Ž .Brain Research 776 1997 230–234

Research report

Expression of the glial cell line-derived neurotrophic factor gene in rat brainafter transient MCA occlusion

K. Abe ) , T. HayashiDepartment of Neurology, Tohoku UniÕersity School of Medicine, Sendai, Japan

Accepted 12 August 1997

Abstract

Ž .Change of the glial cell line-derived neurotrophic factor GDNF gene expression in rat brain was examined after transient middleŽ .cerebral artery MCA occlusion of adult rats. Northern blot analysis showed that the mRNA began to be induced in the occluded MCA

from 1 h of reperfusion with a peak at 3 h, and almost diminished by 1 day of reperfusion. Immunohistochemical analysis with brainsections showed an expression of GDNF-like immunoreactivity in neurons of the cerebral cortex and caudate after 90 min of ischemia ina similar way to the mRNA, but the staining was more disseminated and stronger in the cerebral cortex than the caudate. No glial cell wasstained in the brain sections. The present results indicate that the GDNF gene was expressed in an early stage of reperfusion in neuronalcells of the MCA territory, but that the staining property was different between in the cerebral cortex and caudate. q 1997 ElsevierScience B.V.

Keywords: GDNF; Ischemia; Rat; Middle cerebral artery

1. Introduction

Growth factors are essential for the survival of manypopulations of developing neurons, and could profoundlyinfluence the differentiation of neurons during normaldevelopment. However, they also play important roles inthe recovery of mature neurons under pathological condi-

w xtions 12 . Growth factors are recently classified into sev-eral groups such as neurotrophins, cytokines, fibroblast

Ž .growth factor FGF family, insulin family and transform-Ž . w xing growth factor TGF -b superfamily, or so 12 .

The members of TGF-b superfamily promote survivalof midbrain dopaminergic neurons and protect them against

w xinjury in vitro and in vivo 9,15 . Glial cell line-derivedŽ .neurotrophic factor GDNF has recently been isolated as a

peptide protein for survival and neurite outgrowth ofdopaminergic and motor neurons as well as peripheral

Abbreviations: BDNF, brain-derived neurotrophic factor; GDNF, glialcell line-derived neurotrophic factor; HE, hematoxylin and eosin; MCA,middle cerebral artery; MCAO, middle cerebral artery occlusion; NMDA,N-methyl-D-aspartate; PCR, polymerase chain reaction

) Corresponding author. Department of Neurology, Tohoku University,School of Medicine, 1-1 Seiryo-machi, Aobaku, Sendai 980-77, Japan.

Ž .Fax: q81 22 717-7192; E-mail: abeksedi@mail.cc.tohoku.ac.jp

w xsensory and sympathetic neurons 5,10,18 , and is structu-w xally related to members of the TGF-b superfamily 11 .

Recent reports discovered that the product of the c-retproto-oncogene, Ret, is a member of receptor tyrosinekinase superfamily, and is the functional receptor of GDNF

w xfor its cell-to-cell signalling 4,19 .Induction and the possible protective roles have been

reported in many growth factors after cerebral ischemiaw x3,8,14,17 . Hippocampal neurons expressed GDNF mRNA

w xin response to kainate-induced excitation 7 . However, therole of GDNF in cerebral ischemia has not been wellstudied. Therefore, change of the GDNF gene expressionin rat brain was examined with Northern blot and immuno-histochemical analyses after transient middle cerebral arteryŽ .MCA occlusion.

2. Materials and methods

2.1. Animal model

The animal model employed the occlusion of the MCAby using a microembolus of nylon thread according to our

w x Žprevious report 2 . Adult male Wistar rats body weight:

0006-8993r97r$17.00 q 1997 Elsevier Science B.V. All rights reserved.Ž .PII S0006-8993 97 01041-X

( )K. Abe, T. HayashirBrain Research 776 1997 230–234 231

.250–280 g were anesthetized for surgery by inhalation ofŽ .a nitrous oxideroxygenrhalothane 69% : 30% : 1% mix-

Ž .ture. Occlusion of MCA MCAO or reperfusion wasperformed by an insertion or removal of the nylon thread.Body temperature was maintained at 37"0.38C duringsurgical procedure. With removal of nylon thread after acertain period of ischemia, the animals recovered at ambi-

Ž .ent temperature 21–238C , and decapitated at 1 h, 3 h, 1,3 or 7 days of reperfusion. Sham control brains wereobtained in the similar way without MCAO. The experi-mental protocol and procedures were approved by theAnimal Committee of the Tohoku University School ofMedicine.

2.2. Northern blot analysis

To investigate how the duration of ischemia influencesthe temporal profile of mRNA induction for GDNF, 1 or 2

Ž .h of ischemia was imposed on each animal ns3 . Ateach time point of reperfusion, cerebral cortices of theright MCA territory were quickly taken and frozen. TotalRNAs were extracted from the samples, and Northern blotanalysis was performed according to our previous reportw x1 . Signals for GDNF mRNA were detected with an

Ženhanced chemiluminescence method Direct Nucleic Acid.Labeling and Detection System, Amersham, Tokyo . To

generate a probe for Northern hybridization of GDNF,total RNA was extracted from caudate samples of 1 day ratafter birth, and 700 bp cDNA was obtained by polymerase

Ž . w xchain reaction PCR 16 . Actin mRNA was also detectedŽin the same blot as a control using a cDNA probe 400 bp

.size, Cat. a015-10701, Wako Chemicals, Tokyo .

2.3. Immunohistochemical and histological analyses

For immunohistochemical analysis, rat brains were per-fused, and then fixed with 4% paraformaldehyde in phos-phate buffered saline at each sampling time of reperfusion

Ž .after 90 min of MCAO ns2 . Coronal sections at thecaudate level were cut with 1 mm thickness from theparaffined samples, and were reacted with a rabbit poly-

Žclonal antibody against human GDNF Santa Cruz. ŽBiotechnology, Santa Cruz, CA at 1 : 50 dilution final 2

. w xmgrml according to our previous report 1 . The immuno-reactivity was developed with streptoavidin-biotin using

Žaminoethylcarbazol as a color substrate Histofine SAB.and AEC kits, Nichirei, Tokyo . To ascertain specific

binding of the antibody for the protein, a set of brainsections were stained in a similar way without the firstantibody. The staining was categorized into 5 grades in the

Ž . Ž .following: no staining y , partially and weakly " , fullyŽ . Ž . Ž .q , densely 2q or diffuse and densely 3q , stained.

Ž .Hematoxylin-eosin HE staining was performed to evalu-ate brain cell damage at 3 and 7 days of reperfusion. Thescoring of immunohistochemistry and eÕaluation of HE

Ž .Fig. 1. Representative Northern blot analysis of GDNF upper panel andŽ .actin lower panel mRNAs in rat cerebral cortices of the ipsilateral MCA

Ž . Ž . Žterritory of sham control C and post-ischemic brain at 1 1 h and 3 3. Ž .h h, and 1 day 1 d after 2 h of transient MCA occlusion. Note

Ž .induction and recovery of GDNF mRNA large arrowhead in contrast toŽ .actin mRNA small arrowhead .

staining were performed in a blind fashion by two analystsother than the authors.

3. Results

Northern blot analysis with cerebral cortical samplesshowed that GDNF mRNA was not detectable in the shamcontrol cerebral cortex. After 2 h of transient MCAO, the

Ž .mRNA 0.8 kb began to be induced in the MCA territoryfrom 1 h of reperfusion with a peak at 3 h, and then

Ž .diminished by 1 day Fig. 1, upper panel . Data after 1 h ofMCAO showed an essentially similar pattern with slightdecrease of the signal intensities. No signal was found inbrain samples at 3 and 7 days of reperfusion after 1 or 2 hof MCAO. Actin mRNA did not show a significant change

Table 1Immunohistochemical analyses of GNDF after transient MCA occlusion

Cerebral cortex Caudate

C-1 – –C–2 – –1 h-1 2q "

1 h-2 q -3 h-1 3q q3 h-2 2q "

1 d-1 " -1 d-2 " -3 d-1 - -3 d-2 - -7 d-1 - -7 d-2 - -

Ž .At each time point, two animals were studied -1 and -2 . The stainingŽ .was categorized into 5 grades in the following way: no staining - ;

Ž . Ž . Ž .partially and weakly " , fully q , densely 2q , or diffuse andŽ .densely 3q , stained.

( )K. Abe, T. HayashirBrain Research 776 1997 230–234232

Ž .after 1 or 2 Fig. 1, lower panel h of transient MCAO.Triplicated experiments showed similar results.

GDNF-like immunoreactivity was not detectably pre-Ž .sent in brain sections of sham control Fig. 2a . However,

it began to be expressed in neuronal cytoplasm of theŽ .cerebral cortex Fig. 2b, arrowheads and caudate in the

MCA territory at 1 h of the reperfusion, and reached theŽ .maximum at 3 h Fig. 2c,d, arrowheads with a gradualŽ .decrease by 1 day Fig. 2e, arrowheads . Regional distribu-

tion of the neurons expressing GDNF-like immunoreactiv-ity at 3 h is illustrated in Fig. 2f. The staining was no

longer evident at 3 or 7 days. Brain cells positive forGDNF-like immunoreactivity was mainly large pyramidal

Ž .neurons in layers 4–6 of cerebral cortex Fig. 2c , andlarge neurons in the dorsal portion of the caudate nucleusŽ .Fig. 2d,f . Duplicated experiments showed similar results.The number of staining cells was larger and also thestaining was stronger in the cerebral cortex than in the

Ž .caudate Fig. 2c vs. d . No glial cell was stained in thebrain sections. Sections without first antibody showed nostaining. Temporal profile of the immunostaining is sum-marized in Table 1. Histopathological study with HE stain-

Ž .Fig. 2. Representative photographs of immunostaining against GDNF protein panels a–e and an illustrative demonstration of spatial distribution ofŽ . Ž .GDNF-like immunoreactivity expressing cells at 3 h of reperfusion panel f . Panels a, b, c, e and g represent cerebral cortical sections of sham control a

Ž . Ž . Ž . Ž . Ž . Ž .and at 1 b and 3 c h, 1 e and 7 days g of reperfusion, respectively. Panels d and h represent dorsal portion of caudate at 3 h d and 7 days h ofreperfusion. Shaded area in panel f represents region of MCA territory, and open, dotted, or filled symbols correspond to the grade of staining as ", q,2q to 3q , respectively, as in Table 1. Magnification of panels a–e and g–h is =200 or =300, and bars in panels a–e and g–h represent 0.125 mm.

( )K. Abe, T. HayashirBrain Research 776 1997 230–234 233

Ž .ing showed no evident cell damage at 3 and 7 Fig. 2g,hdays of reperfusion.

4. Discussion

The present results indicate that GDNF gene was ex-pressed in an early stage of reperfusion in neuronal cells ofthe MCA territory. The induction reached the maximum at3 h, and almost diminished by 1 day. Although GDNF was

w xoriginally isolated from cultured glial cells 11 , the geneŽwas expressed only in neurons in the present study Fig.

.2 . Messenger RNA for a neurotrophin, brain-derived neu-Ž .rotrophic factor BDNF , was increased in neurons from 2

w xto 18 h following the focal ischemia 3 . Transient focalischemia caused a significant increase of basic FGF im-munoreactivity in astrocytes that sustained for 30 daysw x17 . The level of TGF-b mRNA and the protein becamemaximum at 2–3 days and decreased by 7–14 days after afocal cerebral injury in rat, and the localization of TGF-b

w xwas predominantly in astrocytes or macrophages 14 .Thus, such a rapid and neuronal expression of GDNF afterfocal cerebral ischemia may be characteristics amonggrowth factors, and suggests a specific role of GDNF inthe focal ischemic injury of rat brain. On the other hand,BDNF and neurotrophin-3 mRNAs were quickly inducedin hippocampal dentate granule cells after transient globalbrain ischemia in rat with a peak at 2–4 h with subsequent

w xdecrease by 1 day 13 . Further study of GDNF on caseswith transient global ischemia may reveal whether if atemporal profile of GDNF gene expression is differentfrom the cases of BDNF and NT-3.

Different staining property between the cerebral cortexand caudate is interesting by taking into account a contrastcase with basic FGF gene expression. A shorter duration oftransient MCAO showed a sustained expression of basic

w xFGF in the dorsal caudate neurons in rat 8 . However,GDNF gene expression was more disseminated and

Žstronger in the cerebral cortex than the caudate Fig. 2c vs..d although the temporal profile was similar between the

Ž .regions Table 1 . This discrepancy suggests different rolesof the growth factors after brain ischemia. Although GDNFgene expression may be regulated by glutamatergic N-

Ž .methyl-D-aspartate NMDA neurotransmission in as-w xtroglial cells of the caudate 6 , the present results suggest

that glutamatergic neurotransmission did not affect theGDNF gene expression in astroglial cells of caudate. Be-cause the present animal model did not result in a patho-

Ž .logical neuronal death Fig. 2g,h , the expression of GDNFgene in the present study may be related to a functionalrecovery of neurons after ischemia.

Acknowledgements

This work was partly supported by Grant-in-Aid forŽ .Scientific Research on Priority Areas Kanazawa I.

07264204 and 08256204, and Grant-in-Aid for ScientificŽ . Ž .Research B 09470151 and C 06807055 from the Min-

istry of Education, Science and Culture of Japan.

References

w x1 K. Abe, M. Aoki, J. Kawagoe, T. Yoshida, A. Hattori, K. Kogure,Y. Itoyama, Ischemic delayed neuronal death: a mitochondrial hy-

Ž .pothesis, Stroke 26 1995 1478–1489.w x2 K. Abe, T. Araki, K. Kogure, Recovery from edema and of protein

synthesis differs between in the cortex and caudate following tran-Ž .sient focal cerebral ischemia in rats, J. Neurochem. 51 1988

1470–1476.w x3 M.C. Comelli, D. Guidolin, M.S. Seren, R. Zanoni, R. Canella, R.

Rubini, H. Manev, Time course, localization and pharmacologicalmodulation of immediate early inducible genes, BDNF and trkBmRNAs in the rat brain following photochemical stroke, Neuro-

Ž .science 55 1993 473–490.w x4 P. Durbec, C.V. Marcos-Gutierrez, C. Kilkenny, M. Grigoriou, K.

Wartiowaara, P. Suvanto, D. Smith, B. Ponder, F. Costantini, M.Saarma, H. Sariola, V. Pachinis, GDNF signalling through the Ret

Ž .receptor tyrosine kinase, Nature 381 1996 789–793.w x5 C.E. Henderson, H.S. Phillips, R.A. Pollock, A.M. Davies, C.

Lemeulle, M. Armanini, L.C. Simpson, B. Moffet, R.A. Vandlen,V.E. Koliatsos, A. Rosenthal, GDNF: a potent survival factor formotoneurons present in peripheral nerve and muscle, Science 266Ž .1994 1062–1064.

w x6 A. Ho, A.C. Gore, C.S. Weickert, M. Blum, Glutamate regulation ofGDNF gene expression in the striatum and primary striatal astro-

Ž .cytes, Neuroreport 6 1995 1454–1458.w x7 C. Humpel, B. Hoffer, I. Stromberg, S. Bektesh, F. Collins, L.¨

Olson, Neurons of the hippocampal formation express glial cellline-derived neurotrophic factor messenger RNA in response to

Ž .kainate-induced excitation, Neuroscience 59 1994 791–795.w x8 T. Kato, S. Nakano, K. Kogure, H. Sasaki, K. Koiwai, Y. Yamasaki,

T. Katagiri, H. Sasaki, The binding of basic fibroblast growth factorto ischemic neurons in the rat, Neuropathol. Appl. Neurobiol. 18Ž .1992 282–290.

w x9 K. Krieglstein, C. Suter-Crazzolara, W.H. Fisher, K. Unsicker,TGF-b superfamily promote survival of midbrain dopaminergicneurons and protect them against MPPq toxicity, EMBO J. 14Ž .1995 736–742.

w x10 L. Li, W. Wu, L.-F.H. Lin, M. Lei, R.W. Oppenheim, L.J. Houenou,Rescue of adult mouse motoneurons from injury-induced cell deathby glial cell line-derived neurotrophic factor, Proc. Natl. Acad. Sci.

Ž .USA 92 1995 9771–9775.w x11 L.-F.H. Lin, D.H. Doherty, J.D. Lile, S. Bektesh, F. Collins, GDNF:

a glial cell line-derived neurotrophic factor for midbrain dopaminer-Ž .gic neurons, Science 260 1993 1130–1132.

w x Ž .12 R.M. Lindsay, Neuron saving schemes, Nature 373 1995 289–290.w x13 O. Lindvall, P. Ernfors, J. Bengzon, Z. Kokaia, M.L. Smith, B.K.

Siesjo, H. Persson, Differential regulation of mRMAs for NGF,¨BDNF, and NT-3 in the adult rat brain following cerebral ischemia

Ž .and hypoglycemic coma, Proc. Natl. Acad. Sci. USA 89 1992648–652.

w x14 A. Logan, S.A. Frautschy, A.M. Gonzalez, M.B. Sporn, A. Baird,Enhanced expression of transforming growth factor beta 1 in the rat

Ž .brain after a localized cerebral injury, Brain Res. 587 1992 216–225.

w x15 K.T. Poulsen, M.P. Armanini, R.D. Klein, M.A. Hynes, H.S. Phillips,A. Rosenthal, TGF-b2 and TGF-b3 are potent survival factors for

Ž .midbrain dopaminergic neurons, Neuron 13 1994 1245–1252.w x16 D.G. Schaar, B.-A. Sieber, C.F. Dreyfus, I. Black, Regional and

( )K. Abe, T. HayashirBrain Research 776 1997 230–234234

cell-specific expression of GDNF in rat brain, Exp. Neurol. 124Ž .1993 368–371.

w x17 K. Takami, M. Iwane, Y. Kiyota, M. Miyamoto, R. Tsukuda, S.Shiosaka, Increase of basic FGF immunoreactivity and its mRNAlevel in rat brain following transient forebrain ischemia, Exp. Brain

Ž .Res. 90 1992 1–10.¨w x18 A. Tomac, E. Lindqvist, L.-F.H. Lin, S.O. Ogren, D. Young, B.J.

Hoffer, L. Olson, Protection and repair of the nigrostriatal dopamin-Ž .ergic system by GDNF in vivo, Nature 373 1995 335–339.

w x19 M. Trupp, E. Arenas, M. Fainzilber, A.-S. Nilsson, B.-A. Sieber, M.Grigoriou, C. Kilkenny, E. Salazar-Grueso, V. Pachnis, U. Arumae,¨H. Sariola, M. Saarma, C.F. Ibanez, Functional receptor for GDNF

Ž .encoded by the c-ret proto-oncogene, Nature 381 1996 785–789.