Among the signals necessary to generate CD4+CD25+FOXP3+ T cells from CD4+CD25−FOXP3− T cells, a pivotal role is played by CD28. However, in humans, it is not known whether CD28 signaling independently of TCR promotes forkhead box protein 3 (FOXP3) expression and regulates CD4+CD25+FOXP3+ T cell functions. To address this issue, starting from our previous experience, we analyzed the unique signals delivered by CD28 following stimulation by its natural ligand B7. Our results show that, in primary CD4+CD25− T cells, CD28 signals independent of TCR-mediated stimulatory pathways are sufficient to induce the transcription of FOXP3 in a small number of CD4+CD25− T cells committed to express FOXP3. These signals are dependent on CD28-derived PI3K/Akt pathways and resistant to cyclosporin A. In addition, we demonstrated that translated FOXP3 was recruited to CD25, Il-2, and Ctla4 target promoters. CD28-mediated FOXP3 expression was transient and correlated with CD25 expression. The presence of FOXP3 in CD28-activated CD4+CD25− T cells correlated with a transient unresponsiveness to antigenic stimuli. The addition of exogenous IL-2 did not influence either FOXP3 or CD25 expression but rescued CD28-activated T cells from apoptosis. Our results, demonstrating that FOXP3 expression driven solely by the CD28/B7 interaction inhibited T cell activation, support the role of CD28 in the regulation of peripheral tolerance and suggest a new mechanism through which it could occur.
Although for many years the receptor CD28 has been considered essential for mediating costimulatory signals in concert with TCR signaling necessary in enhancing T cell activation, proliferation, cytokine production, and survival (1), more recently two novel aspects of CD28 functions have been disclosed. First, many studies from our and other laboratories support the hypothesis that CD28 activates signaling pathways that are distinct from those induced by the TCR. It has been reported that CD28 cytoplasmic tail binds important signaling molecules (1, 2) and can induce cytoskeleton rearrangements and up-regulate IL-2/4 independently of TCR (3, 4). We have shown that CD28 engagement by its ligand B7 can generate TCR-independent signals leading to I-κB kinase complex and NF-κB activation (5, 6). Furthermore, CD28 engagement by B7 induces the transcriptional activation of Il-8, Bcl-xL, and BAFF genes through the recruitment of RelA and p52 on their respective promoters (7). Second, perhaps the most paradoxical novel function of CD28 is its involvement in the development and homeostasis of CD4+CD25+ T cells known as regulatory T cells (Treg).3 Tregs are negative regulators of T cell signaling which contribute to T cell “anergy” and to maintenance of self-tolerance (8, 9, 10). Efficient generation of Treg in the thymus requires CD28 (11, 12) and in periphery its ligand B7 concurs to limit T cell activation by sustaining a population of CD4+CD25+ Treg (13). This supports the hypothesis that the presence of B7 in a mouse, even on cells that are not displaying the cognate Ag, controls adaptive immune responses (14). It is also clear that CD28 deficiency can lead to a reduced disease potential as well as to enhanced susceptibility to autoimmune disease by altering T cell effector and Treg compartments (11, 13, 15). Furthermore, CD4+CD25− T cells can be converted in vivo to become CD4+CD25+ T cells with regulatory function only in the presence of B7 (16). The importance of CD28 signaling in maturation and proliferation of Treg has been also characterized in vitro and both mouse and human Tregs can be expanded using anti-CD3 and anti-CD28 Abs in the presence of high concentrations of IL-2 (13, 17, 18).
The FOXP3 gene, that encodes the forkhead box protein 3 (FOXP3) transcription factor, is thought to be the “master gene” of Treg (19, 20). FOXP3 acts as both a gene transcriptional activator and repressor (21). Lack of FOXP3 leads to development of autoimmune lymphoproliferative diseases and ectopic FOXP3 expression can phenotypically convert effector T cells to Treg cells (22). However, FOXP3 expression does not perfectly correlate with Treg suppressive function (23). Several lines of evidence suggest that CD28 is required for FOXP3 induction. It has been shown that CD28-mediated signals induce developing thymocytes to express FOXP3 and to initiate the Treg cell differentiation program (12); CD28 is also primarily required for the survival/homeostasis of TGF-β-converted thymic CD4+CD25+ Tregs (24). It is widely accepted that FOXP3 expression in TCR-activated CD4+CD25− T cells requires costimulation by CD28, also in humans (18, 23, 25, 26, 27).
Altogether the above reported data strongly suggest that CD28 is capable of delivering signals that promote CD4+CD25+FOXP3+ T cells. However, for several reasons such as: 1) the complexity of signals that in vivo regulate FOXP3 expression, 2) the costimulatory nature of CD28 that amplifies TCR signaling, and 3) the difficulty to analyze in vitro the unique signal transduction pathway that specifically targets FOXP3 promoter, the effects of CD28 signaling independent of TCR-mediated signals in CD4+CD25− T cells are still lacking. We explored this function by using our previously described CD28/B7 selective stimulation system (5, 7, 28).
In this study, we show that CD28 engagement by B7, in the absence of TCR triggering, is sufficient to induce the transcriptional activation of FOXP3 in primary CD4+CD25− T cells. Moreover, we present evidence that in chromatin immunoprecipitation (ChIP) experiments FOXP3 was recruited to CD25, Il-2, and Ctla4 target promoters. CD28-mediated FOXP3 expression was transient and correlated with CD25 expression. The addition of exogenous IL-2 did not influence both FOXP3 and CD25 expression but rescued CD28-activated T cells from apoptosis. Furthermore, FOXP3 expression in CD28-activated CD4+CD25− T cells regulates unresponsiveness to TCR signaling. In conclusion, the above reported data identify CD28 as a key regulator of FOXP3 in primary CD4+CD25− T cells. Moreover, since in these cells CD28-induced FOXP3 regulates CD25 expression, anergy, and apoptosis, our results suggest a new mechanism through which CD28 may regulate peripheral tolerance.
Materials and Methods
Cell lines and reagents
Murine L cells (Dap3) and murine L cells expressing human B7.1 (Dap3/B7) were cultured in complete DMEM (Life Technologies) supplemented with 50 μg/ml hygromycin B (Sigma-Aldrich). Inhibitors used were N-p-tosyl-l-phenylalanine chloromethyl ketone (TPCK; Sigma-Aldrich), 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002; Calbiochem), 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine (PP2; Calbiochem), and cyclosporin A (CsA; Sigma-Aldrich). Anti-CD3 (X-35) was obtained from Immunotech), anti-human CD28.2 was from BD Biosciences, and goat anti-mouse (Poly4053) was from Biolegend. Anti-FOXP3 (H-190) and anti-α-tubulin were purchased from Santa Cruz Biotechnology. PE-conjugated anti-human FOXP3 Abs were either PCH101 and rat IgG2a isotype control was purchased from eBioscience or 259D/C7 and mouse IgG1 isotype control were purchased from BD Pharmingen. FITC-conjugated anti-human CD4 (MEM-241) and anti-human CD25 (MEM-181) were obtained from Immunological Science. rIL-2 was purchased from Roche.
Purification of CD4+CD25+and CD4+CD25− T cells
Human peripheral blood was obtained from a healthy donor blood bank after informed consent in accordance with procedures approved by the local human ethical committee. PBMC were prepared by centrifugation over Lympholyte-H (Cederlane Laboratories) gradients. Isolation of human CD4+CD25+ and CD4+CD25− T cells was performed using a human Treg cell isolation kit (Miltenyi Biotec) according to the manufacturer’s instructions. Briefly, CD4+ T cells were first isolated through negative selection by removing all other cell types. Preisolated CD4+ T cells were incubated with 10 μl of magnetic beads conjugated with anti-CD25 Ab (for 107 cells) to separate the CD4+CD25− T cell population from CD4+CD25+ T cells. The purity of the CD4+CD25− T cell population was confirmed to be >95% by flow cytometry. CD4+CD25− T cells were cultured in RPMI 1640 (Life Technologies) supplemented with 2% human serum (Euroclone), l-glutamine, penicillin, and streptomycin (Sigma-Aldrich).
CD28-mediated CD4+CD25− T cell activation
CD4+CD25− T cells (2 × 106/ml) were activated either with Dap3 (used as control) or Dap3/B7 as previously described (7). Dap3 and Dap3/B7 were previously fixed for 30 min at room temperature with 0.05% glutaraldehyde on culture plate. Alternatively, CD4+CD25− T cells were activated with 1 μg/ml anti-CD28 cross-linked with 10 μg/ml goat anti-mouse.
Quantitative real-time PCR
Total RNA was extracted using TRIzol reagent (Invitrogen) from 3 × 106 purified T cells and was reverse-transcribed into cDNA by using Moloney murine leukemia virus reverse transcriptase (Invitrogen). Quantitative real-time PCR was performed on an Applied Biosystems PRISM 7300 detection system. FOXP3 and CD25 message expression was determined by the TaqMan method of real-time PCR with GAPDH as endogenous control. TaqMan Universal PCR Master Mix and the FOXP3 primer/probe set (part no. Hs00203958_m1), the CD25 primer/probe set (part no. Hs00907778_m1), and the GAPDH primer/probe set (part no. Hs99999905_m1) were purchased directly from Applied Biosystems. Relative quantification was performed using the comparative cycle threshold method as described in User Bulletin edit by Applied Biosystems.
ChIP assays were performed as previously described (29). Briefly, after fixing in 1% formaldehyde, 10 × 106 cells were lysed for 5 min in 50 mM Tris (pH 8.0), 2 mM EDTA, 0.1% Nonidet P-40, and 10% glycerol supplemented with protease inhibitors. Nuclei were resuspended in 50 mM Tris (pH 8.0), 1% SDS, and 5 mM EDTA. Chromatin was sheared by sonication, centrifuged, and diluted 10 times in 50 mM Tris (pH 8.0), 0.5% Nonidet P-40, 0.2 M NaCl, and 0.5 mM EDTA. After preclearing with a 50% suspension salmon sperm (Sigma-Aldrich)-saturated protein A (Amersham Biosciences), lysates were incubated at 4°C overnight with anti-FOXP3 Ab (eBioscience). Immune complexes were collected with sperm-saturated protein A, washed three times with high salt buffer (20 mM Tris (pH 8.0), 0.1% SDS, 1% Nonidet P-40, 2 mM EDTA, and 500 mM NaCl) and five times with 1× Tris/EDTA. Immune complexes were extracted in 1× Tris/EDTA containing 1% SDS and protein-DNA cross-links were reverted by heating at 65°C overnight. DNA was extracted by phenol-chloroform and ∼
Protein extracts were obtained by lysing 106 cells for 30 min at 4°C in lysis buffer (20 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 1% Nonidet P-40) in the presence of protease and phosphatase inhibitors. Proteins were resolved by 10% SDS-PAGE and blotted onto nitrocellulose membranes. Blots were incubated with anti-FOXP3 (H-190) and anti-α-tubulin, extensively washed, and after incubation with HRP-labeled goat anti-rabbit or goat anti-mouse Abs, developed with the ECL detection system (Amersham Biosciences).
Immunofluorescent staining and flow cytometry analysis
Cells (2 × 105) were first surface stained with FITC-conjugated anti-CD4 or FITC-conjugated anti-CD25 Abs; after fixation and permeabilization, cells were incubated with PE-conjugated anti-human FOXP3 Abs or isotype controls according to the manufacturer’s instructions. BD Biosciences cytometric bead array human cytokine kit was used to measure IL-2 levels. Data acquisition and analysis were conducted on a flow cytometric platform using BD Biosciences CellQuest and cytometric bead array software.
Detection of apoptotic T cells
Apoptosis was measured by staining with FITC-conjugated annexin V and propidium iodide according to the manufacturer’s instructions. Flow cytometric analysis was performed on a BD Biosciences FACSCalibur.
Proliferation assays were performed on CD4+CD25− T cells stimulated or not with anti-CD28. CD4+CD25− and CD4+CD25+T cells were plated at 5 × 104 cells/well in 96-well plates with 5 × 104 cells/well CD4+- depleted PBMC as feeders in the presence or absence of 40 IU/ml IL-2 and activated with anti-CD3 and anti-CD28 Abs. Cells were pulsed with 1 μCi/well [3H]thymidine on day 3 and proliferation was assessed 18 h later using a liquid scintillation counter.
Fresh autologous CD4+CD25− responder T cells were stimulated with anti-CD3 and anti-CD28 Abs in the presence of 5 × 104 feeders cells/well and plated at 5 × 104/well either alone or in combination with fresh CD4+CD25+ T cells or in combination with CD28-activated CD4+CD25− T cells. The cells were cocultured at a different ratio responder:suppressor. Suppression was assessed by CFSE (2.5 μM) cell division profiles.
FOXP3 and CD25 are up-regulated in CD4+CD25− T cells by CD28
To study the effect of CD28 signaling on FOXP3 expression, we analyzed the interaction of CD28, expressed on CD4+CD25− T cells, with its natural ligand B7. To this purpose, we measured cell growth, FOXP3 mRNA levels, and FOXP3 protein expression (Fig. 1) in CD4+CD25− T cells cultured with Dap3 or Dap3/B7 murine fibroblasts. The levels of FOXP3 mRNA and protein in freshly isolated CD4+CD25+ T cells were also determined. A time course of FOXP3 mRNA quantification in CD4+CD25− T cells, representative of five performed experiments with different donors, is reported in Fig. 1,A. The shown histogram describes a classical Gaussian-like shape with a maximum of FOXP3 mRNA synthesis at 12 h. However, depending on the donor, maximum synthesis of FOXP3 mRNA could also occur at 24 and 48 h (data not shown). Fig. 1,E shows the expression of FOXP3 protein in two different donors, who reached FOXP3 maximum synthesis at 24 and 48 h after stimulation, respectively. FACS analysis of intracellular FOXP3 stained with anti-FOXP3 mAb PCH101 (Fig. 1, C and D), performed to analyze FOXP3 expression at the single-cell level, shows an increase in the frequency of CD4+FOXP3+ T cells such as mean fluorescence intensity (MFI) after 48 h of culture in the presence of Dap3/B7, and ∼20% (range, 5–40%) of CD28-activated cells stained positive for FOXP3. On the contrary, no increase of intracellular FOXP3 was observed in cultures without B7. To exclude that CD28-activated cells were stained nonspecifically with PCH101, 259D/C7 mAb was also used (30); T cells stained in parallel with PCH101 and 259D/C7 gave similar results (data not shown). Although at t0 FOXP3 was already present in ∼4% (range, 0.5–5%), the possibility that CD28-activated FOXP3+ T cells expanded from the unstimulated FOXP3+ T cells is quite impossible in the absence of cell division. Indeed, CFSE-labeled CD4+CD25− T cell measures demonstrated that the interaction of CD28 costimulatory molecules with B7 did not influence T cell division (data not shown). Moreover, unstimulated CD4+CD25+ T cells expressed even higher levels of FOXP3 in comparison to CD28-activated CD4+CD25− T cells (Fig. 1, A and C). This suggests that after CD28 signaling CD4+FOXP3+ T cells predominantly arose from CD4+FOXP3− T cells, although with levels of FOXP3 comparable to that observed in CD4+CD25+ T cells lacking Treg function (27). To characterize the CD28-mediated biochemical pathways involved in FOXP3 transcription, we used different inhibitors. The results in Fig. 1 B show that LY294002, a specific inhibitor of PI3K/Akt, and TPCK, an NF-κB inhibitor, significantly inhibited FOXP3 mRNA synthesis, consistent with the requirement of PI3K/Akt in transducing the CD28 signal and favoring NF-κB translocation to the nucleus. On the contrary, PP2, an inhibitor of src tyrosine kinases, and CsA, an inhibitor of calcineurin, did not show any activity on FOXP3 transcription, confirming the resistance of the CD28 pathway to CsA (31). Altogether these results strongly support that a unique signal mediated by CD28 may activate FOXP3 transcription and translation. The influence of TGF-β on FOXP3 expression, known to be present in human serum, has been excluded since the stimulation of CD4+CD25− T cells for 48 h with Dap3/B7 in serum-free synthetic medium gave results similar to those obtained with human serum (data not shown).
It is common opinion that agonistic Abs do not always mimic the triggering signals mediated by natural ligands. Discordant results may be obtained depending on many conditions such as Ab concentration, affinity for the recognized epitopes, and position of the binding site on the target molecules (32). However, most of the published results on CD28 signaling have been obtained with mAbs. Therefore, to compare the effect of the natural ligand B7 with that obtained by mAbs, we analyzed in time course experiments the effect of cross-linked anti-CD28 mAb on FOXP3 expression in CD4+CD25− T cells and in parallel the effects of anti-CD3 and anti-CD3 plus anti-CD28 mAbs (Fig. 2). Moreover, anti-CD28 mAb instead of CTLA-4-Ig could also represent a control for excluding the involvement of CTLA-4 in FOXP3 activation (7). An anti-CD3 mAb (X-35) with a specific medullary thymocyte reactivity has been used instead of the more common anti-CD3 mAb UCHT1 that present a pan thymocyte reactivity (33). Indeed, it has been described that FOXP3-expressing thymocytes are largely localized to the thymic medulla (34). Moreover, X-35 is unable to trigger the TCR-related PI3K-dependent pathway. As anti-CD28 mAb, we used clone CD28.2. CD4+CD25− T cells activated in presence of anti-CD3, anti-CD28, and anti-CD3 plus anti-CD28 were analyzed for FOXP3 expression in immunoblotting at different times. It was interesting to note that both anti-CD3 and anti-CD28 alone in culture induced FOXP3, although with different profiles (Fig. 2,A). The comparative analysis of the three different activations at 12 h (Fig. 2,B) shows that the FOXP3 increase was only dependent on CD28 signaling, while at 24 and 48 h CD28 engagement failed to amplify TCR/CD3-mediated FOXP3 synthesis (Fig. 2,A). As control, FOXP3 expression in unstimulated CD4+CD25+ T cells was always higher than that observed in CD4+CD25− cultures stimulated with anti-CD28 mAb (Fig. 2,B). Finally, given that CD28 signaling has been described to be CsA resistant or sensible depending on the absence or presence of TCR ligation, respectively (35), we tested whether the levels of FOXP3 protein induced by anti-CD28 and anti-CD3 plus anti-CD28 could be similarly affected by CsA. This is an important analysis for at least two different reasons: 1) we have shown that FOXP3 transcription induced by B7.1 was completely CsA resistant (Fig. 1,B) but we do not know the sensitivity to CsA of FOXP3 induced by cross-linked anti-CD28 mAb and 2) we can compare our experiments with those performed using TCR plus CD28 that reported the sensitivity of FOXP3 to CsA (25). Therefore, we measured FOXP3 expression in anti-CD28 and anti-CD3 plus anti-CD28 T cell cultures activated for 12 and 48 h in the presence or absence of CsA (Fig. 2, C and D). Interestingly, although FOXP3 in anti-CD28 cultures were CsA resistant at both time points, FOXP3 expression in anti-CD3 plus anti-CD28 cultures was CsA resistant at 12 h and CsA sensitive at 48 h, suggesting that CD28 and TCR influenced FOXP3 expression by different pathways.
Recent evidences support that FOXP3 regulates the expression of CD25 in mice (36). To verify this, we performed a set of experiments similar to those described in Fig. 1. It was interesting to observe, in Fig. 3,A, that the profile of the histogram of CD25 mRNA expression is completely different from that of FOXP3, in that the maximum synthesis of CD25 was delayed with respect to FOXP3. Moreover, the addition of FOXP3 mRNA inhibitors LY294002 and TPCK strongly suppressed CD25 expression. PP2 gave a less marked effect (Fig. 3,B). Altogether these data support the role of FOXP3 in the transcriptional regulation of CD25 and suggest that the pathway CD28/PI3K/NF-κB is deeply involved in the transcriptional activation of FOXP3 and consequently CD25. The analysis of the frequency and MFI of CD4+CD25+ T cells at t0 and t48 confirms that FOXP3 expression correlates with cell surface CD25 expression (Fig. 3 C).
IL-2 rescues CD4+CD25+FOXP3+ T cells from apoptosis but does not influence FOXP3+ and CD25+ expression
Although time course experiments showed a decrease in FOXP3 expression after 48 h from the beginning of CD28 stimulation, we could not analyze its expression at later times because CD28-activated CD4+CD25− T cells became susceptible to apoptosis. Indeed, Fig. 4, A and B, shows the increase of annexin V+-stained T cells in CD28-activated CD4+CD25− T cells cultured for 72 h with Dap3/B7 cells. However, addition of exogenous IL-2 significantly prevented apoptosis (Fig. 4,C) in a dose-dependent manner (Fig. 4,D). The evidence that IL-2 favored T cell viability suggested to us to measure FOXP3 and CD25 following the addition of rIL-2 during and after 48 h of culture (Fig. 5, A and B, respectively). We demonstrated that IL-2 did not modify FOXP3 protein expression. Moreover, it is notable that the decrease of FOXP3 was accompanied with the decrease of the frequency of CD4+CD25+ T cells (Fig. 5 C), suggesting that the CD28-mediated expression of both FOXP3 and CD25 in CD4+CD25− T cells is correlated.
FOXP3 is recruited on CD25, Il-2, and Ctla4 promoters after CD28 activation
It has been shown by ChIP experiments that FOXP3 could occupy CD25, Il-2, and Ctla4 promoters in both murine CD4+CD25+ Treg cells and Jurkat T cells retrovirally transduced with FOXP3 and stimulated by TCR plus CD28 and/or ionomycin (37). We verified this phenomenon in primary CD4+CD25− T cells activated for 24 and 48 h by Dap3/B7. The results reported in Fig. 6 show for the first time, to the best of our knowledge, that the activatory pathways mediated by CD28 in CD4+CD25− T cells favor not only the transcription and translation of FOXP3, but also its occupancy of CD25, Il-2, and Ctla4 promoters. Interestingly, after 24 h, FOXP3 was recruited on all studied promoters, but the occupancy of CD25 and Il-2 promoters (Fig. 6, A and C) was maintained for all of the observation time, whereas FOXP3 exited the Ctla4 promoter between 24 and 48 h from the activation (Fig. 6,B). To analyze the sensitivity of FOXP3 recruitment on Il-2 promoter to CsA, we stimulated CD4+CD25− T cells with anti-CD28 mAbs for 24 h in the presence and absence of CsA (Fig. 6 D). CsA induced a slight decrease of FOXP3 recruitment on the Il-2 promoter, suggesting that, although CD28 signaling regulates FOXP3 expression in a CsA-resistant manner, signals from calcineurin may have a role in this system (38).
CD28-activated CD4+CD25− T cells are anergic but not suppressor
The occupancy of CD25, Il-2, and Ctla4 promoters by FOXP3, an event described as typical of Treg, suggested to us to verify whether CD28-activated T cells expressing FOXP3 and CD25 had acquired suppressive activity. Therefore, CD28-activated or CD4+CD25+-unstimulated T cells (natural Treg), purified using a Treg cell isolation kit, were added to autologous CD4+CD25− T cells (responder cells) and cultured in anti-CD3-coated wells with anti-CD28 for 5 days. Fig. 7,A, where the effects of CD28 on cell growth of responder cells labeled with CFSE were analyzed, shows that CD28-activated CD4+CD25− failed to mediate suppressive activity while autologous natural Treg, used as control, induced a dramatic inhibition of cell growth. It has been described that the transient expression of FOXP3 in anti-CD3-activated nonregulatory CD4+ T cells is strongly associated with hyporesponsiveness (27). We investigated whether this could also occur after CD28 activation. Therefore, we purified CD4+CD25+ T cells from CD28-activated CD4+CD25− T cells after 48 h of culture. These cells were either immediately stimulated or allowed to rest in IL-2 culture medium for 7 days until FOXP3 was completely down-regulated and stimulated. The release of IL-2 in the culture medium and the effect of exogenous IL-2 have been also tested. Fig. 7,B shows that CD4+CD25+ T cells purified from CD28-activated CD4+CD25− T cells and expressing FOXP3 are completely unresponsive to mitogenic stimuli. Indeed, these cells were unable to proliferate and release IL-2 (inset, Fig. 7,B). On the contrary, the same cells, which become FOXP3− after 7 days in IL-2 culture medium (data not shown), were similar to the non-CD28-activated CD4+CD25− T cells in that they proliferated when stimulated (Fig. 7 C).
Both in mice and in humans CD4+CD25+ T cells with regulatory functions are important components of peripheral tolerance that is responsible in the control of autoreactive T cells in normal individuals. FOXP3 transcription factor is a critical regulator of these cells (22). Among the signals necessary to generate CD4+CD25+FOXP3+ T cells, a pivotal role is played by CD28. However, while in mice it was possible to study the role played by the interaction of CD28 with its ligand B7 in the activation and maintenance of CD4+CD25+FOXP3+ T cells, in humans similar results are lacking. Taking advantage from our previous results that allowed us to characterize a human model system for the analysis of the CD28/B7-dependent pathways in T cells, we have approached two important issues: the effect of the unique signal delivered by CD28 in CD4+CD25− T cells and how FOXP3 influences the phenotypic and functional properties of CD28-activated CD4+CD25− T cells.
The TCR and CD28 are independent signaling units and a biophysical and biochemical comparison shows differences in their signaling scope. However, the TCR and CD28 share many redundant downstream signaling pathways and often CD28 acts as an amplifier (1). As a consequence, it is very difficult to discriminate the effects of the two independent signaling units. In previous studies, aiming at identifying the CD28-dependent pathways in a TCR-independent manner in primary CD4+ T cells, we found that the PI3K pathway and NF-κB are the most relevant CD28 biochemical targets (5, 6, 28) and that Bcl-xL, Il-8, and BAFF are some of the CD28-mediated NF-κB-regulated genes. The present work on primary human CD4+CD25− T cells provides for the first time the experimental evidence that CD28 signals independent from TCR and dependent on the PI3K/Akt pathway are sufficient to induce the transcription of FOXP3. We reached this conclusion by activating CD28 either by the natural ligand B7 or by cross-linking with specific mAb. However, despite that FOXP3 was rapidly induced, <30% of the CD4+CD25− T cells expressed FOXP3 after CD28-mediated activation, supporting the view that among CD4+CD25− T cells only a small number of cells are precommitted to express FOXP3 (39).
Recent evidence demonstrates that T cells with lower FOXP3 expression showed surface expression of Treg-associated molecules, such as CD25, intermediate between that of wild-type Tregs and that of conventional T cells, and had greatly compromised suppressive function in vitro and in vivo (39, 40). These authors suggest that rather than initiating a de novo developmental program in self-reactive T cells, FOXP3 takes advantage of preceding and coincidental features of Treg precursor cells, probably facilitated by TCR signaling. Our data support this hypothesis and add a further suggestion: CD28 could be another facilitating signal. Indeed, without cell expansion, CD28-activated culture of CD4+CD25− T cells presented an increase of CD4+FOXP3+ T cells and FOXP3 recruitment on genes normally activated by TCR and capable of negative feedback regulation of T cell activation, such as CTLA-4. The prominent role of TCR could be the amplification and stabilization of FOXP3 (41). Indeed, our results obtained with anti-CD3 and anti-CD28 mAbs show that both TCR/CD3 and CD28 activation mediated up-regulation of FOXP3. However, although the effect of TCR/CD3 is more than two times higher than that mediated by CD28, the lack of synergy between the two stimulators strongly suggests that in FOXP3-committed T cells the two pathways are independent. The evidence that in primary human T cells TCR/CD3 and CD28 have unique roles in the activation of MEK1/2 and PI3K/Akt-dependent pathways support this hypothesis (42).
Another observation supporting the hypothesis of the two independent pathways is the differential sensitivity to CsA of TCR/CD3- and CD28-mediated FOXP3 expression. One of the hallmarks of CD28 stimulation is its insensitivity to the immunosuppressive drugs CsA and FK506; therefore, the lack of effect of CsA on CD28-mediated activation of FOXP3 is consistent with previously published reports (31, 43). The differential sensitivity to calcineurin inhibitors of CD28 signal triggered alone or in association with TCR has been also demonstrated in genomic expression programs (35). However, although CsA is an inhibitor of the calcineurin/NFAT pathway, the CsA-resistant CD28 signal has been described to result in the activation and binding of NFAT to DNA (44) and in enhancing nuclear occupancy of NFAT proteins by CD28-mediated inhibition of GSK3 (35, 45). Since many experimental evidences support the need of NFAT for favoring either the transcription of FOXP3 or its binding to DNA (25, 37, 38), the binding of NFAT to FOXP3 and Il-2 genes in CD28-activated CD4+CD25− T cells cannot be excluded. This idea is currently being tested.
Downstream of PI3K/Akt, the effects of the CD28 unique signal include the activation of NF-κB (7), indicating a possible role for NF-κB in FOXP3 expression in CD4+CD25− T cells. It has been described that FOXP3 physically associates with NF-κB and blocks its ability to induce the expression of a NF-κB-dependent gene (46), but a role of NF-κB in FOXP3 expression is lacking. The observation that inhibitors of NF-κB suppress FOXP3 expression and specific NF-κB units are recruited on the FOXP3 promoter (M. Soligo, C. Camperio, C. Scottà, and E. Piccolella, manuscript in preparation) would be consistent with the view that CD28-mediated activation of NF-κB may represent another decisive signal.
It has been reported that mature peripheral CD4+CD25− T cells can convert to CD4+CD25+ Treg cells spontaneously in vivo in a thymus-independent but B7-dependent manner (16). The converted cells failed to proliferate after stimulation and expressed high levels of FOXP3 mRNA. Our data allow us to speculate that this may also occur in humans, although in a transient way. Indeed, we have demonstrated that CD28-mediated activation of FOXP3 and its recruitment on CD25, Il-2, and Ctla4 promoters converted a small subpopulation of CD4+CD25− T cells to CD4+CD25+ T cells and acquired an anergic phenotype. Indeed, in these cells, T cell proliferation and IL-2 synthesis were impaired. However, we present evidence that this unresponsiveness was not maintained in the absence of FOXP3, and T cell proliferation could be easily reconstituted upon subsequent exposures to Ag. This suggests that the induction of FOXP3, at least in a small number of CD4+CD25− T cells committed to express FOXP3, in the absence of TCR signaling, represents a new mechanism by which FOXP3 can mediate a transient shut down of the TCR pathways. We have also shown that the delivery of the CD28 signal in CD4+CD25− T cells led to the propensity for apoptosis. However, the evidence that exogenous IL-2 rescued CD4+CD25+ T cells from apoptosis but did not affect the decrease of FOXP3 and CD25 supports the view that FOXP3-driven enhancement of CD25 is important for T cell survival (47, 48).
In conclusion, we have not only confirmed in humans that CD28 provides an unique signal to promote FOXP3 expression necessary to convert human CD4+CD25− T cells to CD4+CD25+ T cells, but we went on demonstrating that FOXP3 up-regulation occurred independently of TCR. This suggests a scenario where the expression of B7 on professional APCs may contribute significantly to the homeostasis of CD4+CD25− T cells expressing or committed to acquire FOXP3 and to induce in these cells a transient nonresponsiveness important for the maintenance of peripheral tolerance.
We thank the unknown reviewers of this manuscript for their constructive comments. We thank Paola Del Porto for useful suggestions and Antonio Costanzo for helpful discussion and critical reading of this manuscript.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by grants from the Institute Pasteur Fondazione Cenci Bolognetti, University Sapienza of Rome, from the Ministry of Universities and Scientific and Technological Research, and from Agenzia Spaziale Italiana, MOMA project.
Abbreviations used in this paper: Treg, regulatory T cell; FOXP3, forkhead box protein 3; ChIP, chromatin immunoprecipitation; TCPK, N-p-tosyl-l-phenylalanine chloromethyl ketone; PP2, 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine; CsA, cyclosporin A; LY294002, 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one; MFI, mean fluorescence intensity.