CD3- and CD28-Dependent Induction of PDE7 Required for T Cell Activation

See allHide authors and affiliations

Science  05 Feb 1999:
Vol. 283, Issue 5403, pp. 848-851
DOI: 10.1126/science.283.5403.848


Costimulation of both the CD3 and CD28 receptors is essential for T cell activation. Induction of adenosine 3′,5′-monophosphate (cAMP)–specific phosphodiesterase-7 (PDE7) was found to be a consequence of such costimulation. Increased PDE7 in T cells correlated with decreased cAMP, increased interleukin-2 expression, and increased proliferation. Selectively reducing PDE7 expression with a PDE7 antisense oligonucleotide inhibited T cell proliferation; inhibition was reversed by blocking the cAMP signaling pathways that operate through cAMP-dependent protein kinase (PKA). Thus, PDE7 induction and consequent suppression of PKA activity is required for T cell activation, and inhibition of PDE7 could be an approach to treating T cell–dependent disorders.

Activation of peripheral T cells in vivo by an antigen-presenting cell is a result of the engagement of both the T cell receptor–CD3 complex (TCR-CD3) and the CD28 costimulatory receptor. When both receptors are occupied by their appropriate ligands, T cells are stimulated to proliferate and produce interleukin-2 (IL-2), whereas occupation of the T cell receptor alone favors T cell anergy or apoptosis (1). Occupation of the CD28 receptor alone appears to have no obvious effect on T cells; nevertheless, CD28 costimulation is required for full activation of CD4 T helper cells, if not all T cells (2). Why is CD28 costimulation required for T cell activation? One possible reason has been suggested by the observation that ligation of TCR-CD3 initiates not only stimulatory signal transduction pathways, such as the MAP kinase (mitogen-activated protein kinase)–dependent signaling pathways (3) and the NFAT (nuclear factor of activated T cells)–dependent signaling pathways (4), but also inhibitory signal transduction pathways such as the PKA (the cAMP-dependent protein kinase)–dependent signaling pathway (5). PKA can inhibit proliferation in many cell types, including fibroblasts (Rat-1 and NIH 3T3 cells), smooth muscle cells, and adipocytes (6), as well as T cells (7). Phosphorylation of NFAT by PKA can abolish the translocation of NFAT from the cytoplasm to the nucleus (8); and phosphorylation of Raf kinase, again by PKA, can block the MAP kinase–dependent signaling pathway (9). Both the translocation of NFAT and the MAP kinase–dependent signal are essential for IL-2 gene expression (3, 4). It is plausible, therefore, that to become fully activated, T cells require additional TCR-independent signals to overcome PKA inhibition.

Because phosphodiesterases (PDEs) play a major role in down-regulation of PKA activity by hydrolyzing cAMP in many cell types (10), and PDE7 protein appears to be detected only in lymphocytes (11), we reasoned that PDE7 might play a role in regulating T cell activation. To test this hypothesis, we first analyzed Northern (RNA) blots to determine the tissue distribution of PDE7. Consistent with previous work demonstrating that PDE7 contributed up to 40% of total PDE activity in several T cell lines (11), PDE7A1, one of the two RNA splicing variants of PDE7, is predominantly expressed in human lymphoid tissues (Fig. 1B). However, the PDE7 protein was barely detectable in isolated peripheral T cells (Fig. 2A), but PDE7 activity was increased by costimulation of T cells with antibodies to CD3 and CD28 (Fig. 2B); whereas the activity of PDE4, the other major isoform of cAMP-specific PDE in T cells, remained unchanged (12). Very high doses of antibody to CD3 (anti-CD3) or antibody to CD28 (anti-CD28) alone could induce PDE7 (12). However, at lower doses that are more likely to mimic the normal physiological state (0.2 ng of anti-CD3 per milliliter and 0.2 μg of anti-CD28 per milliliter), only the combination of the two antibodies increased the amount of PDE7 protein (Fig. 2A) and its activity (Fig. 2B), decreased cAMP levels (Fig. 2C), increased IL-2 expression (Fig. 2D), and promoted proliferation (Fig. 2E), in isolated peripheral T cells.

Figure 1

Northern blot analysis of PDE7A1 and PDE7A2. (A) PDE7A2 is expressed in heart and skeletal muscle, whereas PDE7A1 is expressed in the pancreas and the placenta. (B) PDE7A1 is predominantly expressed in lymphoid tissues. Northern blots of multiple human tissues [sample number 7760-1 in (A) and 7768-1 in (B)] (Clontech, Palo Alto, CA) were hybridized with a 32P-labeled, 1.2-kb, Not 1–Eco RI fragment of human PDE7 cDNA, which can hybridize with both PDE7A1 and PDE7A2. ExpressHyb solution (Clontech) was used for both prehybridization and hybridization performed at 60°C. The32P-labeled probe, together with salmon sperm DNA (0.1 mg/ml), was added directly to the pre-hybridization solution for hybridization. Membranes were washed twice in 2× SSC [3 M sodium chloride and 0.3 M sodium citrate (pH 7.0)] and 0.1% SDS at room temperature for 10 min, then once in 0.1× SSC and 0.1% SDS at 60°C for 20 min. β-actin cDNA (Clontech) was used as a control probe for the mRNA quantity of the Northern blots. Relative molecular weights are indicated. 7A1 stands for the 4.2-kb mRNA of PDE7A1 (15) and 7A2 for the 3.8-kb mRNA of PDE7A2 (11, 15). 7A3? indicates a possible third splice variant, PDE7A3.

Figure 2

Induction of PDE7 in peripheral T cells by anti-CD3 and anti-CD28 costimulation. (A) Protein immunoblot analysis of PDE7 protein. The 57-kD band (lane 4) is the predicted size of the PDE7A1 protein. The 80-kD band, which cross-reacts with PDE7 antibodies, is an unknown protein present in the particulate fraction and appears to have no PDE activity. (B) PDE activity in the immunoprecipitated protein of PDE7-specific antibodies was measured using 1 μM cAMP as a substrate. (C) A decrease in cAMP levels (lane 4) upon costimulation was measured by an RIA kit (15). (D) The increase in IL-2 expression (lane 4) and in (E) cell proliferation (lane 4) after costimulation was correlated with the PDE7 induction. The lane numbers stand for isolated peripheral T cells incubated with the following: lane 1, no antibody; lane 2, anti-CD3 (0.2 ng/ml); lane 3, anti-CD28 (0.2 μg/ml); lane 4, both anti-CD3 (0.2 ng/ml) and anti-CD28 (0.2 μg/ml). Lymphocytes were isolated with a Ficoll-Paque isolation system (Pharmacia, Uppsala, Sweden). B cells were depleted with anti-CD19 conjugated Dynabeads M-450 according to the protocol provided (Dynal, Lake Success, New York). Isolated peripheral T cells were plated into a six-well plate precoated with goat antibodies to mouse IgG in RPMI 1640 containing 10% fetal calf serum (5 × 106 cells per well). The cells were then stimulated with anti-CD3 alone, anti-CD28 alone, or a combination of both. Eight hours after incubation, the supernatant of the cell culture was collected for assay of IL-2 with an enzyme-linked immunosorbent assay (ELISA) kit (Endogen, Woburn, Massachusetts), and the cells were harvested for protein immunoblot analysis, immunoprecipitation, radioimmunoassay of cAMP, and cell proliferation (16).

To investigate whether induction of PDE7 was not merely correlated with but essential for T cell proliferation, we tested the effect of blocking PDE7 expression on T cell proliferation as well as on IL-2 production. Because no PDE7-specific inhibitor is yet available, we used PDE7 antisense oligonucleotides to block PDE7 expression. Several nonoverlapping antisense oligonucleotides were tested. In addition, a series of control oligonucleotides that shared the same nucleotide composition with their corresponding antisense oligonucleotide but had a reversed sequence order were also tested. As shown in Fig. 3, 12 μM of the AS-0 oligonucleotide inhibited cell proliferation by 70% (Fig. 3A, lane 1) and attenuated IL-2 production (Fig. 3E, lane 1) in Hut78 T cells, which constitutively express high levels of PDE7 activity (11); incubation with the other two antisense oligonucleotides resulted in 20% inhibition or less (12). None of the control oligonucleotides had any effect on T cell proliferation and IL-2 production at this concentration (Fig. 3, A and E, lane 4).

Figure 3

The effect of a PDE7 antisense oligonucleotide on proliferation and IL-2 production in Hut78 T cells. (A) The PDE7 antisense oligonucleotide, AS-0, inhibits T cell proliferation. (B) PDE7 mRNA was reduced in the AS-0–treated cells analyzed by RT-PCR. (C) AS-0 reduced the 57-kD PDE7 protein band but did not affect the 80-kD band. (D) AS-0 increased cAMP levels in the cell. (E) AS-0 attenuated IL-2 production measured by an ELISA assay. The lane numbers indicate the following: lane 1, the PDE7 antisense oligonucleotide AS-0 (12 μM); lane 2, AS-0 along with Rp-cAMP (0.5 mM); lane 3, RPMI-1640 media; lane 4, the reversed control oligonucleotide (12 μM); lane 5, Rp-cAMP alone (0.5 μM). The three antisense oligonucleotides correspond to human PDE7A1 cDNA sequences as described (17). Cells (5 × 106 per well) were treated with 12 μM of PDE7 antisense oligonucleotides or with the reversed control oligonucleotides in RPMI 1640 media containing 10% fetal calf serum. To test the effect of PKA inhibitors on cell proliferation, 0.5 mM Rp-cAMP (BioLOG, La Jolla, California) was added together with the oligonucleotide. Twenty-four hours after incubation, the cells were labeled with 1 μCi of [3H]thymidine per well for 20 hours and then harvested for scintillation counting. For RT-PCR analysis, total RNA was isolated with a TRIZOL RNA isolation kit (Gibco-BRL). cDNA was synthesized with the SuperScript II reverse transcription system (Gibco-BRL). One-tenth the volume of the reverse-transcribed cDNA was amplified with PDE7- or PDE4-specific primers (17). The products were subjected to electrophoresis on 2% agarose gels.

To determine whether AS-0 treatment indeed caused a decrease in PDE7 expression, we performed reverse transcription polymerase chain reaction (RT-PCR) analysis of PDE7 mRNA and protein immunoblot analysis of PDE7 protein from the Hut78 cells treated with antisense oligonucleotides or with the control oligonucleotides. Both the mRNA (Fig. 3B, lane 1) and the protein (Fig. 3C, lane 1) of PDE7 were decreased in the AS-0–treated cells but were unchanged in the cells treated with control oligonucleotide (Fig. 3, B and C, lane 4). The inhibition by the antisense oligonucleotide was selective for PDE7, because the PDE4 mRNA remained unchanged in all the cells tested (Fig. 3B, right panel). cAMP concentration in the AS-0–treated cells was increased 2.5 times as compared with cells treated with the control oligonucleotides (Fig. 3D). Therefore, inhibition of T cell proliferation and IL-2 production by PDE7 antisense oligonucleotide appeared to be mediated through the targeting of PDE7 gene expression and consequent increases in cAMP and PKA activity. If this were the case, one would predict that suppression of PKA activity should reverse the inhibition of proliferation and IL-2 production by the PDE7 antisense oligonucleotide. As shown in Fig. 3, Rp-cAMP (0.5 mM Rp-adenosine 3′,5′-cyclic monophosphorothioate), a PKA-specific inhibitor, completely reversed both the inhibition of cell proliferation (Fig. 3A, lane 2) and the attenuation of IL-2 production (Fig. 3E, lane 2) in Hut 78 cells that had been caused by treatment with AS-0, whereas Rp-cAMP itself had no obvious effect on the cell (Fig. 3, A and E, lane 5).

Conclusions based on studies with growth factor–independent T cell lines do not necessarily hold true in primary T cells. To confirm the role of PDE7 in T cell activation, we tested the effect of the PDE7 antisense oligonucleotide on isolated peripheral CD4 T cells that had been activated by anti-CD3 and anti-CD28. As expected, the PDE7 antisense oligonucleotide AS-0 inhibited T cell proliferation by 80%, whereas the reverse control oligonucleotide showed no inhibition (Fig. 4). The PKA inhibitor Rp-cAMP blocked the inhibition of proliferation in the PDE7 antisense AS-0–treated cells (Fig. 4). Therefore, induction of PDE7 is essential for T cell activation.

Figure 4

Inhibition of primary CD4 T cell proliferation by PDE7 antisense oligonucleotides. Peripheral T cells were isolated with anti-CD4–conjugated Dynabeads M-450 according to the provided protocol (Dynal). The isolated CD4 T cells were activated by anti-CD3 (0.2 ng/ml) together with anti-CD28 (0.2 μg/ml), with or without 12 μM of PDE7 antisense AS-0 or the reverse control oligonucleotides, in a 96-well plate precoated with goat antibodies to mouse. Rp-cAMP (0.5 mM ) was added along with the oligonucleotides in the defined wells. After 24 hours of incubation, [3H]thymidine was added and the cells were incubated for another 24 hours. Cell proliferation was measured as described above.

It is still not clear how PDE7 protein and activity are increased upon activation of T cells. Easily detectable amounts of PDE7 mRNA were constitutively expressed in resting T cells (Fig. 1B). Therefore, the increase in PDE7 protein amount and activity may result from an up-regulation of protein translation or a decrease in degradation.

It has been reported that PDE4 inhibitors can also be effective in blocking T cell proliferation and that PDE3-inhibitors potentiate this blocking effect (13). None of the inhibitors used in those studies are thought to inhibit PDE7, thus raising the question of how all three of these PDE isozymes could be involved in T cell proliferation. This issue remains unresolved, but it is perhaps worth noting that in other cell types, different PDEs have been shown to be differentially localized within the cell (10). As a result of this localization, the regulation (14) and function of different PDEs may vary accordingly (10). It is also possible that changes in cAMP modulated by different PDE isozymes work at different phases of the cell cycle. In either of these cases, selective inhibitors of specific PDEs might be expected to inhibit the overall process of cell proliferation.

Our data provide evidence for a mechanism of CD28-dependent costimulation during T cell activation. The data also suggest that PDE7 may be a good target for selective therapeutic modulation of T cell responsiveness.

  • * To whom correspondence should be addressed. E-mail: Beavo{at}


View Abstract

Navigate This Article