-eral ventricle bordering the dorsorostral portions ofthe lateral ventricle, and the SVZ of the alveushippocampus, where they formed aggregates of var-ious sizes along the ventricular border of the corpuscallosum (Fig. 2A,B)

Fig. 1. Schematic representation of frontal sections through theanterior (A) and posterior (B) forebrain illustrating the general dis-tribution of bromodeoxyuridine (BrdU)-labeled nuclei (black dots) andthe location of the forebrain structures in which the numerical densityof the labeled nuclei was quantified (open squares). Alv, subventricu-lar zone of the alveus hippocampus; CCa, anterior part of the corpuscallosum; CCp, posterior part of the corpus callosum; ChO, opticchiasma; Cx, cortex; DG, dentate gyrus of the hippocampus; Fb, fim-bria; Hyp, mediobasal hypothalamus; Sept, lateral septum; SVZ, sub-ventricular zone of the lateral ventricle.

221CORTICOSTERONE INHIBITS PROLIFERATION OF OLIGODENDROCYTE PROGENITORS

2. Structures exhibiting moderate numerical densitiesof BrdU-labeled nuclei including the dentate gyrusof the hippocampus, and most white matter regionssuch as the corpus callosum, fimbria and optic chi-asma (Fig. 2C,D)

3. Structures exhibiting low numerical densities of la-beled nuclei, including various gray matter regionssuch as the superficial cortical layers, dorsal hip-pocampus, lateral septum, and ventromedial hypo-thalamus (Fig. 2E,F)

Fig. 2. Peroxidase immunostaining of bromodeoxyuridine (BrdU) indifferent forebrain regions of control rats. High, moderate, and lownumerical densities of BrdU-labeled nuclei are observed, respectively,(1) in the subventricular zones of the lateral ventricle (A) and of the

alveus hippocampus (B), (2) in the subgranular zone of the hippocam-pal dentate gyrus (C) and in the corpus callosum (D), and (3) in thepiriform cortex (E) and the mediobasal hypothalamus (F). CC, corpuscallosum; St, striatum; V, lateral ventricle. Scale bar 5 100 mm.

222 ALONSO

Fig. 3. Peroxidase immunostaining of bromodeoxyuridine (BrdU) inthe corpus callosum (A–C), the optic chiasma (D–F), the lateral sep-tum (G–I) and the mediobasal hypothalamus (J–L) of control (Ct;A,D,G,J), adrenalectomized (Adx; B,E,H,K), and corticosterone-treated (Cort; C,F,I,L) rats. The number of BrdU-labeled nuclei de-tected within both white and gray matter regions is not markedly

modified after adrenalectomy, as compared with controls, whereas itis dramatically decreased in all regions after corticosterone treat-ment. CC, corpus callosum; ChO, optic chiasma; Cx, cortex; Hippo,hippocampus; LV, lateral ventricle; IIIV, third ventricle. Scale bar 5300 mm.

Experimental rats

The effects of corticosteroids on cell proliferationwere determined by comparing the numerical densityof BrdU-labeled nuclei detected in the different fore-brain regions of control rats with that detected in thesame regions after adrenalectomy or after prolonged

administration of corticosterone, testosterone or pro-gesterone (Figs. 3–5).

In adrenalectomized rats, the distribution pattern ofthe BrdU-labeled nuclei throughout the white and grayregions of the forebrain was grossly similar to thatobserved in control rats. The only modification was aslight but significant increase in the numerical density

Fig. 4. Peroxidase immunostaining of bromodeoxyuridine (BrdU) inthe subventricular zones of the lateral ventricle (A–C) and of thealveus hippocampus (D–F), and in the hippocampal dentate gyrus(G–I) of control (Ct; A,D,G), adrenalectomized (Adx; B,E,H) and cor-ticosterone-treated (Cort; C,F,I) rats. The number of BrdU-labelednuclei detected in the three forebrain regions is not markedly modi-

fied in the adrenalectomized as compared with the control rats. In thecorticosterone-treated rats, the number of BrDU-labeled nuclei is notmodified in the subventricular zones of the lateral ventricle (A–C) andthe alveus hippocampus (D–F), whereas it is markedly decreased inthe dentate gyrus (G–I) and the corpus callosum (C,F). CC, corpuscallosum; St, striatum; V, lateral ventricle. Scale bar 5 300 mm.

224 ALONSO

of the labeled nuclei detected in the corpus callosumand the dentate gyrus.

By contrast, in rats treated with corticosterone, spec-tacular modifications were observed in the pattern ofBrdU immunostaining as compared with either thecontrol or the adrenalectomized rats: the number oflabeled nuclei was dramatically decreased in all theforebrain regions examined, except for the SVZ and theSVZ of the alveus hippocampus. Such a decrease in thenumber of labeled nuclei was particularly spectacularwithin the white matter regions that were found toonly contain scarce labeled nuclei.

In animals treated with testosterone or progester-one, no clear modification could be detected in theBrdU immunostaining pattern throughout the differ-ent forebrain regions analyzed.

Double Fluorescence Labeling of BrdU and ofSpecific Cell Markers

In the different forebrain regions analyzed, the im-munostaining patterns obtained with the different celltype markers used fully conformed to previous descrip-tions. In order to identify the proliferating cells de-tected in the different forebrain regions, sections weretreated for double fluorescence labeling of these specificcell type markers and of BrdU. Table 1 shows thequantitative evaluation of the proportion of BrdU-la-beled nuclei associated with the different cell typeswithin the different forebrain regions examined.

Double labeling of BrdU and PSA-NCAM

Numerous cell profiles labeled for PSA-NCAM wereobserved in previously reported locations such as theSVZ (Bonfanti and Theodosis, 1994) and the dentategyrus of the hippocampus (Seki and Arai, 1991), butalso in the SVZ of the alveus hippocampus. BrdU-labeled nuclei were found to be associated with cellbodies immunostained for PSA-NCAM within the SVZand the dentate gyrus, but also along the SVZ of thealveus hippocampus (Fig. 6A,B). Whereas the fre-quency of such an association was very high in the SVZ

Fig. 5. Quantitative evaluation of the effects of adrenalectomy(Adx) and of prolonged administration of corticosterone (Cort), testos-terone (Testo) or progesterone (Prog) on cell proliferation in differentforebrain regions, including germinative zones (SVZ, Alv, DG), whitematter regions (Cca, Ccp, Fb, ChO) and gray matter regions (Cx, Sept,Hyp). As compared with controls (Cont), the numerical density ofbromodeoxyuridine (BrdU)-labeled nuclei is not modified after adre-nalectomy within all the structures analyzed, except for a slightincrease in the anterior portions of the corpus callosum. In corticoste-rone-treated rats, the numerical density of BrdU-labeled nuclei isdramatically decreased in all white and gray matter regions and inthe dentate gyrus, whereas it is not modified in the SVZ of the alveushippocampus. No modification is observed in all the structures ana-lyzed in rats treated with testosterone or progesterone. y-axis valuesrepresent the number of BrdU-labeled nuclei/mm2.Values representmeans 6SEM of data. *P , 0.05, **P , 0.01, Mann-Whitney test:statistically different from control. Alv, subventricular zone of thealveus hippocampus; CCa, anterior part of the corpus callosum; CCp,posterior part of the corpus callosum; ChO, optic chiasma; Cx, cortex;DG, dentate gyrus of the hippocampus; Fb, fimbria; Hyp, mediobasalhypothalamus; Sept, lateral septum; SVZ, subventricular zone of thelateral ventricle.

TABLE 1. Evaluation of the proportion of BrdU-labeled nucleiassociated with specific cell type markers

SVZ AlveusDentate

gyrusWhitematter

Graymatter

PSA-NCAM

111 111 11 0 0

NG2 1 1 11 1111 1111OligoM 0 0 0 0 0IsoB4 0 0 0 0 0GFAP 0 0 0 0 0S100 0 0 0 0 0GR 0 0 0 0 0

The data are based on the observations of colored images of double-labeledsections obtained from three control rats. 0, double-labeled cells not found; 1,about 20%; 11, about 50%; 111, 60–80%; 1111, more than 90%.

225CORTICOSTERONE INHIBITS PROLIFERATION OF OLIGODENDROCYTE PROGENITORS

Fig. 6. Double fluorescence immunostaining of bromodeoxyuridine(BrdU) and of either PSA-NCAM (A,B) or NG2 (C–F). In the subven-tricular zones of the lateral ventricle (A) and of the alveus hippocam-pus (B), most BrdU-labeled nuclei are associated with PSA-NCAM-labeled cells, whereas only a few are associated with NG2-labeled cells(arrows, C,D). By contrast, nearly all BrdU-labeled nuclei detected in

the corpus callosum (E) and the mediobasal hypothalamus (F) areassociated with NG2-labeled cells. Note cells double-labeled for BrdUand NG2 are also detected within the striatum underlying the sub-ventricular zone to the lateral ventricle (arrowheads, C). St, striatum;V, lateral ventricle. Scale bar 5 50 mm.

226 ALONSO

Fig. 7. Double fluorescence labeling of bromodeoxyuridine (BrdU)and of either glial markers (A–D) or glucocorticoid receptor (E,F). Inthe corpus callosum, BrdU-labeled nuclei are never associated withcell bodies labeled for oligoM (oligodendrocyte marker; A), IsoB4 (mi-croglial cell marker; B), or S-100 and glial fibrillary acidic protein(GFAP) (astrocyte markers; C,D). Similarly, BrdU-labeled nuclei lo-

cated in the corpus callosum or the hippocampal dentate gyrus are notassociated with cells immunostained for the glucocorticoid receptor(GR; E,F). The striation observed in A is due to the poor penetrationof the OligoM antibody (IgM) within the section thickness, thus lim-iting the immunostaining to the more superficial (striated) layers ofthe vibratome section. Scale bar 5 50 mm.

227CORTICOSTERONE INHIBITS PROLIFERATION OF OLIGODENDROCYTE PROGENITORS

and along the alveus hippocampus, it was lower in thedentate gyrus where a large number of BrdU-labelednuclei appeared associated with PSA-NCAM negativecells (not shown).

Double Labeling of BrdU and NG2

In agreement with previous descriptions, the markerof oligodendrocyte precursors NG2 was found to beassociated with numerous cell profiles dispersedthroughout the different white and gray matter re-gions, although they always appeared more concen-trated within white matter regions (Nishiyama et al.,1999). In all the white and gray matter regions exam-ined, the large majority of BrdU-labeled nuclei wasassociated with NG2-labeled cell bodies, except for theSVZ and the SVZ of the alveus hippocampus (Fig 6C–F). In these two last regions, the large majority ofBrdU-labeled nuclei were associated with NG2-nega-tive cells, although some BrdU and NG2 double-labeledcells were detected in the surrounding areas (Fig.6C,D).

Double labeling of BrdU and glial markers

The labeling patterns observed with the differentglial cell markers fully conformed to previous descrip-tions: cell profiles labeled with the microglial cellmarker isoB4 (Streit and Kreutzberg, 1987) or for theastrocyte markers GFAP (Bignami et al., 1972) andS-100 (Kuwano et al., 1987) were detected throughoutboth white and gray matter regions, whereas cell pro-files and processes labeled with the oligodendrocytemarker myelin/oligodendrocyte-specific protein (Dyeret al., 1991) were essentially located within the whitematter regions. Throughout all the white and graymatter regions of the forebrain examined, BrdU-la-beled nuclei were never found to be associated with anycell profiles labeled with any of these glial cell markers(Fig. 7A–D).

Double labeling of BrdU and glucocorticoidreceptors

The immunostaining pattern observed throughoutthe forebrain with the antibody anti-GR fully con-formed to previous descriptions (Cintra et al., 1994):GR immunostaining was essentially associated withthe nucleus of a large number of neuron-like profiles ofthe different gray matter regions, and with some glial-like profiles dispersed throughout the white matterregions. Examination of double immunostained sec-tions clearly indicated that, throughout the differentforebrain regions considered, BrdU-labeled nuclei werealways associated with cellular profiles that were GRnegative (Fig. 7E,F).

DISCUSSION

The thymidine analogue BrdU is incorporated withinDNA during the S phase of proliferating cells. SinceBrdU has a short half-life, its detection within thenucleus of proliferating cells is highly dependent on itsavailability within the tissue. In the present study,prolonged administration of BrdU (i.e., two times a dayover a 3-day period) was used in order to be able todetect proliferating cells with low proliferative rate.Under these conditions, regions containing high, mod-erate, or low numerical densities of BrdU-labeled nu-clei could be distinguished throughout the differentforebrain regions examined, which were assumed tocorrespond to regions containing dividing cells withhigh, moderate, or low proliferative rates, respectively.In addition to the high numerical densities detected inthe so-called “germinative” zones of the SVZ of thelateral ventricle and the hippocampal dentate gyrus,moderate to low numerical densities of labeled nucleicould be detected throughout all white and gray matterregions of the adult forebrain. An intriguing observa-tion of the present study is that high concentrations ofBrdU-labeled nuclei were detected along the border ofthe alveus hippocampus, which, to our knowledge, hasnever been mentioned in previous studies. Examina-tion of serial sections cut through the rostrocaudalextension of the forebrain however strongly suggeststhat this germinative zone corresponds to the caudalextent of the SVZ present in the rostral forebrain. Thisis further supported by the present observations that,similar to the SVZ, the SVZ of the alveus hippocampusalways contained high concentrations of cells immuno-stained for PSA-NCAM, that frequently exhibited aBrdU-labeled nucleus.

In a series of previous studies, glucocorticoids havebeen shown to inhibit cell proliferation within the den-tate gyrus of the hippocampus (Gould et al., 1992;Cameron and Gould, 1996; Gould and Cameron, 1996).A prominent finding of the present study is that corti-costerone also dramatically decreases cell proliferationthroughout most white and gray matter regions of theadult rat forebrain. Our data also show that the prolif-eration of these cells is not modified by the prolongedadministration of other steroids such as testosterone orprogesterone, indicating that, these effects are specificto glucocorticoids. Surprisingly, no clear modification ofcell proliferation could be detected after adrenalec-tomy, which induces total withdrawal of circulatingcorticosterone. This may be because the levels of circu-lating corticosterone were particularly low in the intactcontrol rats used in the present study, most likelyattributable to stressless breeding conditions.

The occurrence of proliferating cells throughoutwhite and gray matter forebrain regions of adult ro-dents has already been reported in a series of previousstudies based on the prolonged infusion of tritiatedthymidine (Hommes and Leblond, 1967; Korr et al.,1973; Mares and Loddin, 1974; McCarthy and Leblond,1988). In these studies, such dividing cells present

228 ALONSO

within white and gray matters of the adult rodentbrain were generally thought to be immature glia, al-though their nature was not fully characterized. In thepresent study, double-labeling experiments for BrdUand for specific cell type markers clearly show thatnone of the numerous BrdU-labeled cells detectedthroughout the white or gray matter regions was im-munostained for classical markers of mature oligoden-drocytes, astrocytes, or microglial cells. This indicatesthat these dividing cells are not glial cells or do notexpress any of the proteins that characterize matureglial cells. The occurrence of a population of oligoden-drocyte progenitors within the white and gray matterregions of adult rodent brain is now well documented(Wolswijk and Noble, 1989; Levison and Goldman,1993; Engel and Wolswijk, 1996; Reynolds and Hardy,1997; Shi et al., 1998; Levison et al., 1999). Theseoligodendrocyte progenitor cells present in the adultbrain have been shown to express sulfated antigensrecognized by the antibody directed against the NG2-chondroitin sulfate proteoglycan (Stallcup and Beasley,1987; Nishiyama et al., 1999), although they do notexpress the classic markers of mature oligodendro-cytes. The present finding that most of the BrdU-la-beled nuclei present within the gray or white matterforebrain regions are associated with cell bodies immu-nostained for NG2 provides a strong indication thatthese proliferating cells mostly correspond to oligoden-drocyte progenitors. In addition to cells immunostainedfor NG2, BrdU-labeled nuclei were found to be associ-ated with cells immunostained for PSA-NCAM withinthe SVZ and dentate gyrus. Within these two germina-tive regions, PSA-NCAM has been shown to be essen-tially expressed by neuronal progenitors (Seki andArai, 1993; Bonfanti and Theodosis, 1994; Doetsch etal., 1997; Alonso et al., 1998; Garcia-Verdugo et al.,1998). This indicates that in the adult CNS, proliferat-ing cells essentially include (1) oligodendrocyte progen-itors dispersed throughout the white and gray matterregions, and (2) neuronal progenitors essentially lo-cated within the so-called “germinative” zones, includ-ing the hippocampal dentate gyrus and the SVZ of thelateral ventricle.

It is generally admitted that in the CNS, the effectsof corticosteroids are mainly mediated by the cytosolicreceptors, which act as ligand-activated transcriptionfactors. Two types of receptor have been described inthe CNS: mineralocorticoid receptors (MR) with highaffinity for corticosterone, and glucocorticoid receptors(GR) with lower affinity for corticosterone. Both typesof receptor are expressed by many neurons, as well asa large variety of glial cells, including astrocytes, oli-godendrocytes, and microglia (Ovadia et al., 1984;Ahima et al., 1991; Jung-Testas et al., 1992; Cintra etal., 1994; Tanaka et al., 1997). Glucocorticoid-inducedinhibition of the proliferation of cultured astrocytesand microglia, demonstrated in a number of previousstudies (Kniss and Burry, 1985; Ganter et al., 1992;Castano et al., 1996; Crossin et al., 1997), can thus beassumed to involve the glucocorticoid receptors ex-

pressed by these cells. Because of the high affinity ofMR receptors for corticosterone, it is admitted thatthey are saturated by normal circulating levels of theglucocorticoid and that GR receptors are mainly in-volved in the responses to pharmacological doses ofcorticosterone. However, the present data of doublelabeling experiments clearly indicate that, althoughimmunostaining for GR was associated with numerousneurons and glial cells throughout the different whiteand gray matter forebrain regions considered here, itwas never associated with any of the proliferating(BrdU-labeled) cells detected within these regions.This suggests that the effects of corticosterone on cellproliferation do not involve direct effects via glucocor-ticoid receptors. In the dentate gyrus of the hippocam-pus, it has been established that the effects of glucocor-ticoids on cell proliferation are mediated by N-methyl-D-aspartate (NMDA) receptors (Gould et al., 1994;Cameron et al., 1995). Interestingly, previous studieshave reported that oligodendrocyte progenitors do ex-press functional NMDA receptors (Wang et al., 1996)and that, in vitro, their proliferation is inhibited byglutamate (Yuan et al., 1998). Since it is generallyadmitted that acute or chronic administration of glu-cocorticoid treatment induces an increased release ofglutamate throughout the CNS (Reagan and McEwen,1997; Vereno and Borrell, 1999), one may assume thatthe effects of corticosterone on the proliferation of oli-godendrocyte progenitors described in the presentstudy at least partly result from the stimulation oftheir NMDA receptors.

Interestingly, our data show that corticosteronetreatment decreases cell proliferation in all the fore-brain regions examined, except for the SVZ of the lat-eral ventricle and of the alveus hippocampus. This firstprovides additional support to the idea that these tworegions containing high concentrations of proliferatingcells correspond to the rostrocaudal extension of thesame subventricular germinative zone. It is admittedthat proliferating cells of the SVZ correspond primarilyto neuronal progenitors (Luskin, 1993; Luskin et al.,1997). Our observations clearly show that, in contrastwith all the other forebrain regions examined, the largemajority of the BrdU-labeled nuclei detected in theSVZ and the SVZ of the alveus hippocampus wereimmunostained for the neuroblast marker PSA-NCAM, whereas only a few were immunostained forthe oligodendrocyte precursor NG2. The differentialeffect of corticosterone on cell proliferation within thegerminative SVZ versus the other forebrain regionscould thus result from a differential expression of theNMDA receptors by neuronal versus oligodendrocyteprogenitors. This possibility does not appear to be inagreement with the data of a series of previous studiesindicating that the proliferation of neuronal progeni-tors within the hippocampal dentate gyrus was inhib-ited by glucocorticoids (Gould et al., 1992; Cameronand Gould, 1996; Gould and Cameron, 1996; Nitta etal., 1999). A possible explanation is that corticosteroneinfluences the proliferation of neuronal progenitors in-

229CORTICOSTERONE INHIBITS PROLIFERATION OF OLIGODENDROCYTE PROGENITORS

directly through a population of neighboring cells withNMDA receptors, the differential effects on cell prolif-eration within the SVZ, and the dentate gyrus, result-ing from the differential innervation of these two re-gions by excitatory axonal fibers.

Little is known regarding the precise role played bythe numerous oligodendrocyte precursors presentthroughout white and gray matter regions of the adultCNS. In a series of recent studies, the evidence showsthat in the case of CNS lesions, remyelination compe-tent cells are generated by proliferation and differen-tiation of oligodendrocytes precursor cells (Keirstead etal., 1998, 1999; Levine and Reynolds, 1999; Zhang etal., 1999). During the last few years, glucocorticoidshave been employed extensively in the clinical treat-ment of acute CNS trauma (Braughler and Hall, 1985;Kiwerski, 1993; Ogata et al., 1993; Ducker andZeidman, 1994; Knollema et al., 1997; Segatore, 1999).The rationale for this clinical use of corticoids is thatthey reduce the posttraumatic spinal cord and cerebraledema. Moreover, glucocorticoids have been reported toinduce the maturation of oligodendrocytes by facilitat-ing the synthesis of myelin basic protein (MBP) and ofglutamine synthetase (Jung-Testas et al., 1992; Barreset al., 1994; Melcangi et al., 1997; Baas et al., 1998;Chan et al., 1998). However, the present findings thatprolonged corticosterone treatment inhibits the prolif-eration of oligodendrocytes precursors may have impli-cations for the treatment of CNS trauma. Indeed, sincedemyelinating lesions always induce a rapid andmarked increase in the proliferation of oligodendrocyteprogenitors (Keirstead et al., 1998; Nait-Oumesmar etal., 1999), it can be assumed that postlesional thera-peutic treatments based on massive and prolonged ad-ministration of glucocorticoids may have negative ef-fects on the remyelination process Ongoing studiesusing various demyelinating experimental procedureswill try to determine whether postlesional glucocorti-coid treatment could have any consequence on the pro-cesses of remyelinization.

ACKNOWLEDGMENTS

The author thanks Dr. T. Seki for his generous gift ofthe antibody anti-PSA-NCAM.

REFERENCES

Ahima R, Krozowski Z, Harlan R. 1991. Type I corticosteroid receptor-like immunoreactivity in the rat CNS: distribution and regulationby corticosteroids. J Comp Neurol 313:522–538.

Allen E. 1912. Cessation of mitosis in the central nervous system ofthe albino rat. J Comp Neurol 22:547–568.

Alonso G, Prieto M, Chauvet N. 1999. Tangential migration of youngneurons arising from the subventricular zone of adult rats is im-paired by surgical lesions passing through their natural migratorypathway. J Comp Neurol 405:508–528.

Altman J, Das GD. 1965. Autoradiographic and histological evidenceof postnatal hippocampal neurogenesis in rats. J Comp Neurol124:319–335.

Baas D, Fressinaud C, Vitkovic L, Sarlieve LL. 1998. Glutaminesynthetase expression and activity are regulated by 3,5,39-triodo-L-thyronine and hydrocortisone in rat oligodendrocyte cultures. Int JDev Neurosci 16:333–340.

Barres BA, Lazar MA, Raff MC. 1994. A novel role for thyroid hor-mone, glucocorticoids and retinoic acid in timing oligodendrocytedevelopment. Development 120:1097–1108.

Bignami A, Eng LF, Dahl D, Uyeda CT. 1972. Localization of the glialfibrillary acidic protein in astrocytes by immunofluorescence. BrainRes 43:429–435.

Bonfanti L, Theodosis DT. 1994. Expression of polysialylated neuralcell adhesion molecule by proliferating cells in the subependymallayer of the adult rat, in its rostral extension and in the olfactorybulb. Neuroscience 62:291–305.

Braughler JM, Hall ED. 1985. Current application of “high dose”steroid therapy for CNS injury. A pharmacological perspective.J Neurosurg 62:806–810.

Bryans WA. 1959. Mitotic activity in the brain of the adult rat. AnatRec 133:65–71.

Cameron HA, Gould E. 1996. Distinct populations of cells in the adultdentate gyrus undergo mitosis or apoptosis in response to adrenal-ectomy. J Comp Neurol 369:56–63.

Cameron HA, McKay RDG. 1999. Restoring production of hippocam-pal neurons in old age. Nature Neurosci 2:894–897.

Cameron HA, McEwen BS, Gould E. 1995. Regulation of adult neu-rogenesis by excitatory input and NMDA receptor activation in thedentate gyrus. J Neurosci 15:4687–4692.

Castano A, Lawson LJ, Fearn S, Perry VH. 1996. Activation andproliferation of murine microglia are insensitive to glucocorticoidsin Wallerian degeneration. Eur J Neurosci 8:581–588.

Chan JR Phillips LJ, Glaser M. 1998. Glucocorticoids and progestinssignal the initiation and enhance the rate of myelin formation. ProcNatl Acad Sci USA 95:10459–10464.

Cintra A, Zoli M, Rosen L, Agnati LF, Okret S, Wikstrom AC, Gustafs-son JA, Fuxe K. 1994. Mapping and computer assisted morphome-try and microdensitometry of glucocorticoid receptor immunoreac-tive neurons and glial cells in the rat central nervous system.Neuroscience 62:843–897.

Cowan WM. 1979. The development of the brain. Sci Am 241:112–133.Crossin KL, Tai MH, Krushel LA, Mauro VP, Edelman GM. 1997.

Glucocorticoid receptor pathways are involved in the inhibition ofastrocyte proliferation. Proc Natl Acad Sci USA 94:2687–2692.

Doetsch F, Garcia-Verdugo JM, Alvarez-Buylla A. 1997. Cellular com-position and three-dimensional organization of the subventriculargerminal zone in the adult mammalian brain. J Neurosci 17:5046–5061.

Ducker TB, Zeidman SM. 1994. Spinal cord injury. Role of steroidtherapy. Spine 19:2281–2287.

Dyer CA, Hickey WF, Geisert EE. 1991. Myelin/oligodendrocyte-spe-cific protein: a novel surface membrane protein that associates withmicrotubules. J Neurosci Res 28:607–613.

Engel U, Wolswijk G. 1996. Oligodendrocyte-type-2-astrocyte (O-2A)progenitor cells derived from adult rat spinal cord: in vitro charac-teristics and response to PDGF, bFGF and NT3. Glia 16:16–26.

Ganter S, Northoff D, Mannel D, Gebicke-Harter PJ. 1992. Growthcontrol of cultured microglia. J Neurosci Res 33:218–230.

Garcia-Verdugo JM, Doetsch F, Wichterle H, Lim DA, Alvarez-BuyllaA. 1998. Architecture and cell types of the adult subventricularzone: in search for stem cells. J Neurobiol 36:234–248.

Goldman JE. 1995. Lineage, migration and fate determination ofpost-natal subventricular zone cells in the mammalian CNS. J Neu-rooncol 24:61–64.

Gould E, Cameron HA. 1996. Regulation of neuronal birth, migrationand death in the rat dentate gyrus. Dev Neurosci 18:22–35.

Gould E, Cameron HA, Daniels DC, Woolley CS, McEwen BS. 1992.Adrenal hormones suppress cell division in the adult rat dentategyrus. J Neurosci 12:3642–3650.

Gould E, Cameron HA, McEwen BS. 1994. Blockade of NMDA recep-tors increases cell death and birth in the developing rat dentategyrus. J Comp Neurol 340:551–565.

Hauke C, Ackermann I, Korr H. 1995. Cell proliferation in the sub-ependymal layer of the adult mouse in vivo and in vitro. Cell Prolif28:595–607.

Hommes OR, Leblond CP. 1967. Mitotic division of neuroglia in thenormal adult rat. J Comp Neurol 129:269–278.

Jung-Testas I, Renoir M, Bugnard H, Greene GL, Beaulieu EE. 1992.Demonstration of steroid hormone receptors and steroid action inprimary cultures of rat glial cells. J Steroid Biochem Mol Biol41:621–631.

Keirstead HS, Levine JM, Blakemore WF. 1998. Response of theoligodendrocyte progenitor cell population (defined by NG2 label-ling) to demyelination of the adult spinal cord. Glia 22:161–170.

230 ALONSO

Keirstead HS, Ben-Hur T, Rogister B, O’Leary MT, Dubois-Dalcq M,Blakemore WF. 1999. Polysialylated neural cell adhesion molecule-positive CNS precursors generate both oligodendrocytes andSchwann cells to remyelinate the CNS after transplantation. J Neu-rosci 19:7529–7536.

Kiwerski JE. 1993. Application of dexamethasone in the treatment ofacute spinal cord injury. Injury 24:457–460.

Kniss DA, Burry RW. 1985. Glucocorticoid hormones inhibit DNAsynthesis in glial cells cultured in chemically defined medium. ExpCell Res 161:29–49.

Knollema S, Kemper RHA, Korf J, Wiersma A, Ter Horst GJ, FrugersHJ. 1997. The number of insults and the cerebral damage afterhypoxia/ischemia are altered after acute pretreatment with corti-costerone and metyparone. Neurosci Res Commun 21:203–211.

Korr H, Schultze B, Maurer W. 1973. Autoradiographic investigationsof glial proliferation in the brain of adult mice. I. The DNA synthe-sis phase of neuroglia and endothelial cells. J Comp Neurol 150:169–176.

Kuhn HG, Dickinson-Anson H, Gage FH. 1996. Neurogenesis in thedentate gyrus of the adult rat: age-related decrease of neuronalprogenitor proliferation. J Neurosci 16:2027–2033.

Kuwano R, Usui H, Maeda T, Araki K, Yamakuni T, Kurihara T,Takahashi Y. 1987. Tissue distribution of rat S100a and b subunitmRNAs. Mol Brain Res. 10:291–297.

Landfield PW, Eldridge JC. 1994. The glucocorticoid hypothesis ofage-related hippocampal neurodegeneration: role of disregulatedintraneuronal calcium. Ann NY Acad Sci 746:308–326.

Levine JM, Reynolds R. 1999. Activation and proliferation of endog-enous oligodendrocyte precursor cells during ethidium bromide-induced demyelination. Exp Neurol 160:333–347.

Levison SW, Goldman JE. 1993. Both oligodendrocytes and astrocytesdevelop from progenitors in the subventricular zone of postnatal ratforebrain. Neuron 10:201–212.

Levison SW, Young GM, Goldman JE. 1999. Cycling cells in the adultrat neocortex preferentially generate oligodendroglia. J NeurosciRes 57:435–446.

Lois C, Alvarez-Buylla A. 1993. Proliferating subventricular zone cellsin the adult mammalian forebrain can differentiate into neuronsand glia. Proc Natl Acad Sci USA 90:2074–2077.

Lupien SJ, de Leon M, de Santi S, Convit A, Tarshish C, Nair NP,Thakur M, McEwen BS, Hauger RL, Meaney MJ. 1998. Cortisollevels during human aging predict hippocampal atrophy and mem-ory deficits. Nature Neurosci 1:69–73.

Luskin MB. 1993. Restricted proliferation and migration of postna-tally generated neurons derived from the forebrain subventricularzone. Neuron 11:173–189.

Luskin MB, Zigova T, Soteres BJ, Stewart RR. 1997. Neuronal pro-genitor cells derived from the anterior subventricular zone of theneonatal forebrain continue to proliferate in vitro and express aneuronal phenotype. Mol Cell Neurosci 8:351–366.

Mares VL, Lodin Z. 1974. An autoradiographic study of DNA synthe-sis in adolescent and adult mouse forebrain. Brain Res 76:557–561.

McCarty GF, Leblond CP. 1988. Radioautographic evidence for slowastrocyte turnover and modest oligodendrocyte production in thecorpus callosum of adult mice infused with 3H-thymidine. J CompNeurol 271:589–603.

Melcangi RC, Magnaghi V, Cavarretta I, Riva MA, Martini L. 1997.Corticosteroid effects on gene expression of myelin basic protein inoligodendrocytes and of glial fibrillary acidic protein in type 1 as-trocytes. J Neuroendocr 9:729–733.

Morshead CM, Reynolds BA, Craig CG, McBurney MW, Staines WA,Morassutti D, Weiss S, van der Kooy D. 1994. Neural stem cells inthe adult mammalian forebrain: a relatively quiescent subpopula-tion of subependymal cells. Neuron 13:1071–1082.

Nait-Oumesmar B, Decker L, Lachapelle F, Avellana-Adalid V, BachelinC, Baron-Van Evercooren A. 1999. Progenitor cells in the adult mousesubventricular zone proliferate, migrate and differentiate into oligo-dendrocytes after demyelination. Eur J Neurosci 11:4357–4366.

Nishiyama A, Chang NA, Trapp BD. 1999. NG21 glial cells: a novelglial cell population in the adult brain. J Neuropathol Exp Neurol58:1113–1124.

Nitta A, Ohmiya M, Sometani A, Megumi I, Nomoto H, Furukawa Y,Furukawa S. 1999. Brain-derived neurotrophic factor prevents neu-ronal cell death induced by corticosterone. J Neurosci Res 57:227–235.

Ogata T, Nakamura Y, Tsuji K, Shibata, T, Kataoka K. 1993. Steroidhormones protect spinal cord neurons from glutamate toxicity. Neu-roscience 55:445–449.

Ovadia H, Vlodavsky, I, Abramsky O, Weidenfeld J. 1984. Binding ofhormonal steroids to isolated oligodendria and astroglia grown invitro on a naturally produced extracellular matrix. Clin Neurophar-macol 7:307–311.

Porter NM, Landfield PW. 1998. Stress hormones and brain aging:adding injury to insult? Nature Neurosci 1:3–4.

Reagan LP, McEwen BS. 1997. Controversies surrounding glucocor-ticoid-mediated cell death in the hippocampus. J Chem Neuroanat13:149–167.

Reynolds R, Hardy R. 1997. Oligodendroglial progenitors labeled withthe O4 antibody persist in the adult rat cerebral cortex in vivo.J Neurosci Res 47:455–470.

Sapolsky RM, Krey LC, McEwen BS. 1985. Prolonged glucocorticoidexposure reduces hippocampal neuron number: implication for ag-ing. J Neurosci 5:1222–1227.

Segatore M. 1999. Corticosteroids and traumatic brain injury: statusat the end of the decade of the brain. J Neurosci Nursing 31:239–250.

Seki T, Arai Y. 1991. The persistent expression of a highly polysialy-lated NCAM in the dentate gyrus of the adult rat. Neurosci Res12:503–504.

Seki T, Arai Y. 1993. Highly polysialylated neural cell adhesion mol-ecule (NCAM-H) is expressed by newly generated granule cells inthe dentate gyrus of the adult rat. J Neurosci 13:2351–2358.

Shi J, Marinovich A, Barres BA. 1998. Purification and characteriza-tion of adult oligodendrocyte precursor cells from the rat opticnerve. J Neurosci 18:4627–4636.

Stallcup WB, Beasley L. 1987. Pipotential glial precursor cells of theoptic nerve express the NG2 proteoglycan. J Neurosci 7:2737–2744.

Streit WJ, Kreutzberg GW. 1987. Lectin binding by resting and reac-tive microglia. J Neurocytol 16:249–260.

Tanaka J, Fujita H, Matsuda S, Toku K, Sanaka M, Maeda N. 1997.Glucocorticoid- and mineralocorticoid receptors in microglial cells:the two receptors mediate differential effects of corticosteroids. Glia20:23–37.

Vereno C, Borrell J. 1999. Rapid glucocorticoid effects on excitatoryamino acid levels in the hippocampus: a microdialysis study infreely moving rats. Eur J Neurosci 11:2465–2473.

Wang C, Rougon G, Kiss JZ. 1994. Requirement of polysialic acid forthe migration of the O-2A glial progenitor cell from neurohypoph-yseal explants. J Neurosci 14:4446–4457.

Wolswijk G, Noble M. 1989. Identification of an adult-specific glialprogenitor cell. Development 105–387–400.

Yuan X, Eisen AM, McBain CJ, Gallo V. 1998. A role for glutamateand its receptors in the regulation of oligodendrocyte developmentin cerebellar tissue slices. Development 125:2901–2914.

Zhang SC, Ge B, Duncan ID. 1999. Adult brain retains the potentialto regenerate oligodendroglial progenitors with extensive myelina-tion capacity. Proc Natl Acad Sci USA 96:4089–4094.

231CORTICOSTERONE INHIBITS PROLIFERATION OF OLIGODENDROCYTE PROGENITORS