Journalof NeurochemistryLippincott—RavenPublishers,Philadelphia© 1998 InternationalSociety for Neurochemistry

Neurotrophins RescueCerebellar Granule Neurons fromOxidative Stress-MediatedApoptotic Death: SelectiveInvolvement of Phosphatidylinositol 3-Kinase and the

Mitogen-Activated Protein Kinase Pathway

StephenD. Skaper,MauraFloreani,*AlessandroNegro,Laura Facci,andPietro Giusti

Dipartimento di Farmacologiaand *Dipartimento di Chimica Biologica, CRIBI, Università di Padova, Padova, Italy

Abstract: Cerebellar granule neurons maintained in me-dium containing serum and 25 mM K~reliably undergoan apoptotic death when switched to serum-free mediumwith 5 mM K~.New mRNA and protein synthesis andformation of reactive oxygen intermediates are requiredsteps in K~deprivation-induced apoptosis of these neu-rons. Here we show that neurotrophins, members of thenerve growth factor gene family, protect from K~/serumdeprivation-induced apoptotic death of cerebellargranuleneurons in a temporally distinct manner. Switching gran-ule neurons, on day in vitro (DIV) 4, 10, 20, 30, or 40,from high-K~to low-K~/serum-freemedium decreasedviability by >50% when measured after 30 h. Treatmentof low-K~granule neurons at DIV 4 with nerve growthfactor, brain-derived neurotrophic factor (BDNF), neuro-trophin-3, or neurotrophin-4/5 (NT-4/5) demonstratedconcentration-dependent (1—100 ng/ml) protective ef-fects only for BDNF and NT-4/5. Between DIV 10 and20, K~-deprived granule neurons showed decreasingsensitivity to BDNF and no response to NT-4/5. Cerebel-lar granule neuron death induced by K~withdrawal atDIV 30 and 40 was blocked only by neurotrophin-3. BDNFand NT-4/5 also circumvented glutamate-induced oxida-tive death in DIV 1 —2 granule neurons. Granule neurondeath caused by K~withdrawal or glutamate-triggeredoxidative stress was, moreover, limited by free radicalscavengers like melatonin. Neurotrophin-protective ef-fects, but not those of antioxidants, were blocked by se-lective inhibitors of phosphatidylinositol 3-kinase or themitogen-activated protein kinase pathway, depending onthe nature of the oxidant stress. These observations indi-cate that the survival-promoting effects of neurotrophinsfor central neurons, whose cellular antioxidant defensesare challenged, require activation of distinct signal trans-duction pathways. Key Words: Oxidative stress—Apoptosis—Neurotrophins— Neuroprotection —Signaltransduction— Melatonin—Glutathione.J. Neurochem. 70, 1859—1868 (1998).

Thedeathof centralneuronsis widely recognizedasa normal featureof vertebratedevelopment.Neuronal

programmedcell death(PCD) is thought to servetheremoval of neuronalprecursorsthat fail to establishappropriatesynapticconnections(Oppenheim,1991).Previousstudiessuggestthat someforms of develop-mental PCD in the mammalian brain occur viaapoptosis,a physiologicalmechanismby which a celldies via transcriptionaland translationalactivationofanintrinsic cell deathor suicideprogram(Raff et al.,1993).Inappropriateapoptosishasbeensuggestedtobe involved in pathologicalneuronaldeathin varioushuman neurodegenerativediseases,such as Alzhei-mer’s disease(Cotman and Anderson, 1995), Hun-tington’sdisease(Portera-Cailliauet al., 1995),amyo-trophic lateral sclerosis(Rabizadehet al., 1995), andspinalmuscularatrophy (Roy et al., 1995), in whichneuronalloss is a prominentfeature(Bredesen,1995).

Reactiveoxygen species(ROS) have beenimpli-cated as mediatorsof excitotoxic (Chan, 1996) andapoptotic (Bredesen,1995) nervecell death. In vitromodelsthat recapitulateneuronalapoptosisare essen-tial for studyingthe molecularmechanismsunderlyingcell death,andmayproveusefulfordevelopingneuro-protectiveagentsas well as therapeuticinterventionsfor neurodegenerativedisorders. Cultured cerebellargranule neuronshaveproveduseful for this purpose.Dissociatedcerebellargranuleneuronsdeprivedof se-

ReceivedAugust8, 1997; revisedmanuscriptreceivedDecember5, 1997; acceptedDecember5, 1997.

Addresscorrespondenceandreprint requeststo Dr. S. D. Skaperat SmithKline BeechamPharmaceuticals,NeurosciencesResearchDepartment,New Frontiers SciencesPark North, Third Avenue,Harlow, EssexCMI9 SAW, U.K.

Abbreviations used: BDNF, brain-derived neurotrophicfactor;CNTF, ciliary neurotrophicfactor; DAB

359-NT4, diphtheriatoxin—neurotrophin-4/5fusion protein;Dlv, day in vitro; GDNF,glial cellline-derived neurotrophicfactor; MAP, mitogen-activatedprotein;MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bro-mide; NGF, nervegrowth factor; NT-3, neurotrophin-3;NT-4/5,neurotrophin-4/5;PBN, N-tert-butyl-a-phenylnitrone;PCD, pro-grammed cell death; P1 3-kinase, phosphatidylinositol3-kinase;ROS, reactive oxygenspecies.

1859

1860 S. D. SKAPERET AL.

rum anddepolarizinglevels(25 mM) of extracellularK + reliably undergoapoptosisthat is characterizedbychromatin condensation,pyknosis, andinternucleoso-mal DNA fragmentation,andthat is dependenton mac-romolecular synthesis (D’Mello et al., 1993; Millerand Johnson,1996). ROS areessentialmediators ofK + deprivation-inducedapoptosisof cerebellargran-ule neurons(Schulzet al., 1996).Developingembry-onic cortical neuronssubjectedto cystine deprivationand GSH loss and consequentoxidative stressalsohavemorphological andbiochemicalfeaturescharac-teristic of apoptosis(Ratanet al., 1994).

Theneurotrophinsareafamily of structurallyrelatedmolecules that include, in addition to nervegrowthfactor (NGF), brain-derived neurotrophic factor(BDNF), neurotrophin-3 (NT-3), neurotrophin-4/5(NT-4/5), andneurotrophin-6(Snider, 1994; Ibáñez,1995). The neurotrophinspromote the survival anddifferentiationof specificpartiallyoverlappingpopula-tions of neuronsin both the PNS (Levi-Montalcini,1987)andtheCNS (Lewin andBarde, 1996).Neuro-trophins interactwith two different types of cell-sur-face proteins, the Trk family of receptortyrosineki-nases,and the p75NTh, or low-affinity neurotrophinreceptor. Different membersof the Trk family canserve as signal-transducing receptors for differentmembersof theneurotrophinfamily. TrkA bindsNGF,TrkB binds BDNF andNT-4/5 with high affinity andNT-3 to a lesserextent, and TrkC specifically bindsNT-3 (for review, seeMeakinandShooter,1992;ChaoandHempstead,1995). All neurotrophinsarecapableof binding to ~‘75NTR, although each factor exhibitsslightly different binding characteristicswith p75NTR

(ChaoandHempstead,1995).Theactivities of neuro-trophinshave led to increasinginterestin using themastherapeuticsfor neurologicaldiseases(Hefti, 1997).However, the entire spectrumof the possibleactionsof theneurotrophinsis incompletelyunderstood.Thereis evidencethatneurotrophinsandothergrowth factorsmay play important roles in the brain’s responsetoinjury and in neurodegenerativedisorders resultingfrom excitotoxic insult (Mattson et al., 1993; Hefti,1997). The experimentsdescribedin this article weredirectedat evaluatingthe actionsof neurotrophinsinsettingswhere ROS are mediatorsof apoptotic-likedeathin cerebellargranuleneurons.Ourresultsdem-onstratethatBDNF, NT-4/5, andNT-3 havecytopro-tective effects on granuleneuronsinjured by K + I se-rum deprivationor by glutamate-inducedGSH deple-tion, in a sequentially distinct and Trk receptor-dependent manner. Specific kinase inhibitors wereused to delineatetheparticipationof downstreamsig-nal-transductionpathways in neurotrophin survivalpromotion.

MATERIALS AND METHODS

Neuronal culturesCerebellargranuleneuronswerepreparedfrom 8-day-old

Sprague—Dawleyrat pups (Harlan)as describedpreviously

(Skaperet al., 1990). In brief, cells were dissociatedfromfreshly dissectedcerebellaby mechanicaldisruption in thepresenceof trypsin and DNaseandthen platedin poly-L-lysine-coated60-mm culture dishes (Falcon) or 24-wellplates(16 mm, Falcon). Cells were seededat a densityof 3.0 X i0~cells/cm2in basal modifiedEagle’s mediumcontaining2 mM glutamine,25 mM KC1, 50ptg/mlgentami-cm, and 10%heat-inactivatedfetalbovine serum.Cytosine-/3-D-arabinofuranoside(10 p~M)was added to the culturemediumafter20—22h to preventthegrowth of nonneuronalcells. Cerebellarneuronswerekept alive for up to 40 daysby replenishingthe growth mediumwith D-glucoseon dayin vitro (DIV) 7 andevery fourth day thereafter,and bycompensatingfor lost amountsof waterdueto evaporation.Immunocytochemicalanalysisof thesecultureshasshownthattheycontain>95% granuleneurons(Thangniponet al.,1983). Such primary culturesalso possessthe biochemicalandelectrophysiologicalcharacteristicsof theirin vivo coun-terparts(Cull-CandyandUsowicz,1987;Galloet al., 1987).

Treatment of culturesPotassiumdeprivation-induceddeathwas performedby

usingcerebellargranuleneuronsat DIV 4, 10, 20, 30, and40. Cellswerewashedtwicewith serum-freeculturemediumcontaining5 mM KC1 andsupplementedwith gentamicinasindicatedabove.Control cells werewashedidentically andmaintainedin serum-freemedium supplementedwith KC1at 25 mM. Glutamate-inducedcell deathwasinduced 24—36 h afterseedingthe granulecells by exchangingtheorigi-nalplating mediumwith an equalvolume of freshmediumcontaining 5 mM L-glutamate.Neuronal viability in all ex-perimentswasdetermined30 h later. Trophic factorswereaddedto the cultures from dilutions preparedin medium.Mediacontainingcycloheximide(10,ag/ml)or actinomycinD (1 ~.sg/ml)weremadeby diluting at least1,000-foldcon-centratedsolutions preparedin dimethyl sulfoxide vehicle(0.1%,vol/vol), which hadno protectiveor toxic effect byitself. Unless otherwise indicated, treatment with LY294002,wortmannin,or PD 98059includeda pretreatmentin which the drug wasaddedto the cultures 15 mm beforeexposureto stimuli.

Assessmentof neuronalsurvivalNeuronalinjury wasgrossenoughto be evidentmorpho-

logically when viewedundera phase-contrastmicroscope.Neuronsurvival wasquantifiedby using3-(4,5-dimethylthi-azol-2-yl)-2,5-diphenyltetrazoliumbromide (MTT), whichyields a blueformazanreductionproductin living cellsbutnot in deadcellsor theirlytic debris(Mosmann,1983). Thereactionproduct,solubilizedin dimethyl sulfoxide,is easilymeasuredwith anELISA platereader(Skaperet al., 1990).The MTT assayhasbeenappliedto quantitatethe survival-promotingeffectsof neurotrophins(Manthorpeet al., 1986;Zirrgiebel et al., 1995) and cerebellar granule neuronapoptosis(Shimokeet al., 1997; Taylor et al., 1997).

Glutathione levelsCellularGSH levelsweremeasuredby using a modifica-

tion of themethodofTietze(1969),asdescribedby Floreaniet al. (1997a). Cerebellargranuleneuronsplated on 60-mm culturedishes,24—36 h later, wereexposedto 5 mMglutamate with or without the selected neuroprotectant.Twenty-four hourslater the test solution wasremoved,thedisheswere washedoncewith Locke’s solution (Skaperetal., 1990), and2 ml of 6% metaphosphoncacidaddedper

J. Neurochem.,Vol. 70, No. 5, 1998

NEUROTROPHINSPREVENTCNS NEURONALAPOPTOSIS 1861

dish. After 5 mm on ice, the acid extract was collected,centrifuged, and processed.The cellular content of GSSGwastypically ~5%of theGSH level andwasnotconsidered.

MaterialsTissue culture medium, glutamine, gentamicin, poly-L-

lysine (molecular weight 68,000), cytosine-/~-n-arabmofu-ranoside, L-glutamate, L-aspartate, kainate, N-methyl-D-aspartate,quisqualate,GSH, glutathione monoethylester,N-acetyl-L-cysteine,cycloheximide, actinomycin D, wort-mannin,MTT, Hoechst33258,andN-tert-butyl-a-phenylni-trone (PBN) were obtainedfrom Sigma(St. Louis, MO);/3-N-oxalyl-L-cs,8-diaminopropionicacid was from TocrisCookson(Bristol, U.K.); K252awas from Calbiochem(LaJolla,CA, U.S.A.); LY 294002was from BIOMOL (Plym-outh Meeting,PA, U.S.A.); PD 98059 was from New En-glandBiolabs(Beverly,MA, U.S.A.).Thefollowing recom-binantneurotrophicfactorswerepreparedasdescribedpre-viously: humanciliary neurotrophicfactor(CNTF) (Negroet al., 1991); rat BDNF and NT-3 (Negro et al., 1994);rat NT-4/5 and diphtheria toxin—neurotrophin-4/5fusionprotein (DAB359-NT4) (Negro and Skaper, 1997). Humanglial cell line-derivedneurotrophicfactor (GDNF) wassyn-thesizedin E. coli andrefoldedin vitro asdescribedby Linet al. (1993).

RESULTS

BDNF and NT-4I5 prevent K + deprivation-induced apoptosisof DIV 4 cerebellar granuleneurons

Apoptosis of differentiatedcerebellargranuleneu-rons canbe inducedby removing serumand loweringthe extracellularK~concentrationfrom 25 to 5 mM(D’Mello et al., 1993; Miller and Johnson,1996).Switching DIV 4 culturesto low-K~/serum-freeme-dium decreasedgranule neuron viability by ~60%after 30 h. Exposureof DIV 4 culturesto the proteinsynthesisinhibitor cycloheximide(10 ~tg/ml) or theRNA synthesisinhibitor actinomycinD (1 ,ug/ml) atthe timeof K ~ deprivationprotectedagainstlow K + -

inducedneuronaldeath(87±6 and78 ±7%, survival,respectively,comparedwith 38 ±6% for low K~with-out inhibitors; mean±SD, n = 4). Cerebellargranuleneuronsat DIV 4, switchedto low-K~/serum-freeme-dium, displayed morphologicalfeaturesof apoptoticcells, includingcytoplasmicblebbingand heterochro-matic clumpingrevealedby Nomarskioptics.A char-acteristiccondensationof nuclearchromatin(andag-gregationat thenuclearmembrane)couldalso be seenafter stainingwith the fluorescentdye Hoechst33258(datanot shown).

Neurotrophinswerethen addedto DIV 4 cerebellargranuleneuronsat the time of K~/serumwithdrawal.BDNF and NT-4/5 each produceda concentration-dependentrescueof granuleneuronsafter 30 h (Fig.lA), with EC50 valuesof 3.0 ± 1.2 and2.2 ±0.7 ng/ml, respectively.NeitherNGF, NT-3, nor CNTF, upto 1 ,uglml, was able to reduceK~-inducedneuronaldeathat this age(Table 1). BDNF provideddiminish-ing, but significant protection whenaddedup to 6 h

after switch to low-K~Iserum-freemedium(Fig. lB),consistentwith the commitmentpoint for apoptosisreportedunder theseconditions (Miller andJohnson,1996). K252a,a selectiveinhibitor of the Trk familyof neurotrophin receptors (Tapley et al., 1992),blockedthe neuroprotectiveeffectsof BDNF andNT-415 (Fig. 1A), indicating that TrkB activation wasnecessaryfor actionof the latter.

ROS have been implicated as importanteffectormoleculesin the apoptoticdeathof cerebellargranuleneuronsinduced by K~deprivation (Schulz et al.,1996). We thus testedthe effectsof the neuroprotec-tive antioxidantmelatonin(Giustiet al., 1995;Floreaniet al., 1997b) and the free radical spin trap PBN(KnechtandMason,1993) on K + Iserumdeprivation-inducedapoptosisof DIV 4 granuleneurons.Exposureto melatonin(0.5 mM) or PBN(0.1 mM) raisedcere-bellar granuleneuronsurvival at 30 h to 88 ±3 and87 ±2%, respectively,comparedwith 40 ±9% forlow K~only (mean±SD, n = 3).

Neurotrophin protection of K k/serum-deprivedcerebellar granule neurons is culture age specific

Theability of neurotrophinsto limit granuleneurondeathinducedby K~deprivationasa function of cul-ture agewasnext studied.BDNF, NT-3, andNT-4/5were addedto DIV 10—40cerebellargranuleneurons,andsurvivalevaluatedafter30h of K~/serumdepriva-tion. Theseculturesretaintheir glutamatergiccharac-ter, as evidencedby sensitivity to excitatory aminoacid receptoragonist killing (seeTable 5, and datanot shown) and stimulation of inositol phospholipidhydrolysis.At bothDIV 10 and20,protectionaffordedby BDNF was againevident (Table 2), although in-creasing concentrationsof the factor were needed(EC50 for BDNF, 6.9 ± 2.3 and 13.3 ±2.0 ng/ml atDIV 10 and 20, respectively;p = 0.02, n = 3). It isnoteworthythatat thesein vitro ages,NT-415 failedtorescueK~/semm-deprivedcerebellargranuleneurons.As with DIV 4 cultures,NT-3 was inactiveat DIV 10and20 (Table2). By DIV 30, neitherBDNF nor NT-4/5 wasneuroprotective(Table2). In contrast,NT-3now concentration-dependently(EC20, 7.7 ±2.2 ng/ml) inhibited low-K + -inducedcerebellargranuleneu-ron death(Fig. 2). The action of NT-3 was blockedby K252a (Fig. 2) and was effective (althoughde-creasinglyso) whenaddedup to 4 h after the switchto low-K + medium(datanot shown).NT-3 retainedits efficacy as a neuroprotectantup to DIV 40 (EC50,6.3 ±0.7 ng/ml, n = 3), the longest time tested.NeitherNGF norCNTF(up to 1 ~tg/ml) wasneuropro-tectiveat DIV 10—40. It is interestingthat GDNF, adistantmemberof the transforminggrowth factor /3superfamily (Lindsay and Yancopoulos, 1996), al-thoughnot preventingthe loss of K + -deprivedDIV10 and20 granuleneurons,did promotea modest,andsignificant, rescueof DIV 30 and40 neurons(Table2). K + - and serum-deprivedcerebellargranule neu-ronsat DIV 10—40weresensitiveto the savingactions

.1. Neurochem.,Vol. 70. No. 5, 1998

1862 S. D. SKAPERET AL.

FIG. 1. BDNF and NT-4/5 concentration- and time-dependently maintain survival of K~-depriveddeveloping cerebellar granule neu-rons. Neurons at DIV 4 were switched from culture medium containing 10% fetal calf serum and 25 mM KCI to serum-free mediumcontaining 5 mM KCI (K5) and different concentrations of BDNF or NT-4/5 (A); or 50 ng/ml BDNF, added at different times aftermedium exchange (B). Survival was quantified by the MU assay 30 h after treatment. Data are mean ±SD values (three experiments).BDNF, I; NT-4/5, •; BDNF + 100 nM K252a, A; NT-4/5 + 100 nM K252a, V. For BDNF-treated cultures in B: p <0.01(0, 1, 2,and 3 h) orp < 0.05(6 h), compared with K5 without BDNF at 30 h (35 ±5% survival).

of inhibitors of macromolecularsynthesisandscaven-gers of ROS (e.g., melatonin;datanot shown).

Numerousneuronsubsets expressTrkB, butthe ef-fectsof the TrkBreceptorligandsBDNF andNT-4/5may bedistinct (McAllister et al., 1995;Riddle et al.,1995). We probedthe differential biological respon-sivenessof cerebellargranule neuronsto BDNF andNT-4/5 by usingDAB359-NT4.Thisfusion toxin bindsto TrkB-expressingcells, therebyallowing entry of thecatalytic subunit of diphtheria toxin, which is fatal(Negro andSkaper,1997). Incubationof DIV 10 and30 granule neuronsfor 3 dayswith 1 nM DAB 359-NT4reduced survivalby 44 and52%,respectively(Table3). Granuleneurondeathcausedby DAB359-NT4 wasantagonizedby excessNT-4/5 or BDNF but notby

TABLE 1. Survivalof DIV4 cerebellargranule neuronsafter treatmentwith various trophicfactors in low K~

Addition Survival (%)

NoneKC1 (25 mM)NGF (1 /1g/ml)BDNF (100 ng/ml)NT-3 (1 ~.tg/ml)NT-4/5 (100 ng/ml)CNTF (I ftg/ml)

34±4100 ±4”40 ±574 ±6”42 ±781 ±6”30 ±5

Cultures wereswitchedto low (5 mM)-K~/serum-freemediumatDIV 4, and neurotrophic factorswereadded at theconcentrationsindicated. Survival wasquantifiedby the MU assay at 30h. Dataaremean±SD values (threeexperiments).

“p < 0.01, comparedwith untreatedneuronsswitchedto low K~(one-wayANOVA with Student—Newman—Keulstest).

NT-3 (Table 3), therebyindicating the continuingca-pacity of these cellsto recognizeTrkB ligands.

BDNF and NT-4/5 inhibit death triggered byglutamate-inducedoxidative stressin developingcerebellar granule neurons

Glutamatebindsto bothexcitatoryneurotransmitterbinding sites (Monaghanet al., 1989)and aCl-de-pendentand quisqualate-and cystine-inhibitedtrans-port site onbrainneurons.Glutamateor homocysteateinhibition of cystine uptake and subsequentGSHdepletionin immatureembryoniccortical neuronshasbeenshown toleadto oxidative stress and cell death

TABLE 2. Culture agespecifiesthe ability ofneurotrophinsto protect cerebellargranule

neuronsfrom deathinducedby K~deprivation

Treatment

Surviv al (%)

DIV 10 DIV 20 DIV 30 DIV 40

None 41±7 37±2 38±2 36±10NGF (1 ~ig/ml) 34 ±2 40 ±10 39 ±3 34 ±4BDNF(100 ng/ml) 67 ±2” 83 ±4” 40 ±5 34 ±5NT-3 (1 ~.tg/ml) 44 ±3 32 ±2 82 ± 6” 90 ±11”NT-4/5 (1 ~ig/ml) 42 ± 3 31 ±8 37 ±6 35 ±5GDNF(100 ng/ml) ND ND 60 ±6” 61 ±95GDNF(1 jig/mI) 36 ±5 34 ±8 60 ±6~ 63 ± 7”

Cultures were switched to low (5 niM)-K~/serum-free medium atDIV 10—40, and neurotrophic factors were added at the concentra-tions indicated. Survival was quantifiedby the MTT assay at30 h.Dataaremean±SD values (threeexperiments).ND, not determined.

“p <0.01;b~< 0.05, comparedwith untreatedneurons switchedto low K” (one-way ANOVA with Student—Newman—Keulstest).

J. Neurochem.,Vol. 70, No. 5, 1998

NEUROTROPHINSPREVENTCNS NEURONAL APOPTOSIS 1863

FIG. 2. NT-3 inhibits low K~-induceddeath of DIV 30 cerebellargranule neurons. Neurons at DIV 30 were switched from culturemedium containing 10% fetal calf serum and 25 mM KCI toserum-free medium containing 5 mM KCI and no additives ordifferent concentrations of NT-3 (•). In some cases, K252a (100nM) was included (•). Survival was quantified by the MU assay30 h after treatment. Data are mean ±SD values (three experi-ments).

(Murphy et a!., 1990) byapoptosis(Ratan et al.,1994). Compoundsstructurally relatedto glutamateinducedcytotoxicity in DIV 1—2 and 9—11 cerebellargranule neuronsin correlation,respectively,with theirexpectedpotencyin inhibitingneuronalcystineuptake(Murphy et al., 1990) orin activating glutamatergicreceptors(Monaghanet al., 1989) (Table4). Granuleneurons exposedto a maximally toxic concentration(5 mM) of glutamatefor 14 h and thenstainedwiththeDNA-intercalating,fluorescentdyeacridineorange(Ratanet al., 1994)showedthe presenceof brightlylabeled sphericalfragmentsof nuclei characteristicofapoptosis(datanot shown).Coapplicationof melato-

TABLE 3. BDNFandNT-4/5, but notNT-3, antagonizethe cytotoxicity of DAB

389-NT4for cerebellargranuleneurons

Treatment

Survival (%)

DIV 10 DIV 30

DAB389-NT4DAB389-NT4 + BDNFDAB389-NT4 + NT-4/5DAB389-NT4 + NT-3

56 ±4ND

104 ± 9”48 ±8

48 ±797 ±2”

102 ±4”49 ±7

Cultures at DIV 10 or30 weretreatedwith DAB389-NT4 (1 nM),alone or together with 50 nM BDNF, NT-4/5 or NT-3, in theiroriginal medium. Survival wasquantified by the MTT assay 72 hlater. Data are mean ±SD values (threeexperiments).ND, notdetermined.

“p < 0.01, comparedwith DAB389-NT4-treatedneurons (one-way ANOVA with Student—Newman—Keulstest).

TABLE 4. Cytotoxicityby glutamateanaloguesincerebellargranule neurons

Compound

Cell death, %

DIV 1—2 DIV 9—11

None 7 ±5 11±8L-Glutamate(2 mM) 51 ±6” 87 ±10”Kainate (1 mM) 9 ±6 63 ±8”NMDA(1mM) 4 ±3 71±8”L-Aspartate (1 mM) 5 ±3 90 ±12”Quisqualate (0.S’mM) 59 ±8” 14 ±3BOAA(lmM) 48 ±6” 13±6

Cultures 24—36 h afterplating (DIV 1—2) received an exchangeofcompletemediumcontainingtheindicated compound,andincuba-tion was continued for 30 h. Cultures at DIV 9— II were treatedwith thesamecompoundsin completeorMg

2~-freeLocke’s solutionfor 30 mm (23°C),washed,andreturnedto their originalconditionedmediumfor 24 h (Skaperet al., 1990). Survival was quantifiedbytheMU assay attheendof the respective incubationperiod. Dataare mean±SD values(threeexperiments).BOAA, /3-N-oxalyl-L-a,/3-diaminopropionicacid. The NMDA-selective antagonistMK-801 blocked glutamate toxicity in DIV 9—11, but not in DIV 1—2,granule neurons.

“p < 0.01, comparedwith control (“none”) cultures (one-wayANOVA with Student—Newman— Keuls test).

nm (0.5 mM), cystine (1 mM), the cellular GSHdelivery agentGSHmonoethylester(1 mM), or PBN(0.1 mM) preventedglutamate(5 mM)-inducedcelldeathafter 30 h(87 ± 6, 85 ± 4, 82 ± 7, and 112±7% survival, respectively,comparedwith 35 ±7%survival for glutamateonly, n = 6).

Treatmentof glutamate(5 mM)-exposedDIV 1—2cerebellargranule neuronswith BDNF or NT-4/5 wasefficaciousin limiting cell death30 h later,with respec-tive EC

50values of10.3 ±2.9 and 7.5±2.7 ng/ml.NeitherNGF, NT-3, CNTF, norGDNF (up to I ,uglml) significantly reduced granuleneuron degenerationcausedby glutamate.The neurotrophin-protectiveef-fectsdid not dependon recoveryof intracellularGSHlevels.Incubationof DIV 1—2 cerebellargranuleneu-rons with5 mM glutamatefor 24 h decreasedcellularGSH content to <5% of control (Table 5), whichremaineddepressedin the presenceof BDNF, NT-4/5, or melatonin.NT-4/5 andBDNF (50 ng/ml) werelikewise cytoprotectivewhen DIV 1—2 granule neu-rons were injured with buthioninesulfoximine (datanot shown),a specific andessentially irreversiblein-hibitor of GSH biosynthesis(Griffith and Meister,1979).

Role of intracellular protein kinase pathways inneurotrophin-mediated survival of cerebellargranule neurons

Tyrosinekinasegrowthfactorreceptorsactivatesev-eral intracellular signaling pathways. Of these, thephosphatidylinositol3-kinase(P1 3-kinase)andmito-gen-activatedprotein (MAP) kinasepathwayshavebeen implicatedin survival. Neurotrophinscan activatePT 3-kinase(Soltoff et al., 1992;Shimokeet al., 1997),

J. Neurochem.,Vol. 70, No.5, 1998

1864 S. D. SKAPERET AL.

TABLE 5. Effectof BDNF, NT-4/5, andmelatoninonglutamate-inducedreductionin GSHlevelsin DIV 1—2

cerebellargranuleneurons

Treatment GSH(nmolIlO6 cells)

Control 2.08 ±0.05Glutamate

+ BDNF 0.05 ±0.10”+ NT-4/5 0.06 ±0.06”+ Melatonin 0.02±0.03”

Granule neurons were plated and grown for 24—36 h, and thenwere exposed to glutamate (5 mM) in the presence of BDNF(50ng/ml), NT-4/5 (50 ng/ml), or melatonin (0.5 mlvi). Twenty-fourhours later, the cells were harvested and cellular GSHlevels weremeasured. Data are mean ±SD values (n = 4). GSHvalue in cellschallengedby glutamateonly was belowthe limits of detection.

“p < 0.01, comparedwith control cultures.

and inhibition of PT 3-kinaseblocks the survival-pro-motingeffectsof NGF in PC12 cells(YaoandCooper,1995) and of BDNF in cerebellargranule neurons(Shimokeet al., 1997). Addition of the PT 3-kinaseinhibitorLY 294002(YaoandCooper,1995)(30

1iM)negatedthe ability of NT-4/5 or BDNF (50 ng/ml)to rescueDIV 4—10 granuleneuronsfromdeathcausedby K~/serumdeprivation(Table6). PT 3-kinaseinhi-bition blocked,aswell, thesurvival-promotingactivityof NT-3 (50 ng/ml) for DIV 30 granule neuronsswitchedto low-K~/serum-freemedium. Similar re-sults wereobtainedwith wortmannin,a fungal metabo-lite thatalsoblocksP1 3-kinase(Table6).Comparableconcentrationsof LY 294002(Miller et al., 1997)andwortmannin (Shimokeet al., 1997) inhibit P1 3-kinaseactivity in cerebellargranuleneurons.The activationof P1 3-kinaseis known to induce the activation ofribosomal S6 kinase (p7OS6 kinase). To determinewhetherp70S6kinaseparticipatesin the survival-pro-moting effectsof neurotrophinsdownstreamof P1 3-kinase,weusedrapamycin,apotent inhibitor of p7056kinaseactivation.Althoughrapamycinblocks BDNF-stimulatedp7056kinaseactivationin cerebellargran-ule neurons(Gunn-Mooreet al., 1997), the inhibitor(50 nM) did not reducethe survival responseof K”-deprivedgranule neuronseither to BDNF at DIV 10(80 ±7 vs. 76 ±5%, respectively,n = 4) or to NT-3 at DIV 30 (91 ± 6 vs. 90 ± 4%, respectively,n= 6).

Becauseneurotrophinsstimulate a tyrosine kinasereceptor,the MAPkinasepathwayis anothercandidatefor survival-promotingactivity. PD 98059,a specificinhibitor of MEK (MAP/extracellularsignal-regulatedkinase kinase) reported to have no effect on PT 3-kinase,or other serine/threonineor tyrosine kinases(Alessi et al., 1995),hadno discernibleeffecton neu-rotrophin saving of K + I serum-deprivedcerebellargranuleneurons(Table6). However,whenglutamate(5 mM)-challengedDIV 1—2 granule neuronsweretreatedwith PD 98059 (30 ~tM) in the presenceofBDNF or NT-4/5 (50 ng/ml), the inhibitor signifi-

cantly reversedthe ability of theneurotrophinsto sup-pressneuronalloss(Table6). NeitherLY 294002norPD 98059 blockedmelatonin-mediatedsurvival (datanot shown), providingevidencethat theseinhibitorsare not generallytoxic to cerebellargranuleneurons.

DISCUSSION

Thepresentresultsdemonstratethatmembersof theNGF family of neurotrophinsselectivelylimit cerebel-lar granuleneurondeath,causedby withdrawalof tro-phic concentrationsof K + orby depletionof intracellu-lar GSH stores.This studyalsoprovidesevidencefora link betweenneurotrophinprotectionand signalingpathwaysthat promote the survival of CNS neurons.Apoptosisof cerebellargranuleneuronscanbereliablyinducedby removingserumand loweringtheextracel-lular K~concentration(D’Mello et al., 1993; Millerand Johnson, 1996). Glutamate-inducedoxidativestressof immaturefetal cortical neurons(Murphy eta!., 1990) also leadsto apoptosis(Ratanet al., 1994).Further,ROSareessentialmediatorsof theinjury pro-cessin theseparadigmsof neuronaldeath(Ratanetal., 1994; Schulz et a!., 1996). BDNF and NT-4/5rescuedDIV 1—4 granuleneuronsfrom deathlinkedtoK + deprivationor glutamate-inducedGSHdepletionat concentrationsconsistentwith their interactionwith

TABLE 6. Effectof P13-kinaseandMAP kinaseinhibitors on the survival-promotingeffectsofneurotrophinsfor cerebellargranuleneurons

Survival (%)

DIV 1-2 DIV4 DIV 10 DIV3OTreatment (+ Glu) (low K”) (low K”) (low K~)

None 44±5 32±3 41±2 40±5BDNF (50 ng/ml) 84 ±2 90 ±8 103 ±10 ND

+ LY 294002 80 ±8 41 ±5” 57 ±4” ND+ Wortmannin 75 ±6 36 ±6” 55 ±3” ND+ PD 98059 50 ±2” 78 ± 10 ND ND

NT-4/5 (50 ng/ml) 88 ±4 84 ±9 ND ND+ LY 294002 77 ±8 42 ± 6” ND ND+ Wortmannin 81 ±6 35 ±9” ND ND+ PD 98059 51 ±4” 75 ±5 ND ND

NT-3 (50 ng/ml) ND ND ND 92 ±3+ LY 294002 ND ND ND 48 ±5”+ Wortmannin ND ND ND 45 ±3”+ PD 98059 ND ND ND 90 ±8

Cerebellargranule neuronsat the indicated in vitro ages were

either challengedwith glutamate(Glu; 5 mM) or switchedto low-K”/serum-deprivedmedium(low K*). The cultureswerethentreatedwith oneof theindicatedneurotrophins(50 ng/ml), aloneortogetherwith LY 294002 (30 jiM), wortmannin(0.1 jiM), or PD 98059(30jiM). Becauseof its lability, medium was replacedafter 6 h withmedium containing fresh PD 98059, and wortmannin was addedevery6 h to thecells. Survival wasquantifiedby the MU assayat30 h. Data are mean ±SD values (threeexperiments).ND, notdetermined.

“p < 0.01, comparedwith neurotrophin-treatedneuronsswitchedto low K” (one-wayANOVA with Student—Newman—Keulstest).

J. Neurochem.,Vol. 70, No. 5, 1998

NEUROTROPHINSPREVENTCNS NEURONALAPOPTOSIS 1865

Trk receptors,asconfirmedby the ability of the selec-tive Trk inhibitor K252a to block the neuroprotectiveeffectsof BDNF and NT-415. At later culture times,first NT-4/5, and laterBDNF, lost their survival-pro-moting capabilities for cerebellar granule neurons.Concurrently, NT-3 acquired efficacy in preventingneuronloss triggeredby low K~,again in a K252a-sensitivefashion.Cerebellargranuleneuronsswitchedto low-K ±/serum-freemediumshowedmorphologicalfeaturesof apoptoticcells and were sensitiveto thesaving actionsof inhibitors of macromoleculesynthe-sis and antioxidants,as observedby others(D’Melloet al., 1993; Schulz et al., 1996).

At different stagesof development,neuronsac-quire and lose dependenceon particular neuro-trophins (Davies, 1994). Cultured rat cerebellargranuleneuronsexpressboth functional and trun-catedforms of TrkB andTrkC receptors(Lindholmet al., 1993;Zirrgiebelet al., 1995).BDNF andNT-4/5, butnotNT-3 andNGF,promotethesurvivalofdissociatedcerebellargranule neurons(Lindholmet al., 1993; Zirrgiebel et al., 1995). BDNF alsoreduceslow K~-inducedapoptosisin DIV 4 cere-bellar granuleneurons(Kubo et al., 1995;Shimokeet al., 1997). Changesin granuleneuronrespon-sivenessto neurotrophins(from BDNF to NT-3)during cerebellardevelopmenthavebeendescribed(Sega! et al., 1992, 1995), accompaniedby an in-creasedexpressionof TrkC relativeto TrkB (Sega!et al., 1995).Moreover,cerebellarlevelsof p75NTR

fall drasticallyafter the first week postnatal(Segalet al., 1995). p

7SNTR is suggestedto havea rolein regulating neurona! responsivenessto NT-415(Rydénet a!., 1995), and a diminishedexpressionof p75 NTR after DIV 4 might result in an earlierloss of granuleneuron sensitivity to NT-4/5. Theexpressionof high levelsof truncatedTrkB recep-tor isoformsin developingcerebellargranuleneu-rons (Armanini et al., 1995)could actas anegativeregulatorof BDNF signaling(Ninkina et a!., 1996),therebyleadingto a time-dependentloss of BDNFresponsiveness.It is unlikely that cultureagecom-promisedthe capacityof granuleneuronsto recog-nize BDNF and/or NT-4/5, becausecytotoxicityof the TrkB-selective fusion toxin DAB389-NT4(Negro andSkaper,1997) was blockedby BDNFand NT-415, but not NT-3 up to DIV 30.

NT-415 and BDNF were equipotentin rescuingDIV 1—2 cerebellargranuleneuronsfrom the dele-terious cytotoxic effects of glutamatecoupled toGSH depletion. Molecules structurally related toglutamate killed developing granule neurons,inparallel with the behaviorsobservedfor immatureembryoniccortical neurons(Murphy et a!., 1990).Glutamate-inducedoxidative stress to immatureembryoniccorticalneurons(Ratanet al., 1994)andto developingcerebellargranuleneurons(presentstudy) producesmorphologicalfeaturescharacter-istic of apoptosis.GSH depletionby glutamatewas

not sufficient in itself to causecell lysis, becausebothneurotrophinsandradical scavengerslike mel-atonin failed to reverse the glutamate-induceddepletionin GSH levels, although they protectedagainstneuronaldeath.Neurotrophinscanincreaseantioxidant enzyme activities and attenuateROSformation (Mattson et al., 1995; Dugan et al.,1997). Lindholm et al. (1993) found that BDNFreducedthe deathof DIV 4 cerebellargranuleneu-ronsexposedfor 24 h to 1 mM glutamate.Althoughthese latter authorsdid not describethe entity ofthis cytotoxicity, glutamatereductionof intracellu-lar cystinemay havebeena factor.

On binding to Trk receptors,neurotrophinsacti-vateseveraldownstreamintracellularpathways,in-cluding ras—MAP kinase and PT 3-kinase(Sega!and Greenberg, 1996). PT 3-kinase activity ap-pearedto be necessaryfor survival, becausetwoinhibitors of PT 3-kinase(LY 294002andwortman-nm) that act via distinct mechanismsblocked thesurvival-promotingability of BDNF, NT-4/5, andNT-3 for cerebellar granule neuronsin low-K~/serum-freemedium. Thesefindings are consistentwith observationsthat neurotrophinscan activateP1 3-kinase(Soltoff et al., 1992; Shimokeet al.,1997), andthat inhibition of P1 3-kinaseblocks thesurvival-promoting effects of NGF in PC12 cells(Yao andCooper,1995)andof BDNF againstlowK + -inducedapoptosisin cerebellargranuleneurons(Shimoke et al., 1997). It is surprising that LY294002 failed to affect neurotrophin protectionagainstglutamate-inducedoxidative stressin DIV1—2 granule neurons.Rather, the MEK inhibitorPD 98059provedeffective in blocking BDNF andNT-4IS saving of glutamate-challengedgranuleneurons.In PC12cells,MAP kinaseactivationpro-motes survival in the absenceof NGF (Xia et a!.,1995). However, MAP kinaseis not required forthe survival of NGF-maintainedsympatheticneu-rons (Creedonet al., 1996)or PCI2 cells (Yao andCooper,1995),althoughNGF acutelyblocks ROSformationin neuronsthroughactivationof the MAPkinasepathway(Duganet a!., 1997).

Removalof bothpotassiumandserumfrom dis-sociatedcerebellargranuleneuronstriggers a celldeaththat is morphologically apoptotic,accompa-nied by DNA fragmentation,and dependentonmacromolecularsynthesis(D’Me!lo et al., 1993).This apoptotic cell deathpresumablymimics thenaturally occurring death of 20—30% of granulecells(CaddyandBiscoe,1979),which is importantfor matchingthe numberof granulecells with Pur-kinje cells betweenweeks3 and5 postnatally(Wil-liams and Herrup, 1988). In cortical neurons,thesurvival-promotingeffects of K~are mediatedbyBDNF (Ghosh et al., 1994). That K~-dependentsurvivalof cerebellargranuleneuronsalso involvesspecific growth factors remains to be established.After submissionof our article, a reportby Shimoke

J. Neurochem.,Vol. 70, No. 5, 1998

1866 S. D. SKAPER ET AL.

et al. (1997)appeared,showingthat K”-dependentand BDNF-dependentsurvival of DIV 4 granuleneuronsare independentof eachotherbut are P1 3-kinasedependent.The generalityof this observa-tion will, however,needto be testedfor othercul-ture agesandneurotrophins.Althoughthe survival-promotingactionof high K + for cerebellargranuleneuronsis assumedto result from depolarization-induced Ca2~”influx (Gallo et a!., 1987), neuro-trophin-mediatedsurvival of granule neuronsap-pearsalso to involve increasesin intracellularCa2~(Zirrgiebel et al., 1995).Thesetwo pathwaysmayeachbe of physiologicalrelevance,as functions ofthe context in which they are expressed.The ob-served suppressionof free radical formation, viaTrkA activationandthe MAP kinasecascade(Du-ganet al., 1997), temptsspeculationthat the neuro-protectionaffordedby BDNF and NT-4/5 againstglutamate-inducedoxidative cell death involves amechanismanalogousto TrkB. Different apoptoticstimuli can invoke divergentdeathsignals, assug-gestedby the participationof distinct caspasesinapoptosisof cultured cerebellar granule neuronscaused by K~/serumdeprivation or by stauro-sporine(Tayloret al., 1997). It is, thus,notunrea-sonableto expectthat different downstreamtargetsof TrkB activation might interact with differentapoptotic stimuli, evenin the samecell type.

ROS havebeenimplicatedas mediatorsof acuteand chronic neurodegenerativedisorders(Brede-sen, 1995; Chan, 1996). Deficient suppliesof tro-phic support canleadto ROS formation (Schulzeta!., 1996; Dugan et al., 1997) andnervecell death(Levi-Montalcini, 1987; Lewin andBarde, 1996).In addition, structuralor metabolicinsultsthatcom-promise glutamatesequestration,especiallyearlyin developmentbefore glutamatereceptorslinkedto ion channelsare expressed,could result in im-pairedcystinetransport,oxidative stress,andneu-rona! degeneration.Our resultssuggestthatTrk sig-naling can affect CNS neuronal apoptosiswhereROS are an essential component.Neurotrophinsandrelatedgrowth factorsmayplay importantrolesin how the brainrespondstoinjury andneurodegen-erative processes(Mattson et al., 1993; Hefti,1997), which suggeststhat theseproteinsmay betherapeuticfor neurologicaldiseases(Hefti, 1997).Endogenoussmallmoleculeswith antioxidant/neu-roprotective properties, like melatonin (Giusti eta!., 1996; Floreani et a!., 1997b),may also repre-sent an important facet of the body’s defenseagainstcerebrovascularpathology (Manev et al.,1996).Collectively, thesefindingssuggestthat tro-phic factor—antioxidanttherapycould be a greaterbenefit than eitheragentalone.

REFERENCESAlessi D. R., CuendaA., CohenP., Dudley D.T., andSaltiel A. R.

(1995) PD 098059 is a specificinhibitor of the activationof

mitogen-activatedproteinkinasekinase in vitro and in vivo. J.Biol. Chem.270, 27489—27494.

Armanini M. P., McMahon S. B., SutherlandJ., SheltonD. L., andPhillips H. 5. (1995)Truncatedandcatalytic isoformsof trkBare co-expressedin neuronsof rat and mouseCNS. Eur. J.Neurosci.7, 1403—1409.

BredesenD. E. (1995)Neuralapoptosis.Ann. Neurol. 38, 839—851.Caddy K. W. and BiscoeT. 1. (1979) Structural and quantitative

studieson thenormalC3H andLurcher mutant mouse.Philos.Trans.R. Soc.Lond. 287, 167—201.

ChanP. H. (1996)Roleof oxidantsin ischemicbrain damage.Stroke27, 1124—1129.

ChaoM. V. andHempsteadB. L. (1995) p75 andTrk: atwo-recep-tor system.TrendsNeurosci.18, 32 1—326.

Cotman C. W. and Anderson A. J. (1995) A potential role forapoptosisin neurodegeneratmonand Alzheimer’sdisease.Mol.Neurobiol. 10, 19—45.

Creedon D. J., JohnsonE. M. Jr., and LawrenceJ. C. Jr. (1996)Mitogen-activatedprotein kinase-independentpathwaysmedi-ate theeffects of nervegrowth factorand cAMP on neuronalsurvival. J. Biol. Chem.271, 20713—20718.

Cull-Candy S. G. and Usowicz S. S. (1987) Multiple conductancechannelsactivatedby excitatoryamino acidsin cerebellarneu-rones.Nature 325, 525—528.

Davies A. M. (1994)Role of neurotrophinsin the developingner-vous system.J. Neurobiol. 25, 1334—1348.

D’Mello S. R., Galli C.,Ciotti T., andCalissanoP. (1993)Inductionof apoptosisin cerebellargranuleneuronsby low potassium:inhibition of deathby insulin-like growth factor I and cAMP.Proc. Nail. Acad. Sci. USA90, 10989—10993.

DuganL. L., CreedonD.J., JohnsonE. M. Jr., andHoltzman D.M.(1997) Rapid suppressionof free radical formation by nervegrowth factor involves the mitogen-activatedprotein kinasepathway.Proc. Nail. Acad. Sci. USA 94, 4086—4091.

FloreaniM., PetroneM., Debetto P., and Palatini P. (1997a) Acomparisonbetweendifferentmethodsfor thedeterminationofreducedand oxidized glutathionein mammaliantissues.FreeRadic. Res.26, 449—455.

Floreani M., Skaper S.D., Facci L., Lipartiti M., and Giusti P.(l997b) Melatoninmaintainsglutathionehomeostasisin kainicacid-exposedrat brain tissues.FASEB J. 11, 1309—1315.

Gallo V., Kingsbury A., Balázs R., and Jørgensen0. 5. (1987)The role of depolarizationin the survivaland differentiationofcerebellargranulecells in culture. J. Neurosci.7, 2203—2213.

GhoshA., CarnahamJ., andGreenbergM. E. (1994) Requirementfor BDNF in activity-dependentsurvival of cortical neurons.Science263, 1618—1622.

Giusti P., GusellaM., Lipartiti M., Milani D., ZhuW., Vicini S.,and Manev H. (1995) Melatonin protectsprimary culturesofcerebellargranuleneuronsfromkainatebutnot fromN-methyl-D-aspartateexcitotoxicity. Exp. Neurol. 131, 39—46.

GiustiP., Lipartiti M., FranceschiniD., SchiavoN., FloreaniM., andManevH. (1996)Neuroprotectionby melatoninfrom kainate-inducedexcitotoxicity in rats. FASEBJ. 10, 891—896.

Griffith 0. W. andMeisterA. (1979) Potentand specificinhibitionof glutathionesynthesisby buthioninesulfoximine(S-n-butylhomocysteinesulfoximine).J. Biol. C/tern. 254, 7558—7560.

Gunn-MooreF. J., Williams A. G., Toms N. J., and Tavaré J. M.(1997) Activation of mitogen-activatedprotein kinase andp7056kinaseis not correlatedwith cerebellargranulecell sur-vival. Biochern. J. 324, 365—369.

Hefti F. (1997)Pharmacologyof neurotrophicfactors.Annu.Rev.Pharmacol.Toxicol. 37, 239—267.

Ibáflez C. F. (1995) Neurotrophicfactors:from structure—functionstudiesto designingeffective therapeutics.TIBTECH13, 217—227.

KingsburyA. E., Gallo V., WoodhamsP. L., and BalázsR. (1985)Survival, morphologyandadhesionpropertiesof cerebellarin-terneuronescultured in chemicallydefinedand serum-supple-mentedmedium. Dev. Brain Res.17, 17—25.

Knecht K. T. and Mason R. P. (1993) In vivo spin trapping of

J. Neurochem.,Vol. 70, No. 5, 1998

NEUROTROPHINSPREVENTCNS NEURONALAPOPTOSIS 1867

xenobiotic free radical metabolites.Arch. Biochern. Biophys.303, 185—194.

Kubo T., NonomuraT., EnokidoY., andHatanakaH. (1995)Brain-derivedneurotrnphicfactor (BDNF) can preventapoptosisofrat cerebellargranuleneuronsin culture. Dev. Brain Res. 85,249—258.

Levi-Montalcini R. (1987) The nervegrowth factor 35 yearslater.Science237, 1154—1162.

Lewin G. R. andBarde Y.-A. (1996)Physiologyof theneurotroph-ins. Annu.Rev. Neurosci.19, 289—317.

Lin L.-F. H., Doherty D. H., Lile J. D., BekteshS., and Collins F.(1993)GDNF: a glial cell line-derivedneurotrophicfactor formidbrain dopammnergicneurons.Science260, 1130—1132.

Lindholm D., DechantG., HeisenbergC-P.,andThoenenH. (1993)Brain-derivedneurotrophicfactor is a survival factor for cul-turedrat cerebellargranuleneuronsand protectsthem againstglutamate-inducedneurotoxicity. Eur. J. Neurosci. 5, 1455—1464.

LindsayR. M. andYancopoulosG. D. (1996)GDNF in a bind withknownorphan:accessoryimplicatedin new twist. Neuron 17,57 1—574.

Manev H., Uz T., KharlamovA., and Joo J. Y. (1996) Increasedbrain damageafterstrokeorexcitotoxic seizuresin melatonin-deficientrats. FASEBJ. 10, 1546—1551.

ManthorpeM., FagnaniR., SkaperS. D., and Varon S. (1986) Anautomatedcolorimetric assayfor neuronotrophicfactors.Dev.Brain Res.25, 191—198.

Mattson M. P., Cheng B., and Smith-Swintosky V. L. (1993)Growth factor-mediatedprotectionfrom excitotoxicity anddis-turbancesin calciumand free radical metabolism.Semin.Neu-rosci. 5, 295—307.

Mattson M. P., Lovell M. A., FurukawaK., andMarkesberyW. R.(1995)Neurotrophicfactorsattenuateglutamate-inducedaccu-mulationof peroxides,elevationof intracellular ~ concentra-tion, and neurotoxicity andincreaseantioxidantenzymeactivi-ties in hippocampalneurons.J. Neurochem.65, 1740—1751.

McAllister A. K., Lo D. C., and Katz L. C. (1995) Neurotrophinsregulatedendriticgrowth in developingvisual cortex.Neuron15, 791—803.

Meakin S. 0. and ShooterE. M. (1992) The nervegrowth factorfamily of receptors.TrendsNeurosci.15, 323—331.

Miller T. M. and JohnsonE. M. Jr. (1996) Metabolic andgeneticanalysesof apoptosisin potassium/serum-deprivedrat cerebel-lar granulecells. J. Neurosci. 16, 7487—7495.

Miller T. M., TanseyM. G., JohnsonE. M. Jr., and CreedonD. J.(1997) Inhibition of phosphatidylmnositol 3-kinase activityblocksdepolarization-andinsulin-likegrowth factor-I-mediatedsurvivalof cerebellargranulecells. J. Biol. C/tern. 272, 9847—9853.

MonaghanD. T., BridgesR.J., andCotmanC.W. (1989)The excit-atory amino acid receptors:their classes,pharmacology,anddistinctpropertiesin thefunctionof the centralnervoussystem.Annu.Rev.Pharmacol.Toxicol. 29, 365—402.

MosmannT. (1983)Rapidcolorimetric assayfor cellulargrowthandsurvival: applicationto proliferationand cytotoxicityassays.J.Immunol.Methods65, 55—63.

MurphyT. H., SchnaarR. L.,andCoyle J. T. (1990)Immaturecorti-cal neuronsareuniquely sensitiveto glutamatetoxicity by inhi-bition of cystineuptake.FASEBJ. 4, 1624—1633.

NegroA. and SkaperS.D. (1997) Synthesisand cytotoxicprofileof a diphtheriatoxin-neurotrophin-4chimera. J. Neurochern.68, 554—563.

NegroA., CoronaG., BigonE.,Martini I., GrandiC., SkaperS. D.,andCallegaroL. (1991) Synthesis,purification, andcharacter-izationof humanciliary neuronotrophicfactor from F. co/i. J.Neurosci.Res.29, 251—260.

NegroA., TavellaA., Grandi A., and SkaperS.D. (1994)Produc-tion and characterizationof recombinantrat brain-derivedneu-rotrophicfactorandneurotrophin-3from insectcells. J. Neuro-chern. 62, 471—478.

Ninkina N., Adu J., FischerA., Pifión G. P., BuchmanV. L., and

DaviesA. M. (1996)Expressionandfunctionof TrkB variantsin developingsensoryneurons.EMBO J. 15, 6385—6393.

OppenheimR. W. (1991) Cell deathduring developmentof thenervoussystem.Annu.Rev. Neurosci.14, 453—50l.

Portera-CailliauC., Hedreen J. C., PriceD. L., and Koliatsos V. E.(1995)Evidencefor apoptoticcell deathin Huntingtondiseaseand excitotoxic animalmodels.J. Neurosci. 15, 3775—3787.

RabizadehS., Gralla E. B., Borchelt D. R., Gwinn R., ValentineJ. S.,SisodiaS., WongP., LeeM., HahnH., andBredesenD. E.(1995)Mutationsassociatedwith amyotrophiclateral sclerosisconvertsuperoxidedismutasefrom an antiapoptoticgene to aproapoptoticgene:studiesin yeastand neuralcells.Proc. Nail.Acad. Sci. USA 92, 3024—3028.

Raff M. C., BarresB. A., BurneJ. F., Coles H. S., Ishizaki Y., andJacobsonM.D. (1993)Programmedcell deathandthecontrolof cell survival: lessonsfromthe nervoussystem.Science262,695—700.

RatanR. R., Murphy T. H., and BarabanJ.M. (1994) Oxidativestressinducesapnptnsisin embryoniccortical neurons..1. Neu-rochern. 62, 376—379.

RiddleD. R., LoD. C., andKatzL. C. (1995)NT-4-mediatedrescueof lateralgeniculateneuronsfrom effectsof monoculardepriva-tion. Nature378, 189—191.

Roy N., MahadevanM. S., McLean M., Shutler G., Yaraghi Z.,FarahaniR., Baird S., Besner-JohnstonA., LefebreC., KangX., Salih M., Huguette A., Tamai K., Guan X., IoannouP.,CrawfordT. 0., deJong P. J., SurthL., IkedaJ.-E., KornelukR.G., and MacKenzie A. (1995) The gene for neuronalapoptosisinhibitory protein is partially deleted in individualswith spinal muscularatrophy.Cell 80, 167—178.

RydénM., Murray-Rust J., GlassD., Ilag L. L., TruppM., Yanco-poulos G. D., McDonald N. Q., and Ibáflez C. (1995)Func-tional analysisof mutantneurotrophinsdeficientin low-affinitybinding revealsa role for p75LNGFR in NT-4 signalling. EMBOJ. 14, 1979—1990.

SchulzJ. B., WellerM., andKlockgetherT. (1996)Potassiumdepri-vation-inducedapnptosisof cerebellargranuleneurons: a se-quential requirementfor new mRNA and protein synthesis,ICE-like proteaseactivity, andreactiveoxygenspecies.J. Neu-rosci. 16, 4696—4706.

SegalR. A. andGreenbergM. E.(1996)Intracellularsignalingpath-ways activatedby neurotrophicfactors.Annu.Rev. Neurosci.19, 463—489.

SegalR.A., TakahashiH., andMcKay R. D. G. (1992)Changesinneurotrophmnresponsivenessduringthedevelopmentof cerebel-lar granuleneurons.Neuron 9, 1041—1052.

SegalR.A., PomeroyS. L., andStiles C.D. (1995)Axonal growthand fasciculationlinked to differential expressionof BDNFand NT3 receptorsin developingcerebellargranule cells. J.Neurosci.15, 4970—4981.

ShimokeK., Kubo T., NumakawaT., Abiru Y., Enokido Y., TakeiN., IkeuchiT., and HatanakaH. (1997) Involvementof phos-phatidylinositol 3-kinase in prevention of low K~-inducedapoptosisof cerebellargranuleneurons.Dcv. Brain Res. 101,197—206.

SkaperS. D., Facci L., Milani D., Leon A., and ToffanoG. (1990)Cultureanduseofprimary andclonalneuralcells,in MethodsinNeurosciences,Vol. 2 (Cnnn P. M., ed), pp. 17—33. AcademicPress,San Diego.

SniderW. D. (1994)Functionsof theneurotrophinsduringnervoussystemdevelopment:what the knockoutsare teachingus. Cell77, 627—638.

Soltoff S. P., Rabin S. L., Cantley L. C., and KaplanD. R. (1992)Nervegrowth factorpromotesthe activationof phosphatidyl-inositol 3-kinaseandits associationwith the trk tyrosinekinase.J. Biol. Chern. 267, 17472—17477.

TapleyP., LamballeF.,andBarbacidM. (1992)K252ais a selectiveinhibitor of thetyrosineproteinkinaseactivity of the trk familyof oncogenesand neurotrophinreceptors. Oncogene7, 37 I —

381.

J. Neurochern.,Vol. 70. No. 5, 1998

1868 S. D. SKAPERET AL.

Taylor J., Gatchalian C.L., Keen G., and Rubin L. L. (1997)Apoptosis in cerebellargranuleneurones:involvement of in-terleukin-l/3 convertingenzyme-likeproteases.J. Neurochem.68, 1598—1605.

ThangniponW., Kingsbury A., Webb M., and BalázsR. (1983)Observationson rat cerebellarcells in vitro: influenceof sub-stratum,potassiumconcentrationandrelationshipbetweenneu-ronesand astrocytes.Brain Res.313, 177—189.

TietzeF. (1969) Enzymaticmethod for quantitativedeterminationof nanogramamountsof totalandoxidizedglutathione:applica-tions to mammalianblood and other tissues.Anal. Biochem.27, 502—522.

Williams R.W. and HerrupK. (1988)The control of neuronnum-ber. Annu.Rev. Neurosci.11, 423—453.

Xia Z., DickensM., RaingeaudJ., DavisR.J., andGreenbergM. E.(1995) Opposingeffects of ERK and JNK-p38 MAP kinaseson apoptosis.Science270, 1326—1331.

Yao R. andCooperG. M. (1995)Requirementfor phosphatidylino-sitol 3-kinase in the preventionof apoptosisby nervegrowthfactor. Science267, 2003—2006.

ZirrgiebelU., OhgaY., CarterB., BerningerB., InagakiN., ThoenenH., andLindholmD. (1995)CharacterizationofTrkB receptor—mediatedsignalingpathwaysin rat cerebellargranuleneurons:involvementof proteinkinaseC in neuronalsurvival. J. Neuro-chem. 65, 2241—2250.

J. Neurochem.,Vol. 70, No. 5, 1998