doi:10.1006/cyto.1999.0643 available online at http://www.idealibrary.com on

SHORT COMMUNICATION

MECHANISMS OF CD26/DIPEPTIDYL PEPTIDASEIV CYTOKINE-DEPENDENT REGULATION ON

HUMAN ACTIVATED LYMPHOCYTES

Francisco J. Salgado,1 Elena Vela,2 Margarita Martın,3 Rafael Franco,3

Montserrat Nogueira,1 Oscar J. Cordero1

Among the cellular pathways activated by IL-12, we had previously found that both thepercentage and intensity of CD26+ cells in the PHA-stimulated T cells increased when IL-12was present (independently of its CD4 or CD8 phenotype). Now, we examined the molecularmechanisms of this IL-12-mediated effect. The IL-12 regulation pathway is dependent of de novoprotein synthesis and independent of cytokine secretion. Our results show two transcripts forCD26 in PBMC for the first time and no regulation by ILs at this level. Furthermore, secretionof the serum forms of CD26/DPPIV were not affected by IL-12. Interestingly, assays withneutralizing mAbs against TNF-� suggest that this cytokine negatively modulates CD26expression. The fact that translation and probably translocation of CD26 toward the cell surfacecan be regulated by IL-12 and TNF-� reveals new aspects about the control of this TH1 marker.

� 2000 Academic Press

Dipeptidyl peptidase IV (DPPIV; CD26; EC3.4.14.5) is a 110-kDa cell surface glycoproteinexpressed on different cell types, including T cells. It isan enzyme that cleaves NH2-terminal dipeptides frompolypeptides with either L-proline or L-alanine at thepenultimate position.1 Certain anti-CD26 mAbs areable to transmit an activating signal to the T cell andregulate immune responses.2 Moreover, CD26 is afunctional receptor for collagen and adenosinedeaminase (ADA).3,4

Despite the multifunctionality of CD26, little isunderstood about the involvement of these functions inthe immune response. We have found a strong upregu-lation of CD26 expression and DPPIV function on

1136

human activated lymphocytes by IL-12 and IL-2, butnot by IL-1�, IFN-�, TNF-� or IL-4,5,6 suggesting thatCD26 is a marker of TH1 responses.

In view of the possibilities of IL-12 therapeutic usein cancer, immunodeficiencies, etc.,7 elucidating themolecular mechanisms of this upregulation of CD26on human activated lymphocytes will help to clarifythe immunological role of CD26.

RESULTS

From the 1Department of Biochemistry and Molecular Biology,University of Santiago de Compostela, Santiago de Compostela;2Department of Biochemistry and Molecular Biology, Universityof Barcelona, Barcelona; 3Laboratory of Molecular Biology,Service of Pathological Anatomy, Universitary Hospital‘‘Germans Trias i Pujol’’ of Badalona, Barcelona, Spain

Correspondence to: Dr Oscar J. Cordero, Departamento deBioquımica e Bioloxıa Molecular, Universidade de Santiago deCompostela, Facultade de Bioloxıa, Campus Sur. 15706 Santiagode Compostela, Galicia, Spain; E-mail: bnojcord@usc.es

Received 29 June 1999; received in revised form 29 July 1999;accepted for publication 4 November 1999

� 2000 Academic Press1043–4666/00/071136+06 $35.00/0

KEY WORDS: CD26/dipeptidyl peptidase IV/hydrocortisone/interleukin 12/tumour necrosis factor-�

IL-12 enhances the production of the DPPIVprotein but not CD26 mRNA expression onactivated lymphocytes

We have previously described an IL-12-dependentCD26 upregulation in PHA blasts.5 To test if thisregulation implicates CD26 gene transcription and/ortranslation, Western and Northern blot analyses wereperformed. The expression of CD26 in 5-day blasts,cultured with or without IL-12, was investigated byimmunoblotting using mAb 1F7 (anti-hCD26). Figure1A shows three main bands, the molecular charac-teristics of which are being studied. IL-12-activatedcells expressed enhanced levels of those bands. Insimilar conditions of activation, CD26 mRNA wasquantified by Northern blotting. Interestingly, the two

CYTOKINE, Vol. 12, No. 7 (July), 2000: pp 1136–1141

Mechanisms of CD26/DPPIV regulation on lymphocytes / 1137

3.9 kb and 2.9 kb transcripts described in PHA blasts8

have not been observed before in PBMC (Fig. 1B). Themeaning of this is at present unknown, as only one hadbeen detected by PCR.9 Intensity of these CD26 genetranscripts increased by the action of PHA; however,addition of IL-12, IL-2, IL-4 or TNF-� did not modifythe level of the transcripts in 5 days (Fig. 1C). Analysisof mRNA level at earlier times were also negative (notshown).

In order to know whether expression of cellsurface CD26 is in part due to recruitment fromintracellular pools, experiments with inhibitors wereperformed. To avoid cytotoxicity, 3 day PHA blastswere incubated for 2 additional days with IL-12 in thepresence or absence of cycloheximide (CHX) or actin-omicin D (Act D). As expected, the expression ofCD26 increased in lymphoblasts and increased most ofall with IL-12 (Fig. 2A). Although CHX (10 �g/ml)and Act D (0.5 �g/ml) reduced viability, a number ofviable cells was always obtained sufficient to be stat-istically measured by flow cytometry. Neither CHXnor Act D modified expression of PHA-dependentCD26 (Fig. 2B). As shown, CHX affected, but notcompletely, the IL-12-dependent CD26 expression

(Fig. 2C). Thus, translation of CD26 gene appears tobe partly involved in this IL-12 effect.

CD26

97.4

PHA

PHA

66.2

45.0

IL-12

CD26

A B

C

PBMC

3.9 kb

2.9 kb

3.9 kb

2.9 kb

actina-β

IL-12 IL-2 IL-4 TNF-α

Figure 1. CD26 protein and mRNA expression of 5-day PHA blasts cultured in the absence or presence of different cytokines.

In A, cell extracts were analysed by SDS-PAGE and immunoblotted with 1F7 mAb. An autoradiography of material corresponding to 5�106

PHA blasts cultured with or without IL-12 (2 ng/ml) is shown. The numbers on the left indicate the position of molecular mass (in kDa) proteinmarkers. In B, Northern blot analysis of total RNA (10 �g/lane) from PBMC and long exposure autoradiography is shown (one out of twoperformed). In C, total RNA (10 �g/lane) from 5-day PHA blasts cultured with or without IL-12 (2 ng/ml), IL-2 (30 U/ml), IL-4 (2 ng/ml) andTNF-� (4 ng/ml), was analysed for the expression of CD26 transcripts by hybridizing with a 32P-labelled random-primed CD26 probe. This figureis representative of three experiments with similar results.

The effect of IL-12 on CD26 is not mediated byendogenous cytokines

The presence of hydrocortisone (HC), an inhibitorof cytokine production, reduced PHA as well as IL-12-dependent DPPIV activity. For IL-12, an importantand variable (n=5; range: 22–73%) HC-independentremaining activity was observed (Fig. 3A). Neverthe-less, the CD26/DPPIV upregulation persisted whenIL-12 was added to cells previously activated withPHA plus HC. Therefore, IL-12R expression isunaffected by HC. There are at least three possibilities:(a) endogenous cytokines are needed for a completeCD26 upregulation; (b) CD26 is partially shed into themedium, (c) HC inhibits directly CD26 production. Asa soluble form of DPPIV was described in humanplasma10 we measured the presence of DPPIV activityin culture medium. With regard to PHA, neither IL-12nor HC increased the serum DPPIV levels (not shown).This means that CD26 shedding by HC can be rejectedand that IL-12 do not regulate soluble DPPIVsecretion.

1138 / Salgado et al. CYTOKINE, Vol. 12, No. 7 (July, 2000: 1136–1141)

Analyses of the participation of endogenouscytokines in DPPIV activity were carried out.Unexpectedly, a little enhancement of the activity wasseen when neutralizing anti-IL-12 mAb was added tothe PHA-cultures. In IL-12 cultures, the presenceof anti-IL-12 neutralized a high percentage of theexogenous IL-12 effect, but not completely (Fig. 3B).Then, we evaluated the role of endogenous IL-2 uponthe effect of IL-12 on CD26. Neutralizing anti-IL-2and anti-IL-2R (CD25) Abs and cyclosporin A (CyA),at doses that inhibit IL-2-dependent T-cell prolifer-ation,11 were used. There were no significant reduc-tions in the IL-12-induced expression of DPPIV in thepresence of this antibodies, inhibitors or its combina-tion (Fig. 3C), which indicates that the IL-12 effect isnot mediated by IL-2. Note that CyA, alone or withanti-IL-2, led to a minor but consistent increase inDPPIV activity (Fig. 3C), which led us to hypothesizeon an endogenous factor that might be negativelyaffecting CD26 expression.

TNF-� downregulates CD26 expressionIn our experimental conditions, IL-12 is a

powerful inducer of IFN- �, GM-CSF and TNF�.7 Asexogenous IFN-� did not modify CD26 expression,5

this cytokine was a poor candidate for CD26 down-regulation. Nevertheless, the addition of neutralizingIgG polyclonal anti-TNF-�, but not isotype-matchedcontrol antibodies, enhanced the DPPIV activity ofcells in a dose–response fashion. This happened withIL-12 or IL-2, two TNF-� production inducers. Flowcytometry experiments confirmed that the enhancedDPPIV activity observed with anti-TNF-� corre-sponded to an increase in the CD26 expression (Fig. 4).However, when a relatively high concentration ofexogenous rTNF-� was added, only a partial CD26/

DPPIV downregulation was observed using bothexperimental techniques, which could be due to partialsaturation with endogenous TNF-� (not shown).

Figure 2. CHX and Act D partially inhibit IL-12-induced CD26 upregulation.

Three-day PHA blasts were incubated for 2 days with or without IL-12 (A) and in the presence or absence of CHX (10 �g/ml) and Act D(0.5 �g/ml), respectively (B and C). The samples were analysed on a Becton Dickinson Excalibur cytometer. Histograms show CD26 expressionrevealed with the Ta1-FITC mAb by flow cytometry performed as described in Methods. In B and C, because of cytotoxicity, events wereadjusted in order to make comparisons. One out of three representative experiments is shown.

DISCUSSION

Despite CD26 expression increasing upon lym-phocytes activation, the IL-12 effect suggests subtleroles for this peptidase in T cell activation. We havepresented results indicating that IL-12-induced CD26/DPPIV upregulation requires, at least in part, de novosynthesis of protein, affecting the presence of HC. Thisglucocorticoid could be inhibiting the CD26 expression(in both IL-12 and PHA-activated blasts), as HC didnot enhance serum DPPIV. Analogously, the cytokine-induced expression of ICAM-1 is partially inhibited bydexamethasone-treated eosinophils.12 Taken together,these results highlight a potential role for gluco-corticoid in modulating CD26/DPPIV expression andfunction.

Our data also suggest that CD26 upregulation byIL-12 is a direct effect and cannot be ascribed toproliferative actions of IL-2. Moreover, it is unlikelythat IL-2 was an intermediate in the effect of IL-12,because IL-12 itself does not affect IL-2 production inthese conditions.13 In our assays, in which APCs arepresent, IL-12 has a synergistic effect with B7/CD28and CD40/CD40L interactions inducing proliferationand cytokine production, especially IFN-�, TNF-� andGM-CSF. Neutralization of this TNF-� in IL-12cultures leads to a rise in DPPIV activity. As purifiedDPPIV is known to degrade TNF-�,14 it was plausiblethat the DPPIV catalytic centre could be partiallyblocked by TNF�. Flow cytometry confirmed thatendogenous TNF-� is actually downregulating CD26

Mechanisms of CD26/DPPIV regulation on lymphocytes / 1139

expression. Furthermore, the CyA effect supports thisconclusion, as this inhibitor prevents TNF-� release.15

However, by adding exogenous TNF-�, only a partialCD26 downregulation was observed, which probablymeans that endogenous TNF-� is already affectingCD26 expression.

Both the CD26 mRNA and intracellular CD26pool were detected in freshly isolated CD26+ orCD26� T lymphocytes,9 so it is probable that theintracellular CD26 is maintained by permanent trans-lation. We have observed that cytokines exercise aregulation on CD26 translation and maturation, butnot on CD26 gene transcription. However, as CHXdid not completely abrogate the CD26 upregulation,we hypothesize that cytokines and HC could alsobe regulating the protein transport or translocationtoward the cell surface. This kind of regulation may bea generalized mechanism that should be studied.

MATERIALS AND METHODS

MaterialsRecombinant human TNF-� was purchased from

PeproTech (London, UK), rhIL-2 from Sigma (Madrid,Spain), and rhIL-12 and anti-hIL-12 goat IgG from R&DSystems (Abingdon, UK). Anti-hCD26 mAbs TA5.9/TA5.9-FITC (clone LY12 CC1; IgG1) and Ta1-FITC werefrom Alexis Corp. (Laufelfingen, Switzerland)12 and Coulter(Miami, FL), respectively. The 1F7 and the cyclosporin A(CyA) were a generous gift from Prof. S. F. Schlossman(Harvard Medical School, Boston, MA) and Sandoz(Vienna, Austria), respectively. Hydrocortisone (HC),F(ab�)2 goat anti-mouse (GAM) or anti-rabbit IgG (GAR)and murine anti-goat, labelled with FITC or PE, ascitic fluidcontaining IgG2a and IgG1 isotype control mAbs (UPC10and MOPC21) and neutralizing IgG rabbit anti-hTNF-�were from Sigma (Madrid, Spain). Rabbit anti hIL-2 wasfrom Genzyme (Cambridge, MA). Anti-CD25 was fromBecton Dickinson.

Figure 3. Role of endogenous cytokine production.

In A, PHA blasts (106 cells/ml) were incubated for 2 days with orwithout IL-12 in the presence or absence of HC (0.1 mM) (open). Inthe solid bars, blasts were cultured in the presence of HC, and afterextensive washing the same conditions than in clear were used.DPPIV enzymatic activity assay was measured as described inMethods by incubation of harvested cells with Gly-Pro-pNA sub-strate for 2 h. Formation of pNA in the supernatant was recorded at405 nm in microtitre plates. Results of an experiment representativeof five (see text), are expressed as absorbance data and eachexperimental point is the mean of two measures�SD. In B and C,the role of endogenous IL-12 and IL-2 on DPPIV enzymatic activityinduced by PHA and IL-12 was analysed. Blasts were incubated for2 days in culture medium with (solid bars) or without (open bars,controls) IL-12 (2 ng/ml). In parallel, both culture conditions weresupplemented with (B) neutralizing anti-IL-12 mAb (10 �g/ml)and (C) with anti-IL-2 polyclonal Ab (5 �g/ml), anti-CD25 mAb(1 �g/ml), CyA (1 �g/ml) or a combination as indicated. At the end ofthe culture, cells were collected and DPPIV activity was assayed asabove. Similar results were obtained in at least four experiments.

CulturesHuman PBMC were isolated by Ficoll-Paque

(Pharmacia, Uppsala, Sweden) from buffy coats (kindlyprovided by the Centro de Transfusiones de Galicia,

1140 / Salgado et al. CYTOKINE, Vol. 12, No. 7 (July, 2000: 1136–1141)

Santiago, Spain) and cultured as described.5 PBMC wereactivated with PHA (1 �g/ml; Sigma) for 3 days and then afurther 2 days in the presence or absence of cytokines,inhibitors or antibodies. In the absence of cytokines,20% conditioned medium (CM), which comes from thelymphoblast culture medium, was used. Occasionally,cytokines were added from the first 3 days.

Peptidase assayLymphocytes were incubated for different times at 37�C

in the presence of 0.5 mg/ml Gly-Pro pNA (Sigma) aspreviously described.5 A cell-free control mixture was testedin parallel. The cells were pelleted and the formation of pNAin the supernatant recorded at 405 nm. Results wereexpressed as absorbance data or pkat/106 cells at 37�C.

Flow cytometryImmunofluorescence staining of cultured PBMC was

performed as described5 by using the following anti-CD26mAbs: TA5.9-FITC or Ta1-FITC. The samples wereanalysed on a Coulter EPICS Profile and Becton DickinsonExcalibur cytometers. Isotype matched mAb (MOPC21 andUPC10) were used as control.

ImmunoblotsCells were collected by centrifugation and lysed in

10 mM Tris pH 7.6, 1.5 mM MgCl2, 140 mM NaCl, 0.5%NP-40, 1 mM PMSF (30 min on ice). The same number ofcells were used in each condition. The nuclei were removedby centrifugation (13 000�g for 15 min, 4�C). Crude cellextracts were diluted (1:1) in sample buffer and analysed onSDS/PAGE (7.5%). The anti-hCD26 mAb 1F7 was revealedwith GAM Ig-horseradish peroxidase (ECL�, Amersham,Rainham, UK).

Human CD26 cDNA construct and Northern-blotRT-PCR was performed from total RNA belonging to

human PBMC with the Ultraspec RNA solution (BiotecxLaboratories, Inc., Houston, TX). The first strandreaction was performed with the First Strand Synthesis Kit(Pharmacia) followed by PCR. The oligonucleotide primerswere F-CD26 in exon 18: 5� GCCCTCCAAAAAACTGGACT 3� and R-CD26 in exon 24: 5� TTCAGCTCTGCTCATGACTG 3�. PCR conditions were 95�C, 5 min+94�C, 40 s,55�C, 40 s and 72�C, 35 s for 30 cycles, using Taq polymeraseEurobiotaq (Eurobio, Les Ulis, France). Specific PCRproducts were separated (1.5% agarose gel), visualized byEtBr staining, purified (Geneclean Kit Bio101, La Jolla, CA)and cloned by blunt-end ligation into pUC18 (SureCloneLigation Kit, Pharmacia). Plasmid minipreps were preparedwith Wizard Plus SV Minipreps DNA Purification System(Promega, Madison, USA) and the insert size of each clonedetermined by an EcoRI-BamHI digestion (New EnglandBioLabs, Beverly, MA.) before DNA sequencing with the T7Sequencing Kit (Pharmacia).

Total RNA (isolated by Stratagene Micro RNAisolation kit, La Jolla, CA) was electrophoresed (1% agarose-formaldehyde gel) and transferred to a nylon membrane(Immobilon-s, Millipore, Bedford, UK). The CD26 tran-scripts were detected by hybridizing with �32P-dCTP-labeledrandom-primed probe (NEBlot kit, New England BioLabs;2�106 cpm/ml) in the presence of 6�SSC, 5�Denhardt’sreagent, 0.5% SDS and 100 g/ml denatured salmon spermDNA (all from Sigma) at 68�C (o/n). Blots were washed with0.1�SSC/0.1% SDS (68�C, 15 min) and films exposed for1–3 days. CD26 mRNA size was estimated from rRNA and�-actin mRNA expression used as controls.

Figure 4. TNF-� downregulates CD26 expression.

PBMC were cultured for 5 days with PHA plus IL-12 (2 ng/ml) andIL-2 (100 U/ml) in the presence of (A) isotype-matched IgG, or (B)neutralizing anti-TNF-�. Thereafter, the cells were washed andincubated with anti-CD26 T.A.5.9-FITC mAb for direct flow cytom-etry. An experiment measured with the Coulter EPICS Profilecytometer, representative of several with similar results, is shown.

Acknowledgements

We thank the Centro de Transfusiones de Galiciafor providing us with buffy coats from healthy patients.We also thank Prof. S. F. Schlossman (Dana-FarberCancer Institute, Harvard Medical School, Boston,

Mechanisms of CD26/DPPIV regulation on lymphocytes / 1141

MA) for his kind gift of 1F7 Ab, and J. Trotter(Scripps Institute, LaJolla, CA) for the WinMDIsoftware. This work was supported by grantXUGA20007B96 from the Xunta de Galicia.

REFERENCES

1. Fleischer B (1994) CD26: a surface protease involved inT-cell activation. Immunol Today 15:180–184.

2. Munoz E, Blazquez MV, Madueno JA, Rubio G, Pena J(1992) CD26 induces T-cell proliferation by tyrosine proteinphosphorylation. Immunology 77:43–50.

3. Dang NH, Torimoto Y, Schlossman SF, Morimoto C(1990) Human CD4 helper T cell activation: functional involvementof two distinct collagen receptors, 1F7 and VLA integrin family.J Exp Med 172:649–652.

4. Kameoka J, Tanaka T, Nojima Y, Schlossman SF,Morimoto C (1993) Direct association of adenosine deaminase witha T cell activation antigen CD26. Science 261:466–469.

5. Cordero OJ, Salgado FJ, Vinuela JE, Nogueira M (1997)Interleukin-12 enhances CD26 expression and dipeptidyl peptidaseIV function on human activated lymphocytes. Immunobiology197:522–533.

6. Cordero OJ, Salgado FJ, Vinuela JE, Nogueira M (1998)Interleukin-12-dependent activation of human lymphocyte subsets.Immunol Lett 61:7–13.

7. Chizzonite R, Gubler U, Magram J, Stern AS (1998)Interleukin-12. In Mire-Sluis A, Thorpe R, Page C (eds) Cytokines.Academic Press Limited, London, pp 183–203.

8. Tanaka T, Camerini D, Seed B, Torimoto Y, Dang NH,Kameoka J, Dahlberg HN, Schlossman SF, Morimoto C (1992)Cloning and functional expression of the cell activation antigenCD26. J Immunol 149:481–486.

9. Mattern T, Reich C, Duchrow M, Ansorge S, Ulmer AJ,Flad HD (1995) Antibody-induced modulation of CD26 surfaceexpression. Immunology 84:595–600.

10. Duke-Cohan JS, Morimoto C, Rocker JA, Schlossman SF(1995) A novel form of dipeptidylpeptidase IV found in humanserum. Isolation, characterization, and comparison with T lym-phocyte membrane dipeptidylpeptidase IV (CD26). J Biol Chem270:14107–14114.

11. Cordero OJ, Sarandeses CS, Lopez JL, Cancio E, RegueiroBJ, Nogueira M (1991) Prothymosin-� enhances interleukin 2receptor expression in normal human T-lymphocytes. Int JImmunopharmac 13:1059–1065.

12. Guida L, O’Hehir RE, Hawrylowicz CM (1994) Synergybetween dexamethasone and interleukin-5 for the induction of majorhistocompatibility complex class II expression by human peripheralblood eosinophils. Blood 84:2733–2740.

13. Peng X, Kasran A, Ceuppens JL (1997) Interleukin 12and B7/CD28 interaction synergistically upregulate interleukin 10production by human T cells. Cytokine 9:499–506.

14. Bauvois B, Sanceau J, Wietzerbin J (1992) Human U937 cellsurface peptidase activities: characterization and degradative effecton TNF-�. Eur J Immunol 22:923–930.

15. Vassalli P (1992) The pathophysiology of tumor necrosisfactors. Annu Rev Immunol 10:411–452.