The incidence of stroke is significantly higher in the geri-atric population, with 75% of strokes occurring over the ageof 65.1,2) Oxidative insult after ischemia-reperfusion (I/R) in-jury enhances oxidative modification of lipids and proteins,and modulates apoptotic signaling pathways. Oxidative stressleads to a rise in protein carbonyl content, which increaseswith age,3,4) and thus aged rats are more susceptible to ischemic injury. A high level of glutathione peroxidase(GSHPx), an important antioxidant enzyme, confers greaterprotection against oxidative stress than either superoxide dis-mutase or catalase alone, and GSHPx is usually reported todecrease after ischemic insult.5)

Cerebral ischemia for 85 min followed by reperfusion for45 min results in generation of oxidative stress.6) Free radi-cals have been shown to be associated with both ischemicand post-ischemic injury to the brain.7) Neuronal damage isenhanced after I/R injury and may be manifested as eitherapoptosis or necrosis of neurons.8) Several studies havedemonstrated activation of mitogen activated protein kinase(MAPK) pathways in the enhancement of neuronal cell deathin the cortex and hippocampus regions after global cerebralischemia.9) Some reports also suggest that hydrogen peroxide(H2O2) and nitric oxide (NO), another potentially damagingfree radical, act synergistically to induce neuronal death viaactivation of MAPK, while activation of MAPK may also re-sult in the subsequent elevation of NO.10)

Edaravone (3-methyl-1-phenyl-2-pyrazolin-5-one), a po-tent free radical scavenger, is currently being used in themanagement of acute ischemic stroke.11,12) Edaravone haspotent hydroxyl radical scavenging properties and reduces

nitric oxide production dose dependently.13) Its ability to pre-vent lipid peroxidation is comparable to that of ascorbic acidand a-tocopherol.14) Furthermore, it has recently been re-ported to confer protection against ischemia-induced neu-ronal damage in the neonatal rat brain.15) However, the effectof edaravone on the MAPK pathways after global cerebral is-chemia is still obscure. In the present study, we have investi-gated the effect of edaravone on I/R injury-induced neuronaldamage in aged rats, with a focus on the MAPK familymembers c-Jun NH2-terminal kinase (JNK), extracellularregulated kinase 1/2 (ERK1/2) and p38 MAPK, and c-Jun,which is the key transcription factor of the activator protein(AP)-1 complex.

MATERIALS AND METHODS

Animals Male Sprague–Dawley rats (Charles RiverJapan Inc., Kanagawa, Japan) of about 24 months of age,weighing 500—540 g, were used in the present study. Experi-ments were performed in strict accordance with the guide-lines for animal experimentation of our institute. The ratswere divided into three groups for the study: Sham group(without any treatment; n�6), Vehicle group (rats weretreated with saline after bilateral carotid artery occlusion(BCAO); n�6), and Edaravone group (edaravone was dis-solved in saline and administered intravenously at a dose of1.5 mg/kg at 5 min and 35 min post BCAO; n�6).

BCAO The experimental model was adopted from a pre-vious report.16) During the surgical procedure, the body tem-perature of the animal was maintained at 37.5 °C with a heat-

April 2006 713Biol. Pharm. Bull. 29(4) 713—718 (2006)

Edaravone Inhibits JNK-c-Jun Pathway and Restores Anti-oxidativeDefense after Ischemia-Reperfusion Injury in Aged Rats

Juan WEN,a Kenichi WATANABE,*,a Meilei MA,a Kenichi YAMAGUCHI,b Hitoshi TACHIKAWA,c

Makoto KODAMA,c and Yoshifusa AIZAWAc

a Department of Clinical Pharmacology, Niigata University of Pharmacy and Applied Life Sciences; 265–1 Higashijima,Niigata 956–8603, Japan: b Department of Homeostatic Regulation and Development, Niigata University School ofMedical and Dental Sciences; Niigata, Japan: and c First Department of Medicine, Niigata University School of Medicine;Niigata 951–8510, Japan. Received November 18, 2005; accepted December 27, 2005; published online January 25, 2006

Edaravone, a potent antioxidant, is currently being used in the management of acute ischemic stroke in rela-tively high-aged populations. Mitogen activated protein kinase (MAPK) pathways have been shown to play im-portant roles in neuronal cell death. We examined the role of MAPK pathways and the effect of treatment withedaravone in the brain after cerebral ischemia-reperfusion (I/R) injury in a bilateral carotid artery occlusion(BCAO) model with ischemia for 85 min followed by reperfusion for 45 min in aged rats. Western immunoblot-ting, immunostaining, enzyme-linked immunosorbent assay (ELISA), spectrophotometry, terminal deoxynu-cleotidyl transferase nick end labeling (TUNEL) and triphenyl tetrazolium chloride (TTC) staining were per-formed to evaluate various proteins in the homogenate, c-Jun NH2-terminal kinase (JNK) in the tissue sections,protein carbonyl, glutathione peroxidase (GSHPx), apoptosis and infarct size, respectively. Our results showedthat I/R injury resulted in a reduction of GSHPx, but protein carbonyl content and inducible nitric oxide syn-thase were increased. The activation of JNK and its downstream molecule c-Jun was significantly increased afterinjury, whereas the activities of p38 MAPK and extracellular-regulated kinase 1/2 were slightly but not signifi-cantly increased. Edaravone (3 mg/kg, i.v.) treatment significantly reduced all of these changes. Our findings sug-gest that the JNK pathway differentially mediates neuronal injury in aged rats after BCAO, and edaravone treat-ment significantly reduces the neuronal damage after I/R injury by inhibiting oxidative stress and the JNK-c-Junpathway with concomitant inhibition of overall MAPK activity in the brains of aged rats.

Key words cerebral ischemia-reperfusion; edaravone; mitogen-activated protein kinase; aged rat; neuronal damage; oxidativestress

© 2006 Pharmaceutical Society of Japan∗ To whom correspondence should be addressed. e-mail: watanabe@niigata-pharm.ac.jp

ing pad (Model: TP-401, Gaymar Industries Inc., New York,NY, U.S.A). Anesthesia was maintained with pentobarbital ata dose of 30 mg/kg. Cerebral ischemia was produced by theocclusion of both the right and left common carotid arteriesexposed through a middle skin incision using aneurysm clipsfor 85 min; cerebral ischemia was confirmed by the change incolor of the eyeball from red to white. At the end of the is-chemic period, the carotid arteries were de-clamped to allowblood reperfusion for 45 min. All rats were decapitated underanesthesia and then the whole brain was excised and used foranalysis.

Measurement of Infarct Size After I/R injury, infarctsize was measured macroscopically as follows: The brain tis-sue was removed and sliced into five 2.0-mm-thick coronalsections. Brain slices were stained with 2% 2,3,5-triphenyltetrazolium chloride (TTC) in phosphate buffered saline(PBS) (8 g NaCl, 0.2 g KCl, 2.9 g Na2HPO4·12H2O, 0.2 gKH2PO4 in 1000 ml of distilled water, pH 7.4) at 37 °C for30 min and then fixed with 10% formalin in saline (pH 7.2)for 24 h.17) The viable cells were reddish brown in color,whereas the infarcted area was dull yellow in color. Based onthe color difference, the infarct area was identified and calcu-lated, and was expressed as a percentage. The investigatorwho evaluated the slides was blinded to the status of experi-mental groups.

Brain Tissue Homogenate The whole brain was re-moved, weighed, and immediately chilled in liquid nitrogen.Western blotting analysis was carried out using brain tissuesamples homogenized in lysis buffer [50 mM Tris–HCl (pH7.4), 200 mM NaCl, 20 mM NaF, 1.0 mM Na3VO4, and 1.0 mM

DTT]. Samples were stored at �80 °C until use. For assess-ment of GSHPx activity and protein carbonyl content, braintissue was suspended in cold 0.05 M PBS containing 1.15%(w/v) KCl (9 ml per 1 g of wet tissue), after washing with0.85% NaCl, and then homogenized at 0 °C using a PolytronPT 10-35 homogenizer (Kinematica AG, Littau, Switzer-land). Total protein concentration of samples was measuredusing the bicinchoninic acid (BCA) protein assay usingbovine serum albumin (BSA) as standard.

GSHPx Activity Measurement GSHPx activity was es-timated according to the method of Wendel.18) Briefly, analiquot of brain homogenate (0.4 mg of protein) in 0.05 M

PBS containing 1.15% (w/v) KCl was mixed in a microplatewith 230 m l of coupling solution (containing 33.6 mg of di-sodium EDTA, 6.5 mg of NaN3, 30.7 mg of reduced glu-tathione, 16.7 mg of NADPH and 100 units of glutathione re-ductase in 100 ml of 50 mM Tris–HCl pH 7.6). The volumethen was adjusted to 260 m l with 0.05 M PBS. Kinetic decayof NADPH fluorescence (Ex. 355 nm/Em. 465 nm) wasmeasured after the addition of 40 m l of 1 mM H2O2 as the sub-strate using a micro plate spectrophotometer (Labsystem Flu-oroskan Ascent CF, Osaka, Japan).

Protein Carbonyl Measurement The protein carbonylin brain tissues homogenate was measured by the enzyme-linked immunosorbent assay (ELISA) method.19) Briefly,BSA oxidized by CuSO4/H2O2 (3 mM/5 mM) was used as theprotein carbonyl standard. The carbonyl content of standardoxidized BSA was determined by a colorimetric method re-ported previously.20) The tissue homogenate was centrifugedat 3500 rpm for 10 min to remove the debris. Then the super-natant was diluted 2 fold with PBS and incubated with 10%

streptomycin sulfate (9 : 1, v/w) for 15 min at 4 °C. After cen-trifugation at 10000 rpm for 10 min, the protein concentra-tion in the supernatant was measured by the BCA methodusing BSA as a standard and then the protein concentrationwas adjusted to 1 mg/ml with PBS. The samples (500 m l)were reacted with 10 mM 2,4-dinitrophenylhydrazine (DNPH)in 2.5 M HCl (100 m l) at room temperature for 1 h. Proteins insamples and oxidized BSA were precipitated with 20%trichloroacetic acid (TCA, 500 m l) and the protein concentra-tion was measured again and adjusted to 4 mg/100 m l. Thestandard curve for ELISA was prepared using oxidized BSAdiluted with 40 mg/ml BSA at a defined ratio (0—40%).Aliquots (100 m l) of test samples and standards (4 mg of pro-tein) were loaded into a 96-well immunoplate and incubatedovernight at 4 °C. The plate was washed with PBS containing0.1% Tween 20 (PBST) and then incubated with blockingbuffer (1% BSA in PBST) for 2 h at room temperature. Thesamples were further incubated with primary antibody i.e.,mouse anti-dinitrophenyl (DNP) IgE, (Sigma, St. Louis, MO,U.S.A.) for 4 h at 37 °C, washed with PBST and then incu-bated with secondary antibody, i.e., rat anti-mouse IgE,(Southern Biotechnology Associates Inc., Birmingham, AL,U.S.A.) for 1 h at 37 °C. The peroxidase reaction was per-formed by the addition of 100 m l of 3,3�,5,5�tetramethyl ben-zidine (Sigma, St. Louis, MO, U.S.A) and stopped by adding100 m l of 0.18 M H2SO4. The absorbance was measured at450 nm using a micro plate reader (Model 550, Bio-Rad, CA,U.S.A.).

Protein Analysis by Western Blotting In order to studythe role of the MAPK pathway and inducible nitric oxidesynthase (iNOS) in global cerebral I/R injury, Western blot-ting of rat brain homogenates was performed with antibodieshighly specific for the dual phosphorylated active forms ofJNK, ERK, p38 MAPK and c-Jun, and iNOS. MAPK activa-tion was quantified by normalizing the phospho-MAPK ex-pression level with the total MAPK expression in the samesamples, and iNOS expression was normalized with theGAPDH expression in the same sample. The MAPK activa-tion in the Sham group was taken as 1 arbitrary unit. Proteinswere separated by SDS-PAGE and electrophoretically trans-ferred to nitrocellulose filters. The filters were blocked with5% non-fat dried milk in TBS-T (20 mM Tris, pH 7.6,137 mM NaCl, and 0.5% Tween 20) for 1 h at room tempera-ture. Anti-phospho JNK rabbit polyclonal antibody, anti-JNKrabbit polyclonal antibody, anti-phospho ERK1/2 mousemonoclonal antibody, anti-ERK1/2 rabbit polyclonal anti-body, anti-phospho c-Jun rabbit polyclonal antibody, anti-c-Jun rabbit polyclonal antibody, anti-phospho p38 MAPK rab-bit polyclonal antibody, anti-p38 MAPK rabbit polyclonalantibody (Cell Signaling Technology, Beverly, MA, U.S.A.),anti-iNOS mouse monoclonal antibody (Sigma), and anti-GAPDH goat polyclonal antibody (Santa Cruz Biotechnol-ogy, Santa Cruz, CA, U.S.A.) were used at a dilution of1 : 1000. After incubation with primary antibody, bound anti-body was visualized with the respective horseradish peroxi-dase (HRP)-coupled secondary antibody (Santa Cruz) andchemiluminescence developing agents (ECL Plus, Amer-sham, Piscataway, NJ, U.S.A.).

Histopathology a) Hematoxylin and Eosin (H-E) Stain-ing: After 85 min of ischemia and 45 min of reperfusion, thebrains were removed and fixed in 10% formalin. The brains

714 Vol. 29, No. 4

were then embedded in paraffin and representative coronalsections (6 mm thick), which included the dorsal hippocam-pus, were obtained using a rotary microtome. Tissue sectionswere stained with H-E.

b) Terminal Deoxynucleotidyl Transferase Nick End La-beling (TUNEL) Assay: Paraffin-embedded sections of braintissue were deparaffinized and dehydrated in a descending al-cohol series and then incubated in a 20 mg/ml solution ofproteinase K, washed with PBS, incubated with 3% H2O2 for5 min, and washed with PBS again. TUNEL staining wasperformed as specified in a TUNEL kit (Takara Bio, Shiga,Japan). Sections were mounted and examined using light mi-croscopy. For each animal, five sections were scored region-ally for TUNEL-positive cells located in both the cerebralhemispheres.

c) Immunohistochemistry for Phospho-JNK: Formalin-fixed, paraffin-embedded tissue sections were used for im-munohistochemical staining. After deparaffinization and de-hydration, the slides were washed in Tris-buffered saline(TBS; 10 mmol/l Tris HCl, 0.85% NaCl, pH 7.5) containing0.1% BSA. Endogenous peroxidase activity was quenched byincubating the slides in a solution of 0.6% H2O2 in methanol.To perform antigen retrieval, the sections were pretreatedwith trypsin for 15 min at 37 °C. Blocking was done withnormal serum. After overnight incubation with primary anti-body (phospho-JNK rabbit polyclonal antibody) at 4 °C, theslides were washed in TBS buffer, and then HRP conjugatedanti-rabbit secondary antibody (Santa Cruz) was added andthe slides were incubated at room temperature for 50 min.The slides were washed in TBS buffer and incubated with di-amino benzidine tetrahydrochlioride (DAB) as the substrate,and counterstained with hematoxylin. A negative controlwithout primary antibody was included in the experiment toverify the antibody specificity.

Statistical Analysis Data are presented as means�stan-dard error of the mean (S.E.M.). Statistical analysis of differ-ences between the groups was performed by one-way analy-sis of variance followed by Tukey’s method. Correlations be-tween groups of values were evaluated by calculating the bestfit based on least-squares regression analysis and the coeffi-cient of correlation (r) was determined. Differences wereconsidered significant at p�0.05.

RESULTS

Oxidative Stress after BCAO The protective effect ofedaravone against cerebral oxidative damage induced by I/Rin rats was studied by assessing the oxidative inactivation ofGSHPx enzyme. I/R injury resulted in a reduction of GSHPxactivity by 37�2.6% compared to that in the untreated Shamgroup. The loss of GSHPx activity due to I/R was almostcompletely prevented by treatment with edaravone (Fig. 1A).Oxidative modification of proteins is accompanied by thegeneration of protein carbonyl derivatives. Protein carbonyl,a marker of protein oxidation in brain homogenates, in-creased to 247�5% in our model of I/R injury after BCAOand dropped to 110.6�4.5% in animals treated with edar-avone at a dose of 3 mg/kg compared to the level in the Vehi-cle group (Fig. 1B). By measuring the expression of iNOS,an enzyme responsible for the formation of NO, the ability toproduce NO was evaluated. Western blotting analysis wascarried out using an antibody against iNOS. Densitometricanalysis of iNOS normalized by GAPDH revealed a 2.96-fold increase in the expression of iNOS after I/R injury, andthis increase was significantly attenuated to 2.21-fold in thegroup treated with edaravone (Fig. 1C).

Activation of MAPK Pathways We studied the activa-tion of JNK, c-Jun, ERK and p38 MAPK after BCAO for85 min followed by 45 min of reperfusion in rats. I/R injurysignificantly induced the activation of JNK and c-Jun in theVehicle group (Figs. 2A, B). JNK activation was increasedby 2.22�0.14 fold in the Vehicle group as compared to theSham group ( p�0.01), but only by 1.29�0.07 fold in theEdaravone group (Fig. 2A). c-Jun activation was significantlyincreased by 2.01�0.04 fold in the Vehicle group, but onlyby 1.19�0.05 fold in the Edaravone group (Fig. 2B). ERK1/2 activation increased by 1.3�0.03 fold in the Vehiclegroup compared to the Sham group, but this effect was in-significant. In the Edaravone group, the activation of ERK1/2activity was only 1.18�0.03 fold (Fig. 2C). The activation ofp38 MAPK increased by 1.27�0.05 fold in the Vehiclegroup as compared to the Sham group and by 1.04�0.04fold in the Edaravone group (Fig. 2D). As shown in Figs. 2Eand F, the immunostaining for phospho-JNK in the hip-pocampus CA1 and cortex region was decreased in the Edar-avone group compared to the Vehicle group.

April 2006 715

Fig. 1. Effect of Edaravone on Glutathione Peroxidase (GSHPx) Activity, Protein Carbonyl Content and iNOS Expression after I/R Injury

A significant increase in GSHPx activity (A), and decreases in protein carbonyl content (B) and iNOS expression (C) were seen in the Edaravone group compared to the Vehiclegroup. Edaravone was administered intravenously at a dose of 3 mg/kg. Representative Western blots showing iNOS expression in the brain tissue homogenate are shown. Densito-metric analysis of immunoreactive bands normalized by GAPDH as an internal control was performed using scion image software. Triplicate determinations were done on thebrain homogenate prepared from each rat. The data are shown as the mean signal intensity�S.E.M. of five animals per group. ∗ p�0.05, ∗∗ p�0.01 vs. Sham group; # p�0.05, ## p�0.01 vs. Vehicle group.

Histological Findings of Neuronal Damage after BCAOCharacteristic histological features were seen in the rat brainafter 85 min of ischemia and 45 min of reperfusion. Histo-pathologic changes were obvious in all ischemia-lesioned an-imals (Fig. 3). We observed a significant increase in TUNEL-positive cells in the Vehicle group (0.51�0.03%; a 3.2-foldincrease) after I/R injury, and this increase was inhibited inthe Edaravone group (0.22�0.02%) (Fig. 4A).

In addition, edaravone was found to affect infarct size, asshown in Fig. 4B. Edaravone caused an 8% decrease in in-farct size as compared to that in the Vehicle group ( p�0.01).

DISCUSSION

The incidence of stroke rises markedly with age. However,limited studies have addressed this issue in experimental ani-mal models of stroke. Ischemic injury may be more lethal inolder individuals, since the susceptibility of neurons to is-chemia increases with age.1,2) We found that the administra-tion of edaravone immediately after prolonged ischemia for85 min was effective in preventing the neuronal damage ob-served in our model in aged rats. Early initiation of the treat-ment and re-administration of the drug after 30 min provedbeneficial in the present study. Edaravone protects against I/Rinjury within a therapeutic time window,21) and edaravonetreatment initiated early in patients with acute cerebral in-farction and neurological deficits also leads to marked im-provement and recovery22) and hence the protocol used heremay simulate the actual clinical situation.

We observed a significant reduction in infarct size after

edaravone treatment. Edaravone at a dose of 3 mg/kg hasbeen reported to reduce infarct size in a model of focal is-chemia in adult rats7,23) and to decrease the hydroxyl radicallevel at the ischemic border zone.24) Moreover, Mizuno et al.used the same dose administered twice in rats with middlecerebral artery occlusion, which resulted in a significant de-crease in cerebral damage and the hydroxyl radical level.24)

Furthermore, we demonstrated the effect of edaravone on theoverall oxidative status after I/R injury. GSHPx is an antioxi-dant enzyme which facilitates the breakdown of superoxideto water and oxygen. GSHPx-1 knockout mice show in-creased infarct size with exacerbated apoptosis as comparedto wild-type mice.25) On the other hand, mice overexpressinghuman GSHPx-1 isoform exhibit a reduction in infarct sizeby 48% after I/R injury.26) In the present study, the activity ofGSHPx was significantly decreased after I/R injury, and thisreduction was significantly attenuated by edaravone treat-ment. Our results are in accordance with the findings ofIchikawa and Konishi,16) and Xuejiang et al.6) NO, anotherpotentially damaging free radical, mediates ischemic neu-ronal cell death.10) The level of iNOS expression in the Vehi-cle group was 2.96-fold that in the Sham group and was de-creased significantly in the Edaravone group, as observed byOtani et al.21) Protein carbonyl derivatives, a marker of ox-idative injury, have been reported to increase after ischemicinsult to the brain.3) A substantial upsurge in protein oxida-tion was observed in the present study, as indicated by an in-crease in the protein carbonyl content (2.47-fold) in the Vehi-cle group as compared to the Sham group. Edaravone treat-ment resulted in an overall decrease in the reactive oxygen

716 Vol. 29, No. 4

Fig. 2. Effect of Edaravone on MAPK Activity: Representative Western Blots Showing Significantly Decreased Expression of JNK (A) and Its Down-stream Transcription Factor c-Jun (B) in the Edaravone Group Compared to the Vehicle Group

ERK1/2 (C) and p38 MAPK (D) activation was reduced to the normal level in the Edaravone group. Densitometric analysis of immunoreactive bands was performed using scionimage software. Triplicate determinations were done on the whole brain homogenate prepared from each rat. The data are shown as the mean signal intensity�S.E.M. of five ani-mals per group. ∗∗ p�0.01 vs. Sham group; # p�0.05, vs. Vehicle group. Decreased immunostaining for phospho-JNK was found in the hippocampus CA1 (E) and cortex region(F) of the Edaravone group compared to the Vehicle group. (Magnification: �100; inset �400)

and nitrogen species, and in protein oxidation in our model.In contrast, another report by Nagafuji et al. demonstrated alack of iNOS participation in the ischemic brain damage,27)

and thus these results should be interpreted cautiously.Oxidative stress can trigger the activation of multiple sig-

naling pathways which influence neuronal toxicity, includingpathway involving increased phosphorylation of MAPKs28—30)

JNK and p38 MAPK, which are activated by environmentalstress and are closely associated with cell death, whereasERK 1/2 is activated by growth factors and has an opposingeffect on cell death. Several in vivo studies have demon-strated that the JNK, ERK1/2 and p38 MAPK pathways areactivated after global ischemia in the hippocampus and cor-

tex.9,28) In the present study, we found that oxidative stresscaused by BCAO for a period of 85 min followed by 45 minof reperfusion resulted in a marked increase in the activationof JNK (2.22 fold, p�0.01). Furthermore, immunostainingfor phospho-JNK showed that it was increased in the Vehiclegroup in the hippocampus and cortex, and that this increasewas attenuated in the Edaravone group, indicating that theneuronal injury occurring in our model is JNK-mediated.Slight increases in the activation of both ERK1/2 and p38MAPK were observed in the Vehicle group compared to theSham group, but these increases were not significant. TheMAPK pathways are complex and are not yet fully under-stood: different cell types respond to the same stress in dif-ferent ways, and different types of stress in the same cell ac-tivate different signaling pathways.31) In addition, the activa-tion of MAPKs varies with the time of ischemic injury andshows a differential pattern: a brief period of global cerebralischemia results in peak JNK activity after 30 min, but not asmarked activation of the other MAPK family members.9,29)

Moreover, evidence for the role of JNK is relatively strongerthan that for p38 MAPK with respect to neuronal damage31)

which supports the findings of the present study. On the otherhand, the role of ERK1/2 in brain I/R injury is contro-versial.l9,30,32) The activation of MAPK and the degree of itsactivation vary depending upon the type, degree and time ofthe stimulus. Presently, the molecular mechanisms of the sig-nificant activation of JNK and the moderate activation ofother MAPKs are not known. Further studies will reveal thebasis of the activation of MAPK kinases at different timesafter I/R, and the mechanism of the inhibition by edaravoneof the MAPK pathways.

Among the various substrates phosphorylated by JNK thatcomprise the AP-1 transcriptional activator complex, c-Jun isa crucial component and has been shown to be a point of in-tegration of numerous signals that can differentially affect itsexpression as well as its transcriptional activity.31) JNK activ-ity is essential for the basal level of c-Jun expression and forc-Jun phosphorylation in response to stress.33) In the presentstudy, the expression of phospho c-Jun markedly increased inthe Vehicle group, and this increase was prevented by edar-avone treatment, indicating the involvement of c-Jun in thedevelopment of neuronal damage in the present study. Previ-ously, N-acetyl cysteine, an antioxidant, was shown to signif-icantly inhibit the JNK pathway and thereby to contribute tothe protection against cerebral ischemic injury in rats34,35);however, it is not yet clear whether all antioxidants would actvia the same mechanism. Very recently, activation of theJNK-c-Jun pathway was demonstrated to contribute to early

April 2006 717

Fig. 4. Quantification and Total TUNEL-Positive Cells in Sham Group, Vehicle Group and Edaravone Group (A)

TUNEL-positive staining was observed specifically in the hippocampus, cortex, thalamus and dentate gyrus regions. Staining was particularly strong in the cortex and hippocam-pus (CA1) in the Vehicle group compared to the Edaravone group. A significant reduction in infarct size was observed in the Edaravone group compared to the Vehicle group (B).∗∗ p�0.01 vs. Sham group, ## p�0.01 vs. Vehicle group.

Fig. 3. Representative Photomicrographs of Rat Brain Sections withHematoxylin–Eosin Staining and TUNEL-Positive Insets: Pathologic Fea-tures of Brain Sections after 85 min of BCAO Followed by 45 min of Reper-fusion

Pyknotic nuclei of damaged neuronal cells and protection after edaravone treatmentare obvious. Neuronal perikarya showing characteristic morphologic features of shrink-age, triangulation of the nucleus and cytoplasm, and increased eosinophilia in the Vehi-cle group (left panels) and preserved morphology in the Edaravone group (right pan-els). [H-E staining, �100, TUNEL staining, �400 (inset)]

phase apoptosis36); in that study, pharmacological inhibitionof JNK and overexpression of a dominant-negative form ofc-Jun resulted in reduced neuronal damage, whereas p38MAPK inhibitor failed to do so. Also, treatment with a spe-cific inhibitor of JNK diminished the phosphorylation of c-Jun and thus provided protection from neuronal cell death.37)

A neuroprotective effect of SP600125, a new inhibitor ofJNK, against cerebral ischemia reperfusion injury wasdemonstrated in rats,38) and hence these kinds of inhibitorsare important as future therapeutic options.

Global cerebral ischemia causes neuronal damage in re-gions that are specifically vulnerable to ischemia, such as thecerebral cortex and hippocampus.28) In the present study, weobserved a 3-fold increase in total TUNEL-positive cellsafter 45 min of reperfusion following ischemic insult, prima-rily in the hippocampus, cortex, and thalamus regions of thebrain, and this increase was significantly reduced in the Edar-avone group. However, we consider that TUNEL staining de-tects both apoptotic as well as necrotic phenomena of celldeath.39) Increased sensitivity of aged neurons to ischemia,40)

decreased thresholds of stress-activated kinases41) and longerischemic insult per se may be responsible for the 3-fold in-crease in total TUNEL-positive cells observed in the presentstudy. Moreover, a corresponding increase in infarct size andphospho-JNK immunostaining was also observed in both thehippocampus and cortex of the Vehicle group, which corre-sponded with the altered morphology observed after H-Estaining, indicating obvious neuronal damage in coronal sec-tions of aged rats in our model.

It may thus be concluded that edaravone significantly re-duces infarct size, attenuates oxidative stress and preventsneuronal damage by inhibiting the activation of the JNK-c-Jun pathway in aged rats.

Limitations This study lacks information regardingwhether all three MAPK pathways were activated in the in-jury-prone areas versus the whole brain, and experimentationwith specific pharmacological inhibition of JNK may providefurther clarification of this issue in our animal model.

Acknowledgements This research was supported byYujin Memorial Grant; the Ministry of Education, Culture,Sports, Science and Technology of Japan; and Promotion andMutual Aid Corporation for Private Schools of Japan. Wethank Prof. T. Konishi and Dr. H. Ichikawa for helpful ad-vice. We also thank K. Hirabayashi, K. Shirai, Y. Nagai, M.Wahed, G. Narasimman, P. Suresh, P. Prakash, K. Fadia, R.A.Elbarbary, M. Soga, S. Mito, V. Punniyakoti and M. Kunizakifor assistance in this work.

REFERENCES

1) Kaarisalo M. M., Immonen-Raiha P., Marttila R. J., Lehtonen A.,Torppa J., Tuomilehto J., Arch. Gerontol. Geriatr., 31, 43—53 (2000).

2) Simons L. A., McCallum J., Friedlander Y., Simons J., Stroke, 29,1341—1346 (1998).

3) Sahin E., Gumuslu S., Behav. Brain Res., 155, 241—248 (2004).4) Weisbrot-Lefkowitz M., Reuhl K., Perry B., Chan P. H., Inouye M.,

Mirochnitchenko O., Brain Res. Mol. Brain Res., 53, 333—338(1998).

5) Islekel S., Islekel H., Guner G., Ozdamar N., Res. Exp. Med. (Berl).,199, 167—176 (1999).

6) Xuejiang W., Magara T., Konishi T., Free Radic. Res., 31, 449—455(1999).

7) Abe K., Yuki S., Kogure K., Stroke, 19, 480—485 (1988).8) Nitatori T., Sato N., Waguri S., Karasawa Y., Araki H., Shibanai K.,

Kominami E., Uchiyama Y., J. Neurosci., 15, 1001—1011 (1995).9) TakagiY., Nozaki K., Sugino T., Hattori I., Hashimoto N., Neurosci.

Lett., 294, 117—120 (2000).10) Park S. Y., Lee H., Hur J., Kim S. Y., Kim H., Park J. H., Cha S., Kang

S. S., Cho G. J., Choi W. S., Suk K., Brain Res. Mol. Brain Res., 107,9—16 (2002).

11) Irie H., Kato T., Ikebe K., Tsuchida T., Oniki Y., Takagi K., J. Surg.Res., 120, 312—319 (2004).

12) Zhang W., Sato K., Hayashi T., Omori N., Nagano I., Kato S., HoriuchiS., Abe K., Neurol. Res., 26, 342—348 (2004).

13) Satoh K., Ikeda Y., Shioda S., Tobe T., Yoshikawa T., Redox. Rep., 7,219—222 (2002).

14) Anzai K., Furuse M., Yoshida A., Matsuyama A., Moritake T., TsuboiK., Ikota N., J. Radiat. Res. (Tokyo), 45, 319—323 (2004).

15) Yasuoka N., Nakajima W., Ishida A., Takada G., Brain Res. Dev. BrainRes., 151, 129—139 (2004).

16) Ichikawa H., Konishi T., Biol. Pharm. Bull., 25, 898—903 (2002).17) Chopra K., Singh M., Kaul N., Andrabi K. I., Ganguly N. K., Mol.

Cell Biochem., 113, 71—76 (1992).18) Wendel A., Methods Enzymol., 77, 325—333 (1981).19) Buss H., Chan T. P., Sluis K. B., Domigan N. M., Winterbourn C. C.,

Free Radic. Biol. Med., 23, 361—366 (1997).20) Reznick A. Z., Packer L., Methods Enzymol., 233, 357—363 (1994).21) Otani H., Togashi H., Jesmin S., Sakuma I., Yamaguchi T., Matsumoto

M., Kakehata H., Yoshioka M., Eur. J. Pharmacol., 512, 129—137(2005).

22) Edaravone Acute Infarction Study Group, Cerebrovasc. Dis., 15,222—229 (2003).

23) Kawai H., Nakai H., Suga M., Yuki S., Watanabe T., Saito K. I., J.Pharmacol. Exp. Ther., 281, 921—927 (1997).

24) Mizuno A., Umemura K., Nakashima M., Gen. Pharmacol., 30, 575—578 (1998).

25) Crack P. J., Taylor J. M., Flentjar N. J., de Haan J., Hertzog P., IannelloR. C., Kola I., J. Neurochem., 78, 1389—1399 (2001).

26) Hoehn B., Yenari M. A., Sapolsky R. M., Steinberg G. K., Stroke, 34,2489—2494 (2003).

27) Nagafuji T., Suzuki M., Hosoi Y., Nippon Yakurigaku Zasshi, 111,45—53 (1998).

28) Borsello T., Clarke P. G., Hirt L., Vercelli A., Repici M., Schorderet D.F., Bogousslavsky J., Bonny C., Nat. Med., 9, 1180—1186 (2003).

29) Sugino T., Nozaki K., Takagi Y., Hattori I., Hashimoto N., MoriguchiT., Nishida E., J. Neurosci., 20, 4506—4514 (2000).

30) Harper S. J., Lo Grasso P., Cell Signal, 13, 299—310 (2001).31) Hu B. R., Liu C. L., Park D. J., J. Cereb. Blood Flow Metab., 20,

1089—1095 (2000).32) Noshita N., Sugawara T., Hayashi T., Lewen A., Omar G., Chan P. H.,

J. Neurosci., 22, 7923—7930 (2002).33) Kayahara M., Wang X., Tournier C., Mol. Cell. Biol., 25, 3784—3792

(2005).34) Zhang Q., Zhang G., Meng F., Tian H., Neurosci. Lett., 337, 51—55

(2003).35) Li C. H., Wang R. M., Zhang Q. G., Zhang G. Y., J. Neurochem., 93,

290—298 (2005).36) Carimalo J., Cronier S., Petit G., Peyrin J. M., Boukhtouche F., Arbez

N., Lemaigre-Dubreuil Y., Brugg B., Miquel M. C., Eur. J. Neurosci.,21, 2311—2319 (2005).

37) Gao Y., Signore A. P., Yin W., Cao G., Yin X. M., Sun F., Luo Y., Gra-ham S. H., Chen J., J. Cereb. Blood Flow Metab., 25, 694—712(2005).

38) Guan Q. H., Pei D. S., Zhang Q. G., Hao Z. B., Xu T. L., Zhang G. Y.,Brain Res., 1035, 51—59 (2005).

39) De T. C., Munell F., Ferrer I., Reventos J., Macaya A., Neurosci. Lett.,230, 1—4 (1997).

40) Dorszewska J., Adamczewska-Goncerzewicz, Z., Respir. Physiol. Neu-robiol., 139, 227—236 (2004).

41) O’Donnell E., Vereker E., Lynch M. A., Eur. J. Neurosci., 12, 345—352 (2000).

718 Vol. 29, No. 4