Journal of NeurochemistryLippincott—Raven Publishers,Philadelphia© 1997 InternationalSociety for Neurochemistry

Noradrenergic Activity Differentially RegulatestheExpression of Rolipram-Sensitive, High-Affinity Cyclic AMP

Phosphodiesterase(PDE4) in Rat Brain

Ying Ye, *Marco Conti, l~Mi1esD. Houslay,ShakeelM. Farooqui,~Ming Chen,and JamesM. O’Donnell

Departmentsof Pharmacologyand Therapeuticsand ~Microbiology and Immunology,LouisianaState University MedicalCenter, Shreveport,Louisiana; *Department of Obstetricsand Gynecology,StanfordUniversity Medical Center,Stanford,

California, U.S.A.; and ~Department of Biochemistry, Universityof Glasgow, Glasgow,Scotland

Abstract: In a previous study, it was observed that theactivity of rolipram-sensitive, low-Km, cyclic AMP phos-phodiesterase (PDE4) was decreased in vivo with dimin-ished noradrenergic stimulation. The results of the pres-entexperiments indicated that the reduction in the activitymaybe associated withdown-regulation of PDE4 protein.Immunoblot analysis using PDE4-specific, subfamily-nonspecific antibody (K116) revealed four major bandsof PDE4 in rat cerebral cortex; those with apparent molec-ular masses of 109 and 102 kDa are variants of PDE4A.Diminished noradrenergic activity, produced by intracere-broventricular infusion of 6-hydroxydopamine (6-OHDA)or chronic subcutaneous infusion of propranolol, de-creased the intensities of the protein bands for the 109-and 1 02-kDa PDE4A variants in rat cerebral cortex butnot of the 98- or 91-kDa PDE4 forms. 6-OHDA-inducednoradrenergic lesioning also decreased the content of102-kDa PDE4A in hippocampus as labeled by PDE4A-specific antibody (C-PDE4A). Enhanced noradrenergicstimulation up-regulated PDE4 in cerebral cortex. Thiswas indicated by the finding that repeated treatment withdesipramine increased the intensity of the protein bandfor the 102-kDa PDE4 but not for the other variants ofPDE4. These results suggest that PDE4 subtypes are dif-ferentially regulated at the level of expression, as evi-denced by an apparent change in the amount of PDE4protein, following changes in noradrenergic activity.These observations are consistent with the notion thatPDE4s, especially the PDE4A variants with molecularmasses of 109 and 102 kDa, play an important role inmaintaining the homeostasis of the noradrenergic signaltransduction system in the brain and may be involvedin the mediation of antidepressant activity. Key Words:Phosphodiesterase—Noradrenergic system — I mmuno-blot analysis—6-Hydroxydopamine—Propranolol—De-si pram i ne.J. Neurochem. 69, 2397—2404 (1997).

Phosphodiesterases(PDEs) consistof seven fami-lies of isozymesthat hydrolyzecyclic nucleotidesec-

ond messengers.Rolipram-sensitive,lOW~Km, cyclicAMP phosphodiesterase(PDE4), is composedof atleast four subfamilies:PDE4A, PDE4B, PDE4C, andPDE4D (Beavoet al., 1994; Bolger, 1994; Conti etal., 1995).Eachsubfamily is codedby distinct genes,mostof which producemultiple variantsdueto alterna-tive splicing or different initiation sites.

In ratcerebralcortex,PDE4 is the majorPDEiso-zymehydrolyzingcyclic AMP formedby stimulationof /3-adrenergic receptor-linked adenylyl cyclase(Whalin et al., 1989;Challiss andNicholson,1990;Ye and O’Donnell, 1996). Multiple subfamilies ofPDE4arepresentin ratbrain.Northernblot analysis,RNaseprotectionassays,andPCR studiesrevealthepresenceof transcripts of PDE4A, PDE4B, andPDE4D but minimal or no PDE4C (Swinnenet al.,1989a;Bolgeret al., 1994;Engelset al., 1994;Lob-banet al., 1994;Shakuret al., 1995).Also, the vari-ants of PDE4 subfamilies are presentin rat brain,althoughit is notclearyet whichof thevariantformsof PDE4 are expressed.

PDE4 has beensuggestedto be a targetof a newclassof antidepressants.PDE4-specific inhibitors ex-hibit antidepressanteffectsin bothanimal studiesandclinical trials (WachtelandSchneider,1986;Bobon etal., 1988; Griebel et al., 1991; Fleischhackeret al.,1992;O’Donnell, 1993).Themechanismof action forthis new classof antidepressantsis mostlikely through

ReceivedApril 28, 1997; revised manuscriptreceivedJuly 29,1997; acceptedJuly 29, 1997.

Addresscorrespondenceand reprint requeststo Dr. Y. Ye at De-partmentof PharmacologyandTherapeutics,LouisianaState Uni-versity Medical School, 1501 Kings Highway, Shreveport, LA71130-3932,U.S.A.

Abbreviationsused: PAGE, polyacrylamidegel electrophoresis;NE, norepinephrine;6-OHDA, 6-hydroxydopamine;PDE, phospho-diesterase;PDE4,rolipram-sensitive,low-K,,, cyclic AMP phospho-diesterase;SDS, sodiumdodecyl sulfate.

2397

2398 Y. YE ET AL.

direct inhibition of PDE4 via a postreceptormecha-nism (O’Donnell, 1993),anactionthat is distinctfromthat of the classicalantidepressants.A betterunder-standingof PDE4might leadto insight into thepatho-physiologyof depressionandnovel methodsof thera-peutic intervention.

Variousstudiesusingculturedcells haveshownthatthe activity of PDE4 can be elevatedin vitro withenhancedstimulation of receptor-linkedadenylyl cy-clase.This enhancementcanbe accountedfor, in part,by protein kinaseA-mediatedphosphorylationor byup-regulationof theenzyme.Forexample,the increasein PDE4 activity in FRTL-5 thyroid cells observedshortly after additionof thyroid-stimulatinghormoneis associatedwith phosphorylationof PDE4D (Setteet al., 1994a),whereasthe increasesin PDE4activityin immatureSertolicells and U937 humanmonocytesare mediated by an elevated expressionof PDE4(Swinnenet al., 1989a;Torphy et al., 1995;Vergheseet al., 1995).The elevatedexpressionof PDE4is dueto increasedtranscriptionandpossiblyincreasedstabil-ity of mRNA, depending on the subtype of PDE4(Swinnenet al., 1991).

Studieson culturedcells alsorevealthat subfamiliesof PDE4and their variantsare regulatedby differentmechanismsand that the natureof the regulationde-pendson the type of cells examined.For instance,inU937 monocytes, the expression of PDE4A andPDE4B, but not PDE4Cand PDE4D,is increasedbydibutyryl cyclic AMP. In Jurkatcells, the expressionof PDE4A andPDE4D, butnot PDE4B andPDE4C,is enhancedby dibutyryl cyclic AMP, whereasin SH-SY5Y cells, none of the PDE4 subfamiliesrespondsto such stimulation (Engels et al., 1994; Torphy etal., 1995). In FRTL-5 thyroid cells, oniy PDE4D isphosphorylatedwith enhancedstimulationof adenylylcyclase(Setteet al., 1994a). It also appearsthat alter-native processingproductsfrom an individual PDE4subfamily canbe differentially regulated(Setteet al.,1994b).

In a previousstudy, it was found that diminishednoradrenergicactivity, producedeitherby 6-hydroxy-dopamine(6-OHDA)-inducedlesioningor proprano-lol-mediatedblockade of /3-adrenergicreceptors,re-duced PDE4 activity in vivo by ~40% (Ye andO’Donnell, 1996).However,the mechanismsmediat-ing this reductionin activity are yet to be elucidated;in addition, the PDE4 subfamiliesthat are regulatedremainto beidentified.Becausethereductionin PDE4activity occurs slowly following noradrenergicle-sioning,it is possiblethatdown-regulationis involved.The goalof thepresentstudywasto testthis hypothesisand to identify subtypesof PDE4 that are regulated.Sodium dodecyl sulfate (SDS) .-polyacrymide gelelectrophoresis(PAGE) and immunoblot analysiswere applied to quantify PDE4 following the alter-ationsof noradrenergicactivity.

MATERIALS AND METHODS

AnimalsMale Sprague—Dawleyrats(weighing 225—300 g; Har-

lan,Indianapolis,IN, U.S.A.) wereusedfor theexperiments.Theratswere housedin a temperature(22—24°C)-andlight(on6:00 a.m.—6:00p.m.)-controlledroomandwereallowedfree accessto food pelletsandwater. The studiesreportedin this article havebeencarriedout in accordancewith theGuidefisr the Care and Use of Laboratory’ Animals asadopted and promulgated by the National Institutes ofHealth.

Treatment of animalsThe blockade of central ~-adrenergic receptors was

achievedby chronic, subcutaneousinfusion of propranolol(Eison et al., 1988).Propranolol(30 mg/kg/day, 14 days)dissolved in distilled, deionizedwater wasreleasedfromsubcutaneouslyimplantedosmotic pumps (model 2ML2;Alza Corp., Palo Alto, CA, U.S.A.).

Noradrenergiclesionswereproducedby bilateralintrace-rebroventricularadministrationof 6-OHDA (200 ~ig in 20pJ of 0.2% ascorbic acid/0.9% NaCI) as describedpre-viously (Ye andO’Donnell, 1996).

The enhancementof central noradrenergicactivity wasachievedby repeated,daily injections of desipramine(10mg/kg, i.p., twice adayfor 14 days),anorepinephrine(NE)reuptakeinhibitor.

Preparation of rat cerebral cortical homogenatesfor immunoblot analysis

Fourteendaysfollowing the 6-OHDAinfusionortheiniti-ation of propranololor desipraminetreatment,the ratswerekilled by decapitation.Cerebralcorticesorhippocampiweredissectedand immediately frozen in liquid nitrogen. Eachsamplewashomogenizedin RIPA buffer (150 mM NaCI,0.1% SDS, 0.5% deoxycholate,1% Nonidet P-40, and 50mM Tris-HC1, pH 8) with a Douncehomogenizer.The pro-tein levels of the homogenateswere measuredusing thebicinchoninicacidprotein assay(Smithet al., 1985)(Pierce,Rockford,IL, U.S.A.).Samplesfrom bothcontrol andexper-imental ratswere treatedidentically.

SDS-PAGEand immunoblot analysisWhen the effects of drug treatmentswere determined.

samplesfrom control and experimentalgroupswere exam-ined within single gels.Equalamountsof protein from bothcontrol andtreatedrats, suspendedin 2x Laemmli samplebuffer [10% 2-mercaptoethanol,4%SDS,20%glycerol, 120mM Tris-HCI (pH 6.8), and 0.001%bromophenolblue],were boiled for 3 mm and subjectedto SDS-PAGE (4%stackinggel, 7.5% resolvinggel). The electrophoresiswascarried out at constantcurrent (40 mA per gel) for 3.5 hwith cooling. The separatedproteinswere transferredto anitrocellulosemembraneat constantvoltage(20 V) over-night. Following blocking with 5% bovine serumalbuminfor 2 h, the membranewasblotted with the antibodyKI 16(1:500 dilution in 0.1% bovine serum albumin and 0.5%normal goat serum)for 1 h and subsequentlyfor anotherhourwith goatanti-rabbitantiserumconjugatedwith alkalinephosphatase.After each incubation, the membranewaswashedextensivelywith TBST buffer (100 mM Tris, 0.9%NaC1, and 0.1% Tween 20). Finally, the membranewassoakedin alkaline phosphatasesubstratesolution (Gibco-BRL, Gaithersburg,MD, U.S.A.) to revealstaining.

J. Neurochem.,Vol. 69, No. 6, 1997

REGULATION OF PDE BY NORADRENERGICACTIVITY 2399

Identification of K116-labeledbands inimmunoblots

Thebandslabeledby Ki 16 were identifiedusingsubfam-ily-specific PDE4Aantibodies (AC55 andC-PDE4A).TheantibodyKi 16 is apolyclonalantibodygeneratedin rabbitsagainsta 22-amino-acidpeptide (2224; locatedat aminoacid positions 105—126of rat PDE4D1) from a conservedN-terminal region of PDE4, which is distinct from otherfamilies of PDE but highly homologous amongthe foursubfamiliesof PDE4(Setteet al., 1994a).AC55 wasgener-atedagainst100aminoacidsof thecarboxyl-terminalregionof PDE4A (Naro et al., 1996). C-PDE4A was generatedagainsta 12-amino-acidpeptide(locatedat aminoacidposi-tions 617—628 of PDE4A5) that wascontainedin the Cterminus of both PDE4A1 and PDE4A5 (McPheeet al.,1995). AC55 andC-PDE4A,generatedby differentlabora-tories, recognizeall known forms of PDE4As [PDE4A 1,PDE4A5, and PDE4A8 (Bolger et al., 1996; Naro et al.,1996; S. M. Farooqui and J. M. O’Donnell, unpublisheddata)].

To seewhether the protein bandslabeledby K116 havesame electrophoreticmigration rate as those labeledbyPDE4A antibodies,proteinsfrom rat cerebralcortical ho-mogenateswere separatedby SDS-PAGEand immunoblot-ted using K116as well asAC55 andC-PDE4A.

To identify theseprotein bandsfurther, cerebralcorticalpreparationswereimmunoprecipitatedwith Ki 16 (Setteetal., 1994a),andthe eluatewasblotted by C-PDE4Aanti-body.

Catecholaminelevel measurementandJ3-adrenergic receptor binding

Central /3-adrenergicreceptor blockade producedby con-tinuous,subcutaneousinfusion of propranololwas verifiedby the up-regulationof the receptors.Chronic enhancementin central /1-adrenergicreceptorstimulationproducedby re-peated,intraperitonealinjection of desipraminewas con-firmedby the down-regulationof the receptors.The densityof /3-adrenergicreceptorswasmeasuredby saturationbind-ing of ‘251-pindololto membranepreparationsof thecerebralcortices(O’Donnell andFrazer, 1985;O’Donnell, 1990).

CerebralcorticalNE levelsweremeasuredto verify nor-adrenergiclesioning following intracerebroventricular6-OHDA infusion. The levels were determined usingHPLC(O’Donnell and Seiden,1983; O’Donnell, 1993). Proteincontentwas determinedby the methodof Bradford (1976).

Data analysisThe intensitiesof thebandsin immunoblotswereanalyzed

by densitometryusingNIH ImageAnalysissoftware.Differ-encesin intensitiesbetweensamplesfromcontrolandtreatedrats were tested usingMann—Whitney U tests(Siegel,1956). Two-tailedStudent’st testswereappliedto thecate-cholaminelevel measurementsand ~-adrenergic receptorbindinganalysis(Winer, 1971).

MaterialsSeeBlueprestained molecularmarkerwasobtainedfrom

Novel ExperimentalTechnology(SanDiego, CA,U.S.A.).The alkalinephosphatasesubstrates5-bromo-4-chloro-3-in-dolyl phosphate andnitro blue tetrazoliumwerepurchasedfrom GibcoBRL. The electrophoresissystem,transfercell,andchemicalsused forimmunoblot studies werepurchasedfrom Bio-Rad(Hercules,CA, U.S.A.).‘25I-Pindolol waspur-

FIG. 1. Identification of PDE4 proteins in rat cerebral cortex us-ing immunoblot analysis. A: Immunoblot image labeled by PDE4-specific, subfamily-nonspecific antibody K116 and by immuno-genic peptide 2224-preabsorbed K116. B: Immunoblot imageslabeled by Ki 16 and PDE4A-specific antibodies (AC55 and C-PDE4A). C: Immunoblot image labeled by C-PDE4A for samplesof cerebral cortical homogenate (lane 1), eluate of Ki 1 6-medi-ated immunoprecipitation (lane 2), and eluate of the controlbovine serum albumin (BSA)-mediated immunoprecipitation(lane 3).

chasedfrom DuPont/NEN(Boston,MA, U.S.A.). Therestof the chemicalswere obtainedfrom Sigma ChemicalCo.(St. Louis, MO, U.S.A.) or Fisher Scientific (Dallas,TX,U.S.A.).

RESULTS

Characterization of K116-labeledproteins inimmunoblots

Immunoblot analysis showed that antibody KI 16labeled multiple proteinsfrom rat cerebralcorticalho-mogenates(Fig. 1A). If immunogenicpeptide 2224-preabsorbedKi 16 was usedinstead,the proteinswithmolecularmassesof 109, 98, and 91 kDa were notlabeled, and the102-kDaprotein wasreducedin inten-sity by 95%; the other major bandsremainedun-changed, although severalfaintly labeledbandsof amolecularmass of <91 kDa seemedto disappearaswell (Fig. 1A). Thefour specific bands(109, 102, 98,and91 kDa) also wereobservedin immunoblotsla-beledby Ki 16 following Ki 16-mediatedimmunopre-cipitationof rat brainhomogenate(datanot shown).

When normal rabbit serum was usedas a probe,none of the proteinswith molecularmasses of109,

.1. Neurochem.,Vol. 69, No. 6, 1997

2400 Y YE ET AL.

102, 98, and91 kDa was labeled(Fig. 1B). Thetoptwo bandslabeledby K116 (109 and 102 kDa) havethe sameelectrophoretic migrationratesas thosela-beledby PDE4A-specific antibodies (AC55 and C-PDE4A; Fig. 1B). In addition, thesetwo proteinsalsowere labeledby C-PDE4Afollowing immunoprecipi-tation by K116 (Fig. 1C).

The integrateddensities of the four specific bandslabeledby Ki 16 increasedlinearly asprotein contentwas increasedfrom 20 to 120 ~g (r � 0.97; datanotshown).This linear relationshippermitteddetermina-tion of lesion- anddrug-inducedchangesin PDE4ex-pression.

Effects of chronic propranolol treatment on PDE4detectedby immunoblot analysis

To assesswhetherPDE4 down-regulationwas in-volved in the reduction in PDE4 activity followingdiminishednoradrenergicactivity that wasobservedpreviously (Ye and O’Donnell, 1996), PDE4 wasquantifiedusingimmunoblotanalysisof cerebralcor-tical tissues from propranolol-treatedrats. Equalamountsof protein (80 gig, at which the intensityof

FIG. 2. Effect of propranolol treatment on PDE4. For the pro-pranolol treatment, the drug was administered via a subcutane-ously implanted pump (30 mg/kg/day for 14 days); the controlrats were subjected to a similar surgery but without pump im-plantation. Fourteen days later, cerebral cortical homogenatesfrom these rats containing equal amounts of protein were sub-jected to SDS-PAGE. Separated proteins were transferred to anitrocellulose membrane and probed by antibody Ki 16. A: Im-age of representative immunoblot. B: Densitometry analysis ofthe immunoblot image. Data are mean ± SEM (bars) values forfive or six rats, with control and experimental samples electro-phoresing in a single gel. Values significantly different from thecontrol are indicated: “p < 0.01, ***p < 0.005.

TABLE 1. Effectsof desipramineor propranololtreatmenton binding of ‘

25I-pindolol to ~3-adrenergicreceptorson cerebral cortical membranes

TreatmentB,,,,,,

(fmol/mg)K[)

(nM)

Control for desipramine 148.2 ± 7.6 0.29 ±0.03Desipramine 78.05 ± 5.7° 0.26 ± 0.04Control for propranolol° 116.4 ± 13.9 0.21 ± 0.06Propranolol5 164.4± 12.2” 0.34 ± 0.03

The binding studies fordesipramine-or propranolol-treatedratswereperformedseparately. For desipramine treatment, the ratsre-ceivedeither saline ordesipramine(10 mg/kg, i.p., twice a day,for14 days). After 14 days, therats were killed, and cerebralcorticesweredissected for themeasurementof PDE4 level and‘231-pindololbinding. Data are mean±SEM values from six pairsof rats. Forpropranololtreatment, the rats receivedpropranolol(30 mg/kg/dayfor 14 days) via subcutaneouslyimplanted pumps similarto thoseusedfor the immunoblotanalysis study. After 14 days. thepumpswereremoved,and a 24-h drugwashoutperiod was allowedbeforetherats were killed for themeasurementof 25l-pindolol binding.Data are mean±SEM values from four pairsof rats.

Significantly different from thecorrespondingcontrol, p< 0.05.Data are taken from Ye and O’Donnell (1996).

the bands islinearlyassociatedwith amount ofproteinused for SDS-PAGE) from both control andtreatedgroups wereused.Chronic, subcutaneouspropranololinfusion (30 mg/kg/day for 14 days) decreasedtheintensityof the 109-kDaPDE4by 15% and theinten-sity of the 102-kDaPDE4 by 50%; the intensitiesofthe 98- and91-kDa PDE4isozymes wereunaffectedby propranololtreatment(Fig. 2).

The blockadeof central /3-adrenergic receptorsbysubcutaneousinfusion of propranololwas verified bythe up-regulationof these receptors. TheBmax of the‘251-pindolol bindingwas significantlyincreasedin thecerebralcorticalmembranespreparedfrom proprano-lol-treated rats using anidentical treatmentregimen(Table 1).

Effects of noradrenergic lesioning on PDE4detected by immunoblot analysis

To determinefurther whetherdiminishednoradren-ergic activity down-regulated PDE4, immunoblotanal-ysesusing antibodyKi 16 or C-PDE4A wereappliedto cerebral corticaland hippocampalsamplesfollow-ing 6-OHDA-induced noradrenergiclesioning. Four-teendaysfollowing intracerebroventricularinfusion of6-OHDA, the PDE4swith molecular masses of109and 102 kDawere significantly decreasedin intensityin rat cerebralcortex as detectedby K116 (Fig. 3),but those with molecularmasses of98 and 91 kDaremainedunchanged.In hippocampus,the PDE4witha molecularmass of 102 kDa was significantly de-creasedin intensity as labeledby C-PDE4A, buttheotherforms of PDE4were unchanged(Fig. 4).

Noradrenergic lesioningwas verifiedby measuringcerebral corticalNE levels.Fourteendays following

J. Neurochem.,Vol. 69. No. 6, 1997

REGULATION OF PDE BY NORADRENERGICACTIVITY 2401

FIG. 3. Effect of 6-OHDA treatment on cerebral cortical PDE4.Rats received either vehicle or 200 ~sgof 6-OHDA via intracere-broventricular infusion. Fourteen days later, cerebral cortical ho-mogenates with equal amounts of protein from both control andlesioned rats were subjected to SDS-PAGE. Separated proteinswere transferred to a nitrocellulose membrane and probed byantibody Ki 16. A: Immunoblot image. B: Densitometry analysisof the immunoblot image. Data are mean ± SEM (bars) valuesfor three rats, with control and experimental samples electropho-resing in a single gel. Values significantly different from the con-trol are indicated: *p < 0.05.

intracerebroventricular6-OHDA infusion, NE contentwas reducedfrom the control level of 2.57±0.10to 0.05 ±0.02 ng/mg of protein (2% of control; p<0.005).

Effects of chronic desipramine treatment onPDE4 detectedby immunoblot analysis

To determinewhether increasednoradrenergicstim-ulation producedoppositeeffects from propranololor6-OHDAtreatment, PDE4in cerebral cortexwasquan-tified, andlevelswerecompared betweencontrol anddesipramine-treatedrats using immunoblot analysis.Chronic desipraminetreatment(10 mg/kg/injection,i.p., twice a day for 14 days) significantly increasedtheintensity of the102-kDaPDE4 but not that of thePDE4isozymesof 109, 98, and 91 kDa (Fig. 5).

The enhancementof central ~3-adrenergic stimula-tion by intraperitoneal injection of desipraminewasverifiedby thedown-regulationof /3-adrenergicrecep-tors.TheB,,,,,,, of 251-pindololbinding wassignificantlydecreasedin membranepreparationsfrom therats thathadreceiveddesipramine(Table 1).

DISCUSSION

The presentstudy not only confirmedthe finding ofa previousreportthat PDE4 is regulatedin vivo withchanges innoradrenergicactivity (Ye andO’Donnell,1996),but also suggestedthat analterationin PDE4

expressionmay be involved in thepreviouslyobservedchangein PDE4activity. In addition, thecurrentresultsdemonstratedthat PDE4 subtypes aredifferentiallyregulatedat the levelof expressionwith alterationsofnoradrenergicactivity.

PDE4 was quantified using immunoblot analysiswith antibody Kl 16. This antibody wasgeneratedagainst a peptide sequence,which is homologousamong subfamilies ofPDE4 but distinct from the se-quencesof other families of PDE (Jin et al., 1992).KI 16 hasbeenusedas a PDE4-specificbut subfamily-nonspecific antibody(Setteet al., l994a; Alvarez etal., 1995). This antibodyinteractedspecifically withfour major proteins fromrat cerebralcortexin immu-noblot analysis,with molecularmasses of109, 102,98, and 91 kDa. This is indicatedby the dramaticreductionin the intensitiesof the bands or thedisap-pearanceof the bands if normalrabbit serum orimmu-nogenicpeptide-preabsorbedKl 16 was used. Severalfaintly labeledbandsof a molecularmass <91 kDaalso disappearedfollowing the immunogenic peptidepreabsorbingstep.These faintlylabeledbandsmightbe PDE4saswell, but with low affinity to Kl 16. Thisis consistentto the notion thattherearemultiple sub-typesof PDE4s [>15 subtypes(Beavoet al., 1994)].Subfamily-specific antibodies(AC55 andC-PDE4),which were raised against different epitopes ofPDE4A, also labeled the 109- and 102-kDa bands;thesetwo bandsalso were labeledby C-PDE4A fol-lowing KI 16 immunoprecipitation.Thus, theconver-

FIG. 4. Effect of 6-OHDA treatment on hippocampal PDE4. Therats received either vehicle or 200 ~igof 6-OHDA via intracere-broventricular infusion. Fourteen days later, hippocampal ho-mogenates with equal amounts of protein from both control andlesioned rats were subjected to SDS-PAGE. Separated proteinswere transferred to a nitrocellulose membrane and probed byantibody C-PDE4A. A: Immunoblot image. B: Densitometry anal-ysis of the immunoblot image. Data are mean ± SEM (bars)values for four rats, with control and experimental samples elec-trophoresing in a single gel. Values significantly different fromthe control are indicated: *p < 0.05.

J. Neurocheni.,Vol. 69, NO. 6, 1997

2402 V. YE ET AL.

FIG. 5. Effect of desipramine treatment on PDE4. Rats receivedeither vehicle or desipramine (10 mg/kg, i.p., twice a day, for14 days). At the end of the 14-day treatment period, cerebralcortical homogenates from these rats containing equal amountsof protein were subjected to SDS-PAGE. Separated proteinswere transferred to a nitrocellulose membrane and probed byantibody Ki 16. A: Image of representative immunoblot. B: Den-sitometry analysis of the immunoblot image. Data are mean±SEM (bars) values for six rats, with control and experimentalsamples electrophoresing in a single gel. Values significantly dif-ferent from the control are indicated: * *p < 0.01.

genceof datasuggeststhat the109- and102-kDabandsare PDE4A variants.

The existenceof multiple variantsof PDE4A hasbeenrecognized(Bolger et al., 1996). The 109-kDaPDE4A variantlabeledby KI 16 has the sameelectro-phoreticmigration rate as PDE4A5(McPhee et al.,1995).The 102-kDa proteinappearstobe apreviouslyuncharacterizedvariant of PDE4A. The other twoKI 16-labeledproteins (98 and91 kDa) are probablyPDE4Dvariants,becausethePDE4D-specificantibodyM3S 1 interactedwith two proteinswith similar appar-ent molecularmasses(datanot shown),and PDE4Dvariantswith similar electrophoreticmigration rateshave beenreported (Setteet al., 1994b). PDE4B isnot recognizedwell by KI 16, although the presenceof two PDE4Bvariants (PDE4B1andPDE4B2)in ratcerebral cortex has been observed(Lobban et al.,1994).PDE4Chasbeenreportedto beabsent(Swin-nen etal., 1989a)or presentonly in minimal quantities(Engelset al., 1994) in ratcerebralcortex.

Immunoblotanalysis showedthat theintensitiesofcertain PDE4 bands in rat cerebralcortex were de-creasedfollowing chronic propranololtreatmentor 6-OHDA infusion, whereastheintensitieswere increasedafterchronicdesipraminetreatment.Theseresultssug-gest thatPDE4is up- ordown-regulatedin vivo; thesechanges maycontribute to the changeof PDE4 hy-drolytic activity observedafter similar treatments(Yeand O’Donnell, 1996). The interpretationthat PDE4

undergoesup- anddown-regulationis consistent withthe following findings. First, thereductionin PDE4activity producedby diminished noradrenergicstimu-lation occursslowly (Ye andO’Donnell, 1996).Sec-ond, in vitro studieshavedemonstratedthat the in-creaseof PDE4activity in someculturedcells ismedi-ated by up-regulationof the isozyme.For example,sucheffectsare seen in U937 humanmonocytesstimu-lated by a /3-adrenergicagonistand inimmatureSertolicells stimulated by follicle-stimulating hormone. Intheseexperimentsthe elevation of cyclic AMP hy-drolytic activity is eitherinhibitedby protein synthesisinhibitorsor accompaniedby an increasein Vmax, mes-senger RNAlevel, or protein levels (Swinnenet al.,1989b;TorphyCt al., 1992, 1995).The currentresults,however, do not exclude the contribution of otherforms of regulationin vivo, such asphosphorylation-inducedactivation,following changes innoradrenergicactivity.

Diminishednoradrenergicactivity decreasedthein-tensitiesof both109- and102-kDaPDE4A variants incerebralcortex, but enhancednoradrenergicactivityonly increasedthe intensity of the 102-kDaPDE4A.One possibleexplanationfor this discrepancyis thatthe 109-kDaPDE4 mightbe maximally expressedatnormal levels of noradrenergicactivity and thaten-hancing noradrenergicactivity does not further in-creaseits expression.Also, in hippocampusonly thereductionin intensityof 102-kDaPDE4A variantwasobservedfollowing noradrenergiclesioning, implyingthat in different brain regions,PDE4variantsmay bedifferentially regulated.However, thepossibility thattreatment-inducedchangesfor some PDE4 variantsmight be subtleand difficult to detect cannotbe ex-cluded.

Noradrenergic lesioningor chronic/3-adrenergicre-ceptorblockadedecreasedthe intensitiesof 109- and102-kDaPDE4A but not98- or 91-kDaPDE4s. Thisselectivealterationin theexpressionsuggeststhat sub-types of PDE4 aredifferentially regulatedin vivo bychanges innoradrenergicactivity. There are severalpossible interpretationsfor these findings. First, thePDE4 variants with molecularmasses of98 and 91kDa may not be regulatedin vivo by noradrenergicneuronsand receptors. Second, they may be regulatedby othermeans,e.g.,phosphorylation.Third, the 109-and 102-kDaforms of PDE4A may be more closelyassociatedwith thenoradrenergicsystem than the othertwo bands.This final interpretationis consistentwiththe comparable degreeof reductionin PDE4activityandexpression(Ye andO’Donnell, 1996); followingpropranololtreatment,the hydrolysis of cyclic AMPformed by isoproterenolstimulation was decreased~40—50%, as was thereductionin the intensity of102-kDaPDE4. If this interpretationis true, PDE4A5and the 102-kDaPDE4A variant may be crucial inregulatingcentral /3-adrenergic signaltransduction,es-pecially as it has beenshown thatPDE4is the major

.J. Neurochem.,Vol. 69, No. 6. 1997

REGULATION OF PDE BY NORADRENERGICACTIVITY 2403

or exclusivePDEcatalyzing cyclic AMP formed bystimulation of /3-adrenergicreceptor-linkedadenylylcyclase in rat cerebral cortex (Whalin et al., 1989;Challiss and Nicholson, 1990; Ye and O’Donnell,1996).

The presentstudy appearsto link particularPDE4variants—PDE4A5and a novel, 102-kDa PDE4Avariant— to the /3-adrenergicreceptor-mediatedsignaltransductionsystem. Understandingthis associationmay proveimportantin drug design.PDE4 is a targetof a novel classof antidepressantsandantiinflamma-tory drugs (Wachtel and Schneider, 1986; Fleisch-hackeret al., 1992; Griswoldet al., 1993).Themajordrawbackof thesedrugsis the lack of selectivity.Thedistinct associationof certain subtypesof PDE4 tothenoradrenergicsystemmay providea physiologicalrationalefor the developmentof selectivedrugs fortreatingdepressionor atopicdisorders.

Acknowledgment:This researchwassupportedby grantsMH40694 andMH5 1175 andIndependentScientistAwardMH01231 from theNationalInstitutefor MentalHealth.TheauthorsthankLatis Mclnnis, CharlesDempsey,and GenevaMeachumfor excellenttechnicaland secretarialsupport.

REFERENCES

Alvarez R., SetteC., Yang D., Eglen R.M., Wilhelm R., SheltonE. R., andConti M. (1995)Activation and selectiveinhibitionof a cyclic AMP-specific phosphodiesterase,PDE-4D3. Mol.Pharmacol.48, 616—622.

Beavo J. A., Conti M., and Heaslip R. J. (1994) Multiple cyclicnucleotidephosphodiesterases.Mol. Pharmacol. 46, 399—405.

Bobon D., Breulet M., Gerard-VandenhoveM. A., Guiot-GoffioulF., PlonteuxG., Sastre-y-HernandezM., SchratzerM., Trois-fontainesB., Vov-FrenckellR., andWachtelH. (1988)Is phos-phodiesteraseinhibition a new mechanismof antidepressantaction?A doubleblind double-dummystudybetweenrolipramanddesipraminein hospitalizedmajorand/orendogenousde-pressives.Eur. Arch. PsychiatryNeurol. Sci. 238, 2—6.

BolgerG. B. (1994)Molecularbiology of thecyclic AMP-specificcyclic nucleotidephosphodiesterases:a diversefamily of regu-latory enzymes.Cell. Signal. 6, 851—859.

Bolger G. B., RodgersL., and Riggs M. (1994) DifferentialCNSexpressionof alternativemRNA isoforms of the mammaliangenesencodingcAMP-specific phosphodiesterases.Gene149,237--244.

Bolger G. B., McPhee I., and Houslay M. D. (1996) Alternativesplicing of cAMP-specific phosphodiesterasemRNA tran-scripts. J. Biol. C/acm.271, 1065—1071.

BradfordM. M. (1976)A rapid andsensitivemethodfor thequanti-tation of microgramquantitiesof proteinutilizing theprincipleof protein—dyebinding. Anal. Biochem.72, 248—254.

ChallissR. A. andNicholsonC. D. (1990)Effectsof selectivephos-phodiesteraseinhibition on cyclic AMP hydrolysisin rat cere-bral cortical slices.Br. J. Pharmacol.99, 47—52.

Conti M., NemozG., SetteC., andVicini E. (1995) Recentprogressin understandingthe hormonal regulationof phosphodiester-ases.Endocr. Rev. 16, 370—389.

Eison L. S., Yocca F. D., and GianutsosG. (1988) Noradrenergicdenervationaltersserotonin2-mediatedbehaviorbut not seroto-nin2 receptornumberin rats: modulatoryrole ofbetaadrenergicreceptors.J. Pharmacol.Exp. Ther. 246, 571 —577.

EngelsP., Fichtel K., andLubbertH. (1994)Expressionandregula-tion of human and rat phosphodiesterasetype IV isogenes.FEBSLett. 350, 291—295.

FleischhackerW. W., Hinterhuber H.. Bauer H., Pfiug B., BernerP., SimhandlC., Wolf R., GerlachW., Jaklitsch H., Sastre-y-HernandezM., Schmeding-WiegelH., Sperner-UnterwegerB.,Voet B., and SchubertH. (1992) A multicenterdouble-blindstudyof threedifferentdosesof thenew cAMP-phosphodiester-aseinhibitorrolipram in patientswith majordepressivedisorder.Neuropsychobiology26, 59—64.

GriebelG., Misslin R., Vogel E., andBourguignonJ. (1991)Behav-ioral effectsof rolipram and structurallyrelatedcompoundsinmice: behavioralsedationof cAMP phosphodiesteraseinhibi-tors. Pharmacol.Biochem.Behav.39, 321 —323.

Griswold D. E., WebbE. F., Breton J., White J. R., Marshall P. J.,andTorphy T. J. (1993) Effect of selectivephosphodiesterasetype IV inhibitor, rolipram,on fluid and cellular phasesof in-flammatoryresponse.Inflammation 17, 333—344.

Jin S. L. C., Swinnen J. V., and Conti M. (1992) Characterizationof thestructureof alow K,,,, rolipram-sensitivecAMPphospho-diesterasemappingof thecatalyticdomain. I. Biol. Chem.267,18929—18939.

LobbanM., ShakurY., BeattieJ.,andHouslayM. D. (1994)Identi-fication of two splice variant formsof type IVB cyclic AMPphosphodiesterase,DPD (rPDE-1V

51)andPDE-4(rPDE-IVB2)in brain:selectivelocalizationin membraneandcytosoliccom-partmentsand differentialexpressionin various brain regions.Biochem.J. 304, 399—406.

McPheeI., Pooley L., Lobban M., Bolger G., and Houslay M. D.(1995) Identification,characterizationandregional distributionin brain of RPDE-6 (RNPDE4A5), a novel splice variant ofthePDE4Acyclic AMP phosphodiesterasefamily. Biochem.I.310, 965—974.

NaroF., ZhangR., andConti M. (1996) Developmentalregulationof uniqueadenosine3 ‘,5 ‘-monophosphate-specificphosphodi-esterasevariants during rat spermatogenesis.Endocrinology137, 2464—2472.

O’Donnell J. M. (1990) Behavioral effects of beta adrenergicago-nists and antidepressantdrugsafter down-regulationof beta-2adrenergicreceptorsby clenbuterol.J. Pharmacol.Exp. Ther.254, 147—154.

O’Donnell J. M. (1993)Antidepressant-likeeffectsof rolipram andotherinhibitors of cyclic AMP phosphodiesteraseon behaviormaintainedby differential reinforcementof low responserate.J. Pharmacol.Exp. Ther. 264, 1168—1178.

O’Donnell J. M. and FrazerA. (1985) Effects of clenbuterolandantidepressantdrugson betaadrenergicreceptor/N-proteincou-pling in the cerebralcortexof therat. J. Pharmacol.Exp. Ther.234, 30—36.

O’Donnell J.M. and SeidenL. 5. (1983)Alteredeffectsof desipra-mineon operantperformanceafter6-hydroxydopamine-induceddepletionof brain dopamineor norepinephrine.J. Pharmacol.Exp. Ther. 229, 629—635.

Sette C., lona S., and Conti M. (1994a)The short-termactivationof a rolipram-sensitive,cAMP specific phosphodiesterasebyTSH in thyroidFRTL-5 cells is mediatedby acAMP-dependentphosphorylation.J. Biol. Chem.269, 9245—9252.

SetteC., Vicini E., and Conti M. (1994h)The ratPDE3/IVdphos-phodiesterasegene codes for multiple proteins differentiallyactivatedby cAMP-dependentprotein kinase. J. Biol. Chem.269, 18271—18274.

ShakurY., Wilson M., PooleyL., LobbanM., Griffiths S. L., Camp-bell A. M., BeattieJ.,Daly C., andHouslayM. D. (1995)Iden-tification and characterizationof the type-IVA cyclic AMP-specific phosphodiesteraseRD1 as a membrane-boundproteinexpressedin cerebellum.Biochem.1. 306, 801—809.

Siegel 5. (1956)The caseof two independentsamples,in Nonpara-metric Statisticsfor the BehavioralScience.c(Morgan C. T.,ed), pp. 95—158.McGraw-Hill, New York.

Smith P. K., Krohn R. I., HermansonG. T., MalliaA. K., andGart-ner F. H. (1985) Measurementof protein using bicinchoninicacid. Anal. Biochem.150, 76—85.

Swinnen J. V., JosephD. R.. and Conti M. (1989a)The mRNA

J. Neurocheni.,Vol. 69, No. 6, 1997

2404 V. YE ET AL.

encoding a high-affinity cAMP phosphodiesteraseis regulatedby hormonesand cAMP.Proc. Nati. Acad.Sci. USA86, 8197—8201.

SwinnenJ. V., JosephD. R., andConti M. (1989b)Molecularclon-ing of rat homologuesof the Drosophila melanogasterduncecAMP phosphodiesterase:evidencefor a family of genes.Proc.NatI. Acad. Sci. USA86, 5325—5329.

SwinnenJ.V., TsikalasK. E., andConti M. (1991) Propertiesandhormonalregulationof two structurallyrelatedcAMPphospho-diesterasesfrom therat Sertoli cell. J. Biol. Chem.266, 18370—18377.

Torphy T. J., Zhou H. L., and Cieslinski L. B. (1992) Stimulationof beta adrenoceptorsin a humanmonocytecell line (U937)up-regulatescyclic AMP-specificphosphodiesteraseactivity. J.Pharmacol.Exp. Ther. 263, 1195—1205.

TorphyT. J., ZhouH. L., Foley J. J., SarauH. M., Manning C. D.,andBametteM. S. (1995)SalbutamolupregulatesPDE4activ-ity and inducesaheterologousdesensitizationof U937cells toprostaglandinE2. J. Biol. C/acm.270, 23598—23604.

VergheseM. W., McConnell R. T., LenhardJ. M., HamacherL.,

and Jin S.-L. C. (1995) Regulation of distinct cyclic AMP-specific phosphodiesterase(phosphodiesterasetype 4) iso-zymesin humanmonocyticcells. Mol. Pharmacol. 47, 1164—1171.

WachtelH. and SchneiderH. H. (1986) Rolipram, a novel antide-pressantdrug, reverses the hypothermiaand hypokinesiaofmonoamine-depletedmice by an action beyondpostsynapticmonoaminereceptors.Neuropharmacology25, 1119—1126.

Whalin M. E., Garrett R.L., ThompsonW. J., and Strada S. J.(1989)Correlationof cell-freebrain cyclic nucleotidephospho-diesteraseactivitiesto cyclic AMP decayin intactbrain slices.SecondMessengersPhosphoproteins12, 311—325.

Wirier B. J. (1971) Inference with respectto meansand variance,in StatisticalPrinciplesin ExperimentalDesign(MaythamW.,ShapiroA., and SternJ., eds),pp. 4—54. McGraw-Hill, New

York.Ye Y. andO’DonnellJ. M. (1996)Diminishednoradrenergicstimu-

lation reducesthe activity of rolipram-sensitive,high-affinitycyclic AMP phosphodiesterasein rat cerebralcortex.J. Neuro-

chem. 66, 1894—1902.

J. Neurochem.,Vol. 69, No. 6, 1997