Dexamethasone Downregulates Chemokine Receptor CXCR4 and Exerts Neuroprotection against Hypoxia/Ischemia-Induced Brain Injury in Neonatal Rats

Original Paper

Neuroimmunomodulation 2004;11:404–413DOI: 10.1159/000080151

Dexamethasone Downregulates ChemokineReceptor CXCR4 and Exerts Neuroprotectionagainst Hypoxia/Ischemia-Induced Brain Injuryin Neonatal Rats

Klara Felszeghya,b Ghazal Banisadra William Rostènea Csaba Nyakasb

France Haoura

aINSERM E0350, Hôpital St-Antoine, Paris, France; bHungarian Academy of Sciences and Semmelweis University,Brain Physiology Research Group, Budapest, Hungary

Received: July 23, 2003Accepted: December 15, 2003

Dr F. HaourINSERM 0350184, rue du Faubourg St-AntoineFR–75571 Paris, Cedex 12 (France)Tel. +33 149 26 46 76, Fax +33 143 40 82 70, E-Mail [email protected]

ABCFax + 41 61 306 12 34E-Mail [email protected]

© 2004 S. Karger AG, Basel1021–7401/04/0116–0404$21.00/0

Accessible online at:www.karger.com/nim

Key WordsAstrocytosis W Chemokines W CXCR4 W Dexamethasone W

Neonatal hypoxic/ischemic brain damage

AbstractObjective: Hypoxia/ischemia (H/I) induces rapid andmassive brain damage in neonatal rat brain, resulting inlong-term consequences on structural and functionalmaturation of the central nervous system. Inflammatorymediators contribute to these permanent pathologicalchanges, which are sensitive to corticoid treatments.Since the chemokine receptor CXCR4, specific for theSDF-1·/CXCL12 ligand, regulates both apoptotic andneuroregeneration processes, this receptor was quanti-fied 2 days following H/I in neonatal rat brain in relationwith dexamethasone (DEX) treatment. Methods: Seven-day-old male rats were exposed to a 90-min hypoxia fol-lowing unilateral carotid ligation (H/I) and were sacrificed48 h later. Glucocorticoid-pretreated animals were in-jected subcutaneously 5 h prior to hypoxia with 0.5 Ìg/gDEX. Glial fibrillary acidic protein and cresyl violet stain-ing were used for assessing the extent of brain lesionsubdivided into necrotic and penumbra-like areas. The

density of CXCR4 receptors was determined by quantita-tive autoradiography using [125I]SDF-1· as a ligand. Re-

sults: The H/I resulted in a massive lesion ipsilateral tothe carotid ligation, which was extended to cortical, stria-tal, hippocampal and thalamic areas, while the contralat-eral hemisphere remained apparently unaffected. DEXdecreased the lesion size by reducing mainly the necroticarea. H/I induced a marked increase in CXCR4 receptorbinding in the penumbra-like areas. DEX pretreatmentdecreased CXCR4 receptor density in the penumbra andattenuated astrocytosis. Furthermore, DEX strongly low-ered mortality rate and reduced functional recovery timeright after hypoxia. Conclusion: The rapid enhancementin CXCR4 chemokine receptor binding in the affectedbrain areas suggests that SDF-1·/CXCR4 may play a rolein the hypoxia-induced inflammatory reaction in the neo-natal brain. Attenuation of CXCR4 expression and astro-gliosis could contribute to the neuroprotective effect ofDEX pretreatment via influencing the inflammatory cas-cade induced by H/I in the neonatal brain.

Copyright © 2004 S. Karger AG, Basel

Brain CXCR4 Receptors followingNeonatal Hypoxia and DEX Treatment

Neuroimmunomodulation 2004;11:404–413 405

Introduction

Perinatal hypoxia and reduced cerebral blood flow(ischemia) in the human fetus and newborn are the majorcauses of brain damage resulting in cerebral palsy, mentalretardation and epilepsy [1]. The developing brain is high-ly vulnerable to such a noxious action that evolves both animmediate neuronal damage and a long-term impact onstructural and functional maturation [2]. Underlyingmechanisms of hypoxic/ischemic (H/I) brain injury iscomplex, including energy failure, free-radical-induceddamage, and cytokine-, excitotoxicity- and caspase-3-mediated cell death processes [3–6]. These mechanismscan be investigated in a neonatal animal H/I model ofLevine preparation introduced by Rice et al. [7] on 7-day-old rats. At that age, the pattern of brain injury in rats iscomparable to that of H/I injury in full-term newbornhuman infants [8].

It has been well demonstrated that the central nervoussystem (CNS) responds to different injuries with inflam-matory reactions. Following H/I insult, microglia, macro-phages and neutrophils are activated, which are accompa-nied by a marked expression of pro-inflammatory cyto-kines (IL1-ß, IL-6, TNF-·) [9]. These cytokines subse-quently induce the release of · and ß chemokines in theendothelium as well as inside the neural tissue [10, 11].Recent data implicate that these chemokines are involvedin many pathological processes during inflammation in-duced by brain injury, such as breakdown of the blood-brain barrier, leukocyte recruitment and cytotoxic effects[12, 13].

One of the candidates among the central inflammatorychemokines is the stromal cell-derived factor-1· (SDF-1·/CXCL12). SDF is a CXC4 chemokine produced by a vari-ety of cells including brain astrocytes and neurons [13].SDF-1· could affect neural function through CXCR4receptors. CXCR4 is a Gi-protein-coupled transmem-brane receptor that could alter the normal cell functionvia changes in intracellular Ca2+ mobilization [14] andmodulation of tyrosine phosphorylation [15]. This recep-tor has also been detected in the CNS on cortical neuronsand astrocytes [15, 16]. SDF-1·, by stimulating astrocyteproliferation, could also play an important role in neuro-nal development and/or regeneration as well as in reactiveastrogliosis induced by brain injuries [15]. There areobservations suggesting that CXCR4 signal transductionmediates apoptosis in different brain structures [13, 17].Results of morphological, histochemical and molecularstudies indicate that apoptotic and necrotic mechanismsaccount for neuronal death after cerebral H/I [18]. Con-

sidering the multiple role of SDF-1·/CXCR4 interactionin the brain and the possibility that CXCR4 signal trans-duction could contribute to neuronal death or regenera-tion, it was worth to examine the alteration in brainCXCR4 binding following H/I injury.

Glucocorticoid hormones (GC) are available for clini-cal use and commonly administered to human neonateswith impaired lung maturation [19]. In addition, GCshave been extensively investigated with respect to theirability to reduce brain damage caused by H/I [20, 21]. It isknown that different brain injuries and inflammationinduce adrenal activation and result in GC release intothe circulation [22]. The role of GC during hypoxia and inthe pathological mechanisms following the insult is multi-ple and has not been entirely elucidated. According to theextensive research study by Tuor et al. [23], a syntheticGC, dexamethasone (DEX), applied before H/I in neona-tal rats, resulted in a dramatic reduction in the size ofbrain lesions and mortality. These authors suggested thepreventive role of reduced glucose and phosphate utiliza-tion as one of the cardinal neuroprotective underlyingmechanisms. However, other mediators involved in theprotective action of DEX [24–26] are not yet fully under-stood.

It has been well documented that GCs are also strongimmunosuppressors. They inhibit the expression of pro-inflammatory cytokines [25], diminish astrocytic prolifer-ation [26], but enhance certain immune responses duringthe later phase of inflammation [22]. Nothing is knownabout the possibility that some of the neuroprotectiveeffects of DEX may be mediated by chemokines.

The aim of the present work was to study the chemo-kine system SDF-1·/CXCR4 in the neonatal brain. Theearly role of this system in the neuroprotective effect ofGC pretreatment was investigated in a model of H/I brainlesion.

Methods

Experimental Animals and TreatmentsNewborn male Wistar rats were randomly distributed among fos-

ter mothers within 24 h after birth in a way that all nests contained 8pups. The pups were assigned into four experimental groups by nip-ping the nails of forepaws: (1) sham-operated animals (controlgroup); (2) DEX-pretreated and sham-operated animals (DEXgroup); (3) carotid-ligated and hypoxic animals (hypoxic group), and(4) DEX-pretreated and carotid-ligated hypoxic animals (HpxDEXgroup). Pups within one nest belonged to the same treatment group.Animals were housed under a 12-hour light/dark cycle with freeaccess to food and water. All experimental manipulations concurredwith guidelines established by an institutional animal care and use

406 Neuroimmunomodulation 2004;11:404–413 Felszeghy/Banisadr/Rostène/Nyakas/Haour

committee, and were in accordance with the National Act on the Useof Experimental Animals in Hungary.

At the age of 7 days, the left common carotid artery was isolatedfrom the vagus nerve and ligated under ether anesthesia during asurgical procedure lasting not longer than 5 min. After surgery thepups were allowed to recover for 2 h at their dams. Afterwards, theligated animals were subjected to a hypoxic procedure in a tempera-ture-controlled plastic chamber of 30 cm diameter partially sub-merged into a water bath. Four pups were placed simultaneously inthe chamber and exposed to a mixture of 8% oxygen and 92% nitro-gen atmosphere with a flow rate of 90 l/h for 90 min at 32°C chambertemperature (hypoxia). Control pups were sham ligated and duringthe period of the hypoxic procedure they were isolated from theirmother and kept in an open plastic chamber at 32°C. The DEX andthe HpxDEX groups received 0.5 Ìg/g body weight s.c. DEX (Ora-dexon, Organon) dissolved in physiological saline solution 5 h priorto hypoxia, when maximal protection against H/I injury could beanticipated as shown by previous findings [27, 28]. Controls for DEXinjection received saline solution alone.

The functional recovery of animals after the hypoxic insult wasinvestigated by measuring the reappearance of the righting reflex inseconds [29]. The observation started immediately after removal ofthe pups from the hypoxic chamber. The carotid-ligated and hypox-ia-treated animals were unconscious and needed relatively long(about 0.5 min) to regain consciousness. Measuring the latency ofreappearance of the righting reflex has been used previously to testthe severity of the hypoxic insult [30]. Following recovery, the pupswere returned to their dams. The mortality index was also deter-mined after the hypoxia. Mortality occurred exclusively during the90-min hypoxia. No animal died during the post-hypoxic recoveryperiod when the animals returned to normal atmosphere. For histo-logical examinations, animals were sacrificed 48 h after the hypoxicinsult. Based on our preliminary studies, 48 h proved to be sufficientfor the development of CXCR4 receptor changes and for that of reac-tive astrogliosis. Following decapitation, the brains were quicklyremoved and frozen in isopentane at –42°C. Brains were stored at–80°C until processing.

Histochemical and Immunocytochemical ExaminationsFrozen native brains were used to obtain 20-Ìm-thick coronal

sections on gelatin-coated microscopic slides. The sections were col-lected at several specified levels of the brain according to respectiveareas in the adult brain (described in the atlas of Paxinos and Watson[31]). The frontoparietal cortex (FPC) and caudate putamen (CPu)were examined at the level of interaural 9.7–10 mm (level I); FPCagain, anterodorsal hippocampus, thalamus, nucleus ventralis poste-rolateralis (VPL) and hypothalamus, nucleus ventromedialis hypo-thalami (VMH) at the level of interaural 7.2–6.7 mm (level II); poste-rior hippocampus at the level of interaural 3.8–3.2 mm (level III),and entorhinal cortex and pons at the level of interaural 2.2–1.2 mm(level IV). The levels selected for examination involved all thosebrain areas which are known as susceptible to H/I [7, 18].

The histological evaluation of the magnitude and type of brainlesions was performed by means of cresyl violet histochemistry andglial fibrillary acidic protein (GFAP) immunocytochemical staining.The GFAP staining was carried out as previously described [32].Briefly, thaw-mounted sections on slides were postfixed with 3%phosphate-buffered paraformaldehyde solution, followed by incuba-tion in 0.3% H2O2 and by pre-incubation with 5% normal sheepserum. For primary antibody, monoclonal mouse anti-GFAP (Santa

Cruse, Calif., USA) was used in a 1:100 dilution rate and the incuba-tion lasted for 2 h at 37 °C in a humid chamber. As a secondary anti-body, anti-mouse biotinylated sheep serum was applied in 1:200dilution. After conjugation with streptavidin-horseradish peroxi-dase, the colored precipitate appeared in the presence of 3-3)-diami-nobenzidine. Adjacent sections were processed with cresyl violetNissl staining for visualization of the necrotic core area containingdistracted cells (severe lesion). Tissue areas which showed enhancedGFAP-reactive astrocytes and spared cresyl-violet-stained neuronswere considered as penumbra-like region (intermediate lesion).Therefore, by the combination of the two staining procedures in theadjacent sections, the total injured area could be subdivided intopenumbra-like and necrotic zones.

Autoradiography of SDF-1· Binding to CXCR4The slide-mounted sections were dried with cold air and pro-

cessed for the [125I]-SDF-1· binding assay principally according to amethod described earlier [33]. Briefly, the sections were incubated in300 Ìl MEM medium (Life Technologies) containing 33 pmol [125I]-SDF1· (NEN), 1% bovine serum albumin, 250 KIU/ml Trasylol and0.25 M Tris-HCl, pH 7.4 (binding buffer), for 18 h at 4°C. Addition-al sections were incubated in the presence of 20 nM unlabeled SDF-1· for the determination of nonspecific binding. Following incuba-tion, the slides were washed four times, dried and then exposed toHyperfilm-ßmax (Amersham) for 7 days.

Morphological QuantificationThe total size of brain lesions induced by H/I was quantified fol-

lowing cresyl violet Nissl staining using the computer-based imageanalysis system Biocom Historag (Les Ulis, France) and verified bylight microscopy with a magnification of !40. The confluent necrot-ic areas exposing lack of cell staining with cresyl violet and includingnumerous vacuoles were measured separately and handled as the sizeof necrosis. The size of penumbra-like area was evaluated from paral-lel sections following GFAP immunostaining using the same com-puter-based image analysis system to measure surface brain areas. Inthe GFAP-dense areas, astrocytosis was verified by light microscopywith a magnification of !100. The total size of the lesion, i.e. that ofpenumbra-like and necrotic zones within each brain region, wasexpressed as percent of the surface area of the entire anatomicalstructure selected for studying. At each of the four anteroposteriorlevels, three sections were analyzed, and the average value of thesethree measurements was used for statistical analysis. If the selectedanatomical region covered more than a single anteroposterior level(e.g. levels I–III of the FPC were invaded), the values obtained ateach level were further averaged to obtain a single value for eachanatomical region.

The density of CXCR4 receptor binding on the autoradiogramswas measured using the Biocom Historag computer-based imageanalysis system and corrected for nonspecific background staining ineach area. Since CXCR4 receptor binding density was measured inthe penumbra-like regions, the density of GFAP immunoreactivitywas also quantified in adjacent sections of the same selected areaswith Biocom Historag. The effects of H/I insult on CXCR4 bindingand GFAP immunoreactivity were expressed by calculating the dif-ferences in density values in percent between the ipsi- and contralat-eral sides of the same section. The contralateral side represented100%. The magnitude of CXCR4 binding in the penumbra was alsocompared to the GFAP density in the same penumbra region. Thequotient obtained by this calculation gave only a relative density val-



ue of CXCR4 binding to GFAP-positive astrocytosis. With this com-parison, the question was asked how much the CXCR4 binding wasdue the enhanced astrocytosis. The effects of H/I as well as the effectof DEX pretreatment on CXCR4 density were also evaluated on thecontralateral ‘intact’ side comparing the results to the respective con-trols, i.e. control and DEX animals.

StatisticsThe mortality rate during the hypoxic insult was compared

between hypoxic and HpxDEX groups applying the 2 test. For eval-uation of recovery time after H/I, the Mann-Whitney U test wasused. The planimetric data obtained from groups with different treat-ments were analyzed by repeated-measure ANOVA followed by posthoc t test. For densitometric data for two groups Student’s t test, andfor more than two groups one-way ANOVA followed by the Tukeytest were applied. Comparing the data obtained from the two sides ofthe same brain section, the paired t test was performed. Statisticaltests were carried out by Sigmastat 2 program.

Results

Effects of DEX on H/I LesionIn the present neonatal H/I model, a large brain le-

sion was found ipsilateral to the carotid ligation. Figure 1shows a representative image from an hypoxic animalstained with cresyl violet. Tissue damage (right side)invaded mainly the neocortical areas (FPC), the CPu(dorsolateral part of the corpus striatum), the hippocam-pus (pyramidal cell layers in CA3 and CA1, and dentategyrus) and the dorsolateral thalamus. Occasionally, thehypothalamus, certain brainstem areas, and additionalcortical regions like occipital and entorhinal cortex werealso damaged. On the contralateral side to the carotid liga-tion, no apparent pathological change was observable bythe Nissl staining.

Studying nervous tissue reactions to the lesion, GFAPimmunoreactivity was increased in the injured part of thehemisphere ipsilateral to the carotid ligation (fig. 2A, B).The enhanced GFAP-positive astrocytosis largely ap-peared in patches which could be delineated for densi-tometric measurements as well.

In the penumbra-like zone, i.e. in the GFAP-dense area(fig. 2C), the GFAP density varied according to the brainregion. The increment in GFAP staining was highest inthe FPC (38% increment to the control side; t = 5.45, p !0.001). Statistically significant increases were also de-tected in the CPu (23%; t = 3.19, p ! 0.05), thalamus(20%; t = 2.94, p ! 0.05) and hypothalamus (21%; t =3.38, p ! 0.05). In HpxDEX animals, a significantincrease in GFAP density was measured in the FPC (40%;t = 9.88, p ! 0.001), hippocampus (22%; t = 4.19, p ! 0.05)and thalamus (12%; t = 3.35, p ! 0.05). DEX pretreat-

Fig. 1. Representative brain section indicating the lesion extent 48 hafter ligation of the left carotid artery combined with hypoxia on the7th postnatal day (cresyl violet staining). A large lesion was noted onthe side ipsilateral to the ligation (right side), while the contralateralside remained apparently unaffected. A Level II (anterior hippocam-pus). B Magnification of the anterior hippocampus (!40; bar repre-sents 100 Ìm).

ment did not influence GFAP density in the FPC, CPuand hippocampal regions, i.e. in the overwhelming major-ity of injured areas. Only in the thalamus (t = 2.29, p !0.05) and hypothalamus (t = 2.36, p ! 0.05) was someattenuation noted.

Figure 3 shows the magnitude of the whole lesion, aswell as the necrotic and penumbra-like areas in differentanatomical regions. The total extent of lesions is repre-sented in the upper panel. In the hypoxic rats, the largestlesion occurred in the FPC (35%) and in the thalamus(33%). The CPu (29%), the hippocampus (21%) and the


Fig. 2. Visualization of astrocytosis by GFAP immunostaining fol-lowing neonatal H/I. A Coronal section of a 9-day-old animal isshown 48 h after ligation of the left carotid artery combined withhypoxia. Ipsilateral to the ligation (right side), large dense areas arevisible in the parietal cortex, in the thalamus and partly in the hippo-campus and in the hypothalamus (indicated by arrows). B Astrocyteswith increased GFAP-immunoreactive staining are shown with highmagnification inside the selected area in the neocortex delineated bythe rectangular box. C Percent increment in GFAP density in thepenumbra of the indicated areas compared to the same area in thecontralateral side. FPC = Frontoparietal cortex; CPu = caudate puta-men; Hip = hippocampus; Thal = thalamus; Hthal = hypothalamus.Means B SEM (8 rats per group); * p ! 0.05, vs. hypoxic group (un-paired t test).

60

50

40

30

20

10

0

Perc

ent

incr

emen

t in

den

sity

HypoxHpxDEX

FPC CPu Hip Thal HThal

��

A

B

C

Fig. 3. The size of total brain lesion induced by neonatal H/I in dif-ferent brain regions (A). B Extension of the penumbral area.C Necrotic core area. The extents of the injured areas are expressedin percent of the total areas of the selected brain regions. Values aremeans B SEM (8–10 rats per group). * p ! 0.05, ** p ! 0.01, *** p !0.001, vs. hypoxic group (unpaired t test). FPC = Frontoparietal cor-tex; CPu = caudate putamen; Hip = hip = Hippocampus; Thal = thal-amus; Hthal = hypothalamus; CEnt = entorhinal cortex.

��

50

40

30

20

10

0FPC CPu Hip Thal HThal Pons CEnt

FPC CPu Hip Thal HThal Pons CEnt

FPC CPu Hip Thal HThal Pons CEnt

Hypox HpxDEX

50

40

30

20

10

0

50

40

30

20

10

0

Are

a in

per

cen

tA

rea

in p

erce

nt

Are

a in

per

cen

t

��

��

�

��

�

��

��

A

B

C

hypothalamus (20%) were also largely damaged. Therewas less influence at the pons and entorhinal cortex levels.In the steroid-treated group, only 13% of the FPC wasinjured by H/I, while in other structures less than 10%damage was found. The decreased lesion size followingDEX treatment was significant in the FPC, hippocampusand thalamus (repeated ANOVA, effect of treatment:



F[1, 69] = 15.39; p ! 0.01). There was no sign for lesions inthe pons and in the entorhinal cortex in the DEX-treatedrats.

Figure 3A–C shows the extent of penumbra-like andnecrotic areas within the lesions. In the hypoxic rats (noDEX pretreatment), the ratio between the penumbra andnecrotic regions showed anatomical variations. The sizeof the necrotic area was larger in the FPC, hippocampus,and thalamus, while the size of the penumbra was largerin the CPu and hypothalamus. Necrosis was not observedin the pons. DEX treatment decreased the extent ofnecrotic areas in all brain regions (repeated ANOVA,effect of treatment: F[1, 69] = 62.04; p ! 0.001). The stron-gest, significant decrements were observed in the FPC,hippocampus and thalamus. DEX pretreatment also re-duced the size of the penumbra-like area (repeatedANOVA, effect of treatment: F[1, 69] = 34.53; p ! 0.001). Asignificant decrement was found in the CPu, thalamusand hypothalamus.

CXCR4 Receptor Density Representative autoradiographic images of [125I]-

SDF1· binding site distributions of control and H/I ratsare shown by figure 4. In the control brain (fig. 4A), bind-ing of labelled SDF-1· was observable in cortical areas.Higher receptor binding was found also in the frontal, cin-gular and entorhinal cortex, whereas the striatum andbrainstem areas showed only weak labelling. In the hippo-campus, moderate density could be detected in the CA3and CA4 areas and in the dentate gyrus. In the thalamusand in the hypothalamus, receptor labeling was low exceptfor the ventromedian nuclei. In the thalamus VPL and inthe hypothalamus VMH showed remarkably higher den-sity than the surrounding tissue.

Following H/I insult, a dense labelling occurred withinthe lesion ipsilateral to the carotid ligation (fig. 4B) in-volving the brain areas mostly affected by the insult suchas the ipsilateral neocortex, hippocampus and dorsolater-al thalamus. Following DEX pretreatment (HpxDEX),the enhancement in the SDF-1· binding is still observablein the lesion side, but it was reduced in size and density(fig. 4C).

Quantitative analysis (fig. 5A) of relative binding den-sities of CXCR4 receptors in the penumbra-like zonesas compared to the contralateral side indicates highenhancement in the CPu (66%), FPC (43%), thalamus(49%), hippocampus (30%) and hypothalamus (52%).The increment in receptor binding was strongly attenuat-ed in HpxDEX animals, as indicated by the lower percentincrements in several brain areas: in the FPC (20%), hip-

Fig. 4. Representative autoradiograms (mirror image of fig. 1) ofbrain sections following [125I]-SDF-1· binding to CXCR4 receptorsin 9-day-old rats. A Control, sham-operated rats. B 48 h after H/Iinsult. C 0.5 Ìg/g DEX pretreatment 5 h prior to H/I insult. Arrowsshow the increased receptor density at the site of the lesion.


Fig. 5. Increment in relative density of CXCR4 receptor binding inthe penumbra compared to the control unlesioned side of the brain(A). Columns are means B SEM (10 rats per group); * p ! 0.05,** p ! 0.01, vs. hypoxic group (unpaired t test). B CXCR4 receptordensity related to the actual GFAP density in the penumbra of theH/I lesion. Columns represent the means B SEM of the quotients ofCXCR4 receptor and GFAP optical densities (10 rats per group);* p ! 0.05, ** p ! 0.01, vs. hypoxic group (unpaired t test). FPC =Frontoparietal cortex; CPu = caudate putamen; Hip = hippocampus;Thal = thalamus; Hthal = hypothalamus.

100 A

80

60

40

20

0

CXCR4

CXCR4/GFAP

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Hypox

Rel

ativ

e d

ensi

tyPe

rcen

t in

crea

se in

op

tica

l den

sity

HpxDEX



��

��

��

�

�

�

�

�

B

pocampus (9%), thalamus (17%) and in the hypothalamus(9%).

The GFAP density measured in the penumbra-likezones altered according to the grade of injury induced bythe H/I insult. Therefore, the increment in CXCR4 recep-tor density was also compared to the correspondingGFAP density values in the identical penumbra regions(fig. 5B). The chemokine receptor density related toGFAP density is shown in hypoxic and HpxDEX ani-mals in figure 5B. As a result of DEX pretreatment, theCXCR4/GFAP ratio decreased markedly in most regionsexamined. The decrement reached statistical significancein the FPC, CPu, hippocampus and thalamus.

Contralateral to the LesionThe effects of H/I and DEX pretreatment on CXCR4

receptor density were also examined in the contralateralside to the carotid ligation and compared with the resultsobtained in sham-operated (control) and sham-operatedplus DEX-treated (DEX) animals, respectively. The sin-gle DEX injection alone (DEX) had no effect on receptordistribution in normal tissue (table 1).

Effects of DEX on Mortality and Functional RecoveryH/I induced a 21% mortality among the hypoxic ani-

mals: 10 of 48 died. Mortality rate was significantly lower(5%) among the DEX pretreated rats: 2 of 42 died; (¯2 =3.25, p ! 0.05). Recovery time, the time spent until therighting reflex became reestablished, was significantlyshorter in the HypDEX group compared to the hypoxicgroup (3.5 vs. 37.5 s, respectively; U = 459.5, p ! 0.01).

Table 1. Relative density of CXCR4receptor binding on the side contralateral tothe brain lesion

Treatmentgroup

FPC CPu Hippo-campus

Thalamus Hypo-thalamus

Control 36.4B3.8 14.5B3.6 41.1B5.3 24.9B3.0 23.4B5.5DEX 39.4B5.7 16.2B3.3 42.1B3.0 20.4B3.0 22.6B2.0Hypoxic 42.8B2.7 31.1B2.7a 49.8B3.7 36.7B2.7 32.1B5.2HpxDEX 39.0B3.6 31.0B3.6b 45.0B3.7 33.7B3.9b 34.0B4.2

Data are obtained by autoradiography of 9-day-old rats following H/I insult. Values aremeans B SEM (10 animals per group).

a p ! 0.05, vs. control; b p ! 0.05, vs. the DEX group.



Discussion

In the present study, an increase in CXCR4 chemokinereceptor density was observed in the neonatal rat brainfollowing hypoxia combined with ischemia. The densityof CXCR4 chemokine receptor binding was increased inthe penumbra of the lesion. DEX pretreatment counter-acted the damage induced by H/I, inhibited the enhance-ment in CXCR4 binding, decreased the mortality rateand finally shortened the functional recovery time of ani-mals exposed to H/I.

It has been observed in different cerebral ischemiamodels that the initial injury is primarily necrosis [18, 34]as a result of ischemia-triggered neurotoxicity [35]. Ac-cording to our observation, DEX pretreatment was mosteffective in decreasing the necrotic type of injury. Severalobservations indicate that GC applied shortly before, dur-ing or following neonatal hypoxia exacerbate the braindamage in murine species [20, 36]. On the other hand,GCs have been clearly demonstrated to possess protectiveeffect on H/I-induced neurodegeneration when their ap-plication preceded the insult for many hours or days [21,37]. According to a further study, application of DEX 6 hbefore the insult resulted in a beneficial effect [23]. Evi-dence indicated furthermore that these metabolic re-sponses are mediated by GC receptor [38]. These mecha-nisms could be responsible for the attenuation in the ini-tial injury, especially the lower occurrence of large necrot-ic areas 48 h after H/I.

In the adult CNS, CXCR4 has been shown to beexpressed in glial cells (microglia and astrocytes) as well asin neurons [16, 33, 39]. Astrocytes and neurons also pro-duce SDF-1· [40, 41]. In our neonatal rats, CXCR4 wasfound in higher density than in the adult brain and wasdetectable all over the cortex, mes- and diencephalon andin the examined areas of the brainstem. There were somestructures which expressed high receptor density, e.g. theVPL in the thalamus and the VMH in the hypothalamus.Strong receptor binding was also shown in the dentategyrus, CA1 and CA3 areas in the hippocampus, and in theouter layer of the frontal, parietal and entorhinal cortex.The enhancement in CXCR4 density could be due pri-marily to the increased number of activated astrocytesexpressing CXCR4 in the lesioned areas. DEX pretreat-ment reduced the astrocytosis in several structures af-fected as well as CXCR4 receptor binding. However, thedecrease in chemokine receptor density was stronger thanthe reduction in the astrocytosis. We may assume the exis-tence of an inhibiting type of action of DEX on theCXCR4 expression itself in glial as well as neuronal cells.

It has been well established that DEX inhibits the pro-duction of pro-inflammatory cytokines [25, 42], andamong them, strongly inhibits the IL-1ß release [43, 44].On human fetal astrocytes, the expression of the CXCR4receptor, which is normally low, dramatically increasedby exposure to IL-1ß [13]. This finding supports theassumption that DEX could diminish CXCR4 expressionvia inhibition of pro-inflammatory cytokines promotingthe upregulation of this receptor during inflammation.Further experiments are required to test this hypothesis.

SDF-1· has been suggested to be responsible for reac-tive astrogliosis during brain inflammation. SDF-1· ismarkedly increased in cultured astrocytes followingchronic LPS stimulation [39]. In vitro studies revealedthat SDF-1· stimulated cortical type I astrocytes throughactivation of the ERK signalling pathway and CXCR4-induced Ca2+ mobilization [45]. Thus, the CXCR4 recep-tor could be highly implicated in pathological conditionssuch as reactive gliosis following H/I insult.

The enhanced astrogliosis reflected by intensive GFAPstaining suggests that an intensive inflammatory reactionoccurred with glial cell activation within 2 days after theH/I insult. The role of astrocyte proliferation in triggeringinflammatory activation following hypoxia has been de-scribed earlier both in an in vivo animal experiment andon human astrocytes [46, 47]. Astrocyte proliferation hasbeen described from 48 h to 7 days after ischemia [46].According to our results, DEX was also able to diminishastrocytosis in brain areas with milder ischemic injury.Regarding the data obtained in this study, it is not yetclear, however, whether the decrease in astrocytosis fol-lowing DEX is a consequence of a direct inhibition ofastrocyte proliferation, or the result of a milder initialinjury.

GC have been proved to inhibit astrocyte proliferationfollowing hypoxic insult, but not in the normal brain [26].This is consistent with the present results on CXCR4expression and astrocytes in the neonatal brain (table 1).Astrocytes are reported to play an important role in theregulation of inflammatory responses due to their proxim-ity and interaction with the endothelium that contributesto leukocyte infiltration through the blood-brain barrier[13]. Astrocytes are also responsible for the maintenanceand the prolongation of an inflammatory reaction via pro-ducing inflammatory mediators such as TNF· and IL-1ß[48].

The early brain inflammatory processes induced by theH/I insult are deeply involved in the late neuronal death,and among them the SDF-1·/CXCR4 receptor-ligandactivation can induce apoptotic cell death [48, 49]. Apop-


tosis appears to play a more pronounced role in ischemiain the neonates than in adults [50]; thus, the decrease inCXCR4 receptor availability can reduce the action of thischemokine [51].

Our results suggest that the GC action induces an earlyanti-inflammatory process by decreasing astrocyte prolif-eration and CXCR4 receptor density in astrocytes andneuronal cells. Blocking CXCR4 receptor activity associ-ated with brain inflammation in H/I could represent a sig-nificant therapeutic possibility.

Acknowledgments

This study was partially supported by grants from the MedicalResearch Council (ETT 161/1998) and the National Science Founda-tion (OTKA, T38388) in Hungary, and by the ‘Institut National de laSanté et de la Recherche Médicale’ (INSERM).

References

1 Vannucci RC: Hypoxic-ischemic encephalopa-thy. Am J Perinatol 2000;17:113–120.

2 Marin-Padilla M: Developmental neuropathol-ogy and impact of perinatal brain damage. III.Gray matter lesions in the neocortex. J Neuro-pathol Exp Neurol 1999;58:407–429.

3 Barks JD, Silverstein FS: Excitatory aminoacids contribute to the pathogenesis of perina-tal hypoxic-ischemic brain injury. Brain Pathol1992;2:235–243.

4 Nyakas C, Buwalda B, Luiten PGM: Hypoxiaand brain development. Prog Neurobiol 1996;49:1–51.

5 Ishida A, Ishiva S, Trescher WH, Nakajima W,Lange MS, Blue ME, Johnston MV: Delayedincrease in neuronal nitric oxide synthase im-munoreactivity in thalamus and other brainregions after hypoxic-ischemic injury in neona-tal rats. Exp Neurol 2001;168:323–333.

6 Cheng Y, Deshmukh M, D’Costa A, DemaroGA, Gidday GM, Shah A, Sun Y, Jacquin MF,Johnson MR Jr, Holtzman DM: Caspase inhib-itor affords neuroprotection with delayed ad-ministration in a rat model of neonatal hypox-ic-ischemic brain injury. J Clin Invest 1998;101:1992–1999.

7 Rice JE, Vannucci RC, Brierley JB: The in-fluence of immaturity on hypoxic-ischemicbrain damage in the rat. Ann Neurol 1981;9:131–141.

8 Vannucci RC: Experimental biology of cerebralhypoxia-ischemia: Relation to perinatal braindamage. Pediatr Res 1990;27:317–326.

9 Bona E, Abdersson A, Blomgren K, Gilland E,Puka-Sundvall M, Gustafson K, Hagberg H:Chemokine and inflammatory cell response tohypoxia-ischemia in immature rats. PediatrRes 1999;45:500–509.

10 Giulian D: Microglia, cytokines, and cytotox-ins: Modulators of cellular responses after inju-ry to the central nervous system. J ImmunolImmunopharmacol 1990;10:15–21.

11 Gabellec MM, Crumeyrolle-Arias M, Colin C,Auriou N, Jacque C, Haour F: Expression ofIL-1 and IL-1 receptor gene in the brain afterhippocampal injury. Neurosci Res 1999;33:251–260.

12 Ivacko J, Szaflarski J, Malinak C, Flory C,Warren JS, Silverstein FS: Hypoxic-ischemicinjury induce monocyte chemoattractant pro-tein-1 expression in neonatal rat brain. J CerebBlood Flow Metab 1997;17:759–770.

13 Hesselgesser J, Horuk R: Chemokine and che-mokine receptor expression in the central ner-vous system. J Neurovirol 1999;5:13–26.

14 Gupta SK, Lysko PG, Pilarisetti K, Ohlstein E,Stadel JM: Chemokine receptors in human en-dothelial cells. J Biol Chem 1998;273:4282–4287.

15 Bajetto A, Barbero S, Bonavia R, Piccioli P,Piravi P, Florio T, Schettini G: Stromal cell-derived factor-1alpha induces astrocyte prolif-eration through the activation of extracellularsignal-regulated kinases 1/2 pathway. J Neuro-chem 2001;77:1226–1236.

16 Banisadr G, Fontanges P, Haour F, Kitabgi P,Rostene W, Melik Parsadaniantz S: Neuroana-tomical distribution of CXCR4 in adult ratbrain and its localization in cholinergic anddopaminergic neurons. Eur J Neurosci 2002;16:1661–1671.

17 Zou Y-R, Kottman AH, Koruda M, Taniuchi I,Littman DR: Function of the chemokine recep-tor CXCR4 in haematopoiesis and in cerebel-lar development. Nature 1998;393:595–599.

18 Nakajima W, Ishida A, Lange MS, GabrielsonKL, Wilson MA, Martin LJ, Blue ME, John-ston MV: Apoptosis has a prolonged role in theneurodegeneration after hypoxic ischemia inthe newborn rat. J Neurosci 2000;20:7994–7998.

19 Avery GB, Fletcher AB, Kaplan M, BrudnoDS: Control trial of dexamethasone in respira-tor-dependent with bronchopulmonary dyspla-sia. Pediatrics 1985;106–111.

20 Chen J, Adachi N, Tsubota S, Nagaro T, AraiT: Dexamethasone augments ischemia-in-duced extracellular accumulation of glutamatein gerbil hippocampus. Eur J Pharmacol 1998;347:67–70.

21 Tuor UI: Glucocorticoids and the preventionof hypoxic-ischemic brain damage. NeurosciBiobehav Rev 1997;21:175–179.

22 Sternberg EM: Interactions between the im-mune and neuroendocrine systems. Prog BrainRes 2000;122:35–42.

23 Tuor UI, Yager JY, Bascaramurty S, Del BigioMR: Dexamethasone prevents hypoxia/isch-emia-induced reduction in cerebral glucose uti-lization and high-energy phosphate metabolitesin immature brain. J Neurochem 1997;69:1954–1963.

24 Krugers HJ, Maslam S, Van Vuuren SM, KorfJ, Joels M: Postischemic steroid modulation:Effects on hippocampal neuronal integrity andsynaptic plasticity. J Cereb Blood Flow Metab1999;19:1072–1082.

25 Mathieu M, Gougat C, Jaffuel D, Danielsen M,Godard P, Bosquet J, Demoly P: The glucocor-ticoid receptor gene as a candidate for genetherapy in asthma. Gene Ther 1999;6:245–252.

26 McRae A, Bona E, Hagberg H: Microglia-astro-cyte interactions after cortisone treatment in aneonatal hypoxia-ischemia model. Brain ResDev Brain Res 1996;94:44–51.

27 Tuor UI, Chumas PD, Del Bigio ML: Pre-vention of hypoxic-ischemic damage withdexamethasone is dependent on age and notinfluenced by fasting. Exp Neurol 1995;132:116–122.

28 Odemis V, Moepps B, Gierschilk P, Engele J:Interleukin-6 and cAMP induce stromal cell-derived factor-1 chemotaxis in astroglia by up-regulating CXCR4 cell surface expression. Im-plication for brain inflammation. J Biol Chem2002;277:39801–39808.

29 Jensh RP: Behavioral testing procedures: Areview; in Johnson EM, Kochnar DM (eds):Handbook of Experimental Pharmacology.Berlin, Springer, 1983, vol 65, pp 171–206.

30 Melgar MA, Park H, Rafols JA, Diaz FG: Amodel of global forebrain ischemia/reperfusionin the awake rat. Neurol Res 2002;24:97–106.

31 Paxinos G, Watson C: The Rat Brain in Stereo-taxic Coordinates. Sydney, Academic Press,1986.

32 Abraham I, Veenema AH, Nyakas C, HarkanyT, Bohus BG, Luiten PG: Effect of corticoste-rone and adrenalectomy on NMDA-inducedcholinergic cell death in the rat magnocellularnucleus basalis. J Neuroendocrinol 1997;9:713–720.



33 Banisadr G, Dicou E, Berbar T, Rostène W,Lombet A, Haour F: Characterization and vi-sualization of [125I]-stromal cell-derived factor-1· binding to CXCR4 receptors in rat brainand human neuroblastoma cells. J Neuroim-munol 2000;110:151–160.

34 Li Y, Powers C, Jiang N, Chopp M: Intact,injured, necrotic and apoptotic cells after focalcerebral ischemia in the rat. Neurol Sci 1998;156:119–132.

35 Charriaut-Marlangue C, Remolleau S, Aggoun-Zouaoui D, Ben-Ari Y: Apoptosis and pro-grammed cell death: Role in cerebral ischemia.Biomed Pharmacother 1998;52:264–269.

36 Smith-Swintosky VL, Pettigrew LC, SapolskyRM, Phares C, Craddock SD, Brooke SM,Mattson MP: Metyrapone, an inhibitor of glu-cocorticoid production, reduces brain injuryinduced by focal and global ischemia and sei-zures. J Cereb Blood Flow Metab 1996;16:585–598.

37 Macaya A, Munell F, Ferrer I, de Torres C,Reventos J: Cell death and associated c-juninduction in perinatal hypoxia-ischemia. Effectof the neuroprotective drug dexamethasone.Brain Res Mol Brain Res 1998;56:29–37.

38 Bajetto A, Bonavia R, Barbero S, Piccioli P,Costa A, Florio T, Schettini G: Glial and neu-ronal cells express functional receptor CXCR4and its natural ligand stromal cell-derived fac-tor 1. J Neurochem 1999;73:2348–2357.

39 Bajetto A, Bonavia R, Barbero S, Florio T, Cos-ta A Schettini G: Expression of chemokinereceptors in the rat brain. Ann NY Acad Sci1999;876:201–209.

40 Tuor UI, Del Bigio MR: Protection againsthypoxic-ischemic damage with corticosteroneand dexamethasone: Inhibition of effect by aglucocorticoid antagonist RU38486. Brain Res1996;743:258–262.

41 Banisadr G, Skrzydelski D, Kitabgi P, RosteneW, Melik Parsadaniantz S: Highly regionalizeddistribution of stromal cell-derived factor-1/CXCL12 in adult rat brain: Constitutive ex-pression in cholinergic, dopaminergic and va-sopressinergic neurons. Eur J Neurosci 2003;18:1593–1606.

42 Bourke E, Moynagh PN: Antiinflammatory ef-fects of glucocorticoids in brain cells, indepen-dent of NF-ÎB. J Immunol 1999;163:2113–2119.

43 Lee SC, Liu W, Dickson DW, Brosnan CF, Ber-man JW: Cytokine production by human fetalmicroglia and astrocytes: Differential induc-tion by lipopolysaccharide and IL-1ß. J Immu-nol 1993;150:2659–2667.

44 Santos AA, Scheltinga MR, Lynch E, BrownEF, Lawton P, Chambers E, Browning J, Di-narello CA, Wolff SM, Wilmore DW: Elabora-tion of interleukin 1-receptor antagonist is notattenuated by glucocorticoids after endotox-emia. Arch Surg 1993;128:138–144.

45 Lazarini F, Casanova P, Tham TN, De Clerq E,Arenzana-Seisdedos F, Baleux F, Dubois-Dalcq M: Differential signalling of the chemo-kine receptor CXCR4 by stromal cell-derivedfactor 1 and the HIV glycoprotein in rat neu-rons and astrocytes. Eur J Neurosci 2000;12:117–125.

46 Benjelloun N, Renolleau S, Represa A, Ben AriY, Charriaut-Marlangue C: Inflammatory re-sponses in the cerebral cortex after ischemia inthe P7 neonatal rat. Stroke 1999;30:1916–1924.

47 Stoll G, Jander S, Schroeter M: Inflammationand glial responses in ischemic brain injury.Prog Neurobiol 1998;56:149–171.

48 Bezzi P, Domercq M, Brambilla R, Galli L,Schols D, De Clercq E, Vescovi A, Bagetta G,Kollias G, Meldolesi J, Volterra A: CXCR4-activated astrocyte glutamate release viaTNF·: Amplification by microglia triggers neu-rotoxicity. Nat Neurosci 2001;4:702–710.

49 Pulera MR, Adams LM, Liu H, Santos GD,Nishimura RN, Yang F, Cole GM, WasterlainCG: Apoptosis in a neonatal rat model of cere-bral hypoxia-ischemia. Stroke 1998;29:2622–2630.

50 Hesselgesser J, Taub D, Baskar P, GreenbergM, Hoxie J, Kolson DL, Horuk R: Neuronalapoptosis induced by HIV-1 gp120 and thechemokine SDF-1· is mediated by the chemo-kine receptor CXCR4. Curr Biol 1998;8:595–598.

51 Zheng J, Thylin MR, Ghorpade A, Xiong H,Persidsky Y, Cotter R, Niemann D, Che MH,Zeng YC, Gelbard HA, Shepard RB, SwartzJM, Gendelman HE: Intracellular CXCR4 sig-naling, neuronal apoptosis and neuropathogen-ic mechanisms of HIV-1-associated dementia.J Neuroimmunol 1999, 98:185–200.

Documents

Dexamethasone Downregulates Chemokine Receptor CXCR4 and Exerts Neuroprotection against Hypoxia/Ischemia-Induced Brain Injury in Neonatal Rats