Regulatory T cells facilitate the nuclear accumulation of inducible cAMP early repressor (ICER) and suppress nuclear factor of activated T cell c1 (NFATc1)

Abstract

Inducible cAMP early repressor (ICER) is a transcriptional repressor, which, because of alternate promoter use, is generated from the 3′ region of the cAMP response modulator (Crem) gene. Its expression and nuclear occurrence are elevated by high cAMP levels in naturally occurring regulatory T cells (nTregs). Using two mouse models, we demonstrate that nTregs control the cellular localization of ICER/CREM, and thereby inhibit IL-2 synthesis in conventional CD4⁺ T cells. Ablation of nTregs in depletion of regulatory T-cell (DEREG) mice resulted in cytosolic localization of ICER/CREM and increased IL-2 synthesis upon stimulation. Direct contacts between nTregs and conventional CD4⁺ T cells led to nuclear accumulation of ICER/CREM and suppression of IL-2 synthesis on administration of CD28 superagonistic (CD28SA) Ab. In a similar way, nTregs communicated with B cells and induced the cAMP-driven nuclear localization of ICER/CREM. High levels of ICER suppressed the induction of nuclear factor of activated T cell c1 (Nfatc1) gene in T cells whose inducible Nfatc1 P1 promoter bears two highly conserved cAMP-responsive elements to which ICER/CREM can bind. These findings suggest that nTregs suppress T-cell responses by the cAMP-dependent nuclear accumulation of ICER/CREM and inhibition of NFATc1 and IL-2 induction.

Naturally occurring regulatory T cells (nTregs) are of crucial importance for preventing autoimmunity (1). It is widely accepted that nTregs exert their regulatory capacity via cell contact-dependent and cytokine-independent mechanisms. The inhibition of Il2 gene expression in effector CD4⁺ T cells is a characteristic feature of nTreg-mediated suppression (2). nTreg cells harbor high levels of cAMP (3), and the contact-dependent transfer of cAMP from nTregs to effector CD4⁺ T cells was shown to contribute to nTreg-mediated suppression of effector CD4⁺ T cells (4, 5). Inhibition of cAMP degradation by the phosphodiesterase (PDE)-4 inhibitor rolipram enhanced nTreg-mediated suppression of effector CD4⁺ T cells both in vitro and in vivo (6). Unlike conventional CD4⁺ T cells, which express relatively low ICER/CREM levels, nTregs show a marked increase in inducible cAMP early repressor (ICER)/cAMP response modulator (CREM) mRNA and protein levels because of the enhancing effects of Foxp3 (7). ICER binds specifically to multiple nuclear factor of activated T cell (NFAT)/AP-1 sites within the Il2 promoter (8), which correlates with a strong decrease in the number of IL-2–expressing effector CD4⁺ T cells (7). It was proposed that such composite NFAT/AP-1 binding motifs generate NFAT/Foxp3 inhibitor complexes, which suppress the Il2 gene expression in nTreg cells (9).

In vitro, NFAT and ICER form inhibitory ternary complexes on several composite NFAT/AP-1 DNA binding sites that, in addition to suppression of IL-2 transcription, are essential for the inhibition of other cytokines, such as TNF-α, IL-4, IL-13, and GM-CSF (10). Therefore, NFAT/ICER complexes seem to be instrumental for the transcriptional attenuation of numerous NFAT-driven cytokine promoters in conventional CD4⁺ T cells. A critical role of NFAT factors for inhibitory complex formation is further strengthened by observations indicating that combined NFATc2/c3 deficiency rendered conventional CD4⁺ T cells unresponsive to suppression, although normal nTreg development was detected in those mice (11). Moreover, targeting ICER/CREM in RNAi and antisense RNA approaches antagonized the nTreg-mediated suppression and/or inhibition of IL-2 production in conventional CD4⁺ T cells, rendering these effector T cells refractory to suppression (7, 12).

Activation of effector CD4⁺ T cells results in strong transcriptional induction and nuclear translocation of NFATc1 (13). By contrast, nTregs are unable to induce NFATc1 at the transcriptional level (14). They express relatively low levels of cytoplasmic NFATc1 and do not translocate NFATc1 efficiently to the nucleus upon CD3/CD28 stimulation (15). This property of nTregs is associated with reduced calcium flux, diminished calcineurin activation, and increased activity of the glycogen synthase kinase-3β, a negatively acting NFAT protein kinase. These observations suggest that the signals leading to the generation of suppressive transcription complexes in nTregs differ markedly from those critical for the generation of NFAT/AP-1 and other NFAT complexes that activate cytokine promoters in conventional CD4⁺ T cells.

By using mAbs raised against CD28, including a CD28 superagonistic (CD28SA) Ab (16), we investigated the relationship between the activation of nTregs and ICER-mediated suppression in conventional CD4⁺ T cells. Depletion of nTregs from the T-cell compartment of depletion of regulatory T-cell (DEREG) mice expressing the diphtheria toxin (DT) receptor in Foxp3⁺ T cells before CD28SA stimulation led to cytosolic localization of ICER/CREM in CD28SA-stimulated conventional CD4⁺ T cells. This correlated with an increase in IL-2 expression. Interaction of nTregs with conventional CD4⁺ T cells in vivo resulted in the nuclear localization of ICER/CREM and cessation of IL-2 synthesis. Moreover, contacts of nTregs with B cells led to an increase in nuclear localization of ICER/CREM in a similar fashion as detected for conventional CD4⁺ T cells. One mechanism of ICER/CREM-mediated suppression of conventional CD4⁺ T cells is the binding of ICER/CREM to the inducible Nfatc1 P1 promoter and its suppression in response to increased cAMP levels. This leads to low NFATc1 concentrations, a block in cellular proliferation and IL-2 synthesis and, thereby, to the suppression of CD4⁺ T cells.

Results

Stimulation of Conventional CD4⁺ T Cells with CD3/CD28 mAbs Leads to Cytosolic Localization of ICER/CREM.

Immunohistochemical staining of freshly isolated conventional CD4⁺ T cells and nTregs with Abs specific for ICER/CREM or NFATc1 revealed a predominant nuclear occurrence of ICER/CREM and cytosolic localization of NFATc1 in both cell types (Fig. 1 A and B, “fresh”). A similar cellular distribution was observed in paraffin-embedded sections of murine spleens; almost all follicular lymphocytes showed ICER/CREM in the nucleus and NFATc1 in the cytoplasm (Fig. S1A). Upon priming and/or restimulation with CD3/CD28 mAbs, however, we detected ICER/CREM in the cytoplasm of conventional CD4⁺ T cells, whereas ICER/CREM remained in the nucleus in Tregs (Fig. 1 A and B, “primed”; quantified in Fig. S1 B–G). This cytosolic localization of ICER/CREM in conventional CD4⁺ T cells was partially impaired in T cells from CD28-deficient mice (Fig. S2), suggesting a role for costimulatory CD28 signals in the cytosolic localization of ICER/CREM. In the absence of such signals, upon stimulation with phorbol ester plus ionomycin, ectopically expressed ICER could not be detected in the cytosol of HEK 293 T cells (Fig. S3A). Similar to the function of Foxp3, the nuclear localization of ICER led to inhibition of Il2 promoter linked to a luciferase reporter gene (Fig. S3B). When conventional CD4⁺ T cells restimulated by CD3/CD28 mAbs were also treated with forskolin (which elevates intracellular levels of cAMP through activation of adenylyl cyclase; Fig. 1C), ICER/CREM was detected in the nucleus. A similar effect was observed on blocking of cAMP degradation by 3-isobuthyl-1-methyl xanthine (IBMX), an inhibitor of PDEs (Fig. 1A). In addition, direct contacts between nTregs and conventional CD4⁺ T cells in standard Treg assays led to the increased presence of ICER/CREM in the nucleus of conventional CD4⁺ T cells (Fig. S4).

Fig. 1.

CD3/CD28 mAb stimulation directs ICER/CREM to the cytoplasm in conventional CD4⁺ T cells but not in nTregs, whereas elevated cAMP levels direct ICER/CREM to the nucleus. Conventional CD4⁺ T cells (A, fresh) or nTregs (B, fresh) primed and expanded by CD3/CD28 mAbs (primed) were restimulated either alone (CD3/CD28) or in the presence of forskolin (CD3/CD28 + Forsk) or IBMX (CD3/CD28 + IBMX) (SI Materials and Methods), leading to elevated intracellular cAMP levels (C). Treated cells were analyzed for ICER/CREM and NFATc1 localization by confocal microscopy. Neither ICER nor NFATc1 shuttle in nTregs but in conventional CD4⁺ T cells. (C) Intracellular cAMP levels were assessed in conventional CD4⁺ T cells and nTregs using ELISA. (D) Quantification of endogenous IL-2 mRNA levels in conventional CD4⁺ T cells treated as indicated by quantitative real-time PCR. Data are shown as mean ± SEM (n = 3–4; P < 0.05).

The nuclear localization of ICER/CREM in conventional CD4⁺ T cells corresponded to a marked suppression of endogenous IL-2 mRNA synthesis on forskolin (or IBMX) treatment of conventional CD4⁺ T cells restimulated with CD3/CD28 mAbs (Fig. 1D). Under both conditions, NFATc1 was found in the nuclei of conventional CD4⁺ T cells (Fig. 1A), whereas in nTreg cells NFATc1 was predominantly found in the cytosol and remained unaffected by any of the treatments mentioned previously (Fig. 1B).

Direct Contacts with nTregs Induce cAMP-Dependent Nuclear Localization of ICER/CREM in Conventional CD4⁺ T Cells in Vivo.

nTregs exhibit inherently high levels of intracellular cAMP (Fig. 1C) and are capable of conferring cAMP into conventional CD4⁺ T cells in a contact-dependent fashion (3). Therefore, we asked whether direct contacts between nTregs and conventional CD4⁺ T cells would induce the nuclear localization of ICER/CREM on activation in vivo. To test this, we used the vital dye calcein that spreads from donor to recipient cells via gap junctions (17). RAG-2–deficient CD90.1⁺ OT-II mice, which bear CD4⁺ T cells expressing a transgenic ovalbumin_323–339 (OVA_323–339)-specific T-cell receptor (TCR) but lack nTregs, were immunized with OVA_323–339 peptide in complete Freund's adjuvant by injection into the left hind footpad. Six hours later, calcein-loaded congenic OVA_323–339-specific nTregs expressing CD90.2 were adoptively transferred. After 24 h, CD4⁺ T cells from draining and nondraining lymph nodes were isolated and sorted by FACS (Fig. 2). A substantial portion of the CD4⁺CD90.1⁺ T cells from draining lymph nodes of immunized mice exhibited high calcein staining (Fig. 2B). In contrast, CD4⁺CD90.1⁺ T cells from the nondraining lymph nodes prepared from the same mouse did not exhibit any or low calcein staining (Fig. 2A). After sorting, intracellular cAMP concentration was assessed in calcein-low and calcein-high cells using a cAMP-specific ELISA (Fig. 2C). Total RNA was extracted for determination of IL-2 mRNA levels (Fig. 2D), and the cells were stained for the intracellular localization of ICER/CREM and NFATc1 (Fig. 2E).

Fig. 2.

Direct contacts between nTregs and conventional CD4⁺ T cells in vivo induce the nuclear localization of ICER/CREM and inhibit IL-2 expression. (A and B) Detection of interactions between nTregs and conventional CD4⁺ T cells. CD90.1⁺ RAG-2^−/− OTII mice that contain CD4⁺ T cells, which lack nTregs, were immunized s.c. with OVA_323–339 peptide in the left hind footpad. Six hours later, calcein-loaded congenic OVA_323–339-specific nTregs expressing CD90.2 (B) were injected i.v. At 24 h after immunization, the intracellular levels of calcein and CD90.1 expression of CD4⁺ cells from nondraining (A) and draining (B) lymph nodes were analyzed. (C) Cells were sorted by gating on propidium iodide-negative and CD4⁺ lymph node cells, and their intracellular cAMP concentration was assessed using ELISA in calcein-low and calcein-high cells. (D) Total RNA was extracted from sorted calcein-low and calcein-high cells, respectively, and IL-2 mRNA levels were quantified by real-time PCR. Data are shown as mean ± SEM (n = 3; P < 0.05). (E) Immunofluorescence of calcein-low and calcein-high cells was performed using Abs specific for ICER/CREM and NFATc1 and analyzed by confocal microscopy.

By confocal microscopy, a predominant nuclear staining of ICER/CREM was observed in calcein-high conventional CD4⁺ T cells on interaction with calcein-loaded nTregs, whereas ICER/CREM was observed in the cytosol of T cells that did not contact nTregs (calcein-low cells, Fig. 2E and Fig. S5). Moreover, NFATc1 was prevalently nuclear in draining lymph nodes and colocalized with ICER/CREM in calcein-high cells (Fig. 2E). The calcein-high conventional CD4⁺ T cells showed a significant increase in intracellular cAMP concentration (Fig. 2C), which correlated with the transcriptional attenuation of IL-2 expression (Fig. 2D). These findings indicate a close correlation between the nTreg-mediated increase in intracellular cAMP levels, the nuclear localization of ICER/CREM, and the suppression of IL-2 synthesis in conventional CD4⁺ T cells in vivo.

Ablation of nTregs Drives ICER/CREM into the Cytosol of Conventional CD4⁺ T Cells on CD28SA Administration.

To substantiate these findings, we investigated the cellular distribution of ICER/CREM in lymph nodes from DEREG mice. Those mice express the human DT receptor (as well as GFP) under the control of the Foxp3 locus, and therefore allow the specific ablation of nTregs by DT (18). DT-treated DEREG mice retained only 0.3% of nTregs compared with control mice, which bear approximately 8% of nTregs in their CD4⁺ T-cell compartment (Fig. 3A and Fig. S6F). CD28SA treatment led to a moderate expansion and activation of nTreg cells expressing GFP (Foxp3⁺ CD25⁺ or CTLA4⁺) (16) and, in particular, to up-regulation of CD69 in conventional CD4⁺ T cells (Fig. 3A). Conventional CD4⁺ T cells isolated from lymph nodes of CD28SA-treated DEREG mice revealed the cytosolic localization of ICER/CREM in the absence of nTregs (−nTreg) but nuclear localization of ICER/CREM in their presence (+nTreg) (Fig. 3 B and D). ICER/CREM was mostly cytosolic in CD28SA-treated DEREG mice pretreated with DT (quantified in Fig. S6 A–C). Serum samples obtained from those nTreg-depleted and CD28SA-treated mice after 1.5, 3, and 4 h revealed a significant increase in circulating IL-2 and TNF-α levels (Fig. 3C) accompanied by moderate changes in spleen morphology (Fig. S6 D and E). These data indicate a strong correlation between the cytosolic localization of ICER/CREM in CD4⁺ conventional T cells and cytokine release on activation and depletion of nTregs in DT + CD28SA-treated DEREG mice, implicating ICER/CREM in nTreg-mediated suppression.

Fig. 3.

nTregs direct ICER/CREM to the nucleus on administration of CD28SA in vivo. (A) DEREG mice were treated with DT for 5 d before administration of CD28SA. Lymph node cells were isolated; analyzed for the expression of CD25, GFP/Foxp3, CD69, and CTLA-4; and CD4⁺CD25⁻ cells were sorted 3 d later by FACS and harvested by cytospin. (B) Immunofluorescence was performed with Abs specific for ICER/CREM and NFATc1 and analyzed by confocal microscopy. Conventional CD4⁺ T cells from Treg-depleted and CD28SA-treated DEREG mice (DT + CD28SA) revealed cytoplasmic localization of ICER/CREM (−nTreg). In contrast, ICER/CREM was nuclear in CD4⁺ T cells isolated from lymph nodes of DEREG mice after CD28SA administration (CD28SA) in the presence of nTregs (+nTreg). (C) nTreg depletion results in systemic cytokine release after CD28SA stimulation. Mice were stimulated with CD28SA as indicated. Cytokine measurements of blood serum using cytometric bead array technology are shown for IL-2 and TNF-α. Data are shown as mean ± SEM (n = 3–4; P < 0.05). (D) Representative single-cell analyses showing cytosolic localization of ICER/CREM (blue line) in CD4⁺ T cells after CD28SA and DT treatment (−nTreg). In contrast, ICER/CREM is nuclear on CD28SA administration without prior DT treatment (+nTreg). The nucleus is defined by DAPI staining (cyan blue line), and the cytosolic distribution of NFATc1 is indicated in red.

nTreg–B-Cell Interactions Lead to the Nuclear Presence of ICER/CREM in B Cells.

In the spleen of untreated mice, Foxp3⁺ nTregs are in close proximity to B cells (Fig. S7A). When activated B cells were coincubated with calcein-loaded nTreg cells in vitro, they exhibited increased intracellular levels of calcein (Fig. S7 B–I). This indicates that nTreg cells could confer cAMP into B cells in a similar fashion as into conventional CD4⁺ T cells. B cells also induce ICER mRNA in a cAMP-dependent fashion on forskolin treatment (Fig. S7C). When we examined the subcellular localization of ICER/CREM upon B-cell receptor triggering by anti-IgM mAb (Fig. S7D), ICER/CREM was predominantly cytosolic. As in conventional CD4⁺ T cells, forskolin or IBMX cotreatment of anti-IgM–activated B cells directed ICER/CREM to the nucleus (Fig. S7 E–G), suggesting that contact-dependent transfer of cAMP from nTregs to anti-IgM and/or LPS-activated B cells leads to nuclear localization of ICER/CREM (Fig. S7 H and I), which could modulate B-cell activity. To assess this notion further, in a standard Treg assay, we substituted T cell-depleted splenocytes with purified anti-IgM–activated B cells whose B7 expression was monitored by FACS (Fig. S8). Under these conditions, the proliferation of purified B cells was impaired and the expression of CD80 (B7.1), and to lesser degree CD86 (B7.2), was reduced (Fig. S8). These data suggest that in analogy to conventional CD4⁺ T cells, nTregs communicate with B cells, which leads to nuclear localization of ICER/CREM.

ICER/CREM Binds to the Nfatc1 P1 Promoter and Represses TCR-Mediated NFATc1/αA Induction.

The induction of the Nfatc1 gene in effector lymphocytes is mediated by promoter P1, whose activity is strongly enhanced by immunoreceptor signals. P1 harbors two highly conserved cAMP-responsive elements (CREs), which are denoted as CRE-145 and CRE-640 (or proximal and distal CREs) (13) (Fig. 4A). In conventional CD4⁺ T cells stimulated by CD3/CD28 mAbs, forskolin treatment or cocultivation with nTregs attenuated the induction of the P1 promoter, which directs the generation of NFATc1/α isoforms, particularly the short isoform NFATc1/αA (Fig. 4B). Under the same conditions, no effect was detected on the constitutively active P2 promoter. When we examined ICER binding to the CRE-145 and CRE-640 motifs in EMSAs, the transcriptional attenuation of NFATc1/αA correlated with the binding of recombinant ICER protein to the CRE-145 and, to a lower degree, to the CRE-640 (Fig. 4C). As previously reported for the CD28 responsive element (CD28RE) of the Il2 promoter (8), the CRE-145 motif of P1 is associated with an NFAT binding site. Both the CRE-145 and CD28RE motifs can form inhibitory NFAT/ICER complexes in vitro (Fig. 4C and Fig. S9A). Moreover, NFATc1/αA and ICER interacted via protein–protein interactions (Fig. S9B). After forskolin and CD3/CD28 restimulation of conventional CD4⁺ T cells (prepared in the same way as cells used for confocal imaging in Fig. 1A), we detected a marked increase in ICER/CREM binding to the P1 promoter in ChIP assays (Fig. 4D, lane 4). Only weak ICER/CREM binding was detected after CD3/CD28 restimulation alone (Fig. 4D, lane 3). Cotransfections of a luciferase reporter gene directed by P1 with increasing concentrations of a vector expressing ICER led to the suppression of P1 induction, whereas a leucine-zipper mutant of ICER (ICER-LZ) failed to suppress P1 (Fig. 4E). This shows that dimerization and binding of ICER to DNA play a critical role in the transcriptional attenuation of the P1 promoter. Taken together, these data indicate that nTregs direct the nuclear accumulation of ICER/CREM via elevating cAMP levels, and thereby attenuate the TCR-mediated induction of NFATc1 in CD4⁺ effector T cells.

Fig. 4.

ICER/CREM binds to the Nfatc1 P1 promoter and represses TCR-mediated NFATc1/αA induction. (A) Scheme of the Nfatc1 gene and its P1 promoter. Underlined are the CRE motifs at positions −145 and −640, respectively. (B) (Upper) Evaluation of NFATc1 mRNA expression initiated at the P1 and P2 promoters in unstimulated and CD3 or CD3/CD28 Ab-stimulated conventional CD4⁺ T cells treated without or with forskolin. (Lower) Conventional CD4⁺ T cells stimulated with CD3/CD28 Abs were cultivated alone or in the presence of nTregs. RT-PCR assays are shown detecting transcripts directed by the P1 (NFATc1/α) and P2 (NFATc1/β) promoters. (C) (Left) Recombinant ICERII protein binds specifically to the CRE-145 and CRE-640 motifs of the NFATc1 P1 promoter in EMSAs, as shown in supershift experiments (sICER). (Right) Generation of inhibitory NFAT/ICER (NF/IC) complexes with recombinant proteins at the CREs of P1 promoter is shown (IL-2 promoter is shown in Fig. S9). (D) ChIP assays demonstrate increased binding of ICER/CREM to the NFATc1 P1 promoter on forskolin and CD3/28 mAb restimulation of conventional CD4⁺ T cells treated in the same fashion as cells used for confocal analysis in Fig. 1A. Cross-linked chromatin was immunoprecipitated with an ICER/CREM-specific Ab (CS4). As a control, IgG of normal rabbit serum (NRS) was used in parallel. (E) ICER represses induction of the P1 promoter of NFATc1. A luciferase construct driven by the P1 promoter (NFATc1 P1 Luc) was cotransfected with increasing concentrations of a vector expressing ICER in EL-4 T cells. Cells were stimulated by forskolin and ionomycin (Forsk + Iono) or by phorbol-ester and ionomycin (PMA + Iono). In parallel, a vector encoding a leucin zipper-deficient ICER mutant (ICER-LZ) was transfected. Error bars show SD values of at least three experiments.

Discussion

In this study, we showed that an increase in intracellular cAMP levels in conventional CD4⁺ T cells leads to nuclear localization of ICER/CREM and the repression of NFATc1 and IL-2 induction. This is attributable to direct cell–cell contacts between nTregs and CD4⁺ T cells in vivo. nTregs harbor constitutively high levels of cAMP and are able to confer cAMP through gap junctions into conventional CD4⁺ T cells (3). The increase in cellular cAMP levels enhances the expression and nuclear localization of ICER (19). This results in the binding of ICER to the inducible P1 promoter of the Nfatc1 gene and the repression of NFATc1/αA induction. ICER also binds to the Il2 and other NFAT-driven cytokine promoters and interferes with their activity on T-cell activation.

In antigen-driven immune responses, effector CD4⁺ T cells use IL-2 as an autocrine growth factor and provide IL-2 to nTregs in a paracrine fashion, thereby increasing the number and suppressive activity of nTregs (20). One explanation for the absence of systemic cytokine release in CD28SA-treated mice is the suppression of IL-2 production in CD4⁺ effector T cells by ICER/CREM. We tested this view by depleting nTregs from DEREG mice. The data from these experiments confirmed a direct role for cAMP in nTreg-mediated maintenance of ICER/CREM in the nuclei of CD4⁺ T cells. They correspond with the findings of adoptive transfer experiments using calcein-loaded nTregs. In draining lymph nodes of immunized mice that received calcein-loaded nTregs, a substantial portion of CD4⁺ T cells showed high calcein staining and increased cAMP levels on nTreg transfer. Confocal microscopy revealed nuclear staining of ICER/CREM in calcein-high cells (Fig. 2). This is in striking contrast to the cytosolic localization of ICER/CREM in calcein-low cells, indicating a critical role for enhanced cAMP levels in the expression, localization, and function of ICER during nTreg-mediated suppression.

How nTregs dampen CD4⁺ effector T-cell responses is still a matter of dispute (1). In nTregs, the activity of the Nfatc1 P1 promoter, and therefore the induction of the inducible short NFATc1/αA isoform, is significantly reduced (14). We showed that nTregs could suppress CD4⁺ effector T cells by conferring cAMP via intercellular gap junctions (3), which appears to be instrumental for the induction and nuclear accumulation of ICER. By the binding of ICER alone or together with NFATs to the inducible P1 promoter, ICER interferes with the strong induction of NFATc1/αA, a molecular marker of activation of CD4⁺ effector T cells and splenic B cells (13). Increased nuclear levels of ICER also have a major impact on the transcriptional silencing of the Il2 promoter and other NFAT-driven promoters, leading to impaired CD4⁺ T-cell proliferation and function (21, 22). The experimental findings presented here provide a unique view on molecular mechanism(s) of how nTregs control the activity of CD4⁺ effector T cells (and splenic B cells) in a cAMP-dependent fashion.

Materials and Methods

Mice.

BALB/c, C57BL/6 (CD28^−/−), OT-II mice (DO11.10, CD90.1), and Rag2^−/− mice were obtained from Charles River Laboratories. DEREG mice have been published previously (18).

Preparation of T- and B-Cell Subsets.

CD4⁺ T cells were isolated using a Dynal mouse T-cell negative isolation protocol (Invitrogen). CD25⁺ or CD25⁻CD4⁺ and CD19⁺ B cells were enriched using the MACS system (Miltenyi Biotech).

FACS Staining.

Staining of surface molecules was performed on ice using direct fluorochrome-conjugated mAbs or indirectly by biotin-labeled mAbs and allophycocyanin (APC)-conjugated streptavidin. Samples were acquired on a FACSCalibur (BD Biosciences) or sorted on a FACSAria flow cytometer (BD Biosciences) and analyzed using FlowJo software (Treestar).

nTreg Depletion in DEREG Mice and Treatment with CD28SA.

DEREG mice were treated daily with i.p. injections of 1 μg of DT for 5 d consecutively before administration of 100 μg of CD28SA mAb D665.

Cytokine Measurements Using Cytometric Bead Array Technology.

For detecting cytokines in the serum of DEREG mice, cytometric bead array technology from BD Pharmingen was used.

Cell Culture and Stimulations.

HEK 293 T cells were cultured and stimulated as described previously (14). Priming and restimulation of CD25⁻CD4⁺ T cells as well as CD25⁺ nTregs were performed using plate-bound CD3/CD28 mAbs. After 72 h, the cells were harvested, cultured for an additional 96 h (priming), and restimulated.

Transfection and Luciferase Assays.

HEK 293 T cells were transfected with DNA vectors encoding ICER, NFATc1, or FoxP3 as indicated. Luciferase activity was measured as described previously (14). For immunofluorescence, HEK 293 T cells were transfected with constructs encoding ICER, NFATc1, and FoxP3 using Superfect (Qiagen) according to the manufacturer's protocol.

Quantitative RT-PCR.

Total RNA was extracted using an RNAqueous Kit (Ambion), and first-strand cDNA was synthesized with SuperScript II reverse transcriptase (Invitrogen) and oligo (dT)12–18 primers (Life Technologies). RNA levels were quantified in real-time PCR assays using the ABI/PRISM 7700 system (Applied Biosystems, Inc.) with the primers described in SI Materials and Methods.

cAMP ELISA.

To assess cellular cAMP levels, 1.5 × 10⁶ T cells were snap-frozen in liquid nitrogen before lysis in 0.1 N of HCl. The cAMP-specific EIA Direct cAMP Enzyme Immunoassay (Assay Design) was subsequently performed according to the manufacturer's instructions.

Immunoprecipitations and Western Blot Analysis.

Immunoprecipitations of transfected HEK 293 T cells were performed as described previously (14). For immunodetection, primary Abs were used with peroxidase-coupled secondary Abs developed with an enhanced chemiluminescence system.

ChIP.

ChIP analysis of expanded conventional CD4⁺ T cells was carried out according to the manufacturer's instructions (Cell Signaling).

EMSA.

In the EMSA, binding reactions were performed in a 15-μL reaction as described previously (8) using recombinant proteins. Samples were incubated with ³²P-labeled oligonucleotides (SI Materials and Methods), followed by electrophoresis on a native polyacrylamide gel.

Immunofluorescence of Single Cells or Tissue Sections.

Staining of ICER/CREM, NFAT, and Foxp3 in primary T cells collected on slides or in HEK 293 T cells transfected transiently was performed as described previously (14) (SI Materials and Methods). Formalin-fixed paraffin-embedded tissue samples from spleens of DEREG mice were prepared as described by Akimzhanov et al. (23).

Calcein^AM Staining and in Vivo Calcein Transfer.

nTregs were incubated with 1 mM Calcein^AM (Invitrogen) and stimulated in coculture with CD19⁺ B cells or CD4⁺ T cells for 4 or 20 h, respectively. Calcein transfer was analyzed by FACS. For in vivo calcein transfer, DO11.10, Rag2^−/−, and CD90.2⁺ (or CD90.1⁺, respectively) mice were immunized by s.c. injection of 50 μg of OVA_323–339. After 6 h, 1 × 10⁷ preactivated calcein-loaded nTregs from DO11.10 and CD90.1⁺ (or CD90.2⁺, respectively) mice were injected i.v. into the same mice. At 24 h after immunization, lymph node cells were stained for the expression of CD4 and CD90.1. Calcein-high and calcein-low CD4⁺ T cells were isolated using a FACSAria cell sorter.

Proliferation Assays.

Along with APCs (or purified B cells as indicated) T cells (2.5 × 10⁴ per well in U-bottomed 96-well plates) were cultured for 72 h in the presence of CD3 mAb and evaluated as described in SI Materials and Methods.

Statistical Analysis.

More than 100 cells from at least three independent experiments were counted, and mean fluorescence intensity was calculated using Leica TCS SP5 confocal laser scanning microscopy software. Groups were compared with Prism software (GraphPad) using an unpaired or paired Student's t test. More detailed information is provided in SI Materials and Methods.