Clinical and Experimental Allergy. 1995. Volume 25, pages 616-624

Cyclic nucleotide phosphodiesterases from purified humanCD4+ and CD8+ T lymphocytes

H. TENOR*\ L. STANICIU*, C. S C H U D T " , A. HATZELMANN',A. WENDEL^ R. DJUKANOVIC*, M. K. CHURCH* and J. K. SHUTE*

* Immunopharmacology Group. University of Southampton, Southampton. UK. ^University of Konstanz, Faculty ofBiology, Konstanz, and ' Byk Gulden Pharmaceuticals. Konslanz. Germany

Summary

Background CD4"^ and CDB"̂ T-lymphocytes are suggested to differentially affectairway inflammation in asihma. Agents whieh inerease intracelluiar cAMP levels, suehas PDE inhibitors, have been shown to diminish lymphocyte growth and differentia-tion, and to affeet eytokine expression. DiflTerenees in the PDE isoenzyme profilebetween CD4"^ and CD8^ cells might form a basis to differentially modify theirfunctions by PDE inhibitors.Objective The study investigates and compares the PDE isoenzyme activity profiles ofhuman peripheral blood CDA^ and CDS"̂ T-lymphocytes.Methods CD4"̂ and CD8"^ T-lymphoeytes were puritied (>98%) from peripheralblood mononuciear cells by negative selection. PDE isoenzyme activity profiles wereinvestigated using PDE isoenzyme selective inhibitors and activators.Results In CD4^ and CD8^ T-lymphoeyte homogenates. PDE IV and PDE IIIactivities were the predominant PDE isoenzyme activities at 0-5 ;iM eyelie nueleotidesubstrate concentrations. PDE IV was localized in the soluble fraction whereas PDEIII was membrane bound. Low PDE I. II and V aetivities were detected. About 20% oftotal eAMP hydrolysing capacity at 0 5 / JM C A M P was insensitive to PDE isoenzymeselective inhibitors and activators and therefore could not be assigned to PDE 1-IV.The PDE isoenzyme pattern was not different between CD4"^ and CD8"^ T-lympho-cytes. Moreover, representative inhibitors of PDE III and IV aetivity inhibited cAMPhydrolysis in soluble fractions of both T-lymphocyte subsets with similar potency.Enzyme kinetic analysis similarly did not reveal differences between CD4"^ and CD8"^T-lymphoeytes.

Conclusion Normal CD4"^ and CD8^ T-lymphocytes are likely to be equally sensitivetargets for the effects of PDE inhibitors.

Keywords: phosphodiesterase, T-lymphocytes, airway inflammation, eytokineexpression

Clinical and F.xperimentai Allergy. Vol. 25. pp. 616-624. Submitted 19 October 1993;revised 9 January 1995; accepted 20 January 1995.

, . . ,. response to inhaled allergens in atopie asthmatic indi-Introduetion / i m c . .u u

viduals [1]. For example, an inerease in the numberThere is growing evidence for a pivotal role of CD4^ and of activated CD4^ eells expressing proinfiammatoryCD8"^ T-lymphoeytes in orchestrating the inflammatory eytokines in bronehial biopsies following allergen

challenge has been reported [2]. CD8"^ cells, however.Correspondence: DrJ.K.Shute. ClinicalPharmacologyGroup, Centre . ^ ^ • aQ, I c , I c .u ™ . r- 1 u . 1 c .u . <7r»it are proposed to attenuate the inflammatory response viaBlock, r Level. Southampton General Hospital. Southampton SO16 i r ,6YD. UK. down-regulation of eytokine production by CD4 TH-2-This work was supporied by Byk Gulden Pharmaceuiicals. lymphocytes [3]. Activation of many proinflammatory616 © 1995 Blackwell Science Lid

Cyclic nucelotide phosphodiesterases from purified human CD4^ and CD8^ T lymphocytes 617

cells by cytokines and other mediators of inflammationin asthma is inhibited by targeting the cells with drugswhich enhance intracellular cAMP levels [4]. Agentswhich increase cAMP have been shown to diminishlymphocyte growth and differentiation [5] and expres-sion of the proinflammatory cytokines, lL-2 and IFN7[6]. Intracellular levels of cAMP can be modified bychanging the rate of either biosynthesis or breakdown,of which the latter occurs via the activity of cyclicnucleotide hydrolysing phosphodiesterases (PDE, EC3.1.4.7.).

Seven families of PDE isoenzymes have been describedand characterized by their difl'erent susceptibility towardsactivators and inhibitors [7.8]. The activities of PDE Iand II are stimulated by Ca' ' calmodulin and cGMP.respectively and hydrolyse cAMP and cGMP withsimilar aflinities. The hydrolysis of cAMP by PDE IIIis inhibited by cGMP. PDE IV activity represents the'low KM" cAMP-specific activity which has no knownintracellular regulator, but may be characterized byinhibition by Roiipram. PDE V specifically hydrolysescGMP. PDE VI is the G-protein regulated PDE iso-enzyme found in the retina [7]. PDE VII specificallyhydrolyses cAMP and its activity is not affected byany known inhibitor or activator [8], The regulatorycharacteristics of these enzymes, together with the useof isoenzyme selective inhibitors, form the basis of assaysto determine PDE isoenzyme activity profiles in cells andtissues.

Cyclic nucleotide hydrolysing PDE activities in humanlymphocytes have previously been described [9-14] andthe presence of PDE 111 and IV was recently demonstratedin purified T-lymphocytes [14]. No significant difl'erence intotal cAMP PDE activity in CD4' and CD8^ T cells hasbeen reported, although the PDE isoenzyme activityprofile in these lymphocyte subsets was not analysed [15].

To extend previous studies, we have investigated thePDE isoenzyme profiles in freshly prepared human peri-pheral blood CD4' and CD8^ T-lymphocytes obtainedfrom normal individuals.

Materials and methods

Materials

RPMI 1640; glutamine; heat inactivated fetal calf serum(FCS); sodium pyruvate; streptomycin/penicillin werefrom Gibco Life Sciences BRL {Paisley, UK). Mono-clonal antibodies for lymphocyte isolation (anti-CD33;CD14; CD19; CDIlb; CD16) were purchased fromSerotec Inc (Oxford. UK). Culture dishes. EITC andPE-iabelled monoclonal antibodies for FACS analysis ofthe purity of the T-lymphocyte preparations (anti-CD14:

CD19; CDllb; CD16: CD3; CD4; CDS; CD25) weresupplied by Becton Dickinson (Oxford. UK). Rat anti-mouse IgG was from Dakopatts (High Wycombe. UK).Anti-CD4; anti-CD8 MACS microbeads; MACS col-umns and the MACS separator were from MiltenyiBiotec GmbH (Bergisch Gladbach, Germany). [ H]cAMP and [''H] C G M P was purchased from NENDupont (Stevenage, UK). QAE Sephadex A25 wassupplied by Pharmacia AB (Uppsala, Sweden). UltimaGold scintillation fluid was from Packard (Pangboume,UK). Hepes, /^-mercaptoethanol, MgCK, EGTA, NaCl,KCl, KH2PO4, Na2HPO4 and Tris were purchased fromBDH (Poole, UK). Crotalus atrox snake venom. cAMP,cGMP. pepstatin A. leupeptin. PMSF. soybean trypsininhibitor, benzamidine were from Sigma (Poole, UK).Rolipram was a gift from Schering AG, Berlin,Germany. Zaprinast was a gift from Rhone Poulenc,Dagenham, UK. Motapizone was a gift from Natter-mann, Cologne, Germany. Calmodulin was provided byProfessor Gietzen, University of Ulm, Germany.

Purification of CD4^ and CDS ^ T-lymphocytes fromhuman peripheral venous blood

Several difl'erent methods were combined to a "multi-stepprocedure' [16] in which separation of highly purifiedCD4* and CD8^ T-lymphocytes was achieved withoutactivation of the subpopulations. In brief, mononuclearcells were isolated from lOOmLof heparinized peripheralvenous blood from normal (non-atopic) volunteers bycentrifugation on lymphoprep. In a second step, plateletcontamination of PBMC was reduced as described byPawlowski et al. [17] with modifications: PBMC werewashed with autoiogous plasma containing 5 mM EDTA(four times) and 1 mM EDTA (three times). Cells wererecovered from each wash by centrifugation at 220g for10 min at 20°C. Removal of platelets from PBMC iscrucial because they contain substantial amounts ofsoluble PDE I, II. Ill and V [18]. Monocytes wereremoved by adherence for 1 h at 37'C. B cells, NKcells, remaining monocytes and granulocytes wereremoved by 'panning* as previously described [19].Finally, CD4' and CD8^ cells were isolated by negativeselection using immunomagnetic MACS microbeadsand the MACS separator as previously described [16].FACS analysis was performed to confirm purity of thepreparations.

Preparation of .soluble and particulate fractions fromhomogenates of CD4^ and CDH^ T-lymphocytes

Freshly prepared cells were sedimented and resuspendedin homogenization buffer (lOmM Hepes pH8 2, 1 mM

© 1995 Blackwell Science Ltd, Clinical and E.xperimental Allergy, 25. 616 624

618 H. Tenor c\.a\.

/?-mercaptoethano!. 1 mM MgCL, I mM EGTA, 137mMNaCi.2 7mM KCl. 1-5 mM KH.POj.S-l mM N 25/(M pepstatin A, 10//M leupeptin. 50//M PMSF.soybean trypsin inhibitor. 2mM benzamidine). Aftersonication (Soniprep; 2 microns (8W); 90s) the homo-genate was centrifuged al IHOOOOxj? for I h at 4 C.Supernatant and precipitate were separated and theresidue was resuspended in ;m equal volume of homo-genization buffer. The supernatant and resuspendedresidue were used as soluble and particulate fraction,respectively, for determination of PDE activity.

Measurement of PDE activity

PDE activity was determined as described by Thompsonet al. [20] with some modification [21]. Briefly, theenzyme containing fractions were assayed in a finalvolume of 200/iL containing 40 mM Tris HCl pH K 0,5mM MgCN, 0 01/iM-250/iM cAMP or cGMP,40000cpm [ 'H] C A M P or [̂ H] cGMP, activators andinhibitors. This reaction mixture was incubated lor15 min at 37"C. The reaction was terminated by adding50/iL 0 3N HCl and the assay mixture was left on ice forlOmin. Crotalus atrox snake venom (100//g) was addedfollowed by further incubation for 15 min at 37 C.

Table I. The eoncentralion of PDE inhibitor or activator whichspecifically and completely inhibits or activates ihe appropriateisoenzyme at 0-5 ̂ iM cyclic nucleolide subslrale concenlralions

100 nRA

IV

-9 -8 -7 -6 -5 -4Drug Hog M]

Fig. I. PDE isoenzyme analysis from soluble and particulatefractions of CD4' and CD8' T-Iyniphocyte homogenales.Concentration-dependent inhibition of cAMP-PDE activityfrom soluble fractions of CD4* cells at 05^iM cAMP byrolipram. motapizone and zardaverine is shown (mean ofthree e.\periments±sEM). The minimum concentrations of thePDE inhibitors roiipram (!O/;MJ. motapizone (I//M) andzardaverine (KI/AM) which completely and selectively inhibitthe corresponding PDE isoenzyme. have been derived fromthese concentration-inhibiton curves and are shown in Table 1.The amount ofcAMP hydrolysis inhibited by 10//M rolipram.I /(M motapizone. \OftM zardaverine then corresponds to PDEIV. PDE 111 and the sum of PDE 111 and PDE IV activities,respectively (shown by bars) present in the cellular homoge-nates at 0-5//M cAMP. Residuiil activity (RA) comprises basal(unstimulated) PDE I and PDE II activities as well asundeilnedcAMP hydrolysing PDE activities.

Addition

None1 //M Motapizone (M)IO//M Rolipram (R)IO//M Zardaverine (Zd)IO//M Zaprinast (Zp)250//M CaCl:+ l50nM

Calmodulin (CaCaM)imM EGTA (EGTA)5//,M cGMP(cGMP)

isoenzymespecifity

Control111IVIII/IVV

1I11

Inhibitor(l)Activator(A)

_

IIII

AIA

Finally, the assay mixture was loaded onto QAE-Sephadex A-25 columns (I mL bed volume) and elutedwith 2 ml of 30 mM atnmonium fonnate. pH6 0. Theeluate was counted for radioactivity in a Beckman liquidscintillation counter. All assays were carried out induplicate. Blank values, which were obtained by includ-ing either buffer or denatured enzyme, were below 1 % oftotal radioactivity.

Analysis of PDE isoenzyme activities

PDE isoenzyme activities were determined in soluble andparticulate fractions of human T-lymphocyte hotno-genates using PDE isoenzyme selective inhibitors andactivators at 05//.M cyclic nucleotide substrate con-centrations (cAMP or cGMP). Frotn concentration-inhibition curves for the isoenzyme selective inhibitorsmotapizone (PDE III), rolipram (PDE IV). zaprinast(PDE V) and zardaverine (PDE III/IV) IC50 values andfractional inhibition of basal activities are derived.Figure 1 shows the concentration dependent effects of

Table 2. The formulae for calculating PDE isoenzyme activities.Isoenzyme activities are calculated from the diflerence in PDEactivity in the presence and absence of fixed concentrations ofspecific activators and inhibitors, as shown

Isoenzyme

PDE IPDE 11PDE IIIPDE IVPDE V

Substrate

cGMPcAMPcAMPcAMPcGMP

Calculation of isoenzyme activity

Ca CaMR + M + cGMPControlControlM

EGTA- R-hM- M- R- M+Zp

© 1995 Blackweli Science Ltd, Clinical and Experimenml Allergy. 25, 616-624

Cyclic nucelotide phosphodiesterases from purified human CD4^ and CD8^ T lymphocytes 619

the PDE inhibitors roHpram, motapizone and zardaver-ine on the hydrolysis of 0-5/(M cAMP in soluble frac-tions from CD4^ cells. The concentrations of PDEinhibitors which completely and selectively inhibit thecorresponding isoenzyme are derived from Fig. 1 andsummarized in Table I, These concentrations of PDEinhibitors are applied to calculate PDE isoenzyme activ-ities. PDE III and PDE IV activities are defined as thedifference between cAMP-PDE activity at 0 5 /iM cAMPin the presence and absence of 1 fiM motapizone and10//M rolipram, respectively, PDE V is calculated fromthe difference ofthe motapizone-insensitive cGMP-PDEactivity {0 5/zM cGMP-|-1 pM motapizone) in the pres-ence or absence of 10 //M zaprinast. Motapizone is addedto block cGMP hydrolysis by PDE III. PDE I activity isdefined as the increment of EGTA-insensitive cGMPhydrolysis in the presence of 250/iM Ca^^/150nMcalmodulin. PDE II activity is calculated from theincrease of residual cAMP hydrolysis (insensitive tolO/zM rolipram and 1 //M motapizone) by the additionof 5 /iM cGMP. The concentrations of Ca""^/calmoduiinand cGMP used have been demonstrated to maximallyactivate PDE I and PDE II activities in other systems(our own unpubhshed observations). The completeprocedure for calculation of PDE I-PDE V activities issummarized in Table 2.

Statistical analysis of data

Calculations of mean, standard error of mean (SEM), and/-tests were performed using the Instat program fromGraphpad Software Inc (San Diego. USA), Two-siteanalysis of concentration inhibition curves and calcula-tion of IC50 values were carried out using the InPlot pro-gramme from GraphPad Software Inc (San Diego, USA).

Results

T-lymphocyte preparations obtained by panning were

>98% pure CD3^ cells and contained < 10% neutro-phils. < 0 5% monocytes and < 2% B cells. Separation ofCD4* and CDS' T-iymphocytes by negative immuno-magnetic selection resulted in < 1% contamination withceils ofthe opposite phenotype. PDE activity was thereforedetemiined in T-lymphocyte subsets purified by negativeselection which further ensured no activation of the isolatedcells as a result of binding to antibody coated beads [L,Stanciu et al., manuscript in preparation],

cAMP- and cGMP-PDE activities in subcellularfractionsfrom CD4^ and CD8^ T-lymphocytes

The cAMP-hydrolysing PDE activities (0 5//M cyclicnucleotide as substrate) in the soluble and particulatefractions of CD4^ and CD8^ T-lymphocyte homoge-nates were not significantly different between the T-cellsubsets (Table 3). The cGMP-PDE aetivity at 0 5/iMsubstrate in the soluble fraction of CD8^ cells wassignificantly (P < 0 05) higher than in CD4 ' cells. ThecAMP-PDE activity in soluble fractions was about three-fold higher P < 0 001) than in particulate fractions. ThecGMP-PDE activity was equally distributed betweensoluble and particulate fractions in CD4' cells, but wassignificantly {P < 0 05) higher in the soluble fractioncompared with the particulate fraction of CDS' cells.

Kinetic analysis, using the Eadie-Hofstee-plot. of thesoluble cAMP-PDE activities from CD4+ and CD8+ cellsrevealed high "affinity" (KM = 005/(M and 008//M) andlow "affinity' (K^ — 2 5/(M and 16/iM) components ascalculated by the method of Spears et al. [22] (Table 4). Atypical analysis is shown in Fig, 2. Hydrolysis of cAMPin particulate fractions and cGMP hydrolysis in bothfractions exhibit single-site linear enzyme kinetics andthe kinetic parameters evaluated (KM; VMAX) ^re shownin Table 5. No substantial differences between kineticparameters of CD4"^ and CD8 ' T-lymphocytes werefound. To further investigate the nature of the high"affinity' component of cAMP hydrolysis in soluble

Table 3. Subcellular fractions of human CD4 and CD8* T-lymphocytes were analysedfor PDE activity at 0-5/iM cAMP or 0-5/YM cGMP substrate concentralions. Resultsare the mean ± SEM ofseven lymphocyte preparations, *P < 005 compared with cGMPhydrolysis in soluble fractions of CD4"^ cells

Substrate

PDE activity pmol X min x 10 cells 'CD4+ CD8"

Soluble Particulate Soluble Particulate

0 5 / JM C A M P

05//M cGMP50-6 ± 2 369 ±0-2

17-7±l'35-5 ±0-2

51 0 ± 3 2 ]8-2±2-l6 4±0-6

D 1995 Blackwell Science Ltd. Clinical and Experimental Allergy. 25. 616 624

620 H. Tenor et al.

Table 4. Enzyme activity was measured at 001 /(M-250/(M cAMP or cGMP substrate concentrations in soluble and particulatefractions of peripheral blood CD4* and CD8"̂ T-lymphocyte homogenates from three normal individuals. Data were analysedaccording to Eadie-Hofstee. Kinetic parameters were calculated according to the method of Spears [22] and are given asmeanisEMof three independent experiments

Soluble

Particulate

Soluhle

Particulate

•low''high'

'low''high'

cAMP

[/.M]

25±010-05 ±0 010-27 ±0 03

l-6±010-08 ±0-0050-27 ±0 01

hydrolysis

[pmol/min x 10̂ cells]

246 ± 2610-2 ±1-929-2±0 2

172±718 ±3

26-3±ll

4-8 ±0-9

0-3 ±0-05

5-2 ±0-4

018 ±0 02

cGMP hydrolysis

[pmol/min X 10̂ cells]

80-2 ±10-2

7-6 ±2-4

92-6 ±6-2

8-5 ±1-4

fractions, kinetic analysis was performed in the presenceof either 3//M motapizone (to block PDE III activity) or30//M 2ardaverine (to block PDE III and PDE IVactivity). The presence of 3/iM motapizone did notaffeet the two-component kinetics of cAMP hydrolysisin soluble fractions and its kinetic parameters (K^iVMAX)- Therefore motapizone-inhibitable PDE III isnot present in soluble fractions, an observation whichis further supported by the data in Table 6 showing that

0 50 100 150 200

v/sFig. 2. Kinetics of cAMP-hydroIysis (0-01/iM-250/iM) insoluble fractions of CD4' T-lymphocyte homogenates. Dataare presented according to Eadie-Hofstee. V. Enzyme activity.S. cAMP concentration (//M).

no inhibition is observed in the soluble compartment bymotapizone. In the presence of 30/iM zardaverine alinear kinetic plot, corresponding to the high aflinitycomponent, was found. This was reproduced in twoindependent experiments. This suggests that the high"affinity" component in the soluble fraction is neitherPDE III nor PDE IV. and that PDE IV is present in thelow 'affinity" component.

Inhibition ofcAMP-PDE activity in subcellular fractions

Inhibition of cAMP hydrolysis (0 5^M cAMP substrateconcentration) in the subcellular fractions by theselective inhibitors of PDE IV (rolipram). PDE III(motapizone) and PDE III/IV (zardaverine) was

Tahle 5. IC50 (0-5/iM cAMP) for rolipram. motapizone andzardaverine were calculated from conceniration-inhibitioncurves as shown in Fig. I using the InPlot program. Biphasiccurves were evaluated with a two-site model and IC50 valuescorresponding to specific Inhibition are presented. Numbers Inparentheses represent per cent inhibiton of cAMP hydrolysis at100 /iM inhibitor concentration (corresponding to non-specificinhibition). Values given are mean ± SKM {n = 3)

CD4+SolubleParticulate

SolubleParticulate

Rolipram

0-12 ±0-01(24-5 ±3)

0-13 ±0-005{30-2 ±1)

IC50 [fiM]Motapizone

(48-2 ±5-1)0-013 ±0-001

(42-3 ±3-2)Oil ±0001

Zardaverine

0 51 ±010 58 ±0 09

052±01043 ± 0-06

© 1995 Blackwell Science Ltd. Ctinicat and Experimental Aitergy, 25. 616-624

Cyclic nueelotide phosphodiesterases from purified human CD4^ and CD8^ T lymphocytes 621

Table 6. Effects of PDE isoenzyme-specific activators andinhibitors (Table I) on cyclic nucleotide hydrolysis in sub-cellular fractions of CD4"^ cells. Data are the mean ± SEMfrom seven lymphocyte preparations.

0 5 nM cAMP+R+M+R4-M+Zd+ R + M -FcGMP+ R + M + C a / C a M

0 5/iM cGMP+ M+M+ZpEGTACa/CaM

pmol >

Soluble

50-6±2 314 4 ±0 950'2±2'8121 ±0 912'3±0-713'3±I'O13'6±l-2

6-9 ±0-26 5 ±011-8 ±0-36-8±0 28-6 ±0-3

PDE activity:min"' x 10^cells"'

CD4+

Particulate

17-7±l-316-3 ±3-02-1 ±0-51-7 ±0-61-6 ±0-40-7 ±0-31-5 ±0-4

5-5 ±0-22-1 ±0-91-3 ±0-45-4 ±0-35-6 ±0-2

eel

oX

c1X1

o£Q.>•

ivit

act

LUQQ..

investigated using the inhibitors in the concentrationrange IO''-1O "^M. IC50 values determined accordingto Fig, 1 were not significantly different between CD4"^and CD8+ cells (Table 5). IC50 values of two-phaseconcentration-inhibition-curves as obtained using roli-pram and motapizone are calculated for the first phase ofthe curves representing specific inhibition of PDE IV andPDE III, respectively (for details see Fig. 1).

60 -

40

40

20

(b)

Fig. 3. The PDE isoenzyme activity profile of peripheral bloodCD4+ and CD8^ T-lymphocytes at 0 5 ^ M cAMP/cGMPsubstrate concentration. PDE isoenzyme activities fromsoluble and particulate fractions of CD4"^ and CD8''" eel!homogenates. Results are given as mean ± SEM from sevenlymphocyte preparations. No significant differences (non-paired /-tests) were found between isoenzyme activities in

^ andCD8+ T-lymphocytes.

PDE isoenzyme activity profiles in subeellular fractionsfrom CD4^ and CD8^ cells

PDE isoenzyme activity profiles from human peripheralblood CD4 and CD8' cells were derived from each ofseven normal subjects. Effects of PDE isoenzyme selec-tive activators or inhibitors, at concentrations known tocompletely and selectively activate or inhibit the corre-sponding PDE isoenzymes (Table 1), on cAMP- andcGMP-PDE activities from subcellular fractions ofCD4^ cells are shown in Table 6. Data obtained forCD8' cells (not shown) were essentially the same. Fromthese data PDE isoenzyme activities at 0 5/iM cyclicnucieotide substrate concentration are calculated asdescribed in Table 2 (for details see Methods) andshown in Fig. 3. The predominant PDE isoenzymes inCD4+ and CD8+ cells were identified as PDE IV andPDE III, Whereas PDE IV activity was exclusivelylocated in the soluble fraction., PDE III was membrane-

bound. The activities of PDE I. PDE II and PDE V wererelatively low. The PDE isoenzyme activity profiles fromCD4' and CDS*̂ T-lymphocyte subcellular fractionswere not significantly different (see Fig. 3). About 20%of total cAMP-PDE activity at 0 5 //M cAMP in both T-iymphocyte subsets was insensitive to inhibitors of PDEIII and PDE IV and activators of PDE I and PDE II(Table 6) and therefore could not be assigned to PDE 1-IV activities. Similarly, about 20% of cGMP hydrolysiswas not inhibited by zaprinast and motapizone. Thisactivity is therefore most probably due to a PDEisoenzyme activity different from PDE III or PDE V.

The PDE isoenzyme activity profiles depend on cyclicnucleotide substrate concentrations due to the differentKM values of the PDE isoenzymes. As demonstrated inTable 7, at 5/iM cAMP substrate concentrations PDEIV and PDE III activities increased by approximatelythree-fourfold and 1 4-fold, respectively. Even at

1995 Blackwdt Science Ltd, Clinical and Experimental Allergy. 25. 616-624

622 H. Tenor et al.

Table 7. PDE IV activities in the soluble fractions and PDE 111activities in the particulate fractions of CD4* and CDS' cellhomogenates were determined ai 0 5//M and 5//M c.'\MP asdescribed in Table 2. Results are presented as mean ± SKM fromseven (0-5//M cAMP) and three (5//M cAMP) experiments

CD4'PDEPDE

CD8+PDEPDE

IVIII

!VIII

(soluble)(particulate)

(soluble)(particulate)

PDEpmol X min

0-5 ^M cAMP

37-2 ±3-5I5 5±IO

42-9 ±6-9I7'l :t2-5

activity' xio" cells"'

5/xM cAMP

157-1 ±14-222-8 ±2-7

138-5 ±7-322-9 ±4-5

higher substrate concentrations (3/(M and 20//M cyclicnucleotide) which are more appropriate to their corre-sponding KM values [23] no substantial PDE I or PDE IIactivities were found.

Discussion

The putative different role of CD4" and CD8 ' T-]ymphocytes in the development of chronic airwayintiatnmation in bronchial asthma [1.3] and the poten-tial anti-inflammatory effects of PDE inhibitors [24] gaverise to this investigation of the PDE isoenzyme activityprofi]e of human peripheral blood CD4* and CD8 ' T-lymphocytes. PDE isoenzyme activity profiles wereanalysed in subcellular fractions of CD4' and CDB""T-lymphocytes by investigating the effect of isoenzymeselective inhibitors or activators on cyciic nucleotidehydrolysis at 0 5//M cAMP/cGMP (Fig. 1. Table 1).This approach is a useful allernative to conventionalchromatographic analysis which is hampered by both thelow cell number ( 1 2 x 10 cells) obtained for each sub-ject and technical difficulties in recovery and analysis ofPDE isoenzymes from solubilized particulate fractions.

The predominant PDE isoenzyme activities in hotno-genates of CD4' and CD8' T-Iytnphocytes at 0-5//Mcyclic nucleotide substrate concentrations were PDE IIIand PDE IV activities. These enzyme activities wereclearly compartmentalized, such that PDE IV activitywas almost exclusively locahsed to the soluble fraction,whereas PDE III activity was almost completely mem-brane bound. Relatively low levels of PDE I, PDE II andPDE V activities were detected in both T-lymphocytesubsets. Overall, the PDE isoenzytne activity profileswere not significantly different between CD4' and

CD8' cells. In addition, the IC50 values for inhibitionby selective inhibitors of PDE III and PDE IV activities(motapizone. rolipram. zardaverine) of PDE isoenzymeactivities in soluble and particulate fractions were notdifferent in CD4' and CD8^ celis (Table 5). These IC50values are in the range of the values reported forneutrophils [25]. endothe]ial cells [26], eosinophils (27],alveolar macrophages [28], human bronchi [29] or pul-monary artery [30]. The PDE isoenzyme activity profilein subcellular fractions of CD4'^ and CD8^ cells isreflected in the kinetic analysis of cyclic tiucleotidehydrolysis. The K^ values shown in Table 4 supportthe evidence for the presence of PDE IV and PDE Vactivities in the soluble fractions and PDE III activity inthe particulate fractions. The distinct subceilular locali-zation of PDE III and PDE IV activities in humanperipheral blood T-lymphocytes, although not in sub-sets, has been reported previously [13].

A proportion of the total soluble cAMP-PDE activityat 0-5/iM cAMP could not be assigned to either PDEI-IV. This was in view of the fact that part of the activitywas insensitive to the PDE lII/lV inhibitor zardaverineand the activators of PDE I and PDE II, namelyCa ' '^-calmodulin and cGMP (Table 6). In this respect, itis interesting that a cAMP-specific PDE activity(KM < 0-5/iM). which is insensitive to inhibitors ofPDE III and PDE IV and activators of PDE 11 andPDE I, has recently been isolated from human Tcell lines[31]. Our own data indicate that a similar PDE activity ispresent in the soluble fractions orCD4 andCD8' cells,ie the activator inhibitor in.sensitive cAMP-PDE activitywhich represents about 20% of the total activity. Thissuggestion is further supported by data from Robiscek etal. [13] who previously identified, by chromatographicanalysis, a cAMP-specific PDE activity in the solublefraction of human T-lymphocytes which is insensitive toinhibitors of PDE ill and PDE IV. Interestingly, acAIVlP-specific PDE activity (KM =0-2/iM) which isinsensitive to inhibitors of PDE III/IV has recentlybeen cloned and termed PDE VII [8], but its identity tothat found in lymphocyte preparations is unknown,although PDE VII was recently identified in the humanT cell line HUT 78 [32]. Despite the presence of proteaseinhbitors in the homogenization buffer in our experi-ments, it also cannot be excluded that a product resultingfrom proteolysis of one of the PDE isoenzymes mayaccount for the activator/inhibitor-insensitive cAMP-PDE.

PDE V activity in preparations of mononuclear cellshas previously been shown to originate from contam-inating platelets [10], In our protocol, an extensivewashing procedure reduced platelet contamination to< 1 platelet/100 T-lymphocytes.

1995 Blackwell Science Ltd, Clinical and E.xperimenlal Allergy, 25. 616-624

Cyclic nucelotide phosphodiesterases from purified human CD4^ and CD8^ T lytnphocyles 623

Nevertheless, we also cannot completely exclude thepossibility that the trace amount of PDE V activityfound in our preparations derived from remainingplatelets. About 20% of cGMP hydrolysing PDEactivity could be classified as PDE V or PDE HI. It isconceivable that this inhibitor insensitive cGMP-PDE activity indicates basal PDE I activity whoseactivity cannot be stimulated by Ca' '-Calmoduliii.This may be due to limited proteolysis of PDE I whichhits been described to result in the loss of sensitivity toCa * '-Calmodulin [331. Alternatively zaprinast insen-sitive cGMP-PDH may be the consequence of redticcdsensitivity of PDE V to zaprinast due to phosphory-lation of PDE V [34]. The soluble cGMP-PDE activitywas significantly higher in CD8' than in CD4' T-lymphocytes, although there was no significant differ-ence in PDE V. PDE III and PDE I activities. Also, theinhibitor-insensitive cGMP-PDE activity was not sig-nificimtly higher in CD8^ than in CD4* cells. Ourresults do not. therefore, provide a straightforwardexplanation for the observed difference between CD4'andCD8 ' cells.

The PDE isocn/yme activity profile of human periph-eral blood CD4' and CD8* cells, as described here, isreflected in the results of our previously reported func-tional studies. For example, proliferation of anti-CD3-stimulatcd human T-lymphocytes was almost completely(90%) inhibited by zardaverine. whereas it was onlypartially attenuated by motapizone (30%) and rolipram(40%) at concentrations which selectively inhibitPDH lll and PDE IV [35]. Very similar results wereobtained by Robiscek et al. [13] for the attenuation ofPHA-stimulatcd T-lymphocytc hlastogenesis by selec-tive PDE III and PDE IV inhibitors. More recently,Giembycz et al. [36] investigated effects of the PDE IVseiective inhibitor, rolipram. and the PDE III selectiveinhibitor. SKF 95654. on PHA-stimulated proliferationand cAMP concentrations of CD4' and CDS' T-]ymphocytes. The functiona] effects of the PDE inhibi-tors were not different between CD4' and CDS' cells.SKF 95654 was inactive when given alone, however, itenhanced the antiproliferative effects of rolipram whichwas paralleled by an increase in cAlVIP levels. Thediscrepancy between the presence of PDE III in CD4'and CD8' cells and the apparent functional ineffective-ness iif PDE III inhibition in the lalter study may be dueto the intracellular cAMP concentrations which occur insitu. As shown in the present paper (Table 7) an increasein cAMP concentration augments the PDE IV/PDE IIIactivity ratio and hence, inhibition of PDE III activitybecomes functionally less effective. In addition. PDE IIImay become activated by phosphorylation, a mechanismmediated by a cAMP-activated protein kinase [37]. These

mechanisms should therefore be taken into accountwhen assessing the relative efficacy of either PDE III orPDE IV inhibitors in inhibition of functional responses.

In summary. PDE III and PDE IV activities wereidentified as the predominant PDE isoenzyme activitiesin human peripheral blood CD4' and CDS' T-lympho-cyte subsets. PDE IV was localized in the soluble fractionwhereas PDF III was membrane-bound. Additionally,low levels of PDE I, PDE II and PDE V activities weredetected with 0-5/i,M substrate concentration. At highersubstrate concentrations, the PDH IV PDF III activityratio was elevated but again minimum PDF I and PDF IIactivities were found. Of the total cAMP hydrolytJccapacity in the soluble fractions. 20% could not beassigned to PDF I-IV. since il specifically hydrolysescAMP with a low KM and was not affected by rolipram,motapizone, cGMP or Ca^ ^-calmodulin. The PDFisoenyume pattern, the total activity, and the subcellulardistribution of the enzymes were not different betweenCD4' and CD8" T-lymphocytes.

References

1 Holgate ST. Asthma: past, present and future. Eur Resp J199.1; 6:1507 20.

2 Beniley AM. Meng Q. Robinson DS et al. Increases inactivated T-Iymphocytes. eosinophils, and cytokine mRNAexpression for interleiikin-fi and granulocyte/macrophagecolony-stimulating factor in bronchial biopsies after aller-gen challenge in atopic asthmaties. Am J Respir Cell MolBiol 1993:8:35-42.

3 Noble A. Staynov BZ, Diaz-Sanchez D. Lee TH. KemenyDM. CDK' T cells regulate the production of ThI and Th2cytokines. J Allergy Clin Immunol 1994; 93:261.

4 BarnesPJ. New drugs for asthma. Eur Resp J 1992; 5:1126-32.

5 Kammer GM. The adenyiale cycIase-cAMP-protein kinaseA pathway and the regulation of the immune response.Immunology Today 1988: 9:222-8.

6 Stam WB. Van Oosterhout AJM. Nijkamp FP. Pharmaeo-logic nioduhition ofTHl- and TH2-associated lymphokineproduction. Life Sci 1993; 53:1921-34.

7 Beavo JA. Reifsnyder DH. Primary sequence of cyclicnucleotide pho.sphodiesterase isozymes and the design ofseiective inhibitors. TIPS 1990: 11:150-5.

8 Michaeli T. Bloom TJ. Martins T et al. Isolation andcharacterization of ;i previously undetected human cAMPphosphodiesterase by complementation of cAMP phospho-diesterase-deficient saccharomyees cerevisiae. J Biol Chem1993; 268:12925-32.

9 Wedner HJ. Chan BY. Parker CS. Parker CW. Cyclicnucleotide phosphodiesterase activity in human peripheralblood lymphocytes and monocytes. J Immuno! 1979;123:725 32.

10 Epstein PM, Hachisu R. Cyclic nucleotide phosphodiester-ase in normal and leukemic human lymphocytes and

(j"i 1995 Blackwell Science Ltd, Clinical and E.xperimenlal Allergy, 25. 616-624

624 H. Tenor et al,

iymphoblasts. In: Strada SJ, Thompson WJ, eds. Advancesof Cyclic Nucleotide and Protein Phosphorylation ResearchNew York: Raven Press, 1984:304-21.

U Takemoto DJ, Lee WNP. Kaplan SA, Appleman MM.Cyclic AMP phosphodiesterase in human lymphocytes andIymphoblasts. J Cyclic Nucl Res 1978; 4:123-2.

12 Thompson WJ, Ross CP, Pledger WJ. Strada SJ. Cyclicadenosine 3':5'-monophosphate phosphodiesterase. Dis-tinct forms in human lymphocytes and monocytes. J BiolChem 1976:251:4922-9.

13 Robicsek SA, Krzanowski JJ, Szentivanyi A, Poison JB.High pressure liquid chromatography of cyclic nucleotidephosphodiesterase from purified human T-lymphocytes.Biochem Biophys Res Commun 1989; 163:554-60.

14 Robicsek SA, Blanchard DK. Djeu FY et al. Multiple high-affinity cAMP-phosphodiesterases in human T-lymphocytes.Biochem Pharmacol 1991; 42:869- 77.

15 Chan SC, Reifsnyder D, Beavo JA. Hanifin JM. Immuno-chemical characterization of the distinct monoeyte cyclicAMP-phosphodiesterase from patients with atopie derma-titis. J Allergy Clin Immunol 1993; 91:1179-88.

16 Stanciu L, Shute JK, Janezic A. Holgate ST, Djukanovic R.Enrichment of CD4"^ and CD8"^ lymphocyte subsets fromperipheral blood using a four-step method includingmagnetic cell sorting. J Immunol Meth. Submitted forpublication.

17 Pawlowski NA, Kaplan G. Hamil AL, Conn SA, Scott WA.Arachidonic acid metabolism by human monocytes: studieswith platelet depleted cultures. J Exp Med 1983; 158:393412.

18 Simpson AW, Reeves ML. Rink TJ. Effects of SK&F94120. an inhibitor of cyclic nucleotide phosphodiesterasetype III. on human platelets. Biochem Pharmacol 1988;37:2315-20.

19 Wysocki LJ, Sato VL. 'Panning' for lymphocytes: Amethod for cell selection. Proc Natl Acad Sci, USA 1978;75:2844-8.

20 Thompson WJ. Appleman MM. Assay of cyclic nucleotidephosphodiesterase and resolution of multiple molecularforms ofthe enzyme. Adv Cycl Nucl Res 1979; I0;69-92.

21 Bauer AC. Schwabe U. An improved assay of cyclic 3',5'-nucleotide phosphodiesterases with QAE-Sephadex col-umns. Naunyn-Schmiedeberg's Arch Pharmacol 1980;311:193 8.

22 Spears G. Sneyd JGT, Loten EG. A method for derivingkinetic constants for two enzymes acting on the samesubstrate. Biochem J 1971; 125:1149-51.

23 Murray KJ, England PJ. Inhibitors of cyclic nucleotidephosphodiesterases as therapeutic agents. Biochem SocTrans 1992:20:460-4.

24 Giembycz MA. Could isoenzyme-selective phosphodiester-

ase inhibitors render bronchodilator therapy redundant inthe treatment of bronchial asthma? Biochem Pharmacol1992; 43:2041-51.

25 Schudt C. Winder S. Forderkunz S, Hatzelmann A. UllrichV. Influence of selective phosphodiesterase inhibitors onhuman neutrophil functions and level of cAMP and Cai.Naunyn Schmiedeberg's Arch Pharmacol 1991; 344:682-90.

26 Suttorp N, Weber U, Welsch T, Schudt C. Role ofphosphodiesterases in the regulation of endothelial perme-ability in vitro. J Clin Invest 1993; 91:1421 8.

27 Dent G, Giembycz MA, Rabe KF. Barnes PJ. Inhibition ofeosinophil cyclic nucleotide PDE activity and opsonizedzymosan-stimulated respiratory burst by "type lV-selective"PDE inhibitors. Br J Pharmacol 1991; 103:1339 46.

28 Tenor H, Hatzelmann A. Kupferschmidt R et al. Cyclicnucleotide phosphodiesterases from human alveolar macro-phages. Clin Exp Allergy 1995: 25: 525 63.

29 Rabe KF. Tenor H, Dent G et al. Phosphodiesteraseisozymes modulating inherent tone in human airways:identification and characterization. Am J Physiol 1993;264:L458-L464.

30 Rabe KF, Tenor H, Dent G et al. Identification of phos-phodiesterase-isoenzymes in human pulmonary artery andthe effects of selective PDE inhibitors. Am J Physiol1994:266:L536-L543.

31 ichimura M, Kase H. A new cyclic nucleotide phospho-diesterase isozyme expressed in the T-lymphocyte cell lines.Biochem Biophys Res Commun 1993: 193:985-90.

32 Blooom TJ. Beavo JA. Identification of type VII PDE inHUT T-lymphocytes cells. FASEB J 1994; 8:A372.

33 Kincaid RL, Stith-Coleman IE, Vaughan M. Proteolyticactivation of calmodulin-dependent cyclic nucleotide phos-phodiesterase. J Biol Chem 1995: 260:9009-15.

34 Burns F, Rodge !W. Pyne NJ. The catalytic subunit ofprotein kinase A triggers activation of the type V cyclicGMP-speciiic phosphodiesterase from the guinea pig lung.Biochem J 1992: 283:487-91.

35 Schudt C, Tenor H. Wendel A et al. Effect of selectivephosphodiesterase (PDE) inhibitors on activation of humanmacrophages and lymphocytes. Eur Resp J 1993: 6:367s.

36 Giembycz MA. Corrigan CJ, Kay AB, Barnes PJ. inhibi-tion of CD4 and CD8 T-lymphocyte (T-LC) proliferationand cytokine secretion by isoenzyme selective phosphodies-terase (PDE) inhibitors: correlation with intracellular cyclicAMP (cAMP) concentrations. J Allergy Clin Immunol1994:93:167.

37 Macphee CH. Reifsnyder DH. Moore TA. Lerea K^.Beavo JA. Phosphorylation results in activation of acAMP phosphodiesterase in human platelets. J Bioi Chem1988; 263:10353-8.

1995 Blackwell Science Ltd. Clinical and Experimental Allergy, 25, 616-624