Extracellular Adenine Nucleotides Inhibit the Activation of Human CD4⁺ T Lymphocytes¹

Extracellular nucleotides constitute an ubiquitous family of messengers that exert autocrine or paracrine actions. Their release in fluids results from cell lysis, exocytosis of nucleotide-concentrating granules (synaptic vesicles, platelet-dense bodies), or efflux from cytoplasm through membrane transport proteins (1, 2, 3). They induce a wide spectrum of biological effects that are mediated by P2 purinergic receptors. P2Rs can be divided in two families: the ionotropic P2X, which are ligand-gated cation channels; and the metabotropic P2Y, which are part of the superfamily of G-protein-coupled receptors. At this stage, eight genuine human P2YRs have been identified and characterized: P2Y₁, P2Y₂, P2Y₄, P2Y₆, P2Y₁₁ (4), P2Y₁₂ (5, 6), and P2Y₁₃ (7) as well as the uridine 5′-diphosphoglucose receptor which is structurally related to P2Y₁₂ (8). In particular, extracellular nucleotides exert numerous actions, mediated by several P2Y and P2XRs, on monocytes and macrophages, dendritic cells, lymphocytes, and granulocytes (reviewed in Ref. 3).

Among the P2YRs, the P2Y₁₁ subtype has the unique property of being dually coupled to phospholipase C and adenylyl cyclase activation. Northern blotting analysis revealed a restricted expression of this receptor in spleen and HL-60 cells, supporting a role of ATP in the regulation of immune responses and hemopoiesis (9). Recently, Wilkin et al. (10) showed that ATP and hydrolysis-resistant derivatives of ATP, such as adenosine 5′-O-(3-thiotriphosphate) (ATPγS),³ which are agonists of the P2Y₁₁R, induce the maturation of human monocyte-derived dendritic cells via an increase in cAMP. A coherent picture of lymphoid cells control by extracellular nucleotides has not yet emerged, and it is generally believed that human peripheral T lymphocytes express P2XRs but lack P2YRs (3). In this study, we have investigated the effects of P2Y₁₁R agonists on the activation of human blood-derived CD4⁺ T lymphocytes.

ATP, ATPγS, 2′- and 3′-O-(4-benzoylbenzoyl)-ATP (BzATP), dATP, UDP, UTP, PGE_2, 8-bromo-cAMP, and indomethacin (Indo) were obtained from Sigma (St. Louis, MO). 2-Methylthio-ATP, CGS-21680 (a potent agonist of A₂Rs), and 8-(p-sulfophenyl)theophylline (8-p-SPT) were purchased from Research Biochemicals International (Natick, MA). 2-Propylthio-β,γ-dichloromethylene-d-ATP (AR-C67085) was a gift of Drs. J. D. Turner and P. Leff (Astra Zeneca R&D, Loughborough, U.K.). H-89 was obtained from Biomol Research Laboratories (Plymouth Meeting, PA). The anti-CD3 mAb OKT3 (Orthoclone OKT3) was provided by Janssen-Cilag (Berchem, Belgium), and the anti-CD28 mAb (clone CD28.2) was supplied by BD PharMingen (San Diego, CA). [³H]Thymidine (25 Ci/mmol) was from Moravek Biochemicals (Brea, CA). Rolipram was a gift from the Laboratoires Jacques Logeais (Trappes, France).

PBMC were isolated from buffy coats of healthy blood donors by density gradient centrifugation on Lymphoprep (Nycomed, Oslo, Norway). After three washes in HBSS (Invitrogen, Merelbeke, Belgium). CD4⁺ T cells were isolated from PBMC using the MACS negative depletion system (Miltenyi Biotec, Auburn, CA). No contaminating CD8⁺ T cells, B cells, monocytes, or NK cells were detected.

Purified CD4⁺ T cells were cultured in RPMI 1640 (Invitrogen) supplemented with 10% heat-inactivated FCS from HyClone (Logan, UT), 25 mM HEPES buffer, 2 mM l-glutamine, 1 mM sodium pyruvate, 50 μg/ml gentamicin, and 50 μM 2-ME at 37°C in 5% CO₂. The CD4⁺ T cells (2 × 10⁵/well) were activated in flat-bottom 96-well plates precoated with the anti-CD3 mAb (10 μg/ml) in the presence of soluble anti-CD28 mAb (1 μg/ml) and the presence or absence of different concentrations of nucleotides. Culture supernatants were harvested after 24, 72, or 96 h for measurement of cytokine concentration and the remaining cells were resuspended in PBS to determine CD25 surface expression as well as apoptosis and necrosis by flow cytometry. After 56 h of culture, proliferation was assessed by [³H]thymidine (0.5 μCi/well) uptake during the next 16 h. Each experimental condition was tested in triplicate.

Flow cytometric analysis of surface phenotype of the CD4⁺ T cells was performed by two- or three-color staining using FITC-, PE-, and PerCP-conjugated mouse anti-human mAb. Purified CD4⁺ T cells were stained with mAbs against CD3, CD4, CD8, CD14, CD16, CD19, CD25, CD45, CD56, and CD69 (all from BD PharMingen). The percentage of apoptotic and necrotic CD4⁺ T cells was determined using FITC-conjugated annexin V and propidium iodide, both from BD PharMingen (San Diego, CA). Samples were assayed in duplicate and analyzed using a FACScan flow cytometer (BD Biosciences, Franklin Lakes, NJ) and CellQuest software (BD Biosciences).

Commercially available kits were used for quantification of various cytokine levels: IL-2 (R&D Systems, Oxon, U.K.); IL-10 and IFN-γ from Biosource International (Camarillo, CA). IL-5 levels were measured by two-site sandwich ELISA using Abs from BD PharMingen (San Diego, CA). Each experimental condition was tested in triplicate.

Cells were preincubated for 30 min in complete culture medium with 25 μM rolipram and then incubated in the same medium for 12 min in the presence of the agonists. The incubation was stopped by the addition of 1 ml 0.1 M HCl. The incubation medium was dried up, and the samples were resuspended in water and diluted as required. cAMP was quantified by radioimmunoassay after acetylation as previously described (11). Each experimental condition was tested in triplicate.

Levels of intracellular calcium were measured with the calcium-binding dye fluo-3 (12). Directly after the isolation or after 24 h incubation at 37°C in a 5% CO₂, humidified atmosphere, purified human CD4⁺ T cells were washed twice in calcium- and magnesium-free HBSS and incubated at a concentration of 5 × 10⁶ cells/ml with 0.25 mM sulfinpyrazone (Sigma), 100 μg/ml pluronic acid F-127 (Molecular Probes, Leiden, The Netherlands), and 5 μM fluo-3 (Molecular Probes), for 30 min at 37°C. Cells were then washed twice in complete medium or HBSS with 0.25 mM sulfinpyrazone, resuspended at a final concentration of 5 × 10⁵/ml, and placed in a 37°C water bath. After 10 s acquisition on a FACScan flow cytometer, the cells were stimulated with either 100 nM RANTES or 100 μM nucleotide (ATPγS or BzATP). The FL1 signal for fluo-3-bound calcium was calibrated by transporting in saturating calcium with 100 ng/ml A23187 to obtain the maximum signal (F_max) and then adding 2 mM MgCl₂ to obtain the minimum signal (F_min). The intracellular calcium concentration ([Ca²⁺]_i) was calculated using the formula [Ca²⁺]_i = K_d (F − F_min)/(F_max − F), where K_d = 400 nM for the fluo-3 intracellular dye (13).

Total RNA was isolated using the Tripur Isolation Reagent (Roche Diagnostics, Basel, Switzerland). To avoid DNA contamination, 1 μg total RNA was treated with DNA-free from Ambion (Austin, TX). Reverse transcription was performed with 500 ng RNA using the Superscript II preamplification system with random hexamers (Life Technologies). The primers and TaqMan fluorogenic probe for P2Y₁₁ (GenBank accession number AF030335) were from Eurogentec (Seraing, Belgium): forward (19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39), 5′-CTG CCC TGC CAA CTT CTT G-3′; reverse P2Y₁₁ (76–96), 5′-CAG TAT GGG CCA CAG GAA GTC-3′; probe (45–69), 5′-(FAM)- TGC CGA CGA CAA ACT CAG TGG GTT-(TAMRA)-3′. TaqMan PCR assays were performed in duplicate on cDNA samples or RNA in 96-well optical plates on an ABI Prism 7700 sequence detection system (PE Applied Biosystems, Foster City, CA). For each 25-μl reaction, 1/20 of reverse transcription was mixed with 12.5 μl 2× qPCR Mastermix kit (Eurogentec), primers (300 nM each), and fluorogenic probe (200 nM). PCR parameters were 50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 s and 60°C for 1 min. Data were analyzed with Sequence Detector Software (SDS version 1.6; PE Applied Biosystems).

Purified CD4⁺ T cells were preincubated with 3 μM H-89 for 4 h and then stimulated for 30 min in flat-bottom plates precoated with 10 μg/ml anti-CD3 mAb in the presence of 1 μg/ml soluble anti-CD28 mAb, and in the presence or absence of 100 μM ATPγS. Cells were washed with HBSS and lysed on ice in 120 μl of Laemmli buffer (10% (w/v) glycerol, 5% (v/v) 2-ME, 2.3% (w/v) SDS, 62.5 mM Tris-HCl, pH 6.8) with proteinase inhibitors: 1 μg/ml leupeptin, 60 μg/ml Pefabloc, 1 μg/ml aprotinin, and phosphatase inhibitors (Roche); 1 mM sodium orthovanadate and 10 mM sodium fluoride (Sigma). The protein concentration was determined using the method of Minamide and Bamburg (14). The same amount of protein (10 μg) for each condition was electrophoresed on a 12% SDS-polyacrylamide gel. Proteins were then transferred overnight at 28 V and 4°C onto a nitrocellulose membrane using 20 mM Tris, 154 mM glycine, 20% (v/v) methanol as a transfer buffer. Immunodetection was achieved using an Ab specific for phospho-CREB at a 1/1,000 dilution and the ECL Western blotting detection system (Amersham Pharmacia Biotech) with a biotinylated secondary rabbit Ab (1/50,000). The anti-phospho-CREB Ab was a gift from Dr. M. R. Montminy (The Clayton Foundation Laboratories for Peptide Biology, La Jolla, CA).

Triplicates were obtained for each measurements (i.e., for each agonist, each donor, and each concentration). ANOVA for repeated measures (concentration at three levels) with two intersubject factors (donor and agonist) was performed. Adjustment for multiple comparisons between the three concentrations was performed using the Sidak correction (SPSS 10.0; SPSS, Chicago, IL).

To evaluate the functional expression of P2Y₁₁Rs, we determined the intracellular cAMP accumulation in response to nucleotides. A comparison was made between freshly purified CD4⁺ T cells and CD4⁺ T cells activated during 72 h with the association of immobilized anti-CD3 mAb and soluble anti-CD28 mAb used as an APC-independent and polyclonal T cell stimulus. At 100 μM, ATPγS and BzATP, both agonists of P2Y₁₁R (15), stimulated significantly the cAMP production in both populations (Fig. 1,A). This stimulation was observed in all 16 preparations of CD4⁺ T cells that were tested (6 blood donors for Fig. 1, A and D, 3 for Fig. 1,B, 3 for Fig. 1,C, and 4 for the suramin experiments; data not shown). 2MeSATP was clearly less potent, and uridine nucleotides were inactive (Fig. 1,A). ATPγS (100 μM) produced a 20- ± 11-fold cAMP increase in freshly isolated cells and a 28- ± 15-fold cAMP increase 72 h after activation (mean ± SD of six independent experiments). Concentration-action curves showed that the rank order of agonist potency was ATPγS ≅ BzATP > ATP (Fig. 1,B). Although the profile observed thus far was consistent with the P2Y₁₁R (15), two other P2Y₁₁ agonists, dATP and AR-C67085, had no effect on cAMP accumulation (Fig. 1 C) in all the three preparations tested. At 100 μM, suramin reduced the cAMP accumulation in ATPγS-stimulated CD4⁺ T cells by 51% (mean of four experiments; data not shown).

In three experiments of six, a stimulation of cAMP accumulation could also be detected in response to adenosine (100 μM) or the A_2AR agonist CGS-21680 (100 nM). That response was observed only in activated cells and not in freshly isolated ones (Fig. 1,D). That cAMP response was sensitive to inhibition by 8-p-SPT (300 μM), an A₂R antagonist, whereas the effect of ATPγS was not (Fig. 1 D). In contrast, Indo, an inhibitor of PG release, had a minimal effect on the cAMP response to ATPγS of 25 ± 14% inhibition (mean of three experiments ± SD; range, 11–38%).

To test whether human CD4⁺ T cells exhibit a phospholipase C/inositol 1,4,5-triphosphate/Ca²⁺ response to adenine nucleotides, we performed calcium-binding dye fluo-3 assays. To verify the functionality of the calcium mobilization, we used the chemokine RANTES (100 nM), known to induce an increase of [Ca²⁺]_i (data not shown) (16). In the presence of extracellular calcium, BzATP (100 μM) increased [Ca²⁺]_i in freshly isolated CD4⁺ T cells, whereas ATPγS (100 μM) had no effect (Fig. 2, A and 2B). The effect of BzATP was abolished in a Ca²⁺-free medium (Fig. 2 B).

P2Y₁₁ mRNA was quantified by RT-PCR using TaqMan fluorigenic detection. In human peripheral CD4⁺ T cells, these experiments revealed a number of copies smaller than in dendritic cells and HL-60 cells and comparable with those in 1321N1 astrocytoma cells that do not express functional P2Y₁₁Rs (Table I).

IFN-γ, IL-2, IL-5, and IL-10 production by CD4⁺ T cells obtained from three donors and activated with immobilized anti-CD3 and soluble anti-CD28 mAb (10⁶ cells/ml) was measured in the absence or presence of nucleotide derivatives added at the beginning of the culture. As shown in Fig. 3 and Tables II and III, both ATPγS and BzATP significantly inhibited in a concentration-dependent manner the secretion of the four cytokines tested. The inhibition increased significantly with the concentration (p < 0.001 for each cytokine), but no significant difference could be detected between the agonists, except for IL-10. These effects were insensitive to 8-p-SPT (data not shown), indicating that they are not mediated by adenosine and by activation of A₂Rs. Moreover, the inhibition of IL-2 and IL-10 production by CD4⁺ T cells was not modified by the addition of Indo (5 μg/ml; data not shown), whereas, as previously described (17), the cAMP-elevating agent forskolin (10 μM) suppressed the secretion of the four cytokines (data not shown). No significant amounts of cytokines were detectable in culture supernatants of nonactivated CD4⁺ T cell (<10 pg/ml).

To study the involvement of the protein kinase A (PKA) in the cAMP-dependent responses in CD4⁺ T cells, we tested H-89, which is an inhibitor of PKA that acts by competitive inhibition of the ATP binding site (18, 19). At 3 μM, H-89 was able to prevent the phosphorylation of CREB induced by ATPγS (Fig. 4,B). H-89 per se had a small inhibitory effect on IFN-γ secretion during a 24-h stimulation (31 ± 14% inhibition; mean ± SD of three experiments) and did not reverse the inhibitory effect of ATPγS, BzATP, 8-bromo-cAMP, or PGE₂ (Fig. 4 A).

As shown in Fig. 5, ATPγS and BzATP, added at concentrations ranging from 1 to 100 μM, inhibited the proliferation of CD4⁺ T cells activated by immobilized anti-CD3 and soluble anti-CD28 mAb as assessed by [³H]thymidine incorporation. An average of 41 ± 13 and 46 ± 7% inhibition (mean ± SD of three experiments) of CD4⁺ T cell proliferation were observed with ATPγS and BzATP (both 100 μM), respectively. Moreover, the decrease in proliferation was not reversed by the simultaneous addition of 8-p-SPT (data not shown). Labeling of the CD4⁺ T cells with annexin-FITC and propidium iodide showed that nucleotides did not increase the number of apoptotic and necrotic cells (data not shown).

For further insight into the mechanism of action of the ATP derivatives on CD4⁺ T cell proliferation, CD25 expression induced on these cells during activation by immobilized anti-CD3 and soluble anti-CD28 mAb was evaluated by flow cytometry in the presence or absence of the nucleotides. As shown in Fig. 6, the addition of ATPγS (100 μM) or BzATP (100 μM) at the beginning of the culture inhibited CD25 expression on the CD4⁺ T cells after 72 h of activation by >40%.

In both activated and nonactivated CD4⁺ T cell, cAMP accumulation was observed in response to ATP and ATP derivatives, ATPγS and BzATP, known as agonists of the P2Y₁₁R. The rank order of potency was ATPγS ≈ BzATP > ATP > 2-methylthio-ATP ≫ UTP, UDP. This pharmacological profile is similar to that previously observed in HL-60 cells (20, 21), human dendritic cells (10), or CHO-K1 cells stably expressing the recombinant human P2Y₁₁R (15). However, several arguments do not support the hypothesis that these responses were mediated by the P2Y₁₁R. First, dATP and AR-C67085, two potent agonists of the human recombinant P2Y₁₁R previously found to be active in HL-60 (21) and dendritic cells (10), had no effect on cAMP accumulation in human CD4⁺ T cells. Furthermore, whereas the P2Y₁₁R is dually coupled to adenylyl cyclase and phospholipase C stimulation, neither ATPγS nor BzATP induced a significant mobilization of intracellular calcium in CD4⁺ T cells. As previously shown for total blood T lymphocytes, BzATP, but not ATPγS, induced an influx of extracellular calcium in CD4⁺ T cells, presumably via the activation of a P2XR (22). Finally, quantitative RT-PCR revealed that the number of P2Y₁₁ mRNA copies is lower in CD4⁺ T cells than in HL-60 and dendritic cells and comparable with that found in 1321N1 human astrocytoma cells, a cell line devoid of functional P2YRs. These results suggest that the cAMP response to ATPγS and BzATP is mediated by another receptor than the P2Y₁₁R. These cAMP responses were not due to the degradation of adenine nucleotides into adenosine and the activation of A₂Rs. Indeed, a stimulation of cAMP accumulation in response to adenosine or the A_2AR agonist CGS-21680 was detected in only a few experiments and exclusively in activated cells. That cAMP response was sensitive to inhibition by 8-p-SPT, an A₂R antagonist, whereas the effect of ATPγS was insensitive to 8-p-SPT but partially inhibited by suramin, a nonselective antagonist of P2Rs. The identity of the receptor involved in the effect of ATPγS and BzATP remains thus unknown. Thus far, the P2Y₁₁R is the only P2R directly coupled to adenylyl cyclase stimulation, but several orphan G protein-coupled receptors are structurally close to the P2YRs and constitute potential candidates, some of which are expressed in thymus (23, 24).

According to these data, we have studied the effects of adenine nucleotides on cytokine release by CD4⁺ T cells. Our results demonstrate that ATPγS and BzATP induce an inhibition of the CD4⁺ T cell activation. The release of four cytokines (IL-2, IL-5, IL-10, and IFN-γ) involved in both Th1 and Th2 responses was significantly inhibited in the presence of adenine nucleotides. Simultaneously with inhibition of IL-2 secretion, the level of the IL-2R (CD25) expression was down-regulated; these two effects contribute to the inhibition of proliferation, assessed with the classical [³H]thymidine uptake test. We checked that the inhibition of [³H]thymidine uptake in our experiments was not due to apoptosis or necrosis.

The immunosuppressive effects of ATPγS and BzATP on both Th1 and Th2 responses are reminiscent of the action of PGE₂ on human CD4⁺ T cells, which has been described by many authors (25, 26, 27, 28) and is mediated by cAMP. It is now well established that an increase in cAMP inhibits T cell activation, via several mechanisms more or less proximal to TCR signaling. These mechanisms include inter alia: activation of C-terminal Src kinase (29); inhibition of c-Jun kinase (30); and phosphorylation of NFAT, preventing its nuclear translocation (31). The up-regulation of several phosphodiesterases during T cell activation supports the concept that cAMP exerts a physiological role of T cell response moderator (32, 33, 34). The immunomodulatory effects of cAMP on T lymphocytes are mediated partially by PKA activation and partially via other mechanisms. It has indeed been shown that the inhibitory effect of cAMP-elevating agents on IL-5 release from activated T lymphocytes was not relieved by the PKA inhibitor H-89, although it was confirmed that this agent inhibited the activation of PKA and the resulting phosphorylation of CREB (18, 19, 35). Likewise we have shown that although H-89 inhibited the ATPγS-induced CREB phosphorylation, the inhibitory effect of ATPγS on IFN-γ secretion was not suppressed by this PKA antagonist.

Discrepant effects of adenine nucleotides on lymphocyte proliferation have been reported in the past. Gregory and Kern (36) showed that extracellular ATP stimulated the proliferation of murine thymocytes, whereas Fishman et al. (37) observed that ATP suppressed the proliferation of human peripheral lymphocytes, presumably as a result of its degradation into adenosine. Ikehara et al. (38) observed both inhibitory and stimulatory effects of ATP on murine lymphocyte proliferation, depending on the origin of the cells. ATP had a synergistic effect on the proliferation of human peripheral blood T cells stimulated by PHA or anti-CD3 Ab, an effect apparently mediated by a P2XR (22). Inhibitory effects of ATP on T cell proliferation might be mediated by its degradation into adenosine and activation of A_2ARs which are indeed expressed on some T cell populations (39, 40). It is well known that T cell responses may greatly differ according to the subpopulation studied and the nature of the stimulus used to induce their activation. In our study, in which purified human CD4⁺ T cells were activated by immobilized anti-CD3 and soluble anti-CD28 Abs that closely mimic the physiological activation by the APC, we have observed an inhibitory effect of adenine nucleotides on cytokine secretion and cell proliferation, which is mediated by cAMP independently from adenosine receptors. Interestingly, unlike the P2Y₂ (41) or P2Y₆ (42) receptors that are up-regulated after T cell activation, the cAMP response to adenine nucleotides is present on nonactivated CD4⁺ T cells. The receptor involved in this action might constitute a new interesting target for topical immunosuppression in eye, skin, or airway inflammatory diseases.

We thank Martine Ducarme, Christine Servais and Alain Crusiaux for excellent technical assistance and Dr. Viviane De Maertelaer for the biostatistical analysis.