IGF-I neuroprotection in the immature brain afterhypoxia-ischemia, involvement of Akt and GSK3b?

Katarina G. Brywe,1,2 Carina Mallard,2 Malin Gustavsson,2 Maj Hedtjarn,2 Anna-Lena Leverin,2 Xiaoyang Wang,2

Klas Blomgren,2,4 Jorgen Isgaard3 and Henrik Hagberg1,21Department of Obstetrics and Gynecology,2Department of Physiology and Pharmacology, Perinatal Center, and3Research Center for Endocrinology and Metabolism, Sahlgrenska Academy, Gothenburg, Sweden4Department of Pediatrics, the Queen Silvia Children’s Hospital, Gothenburg, Sweden

Keywords: apoptosis, cell survival, neonatal, rat brain, trophic factor

Abstract

Insulin-like growth factor I (IGF-I) is a neurotrophic factor that promotes neuronal growth, differentiation and survival. Neuroprotectiveeffects of IGF-I have previously been shown in adult and juvenile rat models of brain injury. We wanted to investigate theneuroprotective effect of IGF-I after hypoxia-ischemia (HI) in 7-day-old neonatal rats and the mechanisms of IGF-I actions in vivo. Wealso wanted to study effects of HI and ⁄ or IGF-I on the serine ⁄ threonine kinases Akt and glycogen synthase kinase 3b (GSK3b) in thephophatidylinositol-3 kinase (PI3K) pathway. Immediately after HI, phosphorylated Akt (pAkt) and phosphorylated GSK3b (pGSK3b)immunoreactivity was lost in the ipsilateral and reduced in the contralateral hemisphere. After 45 min, pAkt levels were restored tocontrol values, whereas pGSK3b remained low 4 h after HI. Administration of IGF-I (50 lg i.c.v.) after HI resulted in a 40% reductionin brain damage (loss of microtubule-associated protein) compared with vehicle-treated animals. IGF-I treatment without HI wasshown to increase pAkt whereas pGSK3b decreased in the cytosol, but increased in the nuclear fraction. IGF-I treatment after HIincreased pAkt in the cytosol and pGSK3b in both the cytosol and the nuclear fraction in the ipsilateral hemisphere compared withvehicle-treated rats, concomitant with a reduced caspase-3- and caspase-9-like activity. In conclusion, IGF-I induces activation of Aktduring recovery after HI which, in combination with inactivation of GSK3b, may explain the attenuated activation of caspases andreduction of injury in the immature brain.

Introduction

Insulin-like growth factor I (IGF-I) has an important role in normalbrain development, promoting neuronal growth, cellular proliferationand differentiation (Sara et al., 1993; Ye et al., 1995; D’Ercole et al.,1996; Cheng et al., 2003). IGF-I has previously been shown to beneuroprotective in both adult and immature animal models of cerebralischemia (Guan et al., 1993; Johnston et al., 1996). Several in vitrostudies support a role for IGF-I in cell survival and prevention ofapoptosis (Russell et al., 1998; Delaney et al., 1999; Takadera et al.,1999; Van Golen & Feldman, 2000). Survival-promoting effects areproposed to be elicited by activation of intracellular signallingcascades such as phophatidylinositol-3 kinase (PI3K) pathways(Feldman et al., 1997; Parrizas et al., 1997; Russell et al., 1998;Peruzzi et al., 1999; Zheng et al., 2000). Activation of PI3K leads tophosphorylation and activation of Akt (pAkt) which can inhibit theproapoptotic glycogen synthase kinase 3b (GSK3b) (Dudek et al.,1997; Hetman et al., 2000), caspase 9, Bad and Forkhead transcriptionfactor 1 (Datta et al., 1997; del Peso et al., 1997; Cardone et al., 1998;Brunet et al., 1999; Zheng et al., 2002). Changes in Akt phosphory-lation have been reported after cerebral ischemia in different animalmodels in the adult brain showing a temporal increase of pAkt during

reperfusion (Ouyang et al., 1999; Noshita et al., 2001) or duringpermanent ischemia (Tomimatsu et al., 2001). However, others reporta decrease during post-ischemic reperfusion in the CNS (Ouyanget al., 1999; Tomimatsu et al., 2001; Yoshimoto et al., 2001). Indeed,some studies suggest that inactivation of GSK3b is a principalregulatory target of the PI3K survival pathway (Pap & Cooper, 1998).GSK3b activity is reduced by phosphorylation at Ser9 via Akt. GSK3bactivity can, however, be regulated in several other ways as theactivity is facilitated by phosphorylation of Tyr216, whereas little isknown about possible kinases involved in this process. As for manyother kinases, scaffolding proteins and subcellular redistribution mayrestrict interactions with GSK3b (Jope & Johnson, 2004). Themechanisms for induction of apoptosis via GSK3b are not fullyclarified, but it causes degradation of b-catenin and inhibits mito-chondrial pyruvate dehydrogenase (Hoshi et al., 1996; Zhang et al.,1998). GSK3b may also induce caspase activation either via theintrinsic pathway or via down-regulation of heat shock proteins(McLaughlin et al., 2003).Effects on the phosphorylation status of GSK3b at Ser9 after hypoxia-

ischemia (HI) have not been reported before. Furthermore, IGF-I andIGF-I-related peptides represent one of the most promising targets forneuroprotection at present but their mechanisms of action in theimmature brain in vivo are largely unknown. We therefore investigatedthe effect of IGF-I on brain injury in a model of focal HI in the 7-day-oldrat including effects on caspase activation, and part of the PI3K pathway(including pGSK3b) during control conditions and after HI.

Correspondence: Dr K. G. Brywe, 2Department of Physiology and Pharmacology, asabove.E-mail: katarina.g.brywe@medfak.gu.se

Received 12 January 2004, revised 14 December 2004, accepted 22 December 2004

European Journal of Neuroscience, Vol. 21, pp. 1489–1502, 2005 ª Federation of European Neuroscience Societies

doi:10.1111/j.1460-9568.2005.03982.x

Materials and methods

Animal model

Induction of unilateral HI in a neonatal model that has previouslybeen described was used (Rice et al., 1981; Hagberg et al., 1997).Seven-day-old (PND7) outbred Wistar rats were anesthetized withenflurane (3.5% for induction and 1.5% for maintenance) in a mixtureof nitrous oxide and oxygen on a snout mask. The left commoncarotid artery was cut between two prolene sutures (6-0). Afteranesthesia and surgery the animals were allowed to recover for60 min. They were thereafter exposed to 60 min of hypoxia in ahumidified chamber at 36 �C with 7.77 ± 0.01% oxygen in nitrogen.After hypoxia the pups were returned to and kept with their dams untilthey were killed at times indicated below. Animals were anaesthetizedwith i.p. injections of Brietal and killed by decapitation or perfusionfixed. All animal studies described here were approved by the animalethics committee of Sahlgrenska Academy. All animal studiesdescribed in these protocols were approved by the local ethicalcommittee.The 7-day-old HI rat model has been well characterized and widely

used to study asphyxial brain damage in the neonatal brain. At this agethe rat brain corresponds to 34–36 weeks of gestation in the humanfetus or newborn infant (Vannucci et al., 1999). Tissue injury after HIconsists of a necrotic infarction and ⁄ or apoptotic neuronal cell loss incerebral cortex, hippocampus, striatum and thalamus of the ligatedhemisphere (Towfighi et al., 1991), which resembles the damage seenafter asphyxia in term infants (Volpe, 2001).

Intracerebral injection of IGF-I and vehicle

Animals received three daily intracerebral injections of 5 lL solutioncontaining 5, 25 or 50 lg human recombinant IGF-I (hrIGF-I) (1, 5,10 lg ⁄ lL) (Genentech, Inc.) or vehicle (NaCl 0.9 mg ⁄mL). Injec-tions were aimed at the left ventricle at 1.5 mm lateral to the sagittalsuture, 5.7 mm rostral to lambda and 2.8 mm deep to the skull surfacewith a 27-gauge needle. Animals were under anesthesia with enfluraneon a snout mask while injected with a Hamilton syringe connected toan MCA pump, 1 lL ⁄min.

Evaluation of brain damage

Area measurement

Intact neurons (dendrites and soma) express microtubule-associatedprotein 2 (MAP-2), and infarction in grey matter is associated with adistinct loss of MAP-2 immunoreactivity. MAP-2-positive areas in theipsi- and contralateral hemispheres were outlined by an observerblinded to the study groups and were then calculated using theOlympus Micro Image, analysis software version 4.0 (OlympusOptical, Tokyo, Japan). The proportion of the infarction wascalculated by subtracting the MAP-2-positive area of the ipsilateralhemisphere from the contralateral hemisphere and was expressed as apercentage of the contralateral hemisphere as previously described(Hedtjarn et al., 2002).

Neuropathological scoring

Regional brain injury was also evaluated by an observer blinded to thestudy groups using a neuropathological scoring system in whichcortex injury was graded from 0 to 4, with 0 being no observableinjury and 4 being confluent infarction encompassing most of thehemisphere. The damage in the hippocampus, thalamus and striatumwas assessed regarding both hypotrophy, 0–3, and observable cellinjury ⁄ infarction, 0–3, resulting in a scoring for each region of 0–6.

The total score, 0–22, was the sum score for all the four regions(Hedtjarn et al., 2002).

Study protocol I: effects of HI on pAKT and pGSK3b

After HI animals were killed at 0 min, 45 min, 2 h, 4 h, 8 h and 24 h(n ¼ 5 at each time point). Tissue samples from cortex were dissectedand homogenized and cytosolic fractions were prepared for Westernblot analysis of pAkt, total Akt, pGSK3b and total GSK3 as describedbelow. In addition, following HI, two animals at each time point of0 min, 2 h, 6 h, 24 h, 72 h and controls at PND7 were perfusion-fixedand paraffin sections prepared for immunohistochemical staining forpAkt and pGSK3b as described below. In order to study cell localizationof immunoreactivity, HI (0 min) and control (PND7) sections weredouble stained for immunoflourescence using pAkt or pGSK3b andNeuN antibodies followed by counterstaining with Hoechst.

Study protocol II: effect of IGF-I on brain injury

Immediately after HI, animals were treated with hrIGF-I (5 lg,n ¼ 19) or vehicle (n ¼ 17), rhIGF-I (25 lg, n ¼ 24) or vehicle(n ¼ 25), rhIGF-I (50 lg, n ¼ 24) or vehicle (n ¼ 25) i.c.v. and thenthis dose was repeated once daily for 3 days. At PND10 animals wereperfused, brains were removed from the skull and immersion-fixed at4 �C for 24 h, dehydrated, embedded in paraffin and cut into 5-lmcoronal sections at four anteroposterior levels from the anteriorstriatum to the posterior part of hippocampus. Adjacent sections werestained with thionin ⁄ acid fuschin (Mallard et al., 1993) and forimmunohistochemical MAP-2 (1 : 1000, 1 h) with mouse anti-MAP-2as described below (Gilland et al., 1998).Rectal temperature, which has been shown to correspond to the

brain core temperature (Thoresen et al., 1996), was measured in IGF-I(50 lg, n ¼ 13) and vehicle (n ¼ 13) treated rats at 1, 3, 6, 12, 24, 36,48 and 72 h after HI.

Study protocol III: IGF-I effects on Akt and GSK3bphosphorylation

After i.c.v. injection of hrIGF-I (50 lg) or vehicle without priorexposure to HI, animals were killed at 30 min, 2 h, 8 h and 24 h (n ¼ 5per time point). The cortex was dissected and immediately put into ice-cold homogenization buffer. Tissue samples were homogenized andsubcellular fractions were prepared for Western blot analysis of pAkt,total Akt, pGSK3b and total GSK3, as described below. Additionalanimals were killed 2 h after hrIGF-I (n ¼ 4) or vehicle (n ¼ 4)treatment and prepared for immunohistochemical analysis of pAkt andpGSKb as described below. Further animals were exposed to both HIand hrIGF-I (50 lg) treatment or vehicle and were killed 2 h after HIand the cortex was dissected and immediately put in ice-coldhomogenization buffer. Tissue samples were homogenized and sub-cellular fractions were prepared for Western blot analysis of pAkt, totalAkt, pGSK3b and total GSK3, as described below (n ¼ 7 per group).

Study protocol IV: IGF-I effects after HI on caspase-1, 3, 8and 9 activity

Animals were killed at 8 h (n ¼ 19) and 24 h (n ¼ 20) after HI(IGF-I 50 lg group n ¼ 19 and vehicle n ¼ 20) for measurementof caspase-1-, 3-, 8- and 9-like activity. Brains were separated inleft and right hemisphere and the cortex, were dissected and thetissue was prepared for caspase activity measurement as describedbelow.

1490 K. G. Brywe et al.

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502

Western blotting

Brains were rapidly harvested, and hemispheres were split anddissected. Each cortical sample was immediately homogenized in ice-cold homogenization buffer (15 mm Tris ⁄HCl, pH 7.6, containing3 mm EDTA, 1 mm MgCl2, 320 mm sucrose, 1 mm DTT) proteaseand phosphatase inhibitor cocktail was added to final concentrations of1 and 2%, respectively. The homogenates were centrifuged at 800 gfor 10 min at 4 �C for preparation of nuclear fractions. Supernatantswere centrifuged at 9200 g for 15 min at 4 �C for preparation ofcytosolic fractions. Protein content was quantified using the methodpresented by Whitaker & Granum (1980), adapted for microplates.Samples were mixed with an equal volume of 3· SDS-PAGE bufferand heated (96 �C) for 5 min. One sample containing 20 lg of proteinfrom the cytosolic fraction or 10 lg of the nuclear fraction was appliedin each well of a Novex (San Diego, CA, USA) precast 8–16%Tris ⁄ glycine gel. After electrophoresis proteins were transferred to anitrocellulose membrane (Optitran, 0.2 lm; Schleicher & Schuell,Dassel, Germany). Membranes were blocked in 30 mm Tris ⁄HCl,pH 7, 5100 mm NaCl and 0.1% Tween 20 (TBS-T) containing 5% fat-free milk powder. Incubation with the following primary antibodiesdiluted in TBS-T containing 3% BSA and 9 mm NaN3 for 1 h at roomtemperature was performed: anti-phospho-Akt (Ser473) rabbit poly-clonal (1 : 1000), anti-Akt rabbit polyclonal (1 : 1000), anti-phospho-GSK3b (Ser9) rabbit polyclonal (1 : 1000), anti-GSK3 mousemonoclonal (1 : 500, Cell Signaling Technology, Beverly, MA,USA), anti-LaminB goat polyclonal (1 : 200, Santa Cruz Biotech-nology, Santa Cruz, CA, USA). Blots were washed three times withTBS-T and incubated with the appropriate secondary peroxidase-conjugated antibodies (Vector Laboratories, Burlingame, CA, USA)diluted in blocking buffer. Immunoreactive species were visualizedusing Super Signal Western Dura chemiluminescence substrates(Pierce, Rockford, IL, USA) and a cooled CCD camera (LAS1000;Fujifilm, Tokyo, Japan). Immunoreactive bands were quantified usingImage Gauge software (version 3.3, Fujifilm). The same five controlswere run on every gel to standardize the quantifications. Every samplewas analysed three times and the average value was used as n ¼ 1.Each protein band on the Western blots was derived from one animal.Each immunoblot was performed using the samples from 4–5 animalsin each experimental group. Membranes were stripped for reprobingwith new antibodies by having them incubated in stripping buffer(62.5 mm Tris ⁄HCl, pH 6.7, 100 mm b-mercaptoethanol and 2%SDS) at 55 �C for 30 min.

Fluorometric assay of caspase-1-, 3-, 8- and 9-like activity

The tissue samples were homogenized by sonication in ice-coldhomogenization buffer (50 mm TRIS, pH 7.3, containing 5 mm

EDTA). Lysates were centrifuged at 9000 g for 15 min at 4 �C.Samples were aliquoted and stored at )80 �C. The protein concen-tration was determined according to the method presented by Whitaker& Granum (1980). Samples of homogenate or cytosolic fractions weremixed with extraction buffer [50 mm Tris ⁄HCl, pH 7.3, 100 mm

NaCl, 5 mm EDTA, 1 mm EGTA, 3 mm NaN3, 1 mm PMSF,1 lg ⁄mL pepstatin, 2.5 lg ⁄mL leupeptin, 10 lg ⁄mL aprotinin,0.2% 3-[(3-cholamidoprophylene) dimethylamonio]propane sulphonicacid (CHAPS), protease inhibitor cocktail (P8340; Sigma, St Louis,MO, USA) 1%] 1 : 3, on a microtiter plate (Microflour, Dynatech, VA,USA). After incubation for 15 min at room temperature, 100 lL ofpeptide substrate, 50 lm Ac-Asp-Glu-Val-Asp-aminomethyl couma-rine (Ac-DEVD-AMC, for caspase-3-like activity; Enzyme SystemsProducts, Livermore, CA, USA) in extraction buffer without inhibitorsor CHAPS, but with 4 mm DTT added. Cleavage of the substrate was

measured at 37 �C using a Spectramax Gemini micro plate fluorometer(Molecular Devices, Sunnyvale, CA, USA), with an excitationwavelength of 380 nm and emission wavelength of 460 nm. Degra-dation was determined at 2-min intervals for 2 h, and Vmax wascalculated from the entire linear part of the curve. Vmax is the steadystate maximal rate of catalytic conversion of the peptide substrate in aspecific sample. There is a linear correlation between Vmax and thespecific activity (U ⁄mL), or ‘amount’, of the enzyme assayed.Standard curves with AMC in the appropriate buffer were used toexpress the data in pmoles of AMC formed per minute and permilligram of protein. Cleavage of Ac-YVAD-AMC (ALEXIS Bio-chemicals) for caspase-1-like activity was measured under the sameconditions. Cleavage of Ac-LEHD-AFC for caspase-9-like activityand Ac-IETD-AFC for caspase-8-like activity (all from EnzymeSystems Products, Livermore, CA, USA) were measured under thesame conditions as described above, but with an excitation wavelengthof 400 nm and emission wavelength of 505 nm for caspase-8 and -9,respectively. The activity was expressed as pmol AFC per milligram ofprotein per minute.

Immunohistochemistry of MAP-2, pAkt and pGSK3b

Brains were removed from the skull and immersion-fixed at 4 �C for24 h, dehydrated, embedded in paraffin and cut into 5-lm coronalsections at the hippocampal and striatum levels. The followingprimary antibodies, dilutions and incubation times were used: anti-phospho-Akt (Ser473) rabbit polyclonal 1 : 50 in PBS, 4 �C overnight(O ⁄N) (Cell Signaling Technology); anti-phospho-GSK3b (Ser9),rabbit polyclonal 1 : 50 in 1% BSA PBS, 4 �C O ⁄N (Cell SignalingTechnology); anti-MAP-2, mouse monoclonal 1 : 1000 in PBS, 4 �CO ⁄N (Sigma-Aldrich, Steinheim, Germany). Before immunohisto-chemical staining, sections were deparaffinized and boiled in citricacid buffer (0.01 m, pH 6.0, 10 min) and sections were treatedwith proteinase-K (10 lg ⁄mL; Boeringer Mannheim, Mannheim,Germany) in PBS for 9 min at room temperature. Non-specificbinding was blocked by incubation with appropriate serum (4%). Afterincubation with primary antibodies, sections were washed in PBS andincubated with biotinylated secondary antibodies (Vector Laborator-ies) for 1 h followed by inhibition of endogenous peroxidase (0.6%H2O2 in methanol, 10 min) and incubation with avidin–biotin enzymecomplex (20 lL ⁄mL, 1 h, ABC-Elite; Vector Laboratories). Immu-noreactivity was visualized using DAB (0.5 mg ⁄mL) enhanced withnickel sulphate (15 mg ⁄mL). The specificity of antibodies was testedby omission of the primary antibody.

Immunofluorescence

Sections were deparaffinized and antigen recovery performed as above.After blocking in 4% goat serum in PBS for 1 h at room temperature thesections were incubated with the following antibodies: anti-phospho-Akt (Ser473) rabbit polyclonal 1 : 50 in PBS, 4 �CO ⁄N (Cell SignalingTechnology); or anti-phospho-GSK3b (Ser9), rabbit polyclonal 1 : 50in 1% BSA PBS, 4 �C O ⁄N (Cell Signaling Technology). Afterwashing and with washing procedures in between, sections wereincubated for 1 h at room temperature for each of the following stepswith the subsequent antibodies: biotinylated secondary antibodies, goatanti-rabbit 1 : 250 in PBS (Vector Laboratories); Alexa Fluor 488streptavidin IgG (H + L) conjugate 1 : 100 (10 lL ⁄mL) (green) in PBS(Molecular Probes, Eugene, OR, USA); anti-NeuN 1 : 200 in PBS(Chemicon International Inc, Temecula, CA, USA); Alexa Fluor 594goat anti-mouse IgG (H + L) conjugate 1 : 200 (10 lL ⁄mL) (red) inPBS (Molecular Probes). After washing, sections were placed in

IGF-I neuroprotection and effects on PI3-kinase pathway 1491

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502

Hoechst 33342, 1 lL ⁄mL in PBS for 10 min (Molecular Probes),washed and mounted using Vectashield mounting medium. Sectionswere analysed under an Olympus BX40 flourescence microscopeequipped with an Olympus DP50 cooled digital camera.

Statistics

Data were expressed as means ± SEM. Statistical tests used wereMann–Whitney U-test and anova with Fishers post-hoc test. Valuesof P < 0.05 were considered to be significant. In all cases ncorrespond to the number of animals.

Results

HI decreases phosphorylation of Akt and GSK3b

Immunohistochemical staining demonstrated that pAkt and pGSK3bwere expressed in all regions of control brains (data not shown).Neuronal localization of both pAkt and pGSK3b was confirmed bydouble staining with NeuN (Fig. 1)pAkt expression in the cortex after HI was decreased in both

hemispheres compared with control animals, but was more pronouncedin the ipsilateral hemisphere (Fig. 2A and B). After 45 min ofreperfusion, levels of pAkt were restored and remained unchanged overthe rest of the study period (Fig. 2B). Levels of total Akt wereunchanged throughout the study period (Fig. 2A). Western blot datawere confirmed with immunohistochemical staining of pAkt, whichshowed predominantly a neuronal staining that was most pronounced inpyramidal cells of the cortex and hippocampus. Immediately after HIthe staining was weaker in the ipsilateral hemisphere compared with thecontralateral hemisphere and the staining was generally stronger incontrol animals, in agreement with Western blot data (Fig. 2C–E).Phosphorylation of GSK3b was also decreased in the cytosol in

cortex of the ipsilateral and contralateral hemisphere at 0 min(Fig. 2A) and 45 min after HI. It was still suppressed in the ipsilateralhemisphere but restored contralaterally at 2 h and normalizedbilaterally at 8 h and 24 h after HI (Fig. 3A). No changes were seenin total GSK3 over the period studied (Fig. 2A). Immunohistochem-ical staining was depleted (in agreement with Western blot data) in theipsilateral hemisphere and in parts of the contralateral hemisphere at0 min of reperfusion (Fig. 3B). After 0–2 h of reperfusion, the mainlycytoplasmic staining in control animals and in the contralateralhemisphere decreased (in agreement with Western blot data), andimmunoreactivity shifted to a nuclear localization in the ipsilateralcortex and hippocampus (Fig. 3B, III and IV). This subcellularredistribution of pGSK3b was further confirmed in double stainedsections with pGSK3b, NeuN and Hoechst (Fig. 4A–F). Nuclearstaining peaked at 6 h but was still present at 24 h after HI.

IGF-I reduces brain injury in the neonatal rat

Administration of IGF-I (50 lg) i.c.v. was shown to be neuroprotec-tive after HI in the 7-day-old rat. Area measurements comparing theleft damaged hemisphere to the right undamaged hemisphere demon-strated a reduction of damage in all four studied levels of the brain(Fig. 5A). On average, there was a 40% reduction in brain injury inhrIGF-I treated animals (50 lg) compared with vehicle-treatedcontrols (IGF-I 15.0 ± 1.6% vs. vehicle 25.0 ± 2.5%, P < 0.05). Inthe 5-lg and 25-lg treatment groups there was a tendency to lessdamage compared with vehicle-treated controls but this did not reachstatistical significance (data not shown). In addition, the morpholo-gical scoring analysis showed an overall reduction of brain injury in

hrIGF-I (50 lg) compared with vehicle-treated animals (IGF-I6.1 ± 1.0 vs. vehicle 9.3 ± 0.9, P < 0.05) (data not shown). Scoringof specific brain regions also showed significant protection in thecortex (IGF-I 1.6 ± 0.3 vs. vehicle 2.5 ± 0.2, P < 0.05), hippocampus(IGF-I 1.5 ± 0.3 vs. vehicle 2.4 ± 0.3, P < 0.05) and striatum (IGF-I1.9 ± 0.3 vs. vehicle 2.9 ± 0.3, P < 0.05) (Fig. 5B). There was noprotection in the thalamus (IGF-I 1.0 ± 0.2 vs. vehicle 1.6 ± 0.3, n.s.).There were no differences in body temperature, mortality or bodyweight between the groups (data not shown). Moreover, there were nosignificant differences in brain injury between male and femaleanimals (data not shown).

IGF-I treatment increases Akt phosphorylation

Because the insult is most pronounced in cortex in this model of braininjury, we selected to perform the biochemical analysis preferentiallyin this region. Initially we studied effects of IGF-I without HI toelucidate the IGF-I effect under control conditions, being aware of themarked effects of HI on cellular integrity and intracellular signaling.Administration of hrIGF-I (50 lg) (known to be neuroprotective)

into the left ventricle increased phosphorylation of Akt in the cytosol,peaking at 2–8 h (Fig. 6A and B). Immunoreactivity was normalizedafter 24 h and there were no changes in total Akt over the periodstudied (Fig. 6B). Immunocytochemical staining of pAkt supportedthese data and showed an increased staining of neurons in cortex afterIGF-I treatment compared with vehicle-treated animals (Fig. 6C).

Subcellular differences in phosphorylation of GSK3bafter IGF-I treatment

In the cytosol, phophorylation of GSK3b was decreased 30 min to24 h after IGF-I treatment (Fig. 6D), whereas total GSK3 remainedunchanged (data not shown). In contrast, in the nuclear fraction,pGSK3b was increased at 8 h after IGF-I treatment (Fig. 6E).Immunohistochemical staining of sections 2 h after IGF-I treatment

showed a cellular redistribution, supporting the Western blot data andwith a clearly increased nuclear staining compared with the vehicle-treated animals (Fig. 6F).It appeared contradictory to find similar changes with respect to

pGSK3b after IGF-I as occurred after HI. However, we found thatIGF-I induced a pronounced increase in cytosolic pAkt compared withHI, and that pGSK3b was substantially higher after IGF-I treatmentcompared with the levels found after HI (Fig. 7).

IGF-I increases phosphorylation of Akt and GSK3bafter HI in cytosol and nucleus

After exploring the effect of IGF-I on pAkt and pGSK3b under controlconditions we explored the effect on the PI3K pathway after HI andIGF-I treatment, choosing a time point with known effects on thesekinases both after HI and after IGF-I alone.IGF-I, administered after HI, increased pAkt and pGSK3b in the

cortical cytosolic fraction of the left hemisphere at 2 h of recovery(Fig. 8A and B) compared with vehicle-treated HI animals. Further-more, treatment with IGF-I significantly increased pGSK3b in thenuclear compartment 2 h following the insult vs. HI vehicle-treatedcontrols (Fig. 8C).

IGF-I reduces caspase activity after HI

Caspase activity is known to increase after HI brain damage in theneonatal brain, peaking at 24 h (Wang et al., 2001). We therefore

1492 K. G. Brywe et al.

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502

wanted to see if the reduced tissue loss after IGF-I treatment andincreased Akt and GSK3b phosphorylation was accompanied byeffects on caspase activity. Caspase-3-like and caspase-9-like activitywas significantly reduced in the ipsilateral cerebral cortex at 24 h afterHI and hrIGF-I treatment compared with vehicle-treated controls(Fig. 9A and B), whereas activity was not significantly different at 8 h(data not shown). IGF-I did not affect caspase-8- or caspase-1-likeactivity (data not shown).

Discussion

Effect of HI on the PI3K pathway

Using both Western blot and immunohistochemistry, we found amarked initial depletion of pAkt and pGSK3b after HI, indicating atrophic factor withdrawal response on the PI3K pathway directly afterthe insult, which has not been shown previously but is in line with adecrease in IGF-I mRNA after HI in the immature rat, as previously

Fig. 1. Neuronal localization of pAkt and pGSK3b. pAkt and pGSK3b immunoreactivity were localized in neurons throughout the brain. Double immunolabelingof pAkt (A) and NeuN (B) staining. pAkt immunoreactivity co-localized with the neuronal marker NeuN in cortex of control brains at PND7 (C). The pGSK3b (D)and NeuN (E) staining were also co-localized (F) in cortex of control brains.

IGF-I neuroprotection and effects on PI3-kinase pathway 1493

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502

Fig. 2. Phosphorylation of Akt and GSK3b after HI. (A) Western blots of pAkt, total Akt, pGSK3b and total GSK3 in the cytosolic fraction of the cortex at theend of 60 min of HI. Each band represents a sample from one single animal. Blots are shown for ipslateral (I) and contralateral (C) hemisphere after HI and forcontrols. Total GSK3 and total Akt levels remained unchanged after 60 min of HI. pAkt was markedly decreased in ipsilateral cortex and depressed to a lesser extentin the contralateral cortex. PGSK3b immunoreactivity was abolished in both hemispheres after HI. (B) Densities of pAkt protein bands of the Western blots atdifferent time points after HI were evaluated at 0 min, 45 min, 2 h, 4 h, 8 h and 24 h in cortex in ipsilateral (I) (black bars) and contralateral (C) (grey bars)hemispheres compared with controls (unfilled bars). pAkt was decreased immediately after HI, restored after 45 min and remained at control levels for the rest of thestudy period. Data are expressed as means ± SEM, n ¼ 5 per time point. Data were analysed by anova followed by Fisher’s post-hoc test: **P ¼ 0.01,***P ¼ 0.001. (C–E) Immunohistochemical staining in coronal brain sections of pAkt (C–E) immediately after 60 min of HI; pAkt staining was reduced in theipsilateral hemisphere compared with the contralateral and control brains.

1494 K. G. Brywe et al.

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502

shown (Clawson et al., 1999). The decrease of pAkt was transient,whereas the reduction of pGSK3b was more persistent and levelsremained low at 4 h of recovery. Furthermore, a subcellular redistri-bution from cytosol to the nucleus occurred for pGSK3b during

reperfusion, in agreement with in vitro studies showing a nuclearaccumulation of pGSK3b after heat shock treatment (Bijur & Jope,2001). A similar pattern of increased nuclear and decreased cytosolicpGSK3b was also seen after IGF-I administration; however, the

Fig. 3. Changes in pGSK3b expression after HI. (A) Densities of pGSK3b protein bands of the Western blots at different time points in the cytosol fraction ofcortical samples after HI were evaluated at 0 min, 45 min, 2 h, 4 h, 8 h and 24 h in ipsilateral (I) (black bars) and contralateral (C) (grey bars) hemispheres, and incontrols (unfilled bars). PGSK3b was markedly decreased bilaterally at 0 and 45 min after HI. PGSK3b levels were fully restored by 8 h after HI. Data are expressedas means ± SEM, n ¼ 5 per time point. Data were analysed by anova followed by Fisher’s post-hoc test: **P ¼ 0.01, ***P ¼ 0.001. (B) Immunohistochem-ical staining in coronal brain sections of pGSK3b (I–V). Immediately after 60 min of HI, staining of pGSK3b was abolished in most regions of the left (HI)hemisphere and staining was also weak in the contralateral cortex, whereas immunoreactivity was somewhat more preserved in hippocampus and thalamus (I). After2 h of reperfusion, staining was still reduced in the HI left hemisphere but restored contralaterally (II). In the ipsilateral hippocampus at 0 h and 2 h there was apredominant nuclear staining (III, V) instead of the cytosolic staining seen in contralateral hippocampus (IV).

IGF-I neuroprotection and effects on PI3-kinase pathway 1495

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502

1496 K. G. Brywe et al.

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502

reduction in cytosol was significantly more pronounced after HIcompared with IGF-I treatment, which agrees with the finding thatIGF-I increased cytosolic pGSK3b if administered after HI. The factthat IGF-I activates the PI3K pathway differently in vivo when givenin control brains compared with after HI is not surprising consideringthat the effects of IGF-I are likely to be context dependent and HI hasprofound effects on intracellular signaling cascades. As for many otherkinases, scaffolding proteins and subcellular redistribution may bealtered after HI, which may restrict interactions with GSK3b (Jope &Johnson, 2004).

Such a dramatic loss of anti-apoptotic signals (pAkt and itsdownstream substrates) combined with dephosphorylation of pro-apoptotic GSK3b, probably leading to its activation, may be critical inexplaining the early release of cytochrome C from mitochondria and

activation of caspases (Cheng et al., 1998; Wang et al., 2001) as wellas the translocation of apoptotis-inducing factor (AIF) (Zhu et al.,2003) that occur in the immature brain following HI. Previous studieson pAkt mostly performed in adult animals have demonstrated anincrease rather than a decrease in response to ischemia (Ouyang et al.,1999; Noshita et al., 2001; Shibata et al., 2002). This may partly beexplained in terms of differences between the mature and theimmature brain. Our data are in reasonable agreement with theWestern blot data previously published in the 7-day-old rat model ofHI (Tomimatsu et al., 2001). Although both studies demonstrated aninitial decrease of pAkt, we did not observe the subsequent transientovershoot of pAkt reported by Tomimatsu. This small discrepancymay be explained by different severity of the insult or the fact thatdifferent regions were sampled in the two studies.

Fig. 5. IGF-I reduces HI brain injury. (A) There was a 40% reduction in area loss after IGF-I (50 lg) treatment (IGF-I 15.0 ± 1.6 vs. vehicle 25.0 ± 2.5 P < 0.05),average of four levels. Area measurements were performed in four antero-posterior levels of the brain (MAP-2-stained sections are shown to the right: vehicle, I;IGF-I, II) in hrIGF-I (50 lg) and vehicle-treated animals. Tissue loss was calculated as MAP-2 area deficit expressed as a percentage of the right undamagedhemisphere. Data are expressed as means ± SEM, vehicle (n ¼ 25) and IGF-I (n ¼ 24), analysed by Mann–Whitney U-test: *P £ 0.05, **P £ 0.01.(B) Morphological analysis demonstrated a reduction of brain injury in cerebral cortex, hippocampus and striatum, but not in the thalamus. Values are given asmeans ± SEM, vehicle (n ¼ 25) (unfilled bars) and IGF-I (n ¼ 24) (filled bars) analysed by Mann–Whitney U-test: *P £ 0.05.

Fig. 4. Subcellular redistribution of pGSK3b after HI. Photomicrographs of pGSK3b (A and B) and NeuN (C and D) in cortical neurons 0 min after HI inipsilateral (A and C) and contralateral (B and D) hemispheres. Immunoreactivity of pGSK3b co-localalized with NeuN in both hemispheres (E and F) and a clearnuclear localization was shown after HI in the ipsilateral hemisphere, as demonstrated by double staining of pGSK3b and Hoechst (G), whereas in the contralateralhemisphere it is mainly a cytoplasmic staining of pGSK3b (H).

IGF-I neuroprotection and effects on PI3-kinase pathway 1497

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502

1498 K. G. Brywe et al.

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502

Effect of IGF-I treatment on the immature brain

We found that IGF-I (50 lg administered i.c.v.) reduced HI braininjury by approximately 40% in the 7-day-old rat, which is inagreement with previous results obtained in other models of cerebralischemia in both adult and fetal animals (Guan et al., 1993; Johnstonet al., 1996). There is an on-going pharmaceutical development ofIGF-I-related peptides for neuroprotective treatment (Sizonenko et al.,2001; Brywe et al., 2003). Therefore, it is important to understandmore about the mechanisms of IGF-I action in vivo.

We explored the effects of IGF-I on the immature brain without HI.IGF-I treatment clearly increased pAkt, which is in agreement withprevious studies in vitro showing that IGF-I activates the PI3Kpathway (Parrizas et al., 1997; Peruzzi et al., 1999). PGSK3bincreased in the nucleus as expected but, surprisingly, decreased in thecytosol, which is not the expected response after activation of pAkt(Cross et al., 1995; Moule et al., 1997). It may therefore be suggestedthat IGF-I either promotes translocation of phosphorylated GSK3b tothe nucleus or selectively phosphorylates nuclear GSK3b. Comparingthe effect of IGF-I injection with HI showed that the reduction inGSK3b phosphorylation was more pronounced after HI insult than

after IGF-I treatment. However, when IGF-I was superimposed onpost-HI reperfusion we found a substantial increase of pAkt andpGSK3b (in both the cytosol and the nucleus), accompanied byinhibition of caspase-3- and caspase-9-like activity.Moreover the regulation of GSK3b is a complex topic as it might be

inhibited also by direct effects not involving phosphorylation on theSer9 site (Jope, 2003) and it can interact with p53 in the nucleus andmitochondria independent of phosphorylation status (Watcharasitet al., 2002; Watcharasit et al., 2003). Furthermore, GSK3b activityis facilitated by phosphorylation of Tyr216, and protein complexformation and scaffolding proteins are important regulatory mecha-nisms for GSK3b activity that have not yet fully explored (Grimes &Jope, 2001). Regulation by subcellular relocalization is an emergingtopic given that localization as mentioned above has been shown tochange after apoptotic stimuli in vitro, and in line with IGF-I effectsendothelin-1 has been shown to increase GSK3b in the nucleus ofcardiomyocytes (Grimes & Jope, 2001). Activation of caspase-3 afterHI in the immature brain is likely to be important with respect toneuronal injury as broad-spectrum caspase inhibitors (Cheng et al.,1998) or transgenic over-expression of the endogenous caspaseinhibitor X-linked inhibitor of apoptosis (XIAP) (Wang et al., 2003)

Fig. 7. Changes in cytosolic pAkt and pGSK3b after HI or IGF-I administration. Densities of pAkt (A) and pGSK3b (B) protein bands on Western blots of samples2 h after HI or IGF-I injection in cytosolic fraction of cerebral cortex. pAkt was increased after IGF-I injection compared with after HI. PGSK3b was decreased to agreater extent after HI compared with IGF-I injection. Data are expressed as means ± SEM, n ¼ 5 in each group. Data were analysed by anova followed by Fisher’spost-hoc test: **P £ 0.01, ***P £ 0.001.

Fig. 6. Effect of IGF-I on phosphorylation of Akt and GSK3b in subcellular fractions. (A) Densities of pAkt protein bands on Western blots from cerebral cortexat different time points after IGF-I (50 lg) (filled bars) or vehicle (unfilled bar) injection. pAkt was increased 2 h after IGF-I administration. Immunoreactivity wasnormalized in all regions at 24 h. Data are expressed as means ± SEM, n ¼ 5 per time point. Data were analysed by anova followed by Fisher’s post-hoc test:**P £ 0.01, ***P £ 0.001. (B) Western blots of total Akt and pAkt in the cytosolic fraction of cortex in the ipsilateral (I) and contralateral (C) hemispheres 2 hafter IGF-I or vehicle administration. Each band represents a sample from one animal. Levels of total Akt did not differ between controls and IGF-I-treated rats. pAktlevels were increased in both hemispheres 2 h after IGF-I treatment. (C) Immunohistochemical staining of pAkt in coronal brain sections of 7-day-old rats treatedwith vehicle (I) or IGF-I (50 lg) (II). PAkt staining was increased in pyramidal cells of the cortex 2 h after IGF-I injection. (D) Densities of pGSK3b protein bandson Western blots from cerebral cortex cytosolic fractions at different time points after IGF-I (filled bars) or vehicle (unfilled bars) injection. PGSK3b decreased30 min after IGF-I administration and for the rest of the study period. Data are expressed as means ± SEM, n ¼ 5 per time point. Data were analysed by anova

followed by Fisher’s post-hoc test: *P £ 0.05, ***P £ 0.001. (E) Densities of pGSK3b protein bands on Western blots of samples from 8 h and 24 h after IGF-I(filled bars) or vehicle (unfilled bars) injections in nuclear fractions are shown. PGSK3b was increased 8 h after IGF-I and normalized at 24 h. Data are expressed asmeans ± SEM, n ¼ 5 per time point. Data were analysed by anova followed by Fisher’s post-hoc test. ***P £ 0.001. (F) Immunohistochemical staining ofpGSK3b in coronal brain sections of 7-day-old rats treated with vehicle (I) or IGF-I (50 lg) (II). PGSK3b was subcellularly redistributed after IGF-I treatment. Invehicle-treated rats, the nucleus of pyramidal cells was unstained (I). By contrast, 2 h after IGF treatment, there was a nuclear staining in pyramidal cells (II).

IGF-I neuroprotection and effects on PI3-kinase pathway 1499

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502

reduce injury in the immature brain. A significant attenuation ofcaspase activation exerted by IGF-I may thus partly explain itsprotective effect (Cao et al., 2003). The degree of caspase inhibitionfollowing IGF-I treatment was, however, quite limited as comparedwith the effect of BDNF (Han et al., 2000) or iNOS ⁄ nNOS inhibitors(Peeters-Scholte et al., 2002), suggesting that other mechanisms maybe at play.The finding that IGF-I induced a marked increase of pAkt in vivo

after HI is important as the PI3K–Akt pathway represents a majorpathway mediating neuronal survival (Yellon & Baxter, 1999; Zhenget al., 2000). pAkt is acting on multiple substrates to mediate its effects,including forkhead transcriptional regulators, p53, cAMP-responseelement binding protein, nuclear factor kappa B, Bcl-2 antagonist ofcell death, inhibitors of apoptosis, caspase-9 or other components ofthe apoptosome and GSK3b (Brunet et al., 2001).Our study does not provide information on the relative importance of

these targets in the context of IGF-I protection in the immature brainfollowing HI. We did, however, find a pronounced increase of pGSK3b(leading to its inactivation) after IGF-I treatment; this is a novel finding

that may be critical as GSK3b is strongly pro-apoptotic (Eldar-Finkelman, 2002). Furthermore, inhibitors of GSK3b reduce infarctionafter ischemia in adult animals (Kelly et al., 2003) and improveneuronal survival in vitro (Cross et al., 2001). Activation of GSK3b hasbeen linked to triggering of caspase-dependent apoptosis (Watcharasitet al., 2002) but GSK3b inhibitors given in adult stroke models did notaffect caspase activation (Kelly et al., 2003), implicating an effect oncaspase-independent or necrotic cell death. Caspases do, however, playa more prominent role in HI in immature animals (Hu et al., 2000; Gillet al., 2002) and we cannot exclude the possibility that GSK3b isinvolved in regulating caspase activation in the developing brain as itdoes in some in vitro models (Watcharasit et al., 2003).

Acknowledgements

The work was supported by grants from the Swedish Medical Research Council(09455), the Wilhelm and Martina Lundgren Foundation, the Goteborg MedicalSociety and by grants to researchers in the public health service from theSwedish government.

Fig. 8. PAkt and pGSK3b in response to IGF-I given after HI. Densities of cytosolic pAkt (A), pGSK3b (B) and nuclear pGSK3b (C) protein bands on Westernblots 2 h after HI and IGF-I (filled bars) or vehicle (unfilled bars). There were significant differences between the groups in the ipsilateral hemisphere of all studiedproteins. Data are expressed as means ± SEM, n ¼ 7 in each group. Data were analysed by anova followed by Fisher’s post-hoc test: *P £ 0.05, ***P £ 0.001.

1500 K. G. Brywe et al.

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502

Abbreviations

AMC, aminomethyl coumarine; CHAPS, 3-[(3-cholamidoprophylene) dimeth-ylamonio]propane sulphonic acid; GSK3, glycogen synthase kinase 3; HI,hypoxia-ischemia; IGF-I, insulin-like growth factor I; MAP-2, microtubule-associated protein 2; PND7, postnatal day 7; PI3K, phosphoinositide-3 kinase.

References

Bijur, G.N. & Jope, R.S. (2001) Proapoptotic stimuli induce nuclear accumula-tion of glycogen synthase kinase-3 beta. J. Biol. Chem., 276, 37436–37442.

Brunet, A., Bonni, A., Zigmond, M.J., Lin, M.Z., Juo, P., Hu, L.S., anderson,M.J., Arden, K.C., Blenis, J. & Greenberg, M.E. (1999) Akt promotes cellsurvival by phosphorylating and inhibiting a Forkhead transcription factor.Cell, 96, 857–868.

Brunet, A., Park, J., Tran, H., Hu, L.S., Hemmings, B.A. & Greenberg, M.E.(2001) Protein kinase SGK mediates survival signals by phosphorylating theforkhead transcription factor FKHRL1 (FOXO3a). Mol. Cell Biol., 21, 952–965.

Brywe, G.K., Mallard, C., Leverin, A.L., Gustavson, M. & Hagberg, H. (2003)Neuroprotection by a novel GH-releasing peptide following hypoxic-ischemic brain injury in neonatal rats. Ped. Res., 53, 535A.

Cao, Y., Gunn, A.J., Bennet, L., Wu, D., George, S., Gluckman, P.D., Shao,X.M. & Guan, J. (2003) Insulin-like growth factor (IGF)-1 suppressesoligodendrocyte caspase-3 activation and increases glial proliferation afterischemia in near-term fetal sheep. J. Cereb. Blood Flow Metab., 23, 739–747.

Cardone, M.H., Roy, N., Stennicke, H.R., Salvesen, G.S., Franke, T.F.,Stanbridge, E., Frisch, S. & Reed, J.C. (1998) Regulation of cell deathprotease caspase-9 by phosphorylation. Science, 282, 1318–1321.

Cheng, Y., Deshmukh, M., D’Costa, A., Demaro, J.A., Gidday, J.M., Shah, A.,Sun, Y., Jacquin, M.F., Johnson, E.M. & Holtzman, D.M. (1998) Caspaseinhibitor affords neuroprotection with delayed administration in a rat modelof neonatal hypoxic-ischemic brain injury. J. Clin. Invest., 101, 1992–1999.

Cheng, C.M., Mervis, R.F., Niu, S.L., Salem, N. Jr, Witters, L.A., Tseng, V.,Reinhardt, R. & Bondy, C.A. (2003) Insulin-like growth factor 1 is essentialfor normal dendritic growth. J. Neurosci. Res., 73, 1–9.

Clawson, T.F., Vannucci, S.J., Wang, G.M., Seaman, L.B., Yang, X.L. & Lee,W.H. (1999) Hypoxia-ischemia-induced apoptotic cell death correlates withIGF-I mRNA decrease in neonatal rat brain. Biol. Signals Recept., 8, 281–293.

Cross, D.A., Alessi, D.R., Cohen, P., Andjelkovich, M. & Hemmings, B.A.(1995) Inhibition of glycogen synthase kinase-3 by insulin mediated byprotein kinase B. Nature, 378, 785–789.

Cross, D.A., Culbert, A., Chalmers, K., Facci, L., Skaper, S. & Reith, A.D.(2001) Selective small-molecule inhibitors of glycogen synthase kinase-3activity protect primary neurones from death. J. Neurochem., 77, 94–102.

D’Ercole, A.J., Calikoglu, A.S. & Gutierrez-Ospina, G. (1996) The role of theinsulin-like growth factors in the central nervous system. Mol. Neurobiol.,13, 227–255.

Datta, S.R., Dudek, H., Tao, X., Masters, S., Fu, H., Gotoh, Y. & Greenberg,M.E. (1997) Akt phosphorylation of BAD couples survival signals to thecell-intrinsic death machinery. Cell, 91, 231–241.

Delaney, C.L., Cheng, H.L. & Feldman, E.L. (1999) Insulin-like growth factor-I prevents caspase-mediated apoptosis in Schwann cells. J. Neurobiol., 41,540–548.

Dudek, H., Datta, S.R., Franke, T.F., Birnbaum, M.J., Yao, R., Cooper, G.M.,Segal, R.A., Kaplan, D.R. & Greenberg, M.E. (1997) Regulation ofneuronal survival by the serine-threonine protein kinase Akt. Science, 275,661–665.

Eldar-Finkelman, H. (2002) Glycogen synthase kinase 3: an emergingtherapeutic target. Trends Mol. Med., 8, 126–132.

Feldman, E.L., Sullivan, K.A., Bhumsoo, K. & Russell, J.W. (1997) Insulin-like growth factors regulate neuronal differentiation and survival. Neurobiol.Dis., 4, 201–214.

Gill, R., Soriano, M., Blomgren, K., Hagberg, H., Wybrecht, R., Miss, M.-T.,Hoefer, S., Adam, G., Niederhauser, O., Kemp, J.A. & Loetscher, H. (2002)Role of caspase-3 activation in cerebral ischemia-induced neurodegenerationin adult and neonatal brain. J. Cereb. Blood Flow Metab., 22, 420–430.

Gilland, E., Puka-Sundvall, M., Hillered, L. & Hagberg, H. (1998)Mitochondrial function and energy metabolism after hypoxia-ischemia inthe immature rat brain: involvement of NMDA-receptors. J. Cereb. BloodFlow Metab., 18, 297–304.

Grimes, C.A. & Jope, R.S. (2001) CREB DNA binding activity is inhibited byglycogen synthase kinase-3 beta and facilitated by lithium. J. Neurochem.,78, 1219–1232.

Guan, J., Williams, C., Gunning, M., Mallard, C. & Gluckman, P. (1993) Theeffects of IGF-1 treatment after hypoxic-ischemic brain injury in adult rats.J. Cereb. Blood Flow Metab., 13, 609–616.

Hagberg, H., Bona, E., Gilland, E. & Puka-Sundvall, M. (1997) Hypoxia-ischaemia model in the 7-day-old rat: possibilities and shortcomings. ActaPaediatr. Suppl., 422, 85–88.

Han, B.H., D’Costa, A., Back, S.A., Parsadanian, M., Patel, S., Shah, A.R.,Gidday, J.M., Srinivasan, A., Deshmukh, M. & Holtzman, D.M. (2000)BDNF blocks caspase-3 activation in neonatal hypoxia-ischemia. Neurobiol.Dis., 7, 38–53.

Hedtjarn, M., Leverin, A.L., Eriksson, K., Blomgren, K., Mallard, C. &Hagberg, H. (2002) Interleukin-18 involvement in hypoxic-ischemic braininjury. J. Neurosci., 22, 5910–5919.

Hetman, M., Cavanaugh, J.E., Kimelman, D. & Xia, Z. (2000) Role ofglycogen synthase kinase-3beta in neuronal apoptosis induced by trophicwithdrawal. J. Neurosci., 20, 2567–2574.

Hoshi, M., Takashima, A., Noguchi, K., Murayama, M., Sato, M., Kondo, S.,Saitoh, Y., Ishiguro, K., Hoshino, T. & Imahori, K. (1996) Regulation of

Fig. 9. Effect of IGF-I on caspase-3 and -9 activation after HI. (A) Caspase-3-like activity (DEVD, i.e. Asp-Glu-Val-Asp, cleavage) 24 h after HI and hrIGF-I(n ¼ 10) or vehicle (n ¼ 10) treatment, in the cerebral cortical ipsilateral (I)hemisphere and in the contralateral (C) hemisphere. Reduced activity wasobserved in ipsilateral cortex 24 h after HI in the IGF-I treatment group (filledbars) compared with vehicle (unfilled bars). (B) Caspase 9-like activity (LEHD,i.e. Leu-Glu-His-Asp, cleavage) after HI and hrIGF-I (n ¼ 10) (filled bars) orvehicle (n ¼ 10) (unfilled bars) treatment, in cortex of the ipsilateral (I) andcontralateral hemisphere (C). Reduced activity was observed ipsilateral at 24 hafter HI and IGF-I treatment comparedwith vehicle treatment. Data are expressedas means ± SEM: *P £ 0.05, **P £ 0.01.

IGF-I neuroprotection and effects on PI3-kinase pathway 1501

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502

mitochondrial pyruvate dehydrogenase activity by tau protein kinaseI ⁄ glycogen synthase kinase 3beta in brain. Proc. Natl Acad. Sci. USA, 93,2719–2723.

Hu, B.R., Liu, C.L. & Park, D.J. (2000) Alteration of MAP kinas pathwaysafter transient forebrain ischemia. J. Cereb. Blood Flow Metab., 20, 1089–1095.

Johnston, B.M., Mallard, C., Williams, C. & Gluckman, P. (1996) Inslin-likegrowth factor-1 is a potent neuronal rescue agent after hypoxic-ischemicinjury in fetal lambs. J. Clin. Invest, 97, 300–308.

Jope, R.S. (2003) Lithium and GSK-3: one inhibitor, two inhibitory actions,multiple outcomes. Trends Pharmacol. Sci., 24, 441–443.

Jope, R.S. & Johnson, G.V. (2004) The glamour and gloom of glycogensynthase kinase-3. Trends Biochem. Sci., 29, 95–102.

Kelly, S., Heng, Z., Guo, S. & Danye, C. (2003) Treatment with the GSK3Binhibitor Chir98025 decreases infarct size following MCAo. Stroke, 34,301.

Mallard, E.C., Williams, C.E., Gunn, A.J., Gunning, M.I. & Gluckman, P.D.(1993) Frequent episodes of brief ischemia sensitize the fetal sheep brain toneuronal loss and induce striatal injury. Pediatr. Res., 33, 61–65.

McLaughlin, B.A., Hartnett, K.A., Erhardt, J.A., Legos, J.J., White, R.F.,Barone, F.C. & Aizenman, E. (2003) Caspase 3 activation is essential forneuroprotection in preconditioning. Proc. Natl Acad, Sci. USA, 100, 715–720.

Moule, S.K., Welsh, G.I., Edgell, N.J., Foulstone, E.J., Proud, C.G. & Denton,R.M. (1997) Regulation of protein kinase B and glycogen synthase kinase-3by insulin and beta-adrenergic agonists in rat epididymal fat cells. Activationof protein kinase B by wortmannin-sensitive and -insensitive mechanisms.J. Biol. Chem., 272, 7713–7719.

Noshita, N., Lewen, A., Sugawara, T. & Chan, P.H. (2001) Evidence ofphosphorylation of Akt and neuronal survival after transient focal cerebralischemia in mice. J. Cereb. Blood Flow Metab., 21, 1442–1450.

Ouyang, Y.B., Tan, Y., Comb, M., Liu, C.L., Martone, M.E., Siesjo, B.K. &Hu, B.R. (1999) Survival- and death-promoting events after transientcerebral ischemia: phosphorylation of Akt, release of cytochrome C andactivation of caspase-like proteases. J. Cereb. Blood Flow Metab., 19, 1126–1135.

Pap, M. & Cooper, G.M. (1998) Role of glycogen synthase kinase-3 in thephosphatidylinositol 3-kinase ⁄Akt cell survival pathway. J. Biol. Chem.,273, 19929–19932.

Parrizas, M., Saltiel, A.R. & LeRoith, D. (1997) Insulin-like growth factor 1inhibits apoptosis using the phosphatidylinositol 3¢-kinase and mitogen-activated protein kinase pathways. J. Biol. Chem., 272, 154–161.

Peeters-Scholte, C., Koster, J., van den Tweel, E., Blomgren, K., Hamers, N.,Zhu, C., van Buul-Offers, S., Hagberg, H., van Bel, F., Heijnen, C. &Groenendaal, F. (2002) Effects of selective nitric oxide synthase inhibitionon IGF-1, caspases and cytokines in a newborn piglet model of perinatalhypoxia-ischaemia. Dev. Neurosci., 24, 396–404.

Peruzzi, F., Prisco, M., Dews, M., Salomoni, P., Grassilli, E., Romano, G.,Calabretta, B. & Baserga, R. (1999) Multiple signaling pathways of theinsulin-like growth factor 1 receptor in protection from apoptosis. Mol. CellBiol., 19, 7203–7215.

del Peso, L., Gonzalez-Garcia, M., Page, C., Herrera, R. & Nunez, G. (1997)Interleukin-3-induced phosphorylation of BAD through the protein kinaseAkt. Science, 278, 687–689.

Rice, J.E., Vanucci, R.C. & Brierley, J.B. (1981) The influence of immaturityon hypoxic-ischemic brain damage in the rat. Ann. Neurol., 9, 131–141.

Russell, J.W., Windebank, A.J., Schenone, A. & Feldman, E.L. (1998) Insulin-like growth factor-I prevents apoptosis in neurons after nerve growth factorwithdrawal. J. Neurobiol., 36, 455–467.

Sara, V.R., Carlsson-Skwirut, C., Drakenberg, K., Giacobini, M.B., Hakansson,L., Mirmiran, M., Nordberg, A., Olson, L., Reinecke, M. & Stahlbom, P.A.(1993) The biological role of truncated insulin-like growth factor-1 and thetripeptide GPE in the central nervous system. Ann. NY Acad. Sci., 692, 183–191.

Shibata, M., Yamawaki, T., Sasaki, T., Hattori, H., Hamada, J., Fukuuchi, Y.,Okano, H. & Miura, M. (2002) Upregulation of Akt phosphorylation at theearly stage of middle cerebral artery occlusion in mice. Brain Res., 942, 1–10.

Sizonenko, S.V., Sirimanne, E.S., Williams, C.E. & Gluckman, P.D. (2001)Neuroprotective effects of the N-terminal tripeptide of IGF-1, glycine-proline-glutamate, in the immature rat brain after hypoxic-ischemic injury.Brain Res., 922, 42–50.

Takadera, T., Matsuda, I. & Ohyashiki, T. (1999) Apoptotic cell death andcaspase-3 activation induced by N-methyl-D-aspartate receptor antagonistsand their prevention by insulin-like growth factor I. J. Neurochem., 73, 548–556.

Thoresen, M., Bagenholm, R., Loberg, E.M., Apricena, F. & Kjellmer, I. (1996)Posthypoxic cooling of neonatal rats provides protection against brain injury.Arch. Dis. Childhood, 74, F3–F9.

Tomimatsu, T., Fukuda, H., Endo, M., Watanabe, N., Mu, J., Kohzuki, M.,Fujii, E., Kanzaki, T. & Murata, Y. (2001) Effects of hypothermia onneonatal hypoxic-ischemic brain injury in the rat: phosphorylation of Akt,activation of caspase-3-like protease. Neurosci. Lett., 312, 21–24.

Towfighi, J., Yager, J.Y., Housman, C. & Vannucci, R.C. (1991) Neuropathol-ogy of remote hypoxic-ischemic damage in the immature rat. ActaNeuropathol. (Berl.), 81, 578–587.

Van Golen, C.M. & Feldman, E.L. (2000) Insulin-like growth factor I is the keygrowth factor in serum that protects neuroblastoma cells from hyperosmotic-induced apoptosis. J. Cell Physiol., 182, 24–32.

Vannucci, R.C., Connor, J.R., Mauger, D.T., Palmer, C., Smith, M.B.,Towfighi, J. & Vannucci, S.J. (1999) Rat model of perinatal hypoxic-ischemic brain damage. J. Neurosci. Res., 55, 158–163.

Volpe, J.J. (2001) Neurology of the Newborn. Saunders, Philadelphia.Wang, X., Karlsson, J.O., Zhu, C., Bahr, B.A., Hagberg, H. & Blomgren, K.

(2001) Caspase-3 activation after neonatal rat cerebral hypoxia-ischemia.Biol. Neonate, 79, 172–179.

Wang, S., Zhu, C., Lindholm, D., Sandberg, M., Hagberg, H. & Blomgren, K.(2003) Protective effect of X-linked inhibitor of apoptosis (XIAP) in hypoxic-ischemic brain damage in neonatal rat brain. Pediatric Res., 53, 25A.

Watcharasit, P., Bijur, G.N., Song, L., Zhu, J., Chen, X. & Jope, R.S. (2003)Glycogen synthase kinase-3b (GSK3b) binds to, and promotes the actions of,p53. J. Biol. Chem., 278, 48872–48879.

Watcharasit, P., Bijur, G.N., Zmijewski, J.W., Song, L., Zmijewska, A., Chen,X., Johnson, G.V. & Jope, R.S. (2002) Direct, activating interaction betweenglycogen synthase kinase-3beta and p53 after DNA damage. Proc. NatlAcad. Sci. USA, 99, 7951–7955.

Whitaker, J.R. & Granum, P.E. (1980) An absolute method for proteindetermination based on difference in absorbance at 235 and 280 nm. AnalBiochem., 109, 156–159.

Ye, P., Carson, J. & D’Ercole, A.J. (1995) Insulin-like growth factor-Iinfluences the initiation of myelination: studies of the anterior commissure oftransgenic mice. Neurosci. Lett., 201, 235–238.

Yellon, D.M. & Baxter, G.F. (1999) Reperfusion injury revisited: is there a rolefor growth factor signaling in limiting lethal reperfusion injury? TrendsCardiovasc. Med., 9, 245–249.

Yoshimoto, T., Uchino, H., He, Q.P., Li, A. & Siesjo, B.K. (2001) CyclosporinA but not FK506, prevents the downregulation of phosphorylated Akt aftertransient focal ischemia in the rat. Brain Res., 899, 148–158.

Zhang, Z., Hartmann, H., Do, V.M., Abramowski, D., Sturchler-Pierrat, C.,Staufenbiel, M., Sommer, B., van de Wetering, M., Clevers, H., Saftig, P., DeStrooper, B., He, X. & Yankner, B.A. (1998) Destabilization of beta-cateninby mutations in presenilin-1 potentiates neuronal apoptosis. Nature, 395,698–702.

Zheng, W.H., Kar, S., Dore, S. & Quirion, R. (2000) Insulin-like growth factor-1(IGF-1): a neuroprotective trophic factor acting via the Akt kinase pathway.J. Neural Transm. (Suppl.), 60, 261–272.

Zheng, W.H., Kar, S. & Quirion, R. (2002) Insulin-like growth factor-1-inducedphosphorylation of transcription factor FKHRL1 is mediated by phospha-tidylinositol 3-kinase ⁄ Akt kinase and role of this pathway in insulin-likegrowth factor-1-induced survival of cultured hippocampal neurons. Mol.Pharmacol., 62, 225–233.

Zhu, C., Qiu, L., Wang, X., Hallin, U., Cande, C., Kroemer, G., Hagberg, H. &Blomgren, K. (2003) Involvement of apoptosis-inducing factor in neuronaldeath after hypoxia-ischemia in the neonatal rat brain. J. Neurochem., 86,306–317.

1502 K. G. Brywe et al.

ª 2005 Federation of European Neuroscience Societies, European Journal of Neuroscience, 21, 1489–1502