journal of Neurochemistry.Ipplncou Raven Publishers. Philadelphia

I 99h International Society for Neurochemistry

Cytokine-RegulatedExpressionof Platelet-DerivedGrowthFactorGeneandProtein in Cultured HumanAstrocytes

FrancescaCeccheriniSilberstein,RobertaDe Simone,Giulio Levi, and FrancescaAloisi

Laboratory of Organ and SystemPathophvsiology,Sectionsot Neurophvsiologv and Neurobiology,Istituto Superiore di Sanità, Rome, Ito Iv

Abstract: To elucidate mechanisms regulating the pro-duction of platelet-derived growth factor (PDGF) in theCNS, we analyzed the influence of a panel of cytokineson PDGF mRNA and protein levels in astrocyte-enrichedcultures from the human embryonic brain and spinal cord.Using a specific ELISA, PDGF AB protein was detectedin serum-free astrocyte supernatants and its levels weresignificantly increased after treatment of the cultures withtransforming growth factor-/31 (TGF-/31) or tumor necrosisfactor-a (TN F-a); the largest increase was detected aftercombined treatment with the two cytokines. Interleukin-1/3 (IL-1/3) by itself had little or no effect but synergizedwith TGF-/31 in enhancing PDGF AB production. Superna-tants from human astrocyte cultures stimulated the prolif-eration of rat oligodendrocyte progenitors, and most ofthe mitogenic activity could be accounted for by PDGF.By northern blot analysis, both PDGF A- and PDGF B-chain mRNAs were detected in untreated astrocytes.PDGF B-chain mRNA levels were increased by TGF-/31,TNF-cs, TNF-a/TGF-/31, or IL-1/3/TGF-/31 ,whereas PDGFA-chain mRNA levels were not consistently affected bycytokine treatments. These in vitro data indicate thatTGF-/31, TNF-a, and IL-1/3 are able to stimulate astrocytePDGF production. Thiscytokine network could play a rolein CNS development and repair after injury or inflamma-tion. Key Words: Development—Glia—Inflammation—Transforming growth factor—Tumor necrosis factor—In-terleukin-1J. Neurochem. 66, 1409—1417 (1996).

Platelet-derivedgrowthfactor(PDGF)is apleiotro-pic cytokinethatactson manydifferentcell types(e.g.,fibroblasts,smoothmusclecells,endothelialcells,neu-ral cells) and is ableto stimulatea varietyof cellularresponses,including proliferation, chemotaxis, actinreorganization,and Ca

2~mobilization (reviewed byHeldin and Westermark,1990; Raineset al., 1990).Structurally,PDGFis a 30-kDadimer of two homolo-gous disulfide-linked polypeptide chains (A and Bchains)thatcan combineto form PDGFAA, AB, andBB (reviewedby Raineset at., 1990; 1-leldin, 1992).The variousresponsesto PDGF areelicitedafterPDGFbinding to high-affinity receptors,which themselves

dimerizein thepresenceof ligandandactivateintrinsicprotein tyrosinekinase activity. Two types of PDGFreceptorsubunits,termeda and /3, havebeencharac-terized; whereasthe PDGF a-receptorsubunit bindsall threePDGFisoforms,the PDGF/3-receptorsubunitbindsonly PDGFBB (Heldin andWestermark,1990).

The availableexperimentalevidenceindicatesthatneuronalandglial cells canactboth as a sourceof’ andas a targetfor PDGF, and that PDGFis implicated inseveralphysiologicalandpathologicalprocessesof theCNS. In rodents and nonhumanprimates,PDGF A-and B-chain mRNAs andproteins are presentin neu-rons of the developingandadultCNS (Sasaharaci al.,1991; Yeh et al., 1991), whereasrodent astrocytesexpressmainly PDGF A chain in vitro and in vivo(Richardsonetal., 1988;Mudharetal., 1993).Studieson gliogenesisin the developingrat optic nerve havedemonstratedarole for PDGF in regulatingtheprolif-eration,survival, andchemotaxisof cells belongingtotheoligodendrocytecell lineage (Noble et al., 1988;Raff et al., 1988; Richardsonci al., 1988: Barres cial., 1992, 1993). OligodendrocyteprogenitorsexpressPDGF a receptorsandrespondto all threePDGF iso-forms (Pringleci at., 1989, 1992; Barres et al., 1993;Mudharetal., 1993), whereasPDGF /3 receptorshavebeen detected on differentiated oligodendrocytes(Grinspanci al., 1990). PDGF /3 receptorsare alsoexpressedby rat brain neurons in situ and, at least invitro, PDGFBB is ableto stimulatesurvival andneu-rite outgrowthof distinctneuronalpopulations(Smitset al., 1991, 1993; Nikkhah et al,, 1993).

ReceivedAugust 3. 1995;revisedmanuscript receivedNovember9, 1995; acceptedNovember9, 1995.

Addresscorrespondenceand reprinl requests to Dr. F. Aloisi atNeurophysiologyUnit, Laboratoryof OrganandSystemPathophysi-ology. Istituto Superioredi Sanifa. Vialc ReginaElena 299, 00161Rome. italy.

Abbreviationsused:bFGF; basic fibroblasl growth factor; hUE.epidermal growth factor; 1FN-y. interferon-y; IL. interleukin;PDGF, platelet-derivedgrowth factor; [‘MA. phorbol I 2-myristate13-acetate;rh, recombinanlhuman; SDS, sodium dodecyl sulfate;SSC,saline—sodiumcitrate; TGF—/3, transforminggrowth factor—/i;TNF—a, tumor necrosisfactor-a.

1409

1410 F. CECCHERJNJSILBERSTEINEl’ AL.

Besides its involvement in CNS development,PDGF could play a role in inflammatory and repairprocessesthatareassociatedwith neurologicalinsults.In theexperimentalanimal, increasedlevelsof PDGFprotein havebeendetectedafter mechanicalinjury totheoptic nerve (Lotan andSchwartz, 1992), whereasPDGF B-chain mRNA was induced in neurons andbrain macrophagesby focal brain ischemia(lihara etal., 1994). In human cerebral abscesses(Liu et al.,1994) PDGF immunoreactivityhas beendetectedinreactiveastrocytes.The aboveresults indicate that lo-cal PDGFsynthesismay be stimulatedby lesion-asso-ciated signals..PDGF hasalso beenimplicated in thepathogenesisof brain tumors.Malignant glioma tissueandcell lines expressboth PDGF A andB chainsandPDGF receptors(Nistér et al., 1988; Maxwell et al.,1990;Hermansonetal., 1992), supportingthe notionthat autocrinePDGF stimulation may be involved inpromoting tumor growth.

Elucidationof the mechanismsthat regulatePDGFA- and B-chain expressionin the CNS may help todefine more clearly thepleiotropiceffectsof this cyto-kine in normal, injured, andmalignantCNS tissue.Inthis study,we investigatedthe influenceof a panel ofinflammatory cytokinesand growth factors on PDGFmRNA and protein levels in astrocyte-enrichedcul-tures obtainedfrom the human embryonic brain. Ourresults indicate that a restricted panel of cytokines,which includeTGF-/3, IL-l,8, andTNF-a, areabletoenhancePDGFB chain—specifictranscriptsandsecre-tion of biologically active PDGF AB in humanastro-cytes, suggestingpossiblemeansfor increasingPDGFavailability in brain developmentandpathology.

EXPERIMENTAL PROCEDURES

MaterialsThe following cytokines were used;recombinanthuman

(rh) interleukin-l/3 (IL-l/3), rh IL-6, rh epidermalgrowthfactor (EGF), and rh interferon-y (IFN-y) from JanssenBiochimica; purified humantransforminggrowth factor-/3,(TGF-/31) and rh tumor necrosisfactor-a (TNF-a) fromR&D Systems;rh IL-tO from PeproTech;rh IL-4 from Gen-zyme;rh IL-13, a kind gift of R. de Waal Malefyt, DNAXResearchInstitute,Palo Alto, CA, U.S.A.; fibroblastgrowthfactor(FGF) from bovine pituitary andhumanPDGF fromSigma. Phorbol 1 2-myristate13-acetate(PMA), dibutyrylcyclic AMP, forskolin, actinomycin D, and cycloheximidewere all from Sigma. Antibodiesusedwere rabbit anti-hu-manTNF-aandrabbitanti-humanIL-l/3 polyclonalantibod-iesfromGenzyme;chickenanti-humanTGF-/31,rabbitanti-humanTGF-/3S,andgoatanti-humanPDGFpolyclonalanti-bodies,all purified IgG fractionsfrom R&D Systems.

Cell culturesSixteenhumanembryosof 8—9 weeks of age were ob-

tainedfrom legalcurettageabortionsafterpatientsgavefullyinformed consent. Independentcultures were establishedfrom oneor morebrain areas(myelencephalon,mesenceph-alon, or spinal cord), dependingon theintegrity of theem-bryo (Aloisi et al., l992a). In brief, brain specimenswere

carefullyfreed of’ meninges,choppedinto small pieces(0.4mm thick), and dissociatedinto single-cell suspensionsbya mild trypsinii.ation procedureand gentlemechanicaldis-ruption with aPasteurpipette.Thecells wereresuspendedinminimal essentialmediumcontainingo-valine(GihcoBRL).andsupplementedwith 10 pg/mI gentamicin, 2 mM gluta-mine (Sigma),and 10% fetal calfserum(GibcoBRL); 1.5—2.0 x 10’ cells/cm: were seededon poly-i.-lysine—coatedculture dishesand grown at 37°Cin a 95% air/5% CO:humidifiedatmosphere.In this study, we usedtertiary cul-turesthat contained75—90%glial fibrillary acidic proteincells, <2% fibronectin tibrohlasts, <1% neurofilament-neurons,8—18%vin1entin~neuralprogenitors,no CDI4CDI Ic~microglia, andno 04’ oligodendrocytes(Aloisi etal., l992a). Culturesenrichedin fibronectin~leptomenin-geal cells were establishedand maintainedas previouslydescribed(Aloisi et al., 1992/,).To avoid interferencewithserumfactors, 24 h before cytokinetreatments,the culturemediawere replacedwith chemically defined, serum-freemedium containingthe N2 supplements(BottensteinandStun,1979). Astrocytecultures(--3 X 10: cells/mI/well in12—well Nunc platesor 6 >< I 0 cells/2 nil/35-mm-diameterdish) were treatedfor 4. 8. 24. or 48 h with variousagents.At theendof treatment,supernatantswerecollected,centri-fuged,andstoredat —80°Cuntil use.

Determination of PDGF proteinLevelsof PDGFAB protein in astrocytesupernatantswere

determinedusing a commercially available ELISA that isspecific for PDGFAB anddetectsaslow as 8.4 pg of PDGFAB/mI (R&D Systems). In this ELISA kit a monoclonalantibodyspecific for PDGF AA is used for coating and ahorseradishperoxidase—conjugatedpolyclonalantibodyspe-cific for PDGF BB is used for detection, thus allowingmainly recognitionof human PDGF AB; accordingto themanufacturer,thereis only minimal cross-reactivitywith rhPDGF AA andrh PDGF BB (10 and2%. respectively). rhPDGFAB wasusedas a standard.

Proliferation and neutralizing antibody assaysSecretionof biologically active PDGF was evaluatedby

the ability of astrocytesupernatantsto stimulate the prolifer-ation of oligodendrocyteprogenitors.Oligodendrocytepi’o-genitorswere purified from primary cultures of neonatalratcerebral cortex, as previously described (Agresti et al.,1991), and seeded(1.25—1.5 x l0~cells/well in 96-wellNunc plates)in serum-containingmediLim. After 4 h. themedium was replacedwith serum-free,chemicallydefinedmedium (in the absenceor presenceof human PDGF orcytokines),or with serum-freesupernatantsfrom humanastrocyteculturesthathadbeenmaintainedfir 48 h in theabsenceor presenceof cytokines.In someexperiments,con-trol mediaandastrocytesupernatantswere divided in twoaliquots, oneof which was incubatedfor 30 mm at roomtemperaturewith neutralizinggoat anti-humanPDGF poly-clonal antibody (10 pg/mI); at the end of the incubationperiod, all mediumfractionswere addedto the oligodendro-cyteculturesandassayedin triplicate for mitogenic activity.After 24 h, ~H1methylthymidine (70—85Ci/mmol; Amer-shamInternational;0.5 pCi/well) was addedfor 16 h. Cellswere then harvestedon fiber filters using a 96-well plate,automatedcell harvesterandthe radioactivity incorporatedwasdeterminedby a microplatescintillationcounter(Top-count,PackardInstrumentCo.). In someexperiments,oligo-dendrocytegrowth was evaluatedusing the MTT staining

J. Neuroche,n.,Vol. 66, No. 4, 1996

STIMULATION OF GLIAL PI)GF SYNTHESIS 1411

technique(Mosmann, 1983). In other experiments,humanastrocytes(0.5—I X lO~cells/well in 96-well Nunc plates)were grown for 48 h in the absenceor presenceof humanPDGFor neutralizinganti-PDGFantibody,andtheprolifera-tion wasassayedby [3Hlthymidine incorporation,asabove.

Northern blot analysisTotal cellular RNA wasextractedfrom astrocytes(1.5—

2.0 x 10’ cells/ISO-mm-diameterdish) maintainedfor 8and/or24 h in the absenceor presenceof cytokines,by theguanidineisothiocyanate/cesiumchloride method; 15 pg ofRNA wasrunon agarose-formaldehydegels andtransferredto nylon filters by northerncapillary blot (Thomas, 1980).Thefilters werehybridizedin sequencewith 10: cpm of [a-52P 1 dCTP-labeledfragmentsof eDNA specific for PDGFA(1.3 kb) andB chain (2.0 kb) (a kind gift from Dr. Funa,Ludwig Institute for Cancer Research,Uppsala, Sweden;Hermansonet al., 1988). The hybridizationwasperformedin a solution containing50%formamide, 1 M NaCI, 1.25%sodiumdodecylsulfate(SDS),10% dextransulfate,and0.5mg/ml of salmonspermDNA for 18 h at 42°C.The filterswere washedtwice (2 >< 15 mm) in 2X saline—sodiumcitrate (SSC)/O.l% SDS buffer at room temperature,andonce for 30 mm in O.Ix SSC/O.t% SDS at 65°C.Afterremoval of the last probe,filters were rehybridizedto anoligonucleotideprobefor human28SribosomalRNA(Clon-tech). Theautoradiographicbandswere quantifiedby usingeither a laserdensitomer(UltroscanXL, PharmaciaLKB)or a high-resolutionscanningTM425F Phosphor imager(MolecularDynamics).EachPDGF A- and B-chain bandwas normalizedby using the value for the corresponding28S ribosomal RNA, thus correcting for any variation inamountsof RNA appliedto each lane.

RESULTS

Effects of cytokines on the secretion of PDGF ABprotein by human astrocytes

Thelevelsof PDGFproteinin astrocytesupernatantswereevaluatedusinga commerciallyavailableELISAthat is specific for PDGF AB. As shown in Table I,supernatantsfrom eight independentastrocyteculturescontainedvariable amountsof PDGF AB. In all cul-tures examined, the addition of TGF-/3

1 (0.1, 1, andtO ng/ml) increasedPDGFAB levelsby 1.6—2.6-foldoverbasallevels; themaximal effectwasobtainedwith1 ng/ml TGF-/31 (Table I). TNF-a (1, 10, and 100ng/ml) also elicited a 1.5—2.7-foldincreasein PDGFAB levels in mostcultures analyzed;statistically sig-nificant effectswere observedat the dose of 10 ng/ml (Table I). Additive or slightly more thanadditiveeffects were observedafter simultaneoustreatmentswith TGF-/31 andTNF-a (Table I). Cytokine-stimu-lated increasesof PDGF AB werequantitativelysimi-lar whetherthe amount of PDGF AB was calculatedas picogramsper milliliter of culture medium or aspicogramsper milligram of protein, indicating thatcy-tokine effectswerenot dueto changesin theastrocytenumbers(results not shown from three independentexperiments).

IL-l/3 (10, 50 U/ml) causeda 1.5-fold increasein

PDGF AB in only two of six cultures tested; whenaddedsimultaneously,TGF-/31 andIL-I/3 had eithersynergisticor additive effectsin four of five culturesexamined(Table I).

Kinetics studiesshowedthatthelevelsof PDGFABwere increasedaftera24-h exposureto TGF-/31,TNF-a, TNF-a/TGF-/31,andIL-l/3/TGF-/31 andcontinuedto increaseduring the48-hincubationperiod (Fig. I).

The increasein PDGF AB secretionwascompletelyabolished when astrocyteswere exposed simultane-ously to TGF-/31 or TNF-a, andto cycloheximide(10~tg/ml) or actinomyein D (2 ,ug/ml) for 18—24 h,indicating that protein and RNA synthesis wererequired for PDGF stimulation (results not shownfrom two independentexperiments).Neitherinhibitorcausedsignificantcell death,asjudgedby microscopicinspectionand stainingof astrocytecultureswith try-panblue.

PDGF AB levelswere not increasedafter treatmentwith the T cell—derivedcytokine IFN-y (100 U/mI),which was previouslyshownto synergizewith IL-1/3andTNF-a in stimulatingastrocytereleaseof IL-6 andgranulocytecolony—stimulatingfactor (Aloisi et al.,l992b). IL-b (2 and20 ng/ml). apotent inhibitor ofcytokineproduction in macrophages(deWaal Malefytet al., 1991), did not affect either basal or cytokine-stimulatedPDGF release.No increasein PDGF ABwasdetectedafter treatmentwith IL-4 (100, 1,000U/ml), IL-6 (100, 1,000 U/mI), IL—l3 (5, 50 ng/ml),basicFGF(bFGF) (10, 100 ng/ml), andEGF(10, 100ng/ml). With respectto thesignalpathwayinvolved inPDGF expression,no evidenceemergedfor an activa-tion of either protein kinase A or protein kinase C,becausedibutyryl cyclic AMP (I mM), forskolin (10

tiM), andPMA (10, 100nM) were all unableto affectPDGF AB levels.

Becauseculturedhuman astrocytessecreteboth ac-tive andlatentTGF-/3 (mainly TGF-/32) (Mcml et al.,1994), we testedwhether basal PDGF productioncould dependon an autocrine TGF-/3 loop. Whenaddedto the cultures for 48 h, neither neutralizingrabbitanti-humanTGF-/3, (10,20pg/mI) noracombi-nation of anti-TGF-/32and chicken anti-humanTGF-

/3 (10, 20 ,ug/ml) antibodieshad any effect on thelevels of secretedPDGF AB. Treatmentswith neu-tralizing anti-human IL-I/I and anti-humanTNF-awerealso ineffective.According to our previousstud-ies (Aloisi et al., 1992b, 1995),culturedhumanastro-cytes produce only minimal amounts of IL-I/I,whereasneither TNF-a protein nor TNF-a biologicalactivity hasbeendetectedin astrocytesupernatants.

The possibility that cells other thanastrocytescon-tributed in a significant way to PDGF AB releaseinourcultureswasexcludedby thefinding that very lowlevelsof PDGF AB protein were detectedby ELISAin serum-freesupernatantsof 1-week-oldprimary cul-tures from the human embryonic brain, which con-tainedmainly neuronsand neural progenitors(Aloisiet al., 1992a); 2.5 X lO~cells/2 ml/35-mm-diameter

j. Neurocheni.,Vol. 66, No, 4, 1996

1412 F. CECCHERINI SILBERSTEINFT AL.

TABLE 1. Cytokinestimulation of’ PDGFARsecretionin humanastroc’vtecultures

PDGP AB (pg/mi)p values ss.

Cell treatmenl° Exp. I Exp. 2 Exp. 3 Exp. 4 Exp. 5 Exp. 6 Exp. 7 Exp. 8 controls’

Control 140 155 236 300 350 300 240 380TGF-~I

0.1 ng/ml ND 266 356 NI) 472 ND 380 ND <0.001I ng/mI 370 344 392 520 585 670 450 690 <0.0010 ng/inl ND ND ND ND 545 ND 450 ND

TNF-aI ng/ml ND 308 448 280 425 NI) 300 NDID ng/ml 380 412 470 270 398 400 350 660 <0.01100 ng/ml ND ND NI) ND 500 Nt) 500 ND

TNF-cu 1 ng/ml + TGF’-/~, I ng/ml ND ND 671 700 680 ND 540 ND <0.01TNF-a 10 ng/ml + TGF~8,I ng/ml 980 874 643 1,050 900 715 660 .320 <0.001lL-l~

10 U/mI 160 ND ND 250 285 300 290 54050 U/nil ND 269 NI) ND ND ND ND 545

1L-l~10 U/mI + TGF-~1I ng/nil 720 NI) ND 760 615 ND ND 930 <0.01IL-l/3 50 U/mI + TGF-/~,I ng/ml ND 515 ND ND ND NI) ND 1,050

Exp., experiment;ND. not determined.° Astrocyte culturesobtainedfrom the humanembryonicmyelencephalon/mesencephalonor spinal cord wei’c tncuhatedfor 48 h mu serttmn-

free medium with or without cytokines.Supernatantswere collectedand analyzedfor PDUF AB levels using a specific ELISA. Data fromeight experiments,eachperformedon a different culture, are shown.

“Statistical significancewasevaluatedby pairedStudent’st test.

culture dish were seededand <45 pg PDGF AB/mIwere detectedeven after 48-h treatmentswith TGF-/3, TNF-a, or IL-I/I. PDGFAB wasundetectableinserum-freesupernantantsof unstimulatedand cyto-kine-stimulatedhumanleptomeningealcell cultures.

Effect of PDGF releasedby human astrocytesonthe proliferation of rat oligodendrocyteprogenitors

To determinewhether the increasein astrocytePDGFreleaseelicited by cytokineshadany functionalrelevance,we testedthe ability of astrocytesuperna-tants to stimulatethe proliferation of highly enrichedrat cortical oligodendrocyte progenitors [90—95%GD3-and04-expressingcells,aspreviouslydescribed(Agresti et al., 1991)]. Forty-eight-hoursupernatantsfrom six of 11 culturestestedcauseda 25—100%in-creaseof [‘H]thymidine incorporationinto rat oligo-dendrocyte progenitors. All the mitogenic activitycould be attributedto PDGF, as it was entirely sup-pressedby 30-mm preincubationof astrocytesuperna-tantswith 10 ~g/ml anti-humanPDGFpolyclonal anti-body; at this concentration,the antibody neutralizedthe proliferativeeffect of human PDGF(Fig. 2), butnot that of bFGF (not shown),a potent oligodendro-cyte mitogen(Bogler et al., 1990).

Figure 2 also shows that the mitogenic activity ofsupernatantsfrom TGF-/I1 —, TNF-a—, TNF-a/TGF-/I~—, and IL-I/I/TGF-/I1—treated, but not IL-I/I—treated,astroeytecultureswasgreaterthan that of su-pernatantsfrom untreatedsister cultures. This effectwas detectableonly when the supernatantsfrom un-treatedcultureswerepoor stimulatorsof oligodendro-

cyte proliferation: significant differencesbetweenthemitogenic activity of untreatedand cytokine-treatedastroeytesupernatantswerenot detectedwhen thekr-mer already stimulated [‘H] thymidine incorporationby >50%.Anti-PDGF antibodiesfully or largely neu-tralized the increasein [‘H]thymidine incorporationelicited by cytokine-treatedastrocytesupernatants,in-dicating that most of the mitogenic activity could heaccountedfor by PDGF (Fig. 2). To excludethat thecytokines themselvescould modify oligodendrocytegrowth, the influence of TGF-/I1 (I ng/ml), TNF-a(10 ng/ml), IL-I/I (10 U/mI), TNF-a/TGF-/I1, andIL-l/I/TGF-/I1 was evaluatedafter a 48-h treatmentperiod, using I’H]thymidine incorporation (Fig. 3)and MTT reduction(not shown) assays.Both in theabsenceand presenceof PDGF (I ng/mI). none oftheabovecytokineswasableto affecttheproliferationof oligodendrocyteprogenitors.

Incorporation of [‘H]thymidine into humanastro-cyteswasnot affected by 48-h treatmentswith exoge-nousPDGF (I, 10 ng/mI); in addition, blockadeofendogenousPDGF with anti-PDGFneutralizinganti-body (10 pg/mI) did not affect astrocyteprolifera-tion, indicating that in our experimental conditionsPDGFwas not mitogenic for humanembryonicastro-cytes.

Elevation of PDGF B-chain mRNA, but not ofPDGF A-chain mRNA, in cytokine-treatedastrocytes

The aboveresults indicatedthat a limited panel ofcytokineswereableto enhanceastrocytePDGFproteinsecretion;however, they did not allow to determine

j. Neurochen,.,Vol. 66, No. 4, 1996

STIMULATION OF GLIAL PDGF SYNTHESIS 1413

FIG. 1. Kinetics of PDGF AB production. Human astrocytes were incubated for varying times (4—48 h) in serum-free medium, in theabsence (control) or presence of TGF-

131 (1 ng/mI), TNF-a (10 ng/ml), TNF-cs/TGF-/1i (left), IL-1/3 (10 U/mI), or lL-1/<TGF-~31(right).PDGF AB levels in the culture supernatants were determined by specific ELISA. Mean + SEM values from three independent experimentsare shown.

whetherexpressionof PDGFA chain, PDGFB chain,or both types of chains wasaffected.To elucidatethispoint, we performed northern blot analysisof totalRNA extractedfrom untreated andcytokine-treatedas-trocytes. using eDNAprobesspecific for human PDGFA and B chain.Experimentswereperformedin sevenindependent cultures(five from myelencephalon/mes-encephalonand two from spinal cord). Both PDGFA- and B-chainmRNAs were expressedat variablelevels in all culturesexamined;in two of sevencul-tures, PDGF B-chain mRNA was barely detectable(Fig. 4A, C). Three bandsof 2.8, 2.3,and 1.9 kb andone band of 4.0 kb were detected forPDGF A- andB-chain mRNAs,respectively.TGF-/I1 increasedthelevelsof PDGF B-chain mRNA by I .9—5.3-fold (Fig.4A—C). The increasein PDGF B-chain mRNA die-

FIG. 2. Mitogenic responses of rat oligodendro-cyte progenitors to supernatants from humanastrocyte cultures. Control serum-free mediawithout or with PDGF and 48-h serum-free super-natants from untreated and cytokine [TGF-/11 (1ng/ml), TNF-ce (10 ng/ml), lL-1~3(10U/mI), TNF-a/TGF-/31, IL-i /1/TGF-/11 J-treated astrocyte cul-tures were incubated for 30 mm at room tempera-ture in the absence or presence of goat anti-hu-man PDGF antibody (10 pg/mI) and then addedto freshly isolated rat oligodendrocyte progeni-tors. Proliferation of oligodendrocyte progenitorswas evaluated by [‘H]thymidine incorporation asdescribed in Experimental Procedures. Means+ SEM values from three independent experi-ments are presented. The mean it SEM value ofcpm incorporated in control cultures was 7,169+ 680. “The mean value from two experiments ispresented.

ited by TNF-a rangedbetween2.1- and3.4-fold (Fig.4A). IL- I /3 waseitherineffectiveor causedonly small(I .5—2.0-fold) increasesof PDGF B-chain mRNA(Fig. 4A, C). In most cultures,combined cytokinetreatments (TNF-a/TGF-/I1 or iL-l/I/TGF-/I1) en-hancedPDGFB-chain mRNAto a greaterextent thansinglecytokine treatments(Fig. 4). Dependingon theculture, the stimulationof PDGFB-chainmRNA waseither similarafter 8- and 24-h cytokinetreatments(asin theexperimentshownin panelA of Fig. 4, in whichonly results from 24-htreatmentsare presented)orless prominentat 24 h (as in the experimentshown inpanel B of’ Fig. 4). No consistentmodificationsin thelevelsof PDGF A-chain transcriptsweredetectedafter8- and24-htreatmentswith theabove cytokines.aloneor in combination(Fig. 4A, B). Similar results were

.1. Neu,’,,,’l,enr, Vol 66, No. 4, 1996

1414 F. CECCHERINI SILBERSTEINET AL.

FIG. 3. Mitogenic responses of rat oligodendrocyteprogenitors to cytokines in the presence or absenceof PDGF. Oligodendrocyte progenitors were grown inthe presence of cytokines for 48 h and [

3H]thymidineincorporation was measured as described in Experi-mental Procedures. Mean ± SEM values from threeindependent experiments are presented. “The meanvalue from two experiments is presented.

obtainedwith cultures derivedfrommesencephalonor spinal cord.

DISCUSSION

myelencephalon/

In the adult humanbrain, normal astrocytesdo notexpressPDGF A or B chain (Maxwell et al., 1990;Hermanson etal., 1992).However,PDGFimmunore-activity hasbeendetectedin reactiveastrocytesin non-tumoral brain abscesses(Liu et al., 1994), andbothPDGF A andB chainsare expressedin human glialbraintumorsin situ and invitro (Maxwell et al., 1990;Hermansonet al., 1992). No dataare available onastrocytePDGF expression duringhuman CNSdevel-opment or inother brainpathologies.Studieshave alsobeen lacking on thesolublefactorsregulatingastroglialPDGFsynthesisand release.

The presentin vitro studiesindicatethat astrocytesfrom the humanembryonicbrain andspinalcordex-pressboth PDGFA- and B-chain mRNAsand releasePDGFAB dimers in abiologically activeform in basalcultureconditions.Constitutive PDGFexpressionandreleasecould be related to theembryonicderivationof the cells or tootherstimuli inherentto the cultureconditions. In this respect,human astrocytesbehavelike astrocytesderived from the neonatal rat brain,which were previously shown to secretebioactivePDGFin culture(Raff et al., 1988;Richardsonet al.,1988).However, speciesdifferencesmay explain thatrat astrocytes synthesizeand secretepredominantlyPDGF AA (Richardsonet al., 1988; Pringle et al.,1989).

Similar to what wasobservedfor nonneuralcelltypes (e.g.,endothelialcells,fibroblasts,macrophages)(Starksenet al., 1987;Raines et al.,1989;Shaw etal.,1991), we showherethat a limited panelof cytokinesareable to stimulatePDGF synthesisand secretioninhumanastrocytes.The levels ofPDGFB-chain mRNAincreasedconsistentlyafter exposureof astrocytesto

TGF-/I1 or TNF-a, but notto IL-I/I; the higheststimu-

lations were observed aftercombinedcytokine treat-ments (TNF-a/TGF-/Ij and IL-l/I/TGF-/I1). Thefinding that within the same cultures only PDGFB-chain, but notPDGF A-chain, mRNA accumulatedafter stimulation with the above cytokines providesevidencefor a noncoordinateregulation of the twoPDGFgenes.Despitethe preferential accumulationofPDGF B-chain transcripts, at theprotein level cyto-kines causeda consistent increasein PDGF AB; thismust result from the dimerization of newly formedPDGF B with constitutively producedPDGFA chain,althoughwe cannotexcludethe existenceof aposttran-scriptional regulationof PDGF A chainby cytokines.As in the case ofPDGF B-chain mRNA, PDGF ABsecretion was maximally stimulatedby combinedTNF-a/TGF-/I1 and IL-1/I/TGF-/I1 treatments, andTGF-/3 or TNF-a was more effective than IL-I/I.Othercytokinesand growth factors tested(IFN-y, IL-4, IL-6, IL-b, IL-13, EGF, andbFGF) did not affectPDGFAB protein secretion.By using an ELISAspe-cific for PDGF AB, thetotal amount of PDGF proteinreleasedby astrocyteswas probably underestimated,as no information could beobtainedon the productionof PDGF AA and PDGFBB. However, according topreviousstudies, only a minimal part of PDGF BBis releasedextracellularly, whereasthe major partisretained intracellularlyor in a cell surface—associatedform (Robbinset al., 1985; Ostmanet al., 1992).

Previousstudies in thedevelopingrat optic nervehaveshown that cells of theoligodendrocytecell lin-eageexpressmainlyPDGFa receptors (Pringleet al.,1992; Mudhar et al., 1993) and that PDGF AA andPDGFAB are more potentthan PDGFBB in support-ing survival and inducingproliferationof oligodendro-cyte progenitors (Pringle et al., 1989; Barres et al.,1993). In the developingrat optic nerve, astrocytesexpressthePDGFA-chain mRNA (Richardson etal.,1988;Mudharet al., 1993). This is true alsofor neona-

J. Neurochc’m.,Vol. 66, No. 4, /996

STIMULATION OF GLIAL PDGF SYNTHESIS 1415

FIG. 4. Levels of PDGF A- and B-chainmRNAs in untreated and cytokine-treated astrocytes. Astrocyte culturesderived from the human embryonicmyelencephalon/mesencephalon wereincubated for 24 h (A), 8 and 24 h (B),or 8 h (C) with cytokine-free medium(C, control), TGF-/3, (1 ng/ml), lL-i~l(10 or 50 U/mI), TNF-cs (10 ng/ml),TNF-a (10 ng/ml)/TGF-/~

1(1 ng/ml),or lL-i~3(10 U/ml)/TGF-/31 (1 ng/mI).Fifteen micrograms of total cellularRNA was analyzed by northern blot-ting; the filters were hybridized in se-quence with “P-labeled cDNA specificfor PDGF A or B chain. Normalizationwas performed with “P-labeled oligo-nucleotide probe for 28S RNA. Resultsfrom three independent experiments,each performed on a different culture,are shown. The histograms on the rightside of the figure show values in arbi-trary units of PDGF A- and PDGF B-chain mRNA (normalized to 28S RNA),as obtained from densitometric analy-sis of the corresponding autoradio-grams.

tal rat cerebralcortex astrocytes,which were shownto secretePDGF into the culture medium (Noble Ct

al., 1988;Raff et al., 1988; Richardsonet al., 1988).BecausethesecretedPDGFwas highly mitogenic foroligodendrocyteprogenitors,it hasbeenproposedthatastrocyte-derivedPDGF may playa crucial role in thecontrol of myelination in the developingCNS. Ourpresent experimentsshow that in basal conditionsaboutone-halftheastrocyteculturesof humanembry-onic derivation released sufficient amountsof PDGF(most likely PDGF AB) to stimulatethe proliferation

of rat oligodendrocyteprogenitors.In addition,at leastin somecultures,thePDGF-relatcdmitogenic activitywas enhancedafter treatmentwith TGF-/I1, TNF-a,TNF-a/TGF-/I1, or IL-I/I/TGF-/I1, suggestingthatthis cytokine network could be functionttlly relevantfor oligodendrocyte development.

It has been reportedthat neuronsand endothelialcells expressfunctional PDGF/3 receptors.and thatPDGF BB influencesthegrowth of brain capillaryves-sels and enhancessurvival and neurite outgrowth ofhumanandrat CNS neurons(Smitset al., 1989, 1993;

.1. Neuro, Ii,’,,,. . Vol. 66, No. 4, /996

1416 F. CECCHERINI SILBERSTEINET AL.

Nikkhah et al., 1993).Our finding that PDGF B-chainmRNA is preferentiallystimulatedin culturedhumanastrocytesraisesthe possibility that in vivo neuronsandendothelialcellsrepresentamajortargetfor PDGFBB producedby cytokine-activatedastrocytes.

The datapresentedhere suggestthat IL-I. TNF-a,or TGF-/I, derived from cerebral or blood-derivedcells, may exertapositiveregulationon PDGF synthe-sis in human astrocytes.Macrophages/microgliaarethought to be the main source of IL-I and TNF-a(Hopkins and Rothwell, 1995), whereasTGF-/I canbe producedby severalglial cell types(Constametal.,1992;McKinnonet a!., 1993).Although thepresentinvitro observationscannotbe immediatelyextrapolatedto the in vivo situation, it is possiblethat cytokine-regulatedPDGF expressionis important for enhancingcellular growth in different physiological and patho-logical conditions of the CNS. IL-I/I, TNF-a, andTGF-/I aretransientlyexpressedin thedevelopingro-dentbrain and spinal cord (Giulian et aI., 1988; Flan-derset al., 1991; Gendron et al., 1991). Moreover,their intracerebralexpressionis differentially up-regu-lated in varioushumanneuropathologies,including ac-quired immunodeficiencysyndrome encephalopathy,Alzheimer’s disease,and multiple sclerosis(Benve-niste, 1992;Hopkins andRothwell, 1995). Regulationof cellular growth and differentiation by the abovecytokinesis verycomplexandcouldinvolve thestimu-lation of a variety of astroglialpolypeptidemediators,suchas IL-6, colony-stimulatingfactors,nervegrowthfactor(reviewedby EddlestonandMucke, 1993),leu-kemia inhibitory factor (Aloisi et al., 1994), andPDGF (this study). Intracerebrallevels of PDGF areincreasedafter mechanicalinjury or ischemia in theexperimental animal (Lotan and Schwartz, 1992;liharaet al., 1994),and in neoplasticandnonneoplas-tic human brain abscesses(Liu et al., 1994;Nistér eta!., 1994). Within the damagedCNS, PDGF couldpromotegrowth andchemotaxisof mesenchymalandglial cells, neuronalsurvival, and extracellularmatrixdeposition (reviewed by Heldin and Westermark,1990; Hebdin, 1992), therebycontributing to woundhealingandtissueprotection.Studieson PDGFexpres-sion in inflammatory demyelinatingdiseasesof theCNS should help to understandwhetherastrocyte-de-rived PDGF may play a role in promoting remyelina-lion through its mitogenic, survival, and chemotacticeffectson cells of theoligodendrocytelineage.

Acknowledgment:We are grateful to themembersof theDivision of Obstetricsand Gynecology, Avellino GeneralHospital,inparticular,Dr. G. Russo,for thesupplyof humanembryonic brain tissue, and to Dr. K. Funa, for the giftof human PDGF A- and B-chain eDNAs. This study wassupportedby the Italian National ResearchCouncil, TargetProject on Aging, SubprojectGerontobiology(grant no.94.0O497.PF4O;ref. no. 96.1.643),by the Italian Ministryof Health, Project on AIDS (grant no. 930/H), and by aproject on Multiple Sclerosisof the Istituto SuperiorediSanità,Rome.

REFERENCES

Agresti C..Aloisi F..andLevi G. (1991 ) Heterotypicandhomotypiccellular interactionsinfluencing the growth and differentiationof hipotential oligodendrocyte—type-2astrocyteprogenitors inculture. Dee. Blot. 144, 16—29.

Aloisi F.. BorsellinoG.. SamoggiaP., TestaU., Chelucci C.. RussoG., PeschleC., and Levi G. ( 1992o) Astrocyte cultures fromhuman enibryonic brain: characterizationand modulationofsurfacemoleculesby inflammatory cytokines.J. Neurosci. Rem.32, 494—506.

Aloisi F., Care A., Borsellino G., Gallo P., Rosa S., BassaniA..Cabibbo A., TestaU., Levi G., andPeschleC. (1992h)Produc-tion ofhemolymphopoieticcytokines(IL-6, IL-8, colony-stimu-lating factors) by normal humanastrocytesin responseto IL-l~and tunmornecrosisfactor-a.J. lmmunol. 149, 2358—2366.

Aloisi F., RosaS., TestaU., BonsiP., RussoG., PeschleC.. andLeviG. (1994)Regulationof leukemiainhibitory factorsynthesisincultured humanastrocytes.J. Imnmunol. 152, 5022—5031.

Aloisi F., Borsellino G., Care A.. TestaU., Gallo P., RussoG..PeschleC., andLevi G. (1995)Cytokineregulationof astrocytefunction: in vitro studiesusingcells from thehumanbrain. mu.J. Dee. Neuro,sci.13, 265—274.

BarresB. A., Hart I. K.. Coles H. S. R., Burne J. F.. VoyvodicJ. T., RichardsonW. I).. andRaffM. C. (1992)Cell deathandcontrol of cell survival in the oligodendrocytelineage. Cell 70,3 1—46.

Barres B. A., Schmid R., SendtnerM., and Raff M. C. (1993)Multiple extracellularsignalsarerequiredfor long-termoligo-dendrocytesurvival. Dem’elop,nent118, 283—295.

BenvenisteE. N. (1992) lnfiammatory cytokineswithin thecentralnervoussystem: sources,function, and mechanismof action.Am. J. Phvsiol. 263, 1—14.

Bogler0., Wren D.. Barnett S. C., Land H., and Noble M. (1990)Cooperation betweentwo growth factors promotesextendedself-renewaland inhibits differentiation of oligodendrocytc-type-2astrocyte(0-2A) progenitorcells. Proc. Nat!. Acm).Sc).USA 87, 6368—6372.

BottensteinJ. F. and Sato G. M. (1979) Growth of a rat neuro-blastomacell line in serum-freesupplementedmedium. Proc.A/ott. Acrid. Sri. USA76, 514—5 17.

ConstamD. B.. Philipp J., Malipiero U. V.. tenDijke P.. SchachncrM.. and FontanaA. (1992) Differential expressionof trans-forming growth factor-~

1,-a,, and-~ by glioblastomacells.astrocytes,and microglia. J. Immunol. 148, 1404—1410.

de Waal Malefyt R.. Abrams J.. Bennett B., Figdor C. G., anddeVriesJ. E. (1991) Interleukin-ID (IL-ID) inhibits cytokincsynthesisby humanmonocytes:an autoregulatoryrole of IL-ID producedby monocytes.J. Exp. Mccl. 174, 1209—1220.

Eddleston M. and Mucke L. (1993) Molecularprofile of reactiveastrocytes—implicationsfor their role in neurologic disease.Neuroscience54, 15—36.

FlandersK. C., LudeckeG., EngelsS.. Cissel D. S.. RobertsA. B..KondaiahP.. Lafyatis R., SpornM. B., andUnsickerK. (1991)Localizationand actionsof transforminggrowth factor-asintheembryonicnervoussystem.Development113, 183— 191.

GendronR. L.. Nestel F. P.. andLapp W. S. (1991) Expressionoftumornecrosis factoralpha in the developingnervoussystem.tnt. J. Neuro.sci.60, 129— 136.

Giulian D., YoungD.0., WoodwardI., Brown D. C., andLachmanL. B. (1988) Interleukin-l is anastroglialgrowth factor in thedevelopingbrain. J. Neurosci. 8, 709—714.

GrinspanJ. B.. ReddyU. R., SternJ. L., HardyM., Williams M..Baird L.. and Pleasure D. (1990) OligodendrogliaexpressPDGFa-receptorprotein and are stimulatedto proliferate byPDGF. Ann. NYAcrid. Sc).605, 71—8t).

HeldinC-H. (1992) Structural and functional studieson platelet-derivedgrowth factor.EMBO .1. 11, 4251 —4259.

Heldin C-H. and WestermarkB. (1990) Platelet-derivedgrowthfactor: mechanismof actionandpossiblein i’im’o function. CellRegul.1, 555—566.

HermansonM., Nistdr M., Betsholti.C., Heldin C-H., Westermark

I. Neurochen,..Vot. 66, No. 4, 1996

STIMULATION OF GLIAL PDGF SYNTHESIS 1417

B., andFuna K. (1988) Endothelialcell hyperplasiain humanghioblastoma: coexpression of mRNA for platelet-derivedgrowth factor (PDGF) B chain and PDGF receptorsuggestsautocrinegrowth stimulation. Prc,c. Natl. Acrid. Sc). USA 85,7748—7752.

HermansonM., Funa K., HartmanM., Claesson-WelshL., HeldinC-H., WestermarkB., and NistérM. (1992) Platelet-derivedgrowth factor(PDGF)and its receptorsin humangliomatissue;expressionof mRNA andprotein suggeststhepresenceof auto-crine andparacrineloops.CancerRem. 11, 3213—3219.

Hopkins S. J. andRothwell N. J. (1995)Cytokinesand thenervoussystemI: expressionandrecognition.TrendsNeurosci. 18, 83—88.

lihara K., SasaharaM., Hashimoto N., Uemura Y., Kikuchi H.,and HazamaF. (1994) Ischemiainducesthe expressionof theplatelet-growthfactor-B chain in neuronsand brain macro-phagesin v/vu. .1. Cereb. Blood Flow Metab. 14, 818—824.

Liu H. M., Yang H. B., andChenR. M. (1994)Expressionof basicfibroblast growth factor, nervegrowth factor, platelet-derivedgrowth factorandtransforminggrowthfactor-ain humanbrainabscess.Acta Neuropathol. 88, 143—150.

Lotan M. and SchwartzM. (1992) Postinjurychangesin platelet-derivedgrowth factor-like activity in fish and rat optic nerves..1. Neurochem.58, 1637—1642.

Maxwell M., NaberS. P., Wolfe H. J., GalanopoulosT., Hedley-Whyte E. T., Black P. M., andAntoniadesH. N. (1990)Coex-pressionof platelet-derivedgrowth factor (PDGF) andPDGF-receptorgenesby primary humanastrocytomasmaycontributeto theirdevelopmentandmaintenance.J. Clin. Invest. 86, 131—140.

McKinnon R. D., PirasG., Ida i. A., and Dubois-DalcqM. (1993)A role for TGF—,8 in ohigodendrocytedifferentiation. J. CellBiol. 121, 1397—1407.

Meinl F., Aloisi F., ErtI B., WeberF., deWaalMalefyt R., WekerleH., andHolhfeldR. (1994)Multiple sclerosis:immunomodula-tory effects of humanastrocyteson T cells. Brain 117, 1323—1332.

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

Mudhar H. S., Pollock R. A., WangC., StilesC. D., and RichardsonW. D. (1993) PDGFand its receptorsin thedevelopingrodentretina andoptic nerve.Development118, 539—552.

Nikkhah G., Odin P., SmitsA., TingstromA., OthbergA., BrundinP., Funa K., and Lindvall 0. (1993) Platelet-derivedgrowthfactorpromotessurvivalof rat andhumanmesencephalicdopa-minergic neuronsin culture. Exp. Brain Res.92, 5 16—523.

NistérM., LibermannT. A., BetsholtzC., PetterssonM., Claesson-Welsh L., HeldinC-H., SchlessingerJ., and WestermarkB.(1988) Expression of messengerRNAs for platelet-derivedgrowth factorandtransforminggrowthfactor-aandtheirrecep-tors in human malignant glioma cell lines. Cancer Res. 48,3910—3918.

Nistér M., EnbladP.,BackstromG., SddermanT., PerssonL., HeldinC-H.,andWestermarkB. (1994)Platelet-derivedgrowth factor(PDGF)in neoplasticand non-neoplasticlesionsof the centralnervoussystem and in the cerebrospinalfluid. Br. J. Cancer69, 952—956.

Noble M., Murray K., Stroobant P., Waterfield M. D., and RiddleP. (1988) Platelet-derivedgrowth factorpromotesdivision and

motility and inhibits prematuredifferentiationof the oligoden-drocyte/type-2astrocyteprogenitorcell.Nature 333,560—562.

OstmanA., Thyberg J., WestermarkB., and Heldin C-H. (1992)PDGF-AA and PDGF-BB biosynthesis:proproteinprocessingin theGolgi complex andlysosomaldegradationof PDGF-BBretainedintracellularly. I. Cell Biol. 118, 509—S19.

PringleN., Collarini E. J., Mosley M. J., Heldin C-H., WestermarkB., andRichardsonW. D. (1989) PDGFA chainhomodimersdrive proliferation of bipotential (O-2A) glial progenitorcellsin the developingrat optic nerve.EMBO J. 8, 1049—1056.

Pringle N., Mudhar H .S., Collarini E. J., and RichardsonW. D.(1992) Expression of PDGF-alphareceptorsin the rat CNSappearsto be restricted to glial cells of the oligodendrocytelineage.Development115, 535—551.

Raff M. C., Lillien L. E., RichardsonW. D., BumeJ. F., andNobleM. (1988)Platelet-derivedgrowth factorfrom astrocytesdrivesthe clock that times oligodendrocytedevelopmentin culture.Nature 333, 562—565.

RainesE. W., Dower S. K., and Ross R. (1989) IL-I mitogenicactivity for fibroblasts and smooth musclesis due to PDGF-AA. Science243, 393—396.

RainesE. W., Bowen-PopeD. F., and Ross R. (1990) Platelet-derivedgrowth factor, in Handbookof ExperimentalJ-’harma-cology (Spurn M. B. and RobertsA. B., eds), pp. l73—262.Springer,Heidelberg.

RichardsonW. D., Pringle N., Mosley M. J., WestermarkB., andDubois-DalcqM. (1988) A role for platelet-derivedgrowthfactor in norma] gliogenesisin thecentralnervoussystem.Cell53, 309—319.

Robbins K. C., Leal F., PierceJ. H., and AaronsonS. A. (1985)The v-sis/PDGF-2transforminggene product localizesto cellmembranesbut is not a secretoryproduct.EMBO J. 4, 1783—1792.

SasaharaM., Fries J. W. U.. RainesE. W., Gown A. M., WestrumL. E., Frosch M. P., BonthronD. T., RossR., andCollins T.(1991)PDGFB-chain in neuronsof thecentralnervoussystem,posteriorpituitary, and in a transgenicmodel. Cell 64, 217—227.

ShawR. J., BenedictS. H., Clark R. A. F., and King T. E. (1991)Pathogenesisof pulmonaryfibrosis in interstitial lung disease:alveolarmacrophagePDGF (B) geneactivationandup-regula-tion by interferongamma.Am. Rev.Re.vpir. Dis. 143, 167—173.

SmitsA., HermansonM.. Nistér M., Karnushina1., Heldin C-H.,WestermarkB., and FunaK. (1989) Rat brain capillary endo-thehial cells expressfunctional B-typereceptors.Growth Fac-tors 2, 1—8.

Smits A., Kato M., WestermarkB., Nistdr M., Heldin C-H., andFuna K. (1991) Neurotrophic activity of PDGF: rat neuronalcells expressfunctional PDGF a-receptorsand respond toPDGF. Proc. Nat!. Acad. Sri. USA 88, 8159—8163.

SmitsA., BallagiA. E., andFunaK. (1993)PDGF-BBexertstrophicactivity on cultured GABA interneuronsfrom the newbornratcerebellum.Eur. J. Neuro.rci. 5, 986—994.

StarksenN. F., HarshG. R., GibbsV. C., andWilliams L. T. (1987)Regulatedexpressionofthe PDGFA chaingenein nlicrovascu-lar endothelia]cells. J. B/el. Chem. 262, 1438 1—14384.

ThomasP. S. (1980) Hybridization of denaturedRNA and smallDNA fragmentstransferredto nitrocellulose.Proc. Natl. Ac’ad.Sci. USA 77, 5201—5205.

Yeh H.-J., Ruit K. G., Wang Y.-X., Parks W. C., SniderW. D.,and Deuel T. F. (1991) PDGFA-chain gene is expressedbymammalianneuronsduringdevelopmentand in maturity. Cell64, 209—216.

I. Neurochemn.,Vol. 66, No. 4, 1996