Research ArticleImmunology

Human Regulatory T Cells Rapidly Suppress T Cell Receptor–Induced Ca2+, NF-κB, and NFAT Signaling in Conventional T Cells

See allHide authors and affiliations

Science Signaling  20 Dec 2011:
Vol. 4, Issue 204, pp. ra90
DOI: 10.1126/scisignal.2002179

Abstract

CD4+CD25hiFoxp3+ regulatory T cells (Tregs) are critical mediators of self-tolerance, which is crucial for the prevention of autoimmune disease, but Tregs can also inhibit antitumor immunity. Tregs inhibit the proliferation of CD4+CD25 conventional T cells (Tcons), as well as the ability of these cells to produce effector cytokines; however, the molecular mechanism of suppression remains unclear. Here, we showed that human Tregs rapidly suppressed the release of calcium ions (Ca2+) from intracellular stores in response to T cell receptor (TCR) activation in Tcons. The inhibition of Ca2+ signaling resulted in decreased dephosphorylation, and thus decreased activation, of the transcription factor nuclear factor of activated T cells 1 (NFAT1) and reduced the activation of nuclear factor κB (NF-κB). In contrast, Ca2+-independent events in Tcons, such as TCR-proximal signaling and activation of the transcription factor activator protein 1 (AP-1), were not affected during coculture with Tregs. Despite suppressing intracellular Ca2+ mobilization, coculture with Tregs did not block the generation of inositol 1,4,5-trisphosphate in TCR-stimulated Tcons. The Treg-induced suppression of the activity of NFAT and NF-κB and of the expression of the gene encoding the cytokine interleukin-2 was reversed in Tcons by increasing the concentration of intracellular Ca2+. Our results elucidate a previously unrecognized and rapid mechanism of Treg-mediated suppression. This increased understanding of Treg function may be exploited to generate possible therapies for the treatment of autoimmune diseases and cancer.

Introduction

CD4+ regulatory T cells (Tregs), which are characterized by high abundance of the surface marker CD25 and by the presence of the transcription factor Foxp3, play a pivotal role in the immune system and are involved in the prevention of autoimmune diseases and allergies (1). In addition to the beneficial functions of Tregs in graft-versus-host disease (GVHD) and in the prevention of organ pathology after infection, Tregs also lead to the unwanted dampening of immune responses, for example, in cancer (2). Tregs directly inhibit the proliferation of responder CD4+CD25 conventional T cells (Tcons) and the production of effector cytokines by these cells (3). Some mechanisms of Treg-mediated suppression have been described thus far (4, 5). For example, adenosine 3′,5′-monophosphate (cAMP) inhibits the production of the cytokine interleukin-2 (IL-2) by murine Tcons. This study implied that cAMP was transferred from Tregs to Tcons to induce expression of the gene encoding the transcriptional repressor inducible cAMP early repressor (ICER), which in turn represses expression of the Il2 gene encoding the cytokine IL-2 in Tcons (6). Various factors, such as location, inflammatory milieu, antigen load, cell number, and activation status of the responder cells, influence the mechanism of Treg-mediated immune regulation in vivo (4). Furthermore, species differences regarding Treg function have been found. For example, the immunosuppressive cytokine IL-35 is produced by murine Tregs but is not constitutively produced by human Tregs (7, 8). Therefore, we sought to clarify the inhibitory mechanism(s) in vitro and investigated the direct suppression of primary human Tcons by Tregs.

We hypothesized that Tregs might disrupt a key component of T cell receptor (TCR) signaling and that such disruption could perturb the function of Tcons. Thus far, the influence of Tregs on signaling events in suppressed Tcons has not been investigated extensively. Stimulation of the TCR initially leads to the activation of the Src family kinase leukocyte-specific protein tyrosine kinase (Lck), which phosphorylates tyrosine residues in CD3, ζ chain–associated protein kinase of 70 kD (ZAP-70), and other substrates. This initiates signaling cascades that lead to the activation of transcription factors crucial for the expression of genes involved in T cell activation and proliferation. Central transcription factors that induce the expression of genes encoding cytokines are nuclear factor κB (NF-κB), nuclear factor of activated T cells (NFAT), and activator protein 1 (AP-1) (9). The AP-1 pathway depends on the activation of mitogen-activated protein kinases (MAPKs), such as extracellular signal–regulated kinase (ERK), c-Jun N-terminal kinase (JNK), and p38, which promote the synthesis, phosphorylation, and activation of the Fos and Jun proteins that together comprise the AP-1 transcription factor (10). Activation of both NF-κB and NFAT requires the activity of phospholipase C–γ1 (PLC-γ1), which generates the second messengers diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP3). DAG leads to activation of protein kinase C θ (PKCθ), which, in turn, activates the inhibitor of κB (IκB) kinase (IKK) complex, resulting in the phosphorylation and degradation of IκBα and the translocation of the NF-κB p50:p65 heterodimer to the nucleus (11). IP3 induces an increase in the concentration of cytoplasmic calcium (Ca2+) and activation of the Ca2+-dependent phosphatase calcineurin, which results in the rapid dephosphorylation and activation of NFAT, which is followed by its translocation to the nucleus (12). Ca2+ is critically involved not only in the activation of NFAT but also in the activation of NF-κB, because Ca2+ channel blockers, as well as other agents that prevent Ca2+ influx, inhibit the activation of NF-κB (13, 14). Indeed, Ca2+ may lead to the activation of NF-κB signaling by inducing the degradation of IκBα (15, 16). A previous study showed that murine Tcons deficient in both NFAT1 and NFAT4 are less susceptible to Treg-mediated suppression than are their wild-type counterparts (17); however, TCR signaling and the subsequent activation of the transcription factors NFAT, NF-κB, and AP-1 in suppressed Tcons have not been analyzed so far.

We previously showed that suppression of the expression of cytokine-encoding genes by human Tregs is a rapid process (18). Here, we describe that this suppressive effect was persistent in Tcons even after the removal of Tregs from the coculture. To analyze this rapid suppression in more detail, we compared TCR signaling in suppressed and nonsuppressed Tcons by focusing on the signaling events immediately after TCR stimulation. We found that TCR-proximal signals as well as activation of PLC-γ1 were not altered in suppressed Tcons, whereas Ca2+, NF-κB, and NFAT signaling pathways were strongly inhibited. An increase in intracellular Ca2+ concentrations in Tcons abrogated the Treg-mediated suppression of NFAT and NF-κB activation as well as of the expression of IL2. These results suggest that suppression of Ca2+ signals appears to be a crucial event in the Treg-mediated suppression of Tcons.

Results

Suppression of Tcons is sustained upon removal of Tregs

To study Treg-mediated suppression of Tcons in detail, we analyzed primary human responder Tcons from cocultures with Tregs (suppressed Tcons) and, as a control, responder Tcons from cocultures with Tcons (control Tcons) (Fig. 1A). The frequent human leukocyte antigen (HLA) serotype HLA-A2, which belongs to the major histocompatibility complex (MHC) class I proteins, was used as a cell surface marker to differentiate between the two populations of cells in the coculture. By the use of HLA-A2–disparate cells from HLA-A2–disparate donors in the coculture, we were able to reisolate pure populations of responder Tcons after coculture and, thus, to compare signaling in suppressed Tcons and control Tcons. All experiments were set up as described (Fig. 1A) with pre-activated Tregs, unless stated otherwise, and were controlled by determining the extent of inhibition of expression of IL2 and IFNg [which encodes the proinflammatory cytokine interferon-γ (IFN-γ)] through measurement of the amounts of their mRNAs in suppressed Tcons and comparing them to those of control Tcons. As shown previously (18), we detected the inhibition of IL2 and IFNg expression within 3 hours of TCR stimulation in suppressed Tcons (Fig. 1A). Within that time frame, either pre-activated or resting Tregs suppress cytokine mRNA production to a similar extent (18).

Fig. 1

Suppression of Tcons persists after the removal of Tregs. Quantitative reverse transcription–polymerase chain reaction (qRT-PCR) analysis of the abundances of IL2 and IFNg mRNAs in HLA-A2 Tcons. (A) HLA-A2 Tcons were cocultured with pre-activated HLA-A2+ Tregs (supTcon) or with HLA-A2+ Tcons as a control (cTcon) at a 1:1 ratio and stimulated with cross-linked antibodies against CD3 and CD28 for 3 hours. HLA-A2 Tcons were then isolated and the abundances of IL2 and IFNg mRNAs were analyzed. Results are presented as the fold change in the abundance of the indicated mRNA in stimulated suppressed Tcons or stimulated control Tcons compared to that in unstimulated Tcons, which was set to 1. Data are means ± SD of duplicate samples of a single donor and are representative of >20 donors. The percentages of reduction in the abundances of IL2 and IFNg mRNAs in the suppressed Tcons compared to those in control Tcons are indicated in gray (% Suppression). (B) Cells were cocultured as described for (A) (coculture stimulation, left bars) or were cocultured for 10 to 60 min before undergoing stimulation of isolated responder Tcons (“pre-cocultured”). For the pre-cocultures, HLA-A2 Tcons were left unstimulated in the presence of pre-activated Tregs or Tcons, as a control, at a 1:1 ratio. HLA-A2 Tcons were then isolated and stimulated alone as described in (A), and the abundances of mRNAs for the indicated cytokines were measured. Shown are the percentages of suppression as described for (A). n.s., not significant. (C) Cocultures were established as described in (B), with the exception that after a pre-coculture period of 60 min, HLA-A2 Tcons were isolated and rested for 24 hours before they were stimulated. Data are the means ± SEM from four (B) or three (C) donors. Statistical significance of suppression was calculated with the Student’s one-sample t test.

We were interested in how rapidly Tregs could exert suppression and whether Tcons remained suppressed after removal of the Tregs. Therefore, we used another coculture system (Fig. 1B), in which we cocultured Tcons with pre-activated Tregs (or with Tcons as control) for 10, 30, 45, and 60 min without TCR stimulation and tested whether this short coculture period was sufficient to suppress the Tcons. After the coculture period, Tcons were isolated on the basis of their distinct HLA-A2 and were stimulated separately, without Tregs, for 3 hours through the TCR to induce the expression of cytokine-encoding genes. We observed that IL2 and IFNg expression in suppressed Tcons was inhibited compared to that in control Tcons after a coculture period of 30 to 45 min (Fig. 1B). Suppression of cytokine secretion by Tcons was retained upon removal of the Tregs, whereas inhibition of proliferation was not sustained (fig. S1). Even when the TCR stimulus was provided 24 hours after a short coculture period, the abundances of IL2 and IFNg mRNAs were still reduced in suppressed Tcons compared to those in control Tcons (Fig. 1C). These results implied that Tregs altered the TCR signaling machinery of Tcons within 30 to 45 min of cell contact, which subsequently led to long-lasting suppression of cytokine gene expression in Tcons even in the absence of Tregs.

To analyze the suppression of cytokine gene expression in a more physiological setting, we performed cocultures in the presence of antigen-presenting cells (APCs). Costimulation with APCs and polyclonally stimulating antibody against CD3 (which mimics antigenic TCR stimulation) resulted in the production of lower amounts of cytokines than occurred in cells stimulated through cross-linked antibodies against CD3 and against the costimulatory receptor CD28; however, the extent of suppression of IL2 mRNA generation was similar (fig. S2). Thus, the rapid Treg-mediated suppression of cytokine-encoding gene expression in Tcons also occurred in the presence of APCs, which may be similar to the in vivo situation.

CTLA-4 is not involved in the rapid suppression of cytokine gene expression in Tcons

The presence of the inhibitory receptor cytotoxic T lymphocyte antigen 4 (CTLA-4) on the surface of murine Tregs contributes to the suppression of Tcons by decreasing the abundance of the costimulatory B7 molecules on APCs, thereby limiting the stimulatory capacity of the APCs (1921). Mice with Tregs deficient in CTLA-4 suffer from autoimmune disease (19), suggesting that CTLA-4 is a key component of Treg-mediated suppression. Because Tregs can suppress Tcons directly, we asked whether CTLA-4 was also important for Tcon suppression in the absence of APCs and in the context of human cells, because CTLA-4 on Tregs may bind to B7 molecules that may be expressed on the surface of Tcons. First, we analyzed the kinetics of B7 expression on Tcons because it was reported that stimulation of the TCR for more than 24 hours can lead to the appearance of CD80 (B7.1) and CD86 (B7.2) on the surface of T cells (22). However, we could not detect CD80 before 24 hours of TCR stimulation, and we could not detect CD86 at any time point (fig. S3A). Thus, the involvement of CTLA-4 in the direct suppression of cytokine gene expression in Tcons through these ligands was unlikely.

We could also exclude a role for unknown CTLA-4 receptors (23) in the rapid suppression of cytokine gene expression because we detected no difference in the suppression of IL2 and IFNg expression (as determined by measurement of mRNA abundance) in cocultures treated with a blocking antibody against CTLA-4 compared to that in cocultures incubated with isotype control antibodies (fig. S3B). In addition, in the presence of APCs, blocking CTLA-4 did not affect the rapid suppression of cytokine gene expression (fig. S3B). However, in proliferation assays containing APCs, suppression of proliferation was partly abrogated by blocking CTLA-4 (fig. S3C), as was shown similarly by others (21). Thus, we concluded that CTLA-4 was not involved in the Treg-mediated rapid suppression of cytokine-encoding gene expression in human Tcons in the presence or absence of APCs. Because the rapid suppression of cytokine gene expression occurred to similar extents in the presence and absence of APCs (figs. S2 and S3), this suggested that the suppression of cytokine gene expression occurred through an APC-independent mechanism.

TCR-proximal signaling is not altered in suppressed Tcons

The observed rapid suppression of cytokine gene expression prompted us to investigate whether Tregs inhibited the activation of Tcons at the level of the TCR. If this were true, it would result in the immediate blockage of signaling pathways that lead to TCR-induced expression of cytokine-encoding genes. To explore whether Treg-mediated suppression occurred at the TCR itself or at a TCR-proximal level, we set up cocultures as outlined earlier (Fig. 1A) and determined the extent of phosphorylation of CD3ε and ZAP-70. We found that the rapid tyrosine phosphorylation of CD3ε after TCR stimulation in control Tcons was not decreased in suppressed Tcons (fig. S4A). Similarly, we detected the rapid phosphorylation of ZAP-70 in both control Tcons and suppressed Tcons (fig. S4B). We then tested whether Treg-mediated suppression affected the activation of PLC-γ1 downstream of ZAP-70, which represents the branching point at which the NF-κB and NFAT signaling pathways diverge. Flow cytometric and Western blotting analyses revealed that the extent of PLC-γ1 phosphorylation was similar in control Tcons and suppressed Tcons 5 min after TCR stimulation (fig. S4C). To further analyze PLC-γ activity, we determined the amounts of DAG by lipid mass spectrometry, and we found no difference in the amounts of DAG between control Tcons and suppressed Tcons (fig. S4D). Our results showed that the TCR-proximal signaling machinery was not affected during the rapid Treg-mediated suppression of Tcons. Thus, we further investigated signaling events downstream of PLC-γ1.

NF-κB signaling is impaired in suppressed Tcons

To study the involvement of PKCθ in Treg-mediated suppression, we set up cocultures as described earlier (Fig. 1A) and performed Western blotting analysis with phosphospecific antibodies. We detected no difference between control Tcons and suppressed Tcons in terms of the phosphorylation of PKCθ 5 and 25 min after TCR stimulation (Fig. 2A). PKCθ activates the IKK complex, which consists of IKKα, IKKβ, and IKKγ (11). Despite the phosphorylation of PLC-γ1 and PKCθ being unaffected in suppressed Tcons, we found that IKK phosphorylation was markedly inhibited in suppressed Tcons compared to that in control Tcons (Fig. 2B), which also occurred in the presence of APCs (fig. S2). Because the IKK complex directly affects the phosphorylation and subsequent degradation of IκBα (24), we were interested in determining whether the phosphorylation of IκBα was also diminished in these cells. Indeed, we found that the phosphorylation of IκBα was strongly inhibited in suppressed Tcons compared to that in control Tcons within 5 min of TCR stimulation (Fig. 2C).

Fig. 2

NF-κB signaling is inhibited in suppressed Tcons. (A to C) Western blotting analysis of the phosphorylation of (A) PKCθ (pT538), (B) IKKα (pS180) and IKKβ (pS181), and (C) IκBα (pS32) in HLA-A2 Tcons. Cocultures were established as described for Fig. 1A, and cells were stimulated for the indicated times. Suppressed Tcons or control Tcons were analyzed with phosphospecific antibodies (upper images in each panel). Blots were then incubated with antibodies against the indicated total proteins and with antibody against tubulin, which served as a loading control. Blots are from one experiment representative of 4 (A), 3 (B), or >12 (C) donors. The right panel in (C) shows the mean percentage suppression of IκBα phosphorylation ± SEM from 19 (5 min) or 12 (25 min) donors. Statistical significance of suppression was calculated with the Student’s one-sample t test. (D) Western blotting analysis of the amount of phosphorylated p65 (p-p65). Suppressed Tcons (red) or control Tcons (black) were analyzed with fluorescently labeled antibodies against p-p65 (pS536) and β-actin. One representative experiment from five different donors is shown. (E) NF-κB target genes obtained from gene array data. The gene expression patterns of control and suppressed Tcons were compared after 3 hours of TCR stimulation. D signifies genes whose expression was significantly decreased in suppressed Tcons compared to that in control Tcons. The array was performed with two different donors, and one donor is shown. T, threonine; S, serine.

In T cells, the classical NF-κB pathway depends mainly on the NF-κB heterodimer p50:p65, which translocates to the nucleus upon degradation of IκBα. Similarly to IKK and IκBα, the phosphorylation of p65 was almost completely blocked in suppressed Tcons 30 min after TCR stimulation, and it remained blocked until 2.5 hours after stimulation (Fig. 2D). Furthermore, we performed gene array analysis and compared the mRNA expression profiles of suppressed Tcons and control Tcons 3 hours after TCR stimulation. The expression of several NF-κB target genes (25), including CCL22, ICAM1, IL2, IL8, TNF, and TNFAIP3, was inhibited specifically in suppressed Tcons (Fig. 2E), confirming our Western blotting analysis that NF-κB signaling was impaired in these cells. Because Tregs need to be activated through the TCR to be suppressive (26, 27), we pre-activated Tregs before their coculture with Tcons. However, to exclude any effects of the antibody against CD3 that was used for the pre-activation of the Tregs, we covalently coupled this antibody to plates that were used for the pre-activation of Tregs or, as a control, of Tcons. To minimize the production of the transcription factor Foxp3 in activated human Tcons, which may result in suppressive function (28, 29), we also shortened the time for the pre-activation of Tregs and Tcons. In addition, we used Tregs that had not been pre-activated in our assays of cytokine mRNA suppression, because stimulation in coculture for several hours might have been sufficient for the Tregs to acquire suppressive function. With these controls, we only observed the suppression of NF-κB target genes and of IκBα phosphorylation in responder Tcons when they were cocultured with Tregs (fig. S5A). Thus, Tregs specifically and rapidly blocked the NF-κB pathway in suppressed Tcons.

Dephosphorylation of NFAT1 is inhibited in suppressed Tcons

In T cells, transcription of many activation-associated genes is dependent not only on NF-κB but also on the binding of the cooperative NFAT-Fos-Jun complex to DNA response elements, which results in T cell proliferation and the expression of cytokine-encoding genes. Therefore, we hypothesized that the transcription factors NFAT and AP-1 were additional possible targets in the Treg-mediated suppression of Tcons. The NFAT family consists of five transcription factors, four of which (NFAT1 to NFAT4) are Ca2+-dependent (30). In resting T cells, NFAT1 is the predominant isoform (31). TCR stimulation induces the activation of calcineurin, which dephosphorylates NFAT and thereby unmasks its nuclear localization sequence. This enables NFAT to translocate to the nucleus and induce the expression of cytokine-encoding genes in synergy with AP-1 and NF-κB (32).

We established cocultures and investigated the activation status of NFAT1 by Western blotting analysis by comparing the extent of NFAT1 dephosphorylation in control and suppressed Tcons. We observed a marked inhibition of NFAT1 dephosphorylation in suppressed Tcons after 5 min of TCR stimulation (Fig. 3A). We could not detect NFAT2 in human T cells after short-term stimulation. In addition, gene array analyses revealed significant inhibition of the expression of NFAT-dependent genes (33), such as EGR2, IL2RA, CTLA4, IL3, IL4, and COX2, in suppressed Tcons compared to that in control Tcons (Fig. 3B). As described earlier for NF-κB, only Tregs pre-activated with covalently plate-bound antibody against CD3, but not Tcons pre-activated in the same manner, suppressed NFAT1 dephosphorylation and the expression of NFAT target genes in responder Tcons (fig. S5B). Together, these findings showed that Tregs potently and rapidly suppressed TCR-induced NFAT1 activation in Tcons.

Fig. 3

NFAT1 dephosphorylation is inhibited in suppressed Tcons. (A) Western blotting analysis of NFAT1 dephosphorylation in HLA-A2 Tcons. Cocultures were established as described for Fig. 1A and cells were stimulated for the indicated times. Suppressed Tcons and control Tcons were then analyzed for the presence of phosphorylated and total NFAT1 proteins. The upper bands represent phosphorylated NFAT1 (p-NFAT1), whereas the lower bands represent dephosphorylated NFAT1. Blots were then incubated with antibody against tubulin as a loading control. Data are from one donor representative of 11 (5 min) or 10 (25 min) donors. Lower graph shows the percentage suppression of NFAT1 dephosphorylation as determined by densitometric analysis of Western blots as shown in the upper panels. Data are the means ± SEM of 11 (5 min) or 10 (25 min) donors. Statistical significance of suppression was determined by Student’s one-sample t test. (B) NFAT target genes obtained from gene array analysis as described for Fig. 2E. D, genes whose expression in suppressed Tcons was significantly decreased compared to that in control Tcons; MD, marginal decrease.

AP-1 signaling is not involved in Treg-mediated suppression of cytokine gene expression

Because Treg-mediated suppression of NF-κB and NFAT signaling pathways occurred downstream of TCR-proximal signaling events, the question remained whether Tregs acted through a general blockage of all transcription factors relevant for T cell activation, including AP-1. AP-1 is controlled mainly by phosphorylation of ERK, p38, and JNK and the subsequent activation of Fos and Jun (34). We found that ERK was activated similarly in control Tcons and suppressed Tcons after short-term TCR stimulation as assessed by Luminex analysis (Fig. 4A). Correspondingly, we observed no suppression of the phosphorylation of ERK in suppressed Tcons compared to that in control Tcons, as assessed by flow cytometry (fig. S6A). We monitored the activation status of p38 by Western blotting analysis with phosphospecific antibodies. We did not observe suppression of p38 phosphorylation in suppressed Tcons (Fig. 4B), which was confirmed by flow cytometric analysis (fig. S6B).

Fig. 4

AP-1 signaling is not involved in Treg-mediated suppression of Tcons. (A) Luminex analysis of the phosphorylation of ERK1 and ERK2 (ERK1/2) in HLA-A2 Tcons after TCR activation. Cocultures were established as described for Fig. 1A, and cells were stimulated for the indicated times. Suppressed Tcons and control Tcons were analyzed for the extent of phosphorylation of ERK1/2 (pT185/pY187). Values were normalized to account for the amounts of total protein. Data are expressed as the fold change in the abundances of the indicated phosphorylated protein compared to those in unstimulated control Tcons, which were set at 1. Analyses were performed in duplicate, and mean fold change ± SD is shown. Data are from one donor representative of three different donors. (B) Western blotting analysis of the presence of p-p38 (pT180/pY182) in suppressed Tcons and control Tcons. Blots were analyzed with antibody against p-p38 and then incubated with antibodies against total p38 and tubulin. Data are from one experiment that is representative of five experiments. (C) Analysis of the phosphorylation of c-Jun (pS63) as described for (A). Data are from one donor and are representative of five donors. (D) Analysis of the expression of AP-1 target genes obtained from gene array assays as described for Fig. 2E. NC, genes whose expression in suppressed Tcons was not significantly different from that in control Tcons; Y, tyrosine.

To test events further downstream in the AP-1 pathway, we investigated the phosphorylation of the transcription factor c-Jun, but we did not observe any inhibition of the phosphorylation of c-Jun in suppressed Tcons compared to that in control Tcons up to 60 min after TCR stimulation (Fig. 4C). These data are supported by analysis of the expression of representative AP-1 target genes (35), including MMP9, MMP13, EGFR, FGF7, VEGF, CD69, and FYN, which were not significantly changed in suppressed Tcons compared to that in control Tcons (Fig. 4D). Thus, in contrast to their potent inhibition of NF-κB and NFAT signaling pathways in Tcons, Tregs did not influence the early activation of AP-1 in Tcons.

Ca2+ signaling is abrogated in suppressed Tcons

Whereas NF-κB and NFAT signaling pathways rely on Ca2+ signals, AP-1 activation is largely Ca2+-independent. The Ca2+- and calmodulin-dependent phosphatase calcineurin dephosphorylates numerous sites in NFAT, which enables the nuclear translocation of NFAT and its subsequent binding to DNA target sequences. Furthermore, Ca2+ is involved in the activation of NF-κB, for example, by enhancing the degradation of IκBα (15, 16). Because we observed the suppression of NF-κB and NFAT1 signaling, but not of TCR-proximal events or AP-1 signaling, we speculated that the inhibition of Ca2+ signaling might be the cause of Treg-mediated suppression of Tcons. TCR-induced IP3 stimulates the release of Ca2+ from intracellular stores into the cytoplasm (36). Depletion of Ca2+ stores is sensed by stromal interaction molecule 1 (STIM1), which resides in the endoplasmic reticulum membrane and contributes to the opening of Ca2+ release–activated Ca2+ (CRAC) channels and the influx of Ca2+ from the extracellular space. To measure Ca2+ signals, we loaded responder Tcons with the Ca2+-sensitive ratiometric dye Indo-1 AM. To further distinguish responder Tcons from cocultured Tregs (or Tcons in control assays), we labeled each cell population with different PKH dyes for general cell membrane labeling (Fig. 5A). Hence, we were able to specifically measure changes in Ca2+ concentrations only in responder Tcons even if they were conjugated with Tregs or, as a control, with Tcons. To enable the formation of cell pairs, we established cocultures at least 30 min before TCR-induced triggering of Ca2+ signals. In our flow cytometric analysis, by gating on Tcons that were not conjugated during the time of measurement, we observed an immediate increase in Ca2+ concentrations in the cytosol after TCR stimulation in control Tcons, and a partial inhibition of Ca2+ flux in suppressed Tcons (Fig. 5A, Q4 gate, lower left). By gating on PKH26 and PKH67 double-positive cell conjugates (Fig. 5A, Q2 gate), we detected an almost complete block of Ca2+ influx in suppressed Tcons (Fig. 5A, lower right). This result implied that, upon cell contact, Tregs entirely abrogated Ca2+ signals in Tcons. The partial inhibition of Ca2+ signals in the nonconjugated cell population might be ascribed to suppressed cells within the population that had contact with Tregs before the flow cytometric measurement.

Fig. 5

Ca2+ signaling is abrogated in suppressed Tcons independently of IP3. (A) Tcons were labeled with Indo-1 AM and PKH67 and cocultured with PKH26-labeled, pre-activated allogeneic Tregs (supTcon) or with allogeneic Tcons (cTcon) for at least 30 min. Cells were analyzed by flow cytometry and gated on Indo-loaded responder Tcons, as shown in the dot plots, and cells were then monitored for Ca2+ signaling as shown in the bottom panels depicting Ca2+ traces. After 1 min to establish a baseline, cross-linked antibodies against CD3 and CD28 were added to the cells (arrow) to stimulate Ca2+ signaling. The lower left panel shows Ca2+ influx in unconjugated Tcons (stained only with PKH67, from the Q4 gate in the upper dot plot), whereas the lower right panel shows Ca2+ influx in conjugated Tcons (PKH67- and PKH26-positive cells, from the Q2 gate in the upper dot plot). Data are from one donor and are representative of 20 donors. (B) Cells were labeled and cocultured as described for (A). Cocultured cells were then washed, resuspended in Ca2+-free PBS, and analyzed for Ca2+ signaling. After 1 min, cross-linked antibodies against CD3 and CD28 were added to the cells (arrow), and then medium was added to provide a final extracellular concentration of Ca2+ of 0.75 mM (arrowhead). Ca2+ influx in unconjugated Tcons (from the Q4 gate) is shown. Data are from one donor representative of seven independent donors. (C) Responder Tcons were labeled with [3H]inositol, cocultured with either unlabeled Tcons (cTcon) or unlabeled Tregs (supTcon), and then were stimulated for 5 min or were left unstimulated. Inositol phosphates were extracted from the cells and fractionated by HPLC, and the counts per minute (cpm) values were measured. Data are from one donor and are representative of five donors. OD, optical density.

Our analyses of cytokine gene expression and protein phosphorylation required the coculture of responder Tcons with Tregs (or Tcons, as a control) from different (HLA-A2–disparate) donors. Because of the staining protocol used for Ca2+ analyses, we were also able to analyze the suppression of Ca2+ signals in responder Tcons upon coculture with Tregs isolated from the same donor (autologous). We also observed suppression of Ca2+ signals when autologous Tregs were used (fig. S7A), and, thus, we can exclude the possibility that suppression is an artifact that resulted from coculture with T cells from a different donor. Because we observed suppression of NFAT1 and NF-κB signaling as well as IL2 expression after a 30- to 45-min coculture period with Tregs before TCR stimulation (Fig. 1), we tested whether suppression of Ca2+ signaling also required similar conditions. Indeed, we found that for suppression of Ca2+ influx, it was necessary to coculture Tregs and Tcons for at least 30 min before TCR stimulation (fig. S7B); however, maximal conjugate formation was already reached after about 10 min of coculture (fig. S7C). Almost no conjugates formed between Tcons and control Tcons (fig. S7C). Because conjugates were strongly enriched in the population representing cell pairs according to their size (fig. S7D), they are unlikely to result from the transfer of PKH dyes from one cell to another.

We next tested whether the depletion of intracellular Ca2+ stores was blocked by Tregs or if suppression occurred later, at the level of influx of Ca2+ through the plasma membrane. To this end, cocultures were established and pre-cocultured as described earlier to enable formation of cell pairs. After this pre-coculture, cocultures were washed and resuspended in Ca2+-free buffer to permit measurement of intracellular depletion of Ca2+ stores by flow cytometric analysis. We found that Tregs suppressed Ca2+ store depletion in Tcons during TCR stimulation in Ca2+-free buffer (Fig. 5B). Unfortunately, because of the low number of conjugates formed in Ca2+-free buffer, we could not analyze conjugates thoroughly with respect to Ca2+ store depletion. Because Ca2+ store depletion in T cells is mainly triggered through the binding of IP3 to IP3 receptors (IP3Rs), we investigated whether the amounts of IP3 in suppressed Tcons were reduced compared to those in control Tcons because of either decreased IP3 generation or altered inositol phosphate metabolism. Consistent with earlier analysis of PLC-γ1 phosphorylation and DAG generation, we detected no difference in the amount of IP3 in responder Tcons cocultured with Tregs and that of responder Tcons cocultured with Tcons (Fig. 5C). Similarly, stimulation-induced quantities of IP4 species were not reduced in suppressed Tcons, further implying that IP3 metabolism was unchanged (Fig. 5C).

cAMP has been described as a suppressive molecule involved in the Treg-mediated inhibition of murine T cells (6) and the suppression of Ca2+ signals in mast cells (37). However, we observed that the rapid suppression of Ca2+ signaling in Tcons was independent of cAMP, because a cAMP antagonist was unable to prevent suppression of Ca2+ influx in suppressed Tcons (fig. S8). Because Ca2+ signals were measured immediately after TCR stimulation, this time frame was likely too short to activate the suppressive potential of Tregs; thus, Tregs likely have to be pre-activated. To exclude potential effects of the antibody against CD3 that was used for pre-activation, we pre-activated Tregs or control Tcons with covalently plate-bound antibody against CD3. Tregs that were pre-activated in this way suppressed Ca2+ signals in responder Tcons, whereas pre-activated Tcons did not (fig. S9A). However, the suppressive capacity of Tregs was increased after pre-activation with soluble antibody against CD3. In addition, Tregs pre-activated with covalently plate-bound antibody against CD3 formed substantially more conjugates with responder Tcons than did Tcons that were pre-activated in the same way (fig. S9B).

Suppression of Ca2+ signaling causes the inhibition of NFAT and NF-κB signaling in Tcons

We hypothesized that suppression of Ca2+ signals might lead to inhibition of the activation of NFAT and NF-κB. To test the Ca2+ dependency of NFAT and NF-κB activation in our experimental system, we stimulated Tcons through the TCR in the presence of the Ca2+ chelator EGTA and investigated the phosphorylation status of various signaling molecules. As expected, we found that NFAT dephosphorylation was completely inhibited in the presence of EGTA, whereas the phosphorylation states of the TCR-proximal molecules ZAP-70 and PLC-γ1 were not affected (fig. S10). Phosphorylation of IκBα and IKK was also entirely inhibited in the presence of EGTA, whereas the phosphorylation of PKCθ was unchanged (fig. S10), supporting the notion that the suppression of Ca2+ signals caused the inhibition of NFAT and NF-κB activation in Tcons that are suppressed by Tregs.

To further support this idea, we investigated whether artificial depletion of Ca2+ stores in responder Tcons could abrogate the suppression of Ca2+, NFAT, and NF-κB signaling. To this end, we performed experiments with the specific Ca2+ store–depleting agent thapsigargin. Alternatively, we used the Ca2+ ionophore ionomycin in concentrations (≤0.5 μM) that are described to selectively release Ca2+ from intracellular stores without increasing the Ca2+ permeability of the plasma membrane (38, 39). Although suppression of Ca2+ store depletion was not complete in suppressed Tcons when Ca2+ amounts were compared to those in control Tcons (Fig. 5B), it was completely abrogated by addition of the Ca2+ store–depleting agent thapsigargin, further underlining that the observed partial reduction in Ca2+ signals was a result of suppression of Ca2+ store depletion (Fig. 6A). We also examined whether not only the suppression of Ca2+ store depletion (Fig. 6A) but also the suppression of Ca2+ signals in Ca2+-containing medium was abrogated in the presence of thapsigargin. For this purpose, we added thapsigargin at the peak of Ca2+ influx when suppression already occurs, and indeed, we found that the addition of thapsigargin immediately resulted in an increase in intracellular Ca2+ concentrations in both suppressed and control Tcons, leading to nearly complete abrogation of suppression (Fig. 6, B and C).

Fig. 6

Forced depletion of intracellular Ca2+ stores abrogates Treg-mediated inhibition of Ca2+, NFAT, and NF-κB signals in Tcons. (A to C) Cells were labeled and cocultured as described for Fig. 5. (A) Analysis in Ca2+-free PBS. The arrow indicates the addition of cross-linked antibodies against CD3 and CD28 in the absence or presence of thapsigargin (TG, 1 μM). The arrowheads at the top of the graph indicate the addition of medium to obtain an extracellular Ca2+ concentration of 0.5 mM. (B and C) The arrow indicates the addition of cross-linked antibodies against CD3 and CD28. Dashed arrow indicates the addition of (B) 100 nM or (C) 1 μM thapsigargin. Ca2+ influx in conjugated Tcons is shown. Data are from a single donor representative of four (A) or three (B and C) donors. (D) HLA-A2 Tcons were cocultured with HLA-A2+ Tcons (cTcon) or pre-activated HLA-A2+ Tregs (supTcon) for 60 min, and then cell populations within the cocultures were separated on the basis of HLA-A2 expression. Middle panel shows Western blotting analysis of HLA-A2 Tcons stimulated with cross-linked antibodies against CD3 and CD28 in the absence or presence of ionomycin (500 nM) or thapsigargin (50 nM) for 5 min. Lower panel shows analysis of the abundance of IL2 mRNA in HLA-A2 Tcons after 2 hours of stimulation with cross-linked antibodies against CD3 and CD28 in the absence or presence of ionomycin (100 and 500 nM). Relative mRNA amounts were normalized to those of GAPDH mRNA. Results are presented as the fold change in the abundance of IL2 mRNA compared to that in unstimulated Tcons, which was set to 1. Data are the means ± SD of the analysis of duplicate samples from one donor by qRT-PCR and are representative of three donors.

To further examine whether suppression of Ca2+ signals caused inhibition of NFAT and NF-κB activation, we examined the phosphorylation of signaling molecules and the abundance of cytokine mRNAs in suppressed and control Tcons upon addition of thapsigargin and ionomycin, respectively. To treat only Tcons and not Tregs with thapsigargin or ionomycin, we set up cocultures similar to those described earlier (Fig. 1B): Tcons and Tregs were preincubated and then separated, after which responder Tcons were stimulated through the TCR in the absence or presence of thapsigargin or ionomycin (Fig. 6D). We found that suppression of NFAT1 activation was partially abrogated in the presence of thapsigargin and completely abrogated in the presence of ionomycin (Fig. 6D). In addition, suppression of the phosphorylation of IκBα was abrogated with both substances (Fig. 6D). Accordingly, suppression of IL2 mRNA generation was reversed upon treatment with ionomycin (Fig. 6D). Thus, inhibition of Ca2+ signals largely contributed to Treg-mediated suppression of NFAT and NF-κB signaling and, consequently, of cytokine gene expression. In conclusion, our results demonstrate that pre-activated Tregs need a short period (30 to 45 min) of coculture with Tcons to actively alter and suppress the signaling machinery of Tcons. Suppressed Tcons are unable to activate Ca2+, NF-κB, or NFAT1 signaling pathways upon subsequent stimulation of their TCRs, which consequently prevents the expression of cytokine-encoding genes.

Discussion

Tregs play an important role in the immune system because they can suppress different immune cells, including autoreactive T cells, and thereby prevent autoimmunity (1). Many groups have investigated the features of Tregs, but few data exist from studies of Treg-mediated suppression mechanisms and signaling events in suppressed T cells. Most of the previous studies were conducted with mouse T cells; however, suppression mechanisms in mice and humans can differ (4). Therefore, research on the function of human Tregs is crucial, especially with regard to their clinical application.

We previously showed that inhibition of the expression of cytokine-encoding genes occurs rapidly in Tcons upon coculture with Tregs (18). In the current study, we demonstrated that suppression of the expression of cytokine-encoding genes persisted in Tcons even if the Tregs were removed before the Tcons were stimulated through the TCR. To be suppressed, Tcons had to have cell-cell contact with pre-activated Tregs for only 30 to 45 min. This suggests that Tregs, despite being present in lower numbers than those of Tcons in vivo, might have the capacity to suppress several Tcons in a sequential fashion.

In vivo, several immunosuppressive cytokines, such as transforming growth factor–β (TGF-β) and IL-10, are important for the prevention of inflammation of the colon and of experimental autoimmune encephalomyelitis (which is a mouse model similar to human multiple sclerosis), but they seem to be unnecessary for the prevention of other disorders, for example, autoimmune gastritis (4042). Because various factors might contribute to T cell suppression in different microenvironments in vivo, it might be easier to clarify the inhibitory mechanisms in vitro. Therefore, many groups have investigated the direct suppression of Tcons by Tregs and have revealed a contact-dependent inhibitory mechanism that is independent of soluble cytokines (43, 44). To date, several putative surface markers for Tregs have been described (45). For example, CTLA-4 is predominantly found on the surface of Tregs as well as on activated T cells, and CTLA-4 influences Treg-mediated suppression in mice by limiting the stimulatory capacity of APCs (1921). Our blocking experiments excluded a role for CTLA-4 in the rapid suppression of cytokine gene expression in human Tcons in the presence or absence of APCs, although CTLA-4 was involved in suppression of the proliferation of human Tcons in the presence of APCs. The receptor on Tcons that is recognized by Tregs remains elusive; however, our results indicate that triggering of this unknown receptor must lead to a suppressive signal within about 30 min.

We compared TCR signaling in control and suppressed Tcons and investigated the main signaling pathways that lead to the transcription of cytokine-encoding genes, namely, the AP-1, NF-κB, and NFAT pathways. We found no indication that the TCR itself or its associated proximal signaling molecules were involved in the suppression mechanism, because there was no change in the extent of phosphorylation of CD3ε, ZAP-70, or PLC-γ1 between control Tcons and suppressed Tcons, neither was there any difference in the generation of DAG and IP3. Accordingly, we found no evidence of suppression of the AP-1 pathway by Tregs, as determined by analysis of the phosphorylation of ERK, p38, and c-Jun, as well as of the expression of AP-1 target genes in suppressed Tcons compared to control Tcons. However, we found that the NFAT1 and NF-κB pathways in suppressed Tcons were rapidly and markedly inhibited directly after TCR stimulation. We also detected the immediate suppression of Ca2+ release from intracellular stores in suppressed Tcons, which caused the reduced activation of NFAT and NF-κB and, consequently, the suppression of NFAT- and NF-κB–dependent expression of IL2.

As suggested by others (46), different mechanisms of suppression might be used by Tregs depending on, for example, the site of inflammation, the effector cell types, and the time point of suppression. In addition to suppression of the priming and proliferation of Tcons through the inhibition of APCs by Tregs, direct suppression of CD8+ T cell effector function independently of repressed proliferation has been described (47), underscoring the in vivo relevance of direct suppression of T cells by Tregs. Our data suggest that at least two different mechanisms of suppression of human Tcons might operate. First, rapid suppression of Ca2+, NFAT, and NF-κB signaling might result in inhibition of cytokine production, which is maintained upon removal of Tregs and seems to be independent of CTLA-4 and APCs. Second, suppression of the proliferation of human Tcons might require prolonged contact with Tregs and may be independent of Ca2+ signaling and the suppression of cytokine production. Consistent with this, mice with CD4+ T cells that are doubly deficient in STIM1 and STIM2 display enhanced T cell proliferation compared to that of wild-type cells despite the lack of store-operated Ca2+ influx (48). In addition, a marked decrease in Treg numbers contributes to lymphoproliferation in these mice. However, alternative modes of Ca2+ entry might compensate for the absence of STIM, for example, Cav calcium channels might operate when their inhibition by STIM1 is lacking (49, 50).

Furthermore, NFAT might play multiple roles during the course of suppression. We showed that the suppression of NFAT signaling occurred within minutes of TCR stimulation, which, consequently, led to reduced NFAT-dependent gene expression. Nevertheless, NFAT may also actively contribute to the suppression of gene expression at later time points. In assays in which long-term suppression was assessed, such as by measuring proliferation or IL-2 production 18 hours after TCR stimulation, a study showed that murine Tcons doubly deficient in NFAT1 and NFAT4 are less susceptible to Treg-mediated suppression than are wild-type Tcons (17). NFAT might form inhibitory complexes on the promoters of cytokine-encoding genes in conjunction with transcriptional repressors, as has been proposed for ICER-NFAT complexes (51) and in particular for NFAT2 (52). The amounts of ICER protein and mRNA are increased in suppressed mouse Tcons after 17 to 20 hours of stimulation compared to those in control Tcons, which correlates with an increase in the amount of cAMP in suppressed Tcons (6, 53). In suppression assays with B7-deficient responder T cells, ICER fails to accumulate, but residual suppression is still detectable (53). This argues for additional suppressive mechanisms other than those mediated by CTLA-4, cAMP, and ICER. We did not detect accumulation of ICER mRNA or of mRNAs for other anergy-related genes in suppressed human Tcons up to 5 hours after TCR stimulation, whereas we found that the expression of cytokine-encoding genes was repressed earlier (18). Also, the passive diffusion of large amounts of cAMP through gap junctions from Tregs to Tcons, as has been proposed (6), seems unlikely within the short time period that we investigated.

The exact signaling events leading to the inhibition of Ca2+ release and NFAT1 and NF-κB activation in suppressed Tcons remain to be determined. We did not find any involvement of TCR-proximal signaling molecules, including PLC-γ1, which is crucial for Ca2+ influx and PKCθ activation. A report by Gri and colleagues showed the Treg-mediated suppression of Ca2+ signals in murine mast cells in a PLC-γ2–independent fashion (37). However, suppression of Ca2+ signals in mast cells does not involve the suppression of depletion of Ca2+ stores, but rather involves a cAMP-dependent suppression of Ca2+ influx through the plasma membrane. We now describe a cAMP-independent, rapid suppression of Ca2+ store depletion in human T cells, which occurred downstream of PLC-γ1. Ca2+ store depletion was suppressed without alterations in PLC-γ1 activity. Because the amounts of IP3 in suppressed Tcons were unaltered compared to those in control Tcons, it appears that IP3 metabolism was unchanged. The exact mechanism through which Tregs inhibit store depletion therefore remains elusive. It might involve currently unknown modifications of the IP3R that could hinder its interaction with IP3. Future studies will be needed to address these questions and to decipher the mechanism by which Tregs inhibit Ca2+ signals in Tcons. We propose that the suppression of Ca2+ signals leads to the suppression of both NFAT1 and NF-κB signaling because treatment of Tcons with ionomycin or thapsigargin relieved the suppression of NFAT1 and NF-κB activation. Suppression of IL2 expression was completely abrogated upon treatment with ionomycin. This implies that inhibition of Ca2+ release is sufficient for the suppression of IL2 expression.

In conclusion, we have elucidated the initial molecular processes that occur in primary human CD4+CD25 Tcons upon suppression by CD4+CD25+ Tregs. Our data reveal that Tregs mediate an immediate block of NF-κB and NFAT signaling in Tcons through the inhibition of Ca2+ signals, which then leads to the suppression of cytokine gene expression. In cancer, suppression of effector T cells is deleterious, and breaking this suppressive state is highly desirable, whereas in autoimmunity, the suppression of autoreactive T cells is warranted. Thus, an understanding of the molecules and signaling pathways in Tcons that are affected upon suppression by Tregs is crucial for their future therapeutic manipulation.

Materials and Methods

Preparation of Tregs and Tcons

Human peripheral blood leukocytes were purified from buffy coats by Biocoll (Biochrom) gradient centrifugation followed by adherence to plastic to deplete monocytes. Blood from HLA-A2+ donors was used to isolate Tregs and Tcons, and blood from HLA-A2 donors was used to isolate responder Tcons. We first isolated CD25high cells with CD25-specific magnetic-activated cell sorting (MACS) beads (2 μl per 107 cells, Miltenyi). CD4+CD25 Tcons were isolated from the CD25 fraction with the CD4 Isolation Kit II and were additionally depleted from the CD25+ cells with CD25-specific MACS beads (6 μl per 107 cells). For some experiments, CD4+CD25high Tregs were sorted on a FACS (fluorescence-activated cell sorting) Diva flow cytometer. Tcons were rested overnight in X-VIVO 15 medium (Lonza) supplemented with 1% GlutaMAX. Tregs were pre-activated overnight with antibody against CD3 (1 μg/ml) in X-VIVO 15 medium containing 1% GlutaMAX and IL-2 (50 U/ml), if not stated otherwise.

Coculture assays

HLA-A2+ Tregs (using a pool of cells from several donors when we needed to obtain sufficient cell numbers) and HLA-A2+ Tcons were incubated with fluorescein isothiocyanate (FITC)–conjugated antibody against HLA-A2 and FITC-specific microbeads (Miltenyi), whereas HLA-A2 Tcons were left untreated. Cocultures of HLA-A2 Tcons and either HLA-A2+ Tregs or HLA-A2+ Tcons as a control were set up in a 1:1 ratio. Cells were cultured for 30 to 60 min and then stimulated with soluble antibody against CD3 (0.2 μg/ml), antibody against CD28 (2 μg/ml), and goat antibody against mouse antibody as a cross-linker (2 μg/ml) for the indicated time periods at 37°C. Stimulation was then stopped, and the differing cell populations were separated on the basis of HLA-A2 expression by passing the cells over an LS column (Miltenyi) on ice. HLA-A2 Tcons (flowthrough from the column, >98% pure) were used for subsequent Western blotting, Luminex, and mRNA analyses. The suppressive capacity of Tregs was controlled in every assay by analyzing the abundances of IL2 and IFNg mRNAs in the HLA-A2 Tcons after 2 to 3 hours of coculture with the Tregs. In some experiments, cell populations within cocultures were separated by MACS technology before stimulation of responder Tcons, which had been washed twice with X-VIVO medium to remove MACS buffer before stimulation. In some experiments, ionomycin (0.1 to 0.5 μM) or thapsigargin (0.05 to 1 μM) was added in addition to the stimulating antibodies.

RNA preparation and quantitative reverse transcription–polymerase chain reaction assays

Total RNA was isolated with the RNAqueous Micro Kit (Ambion), and complementary DNA (cDNA) was prepared with oligo(dT)16 primers (Invitrogen). We quantified mRNAs by detection of incorporated SYBR Green with the ABI Prism 5700 sequence detector system (Applied Biosystems). The relative abundance of a given mRNA was determined by normalization to that of GAPDH mRNA, and the results are presented as the fold difference compared to the abundance of mRNA in unstimulated Tcons, which was set to 1. Primer sequences are listed in the Supplementary Methods.

Gene array analysis

HLA-A2+ Tregs and HLA-A2+ Tcons were incubated with FITC-conjugated antibody against HLA-A2 and FITC-specific microbeads. HLA-A2 Tcons were left unstimulated or were stimulated with antibody against CD3 (0.2 μg/ml) and antibody against CD28 (0.5 μg/ml) for 3 hours in the presence of HLA-A2+ Tregs or HLA-A2+ Tcons at a 1:1 ratio. After the stimulation period, cocultures were separated, RNA was isolated as described earlier, and gene array analysis (Affymetrix whole genome U133A 2.0.CHIP) was performed.

Western blotting analysis

Cocultures were established as described earlier, and cells were incubated for 30 to 60 min before stimulation. After separation of the individual populations in the coculture, responder Tcons were washed in tris-buffered saline (TBS) and lysed in Beadlyte Cell Signaling Universal Lysis Buffer (Upstate) supplemented with complete protease inhibitors (Roche Applied Science) and PhosSTOP (Roche Applied Science). Proteins were denatured in SDS sample buffer, resolved by 8 to 12% SDS–polyacrylamide gel electrophoresis (SDS-PAGE), and transferred to Hybond nitrocellulose membranes (Amersham). After the membranes were blocked with 5% nonfat dry milk in TBS containing 0.1% (w/v) Tween 20 (TBST) and incubated with primary and horseradish peroxidase (HRP)–conjugated secondary antibodies, protein bands were developed with Immobilon Western Chemiluminescent HRP substrate (Millipore) in a Vilber Lourmat chemiluminescence acquisition system, and bands were quantified with Bio-1D software (Vilber Lourmat). For the detection of phosphorylated p65 by Western blotting analysis with fluorescently labeled secondary antibodies, cells were lysed by sonication for 1.5 min in sample buffer, Western blotting was performed as described earlier, and blots were scanned with the LI-COR Odyssey infrared imaging system.

Measurement of Ca2+ flux

T cells were incubated with the dyes PKH26 or PKH67 (1:200 in diluent C, 200 μl/106 cells, Sigma) for 5 min at 20°C, and the reaction was stopped with fetal bovine serum (FBS, Gibco). Cells were washed three times with X-VIVO 15 medium and left in X-VIVO 15 for 2 hours at 37°C. Responder Tcons were then labeled with Indo-1 AM (1 μM, Invitrogen) for 30 min at 37°C, washed, and resuspended in X-VIVO 15 medium. In some experiments, responder Tcons were incubated with the cAMP antagonist Rp-cAMPS (1 mM, Calbiochem) for 30 min at 37°C in X-VIVO 15 medium after incubation with PKH and Indo-1 and subsequently washed twice with medium before cocultures were established. Tregs or Tcons were cocultured in X-VIVO 15 medium supplemented with 0.5 mM CaCl2 in a 1:1 ratio with allogeneic responder Tcons at 5 × 106 cells/ml for at least 30 min (if not stated otherwise) at 37°C to enable cell pairs to form. Ca2+ influx in X-VIVO 15 medium supplemented with 0.5 mM Ca2+ (final concentration, 2 mM) was induced after 1 min of measurement of basal Ca2+ flux by the addition of antibody against CD3 (0.2 μg/ml), antibody against CD28 (2 μg/ml), and cross-linking antibodies (2 μg/ml). In some experiments, thapsigargin was added. Measurements were performed on a FACS Diva or LSRII flow cytometer (BD Biosciences). For measurements of Ca2+ store depletion and Ca2+ influx through Ca2+ channels in the plasma membrane, cocultures were incubated for at least 30 min in X-VIVO 15 medium at 37°C to enable conjugate formation and then cells were washed twice with Ca2+-free, phosphate-buffered saline (PBS) and immediately measured in PBS. After 1 min, Ca2+ store depletion was induced by the addition of antibodies against CD3 and CD28 and cross-linking antibodies, as described earlier, in PBS, and after further measurement, extracellular Ca2+ was added in X-VIVO 15 medium to a final concentration of 0.75 mM. Data were analyzed with the kinetics tool of FlowJo software and exported to PowerPoint to adjust baselines (Indo-1 ratio before stimulation) to equal levels for responder Tcons of one donor within one experiment. As an alternative to staining with PKH, responder Tcons were distinguished from allogeneic Tcons or Tregs by staining for surface HLA-A2, and Ca2+ measurements were performed as described above.

Inositol phosphate analysis

Responder Tcons were labeled with myo-[3H]inositol (10 μCi/ml; Perkin Elmer) in inositol-free RPMI medium (Biomol) containing 10% dialyzed FBS (Sigma) for 48 to 50 hours. Tcons were then washed and resuspended in X-VIVO 15 medium for 12 hours. Cocultures were established with unlabeled Tregs or Tcons in a 1:1 ratio; the cells were incubated for 60 min at 37°C and then stimulated as described earlier. After stimulation, cocultures were washed with cold PBS, lysed in 10% trichloroacetic acid (TCA) supplemented with EDTA and sodium fluoride, and subjected to two rounds of freezing and thawing, and then the TCA was removed by diethyl ether extraction. The pH of the sample was adjusted to ~6, and samples were partially lyophilized, frozen, and shipped on dry ice. Inositol phosphate extracts were subjected to metal-dye detection high-performance liquid chromatography (MDD-HPLC), which was performed on a MiniQ PC 3.2/3 column (Pharmacia Biotech) with a Kontron system (BioTEK), as described previously (54), with slight modifications (see the Supplementary Methods). The radioactivity of [3H]-labeled inositol phosphate isomers synthesized from myo-[3H]inositol in the collected fractions was determined by liquid scintillation counting (Flo-Scint IV, Packard). The suppression of cytokine mRNA expression was controlled in parallel experiments with an aliquot of cells cultured under the same conditions, but with nonradioactive inositol. Before the analysis of mRNAs in responder Tcons, cell populations within the coculture were separated as described above.

Antibodies and reagents

Antibody against CD3 (OKT3) and antibody against CD28 (15E8) were purified from hybridoma supernatants. The cross-linking goat antibody against mouse antibody was obtained from Southern Biotechnology. Antibodies for the flow cytometric analysis of human HLA-A2, CD80, CD86, CTLA-4, PLC-γ1 (pY783), p38 (pT180/pY182), and ERK1/2 (pT202/pY204) and the corresponding isotype control antibodies were purchased from BD Biosciences. Antibodies for the detection by Western blotting of IκBα (pS32), IKKα (pS180)/β(pS181), ZAP-70 (pY319)/Syk (pY352), p65 (pS536), and p38α were from Cell Signaling Technology; antibodies against ZAP-70, NFAT1, IκBα (c-21), PLC-γ1, and actin were from Santa Cruz Biotechnology; antibodies against PLC-γ1 (pY783) and PKCθ (pT538) were from BD Biosciences; antibody against tubulin was obtained from Sigma; and antibody against p38 (pT180/pY182) was from Promega. The HRP-conjugated antibodies against mouse immunoglobulin G (IgG), IgG1, IgG2a, IgG2b, or rabbit IgG were purchased from Santa Cruz Biotechnology. Alexa Fluor 680–conjugated antibody against rabbit IgG, Indo-1 AM, and thapsigargin were obtained from Invitrogen. PKH dyes, EGTA, and ionomycin were purchased from Sigma.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/4/204/ra90/DC1

Methods

Fig. S1. Suppression of cytokine secretion is retained after removal of Tregs.

Fig. S2. Treg-mediated rapid suppression of IL2 expression in Tcons also occurs in the presence of APCs.

Fig. S3. CTLA-4 is not involved in the suppression of cytokine gene expression.

Fig. S4. Proximal TCR signaling is not altered in suppressed Tcons.

Fig. S5. Suppression of NFAT and NF-κB activation is independent of the antibody used to activate Tregs.

Fig. S6. Phosphorylation of ERK and p38 is not affected in suppressed Tcons.

Fig. S7. The block in Ca2+ influx in Tcons is not caused by allogeneic responses and requires 30 min of previous coculture with Tregs.

Fig. S8. Suppression of Ca2+ signaling is independent of cAMP.

Fig. S9. Tregs, but not Tcons, pre-activated with covalently plate-bound antibody against CD3 suppress Ca2+ signaling in responder Tcons.

Fig. S10. NFAT and NF-κB activation in Tcons is Ca2+-dependent.

References

References and Notes

Acknowledgments: We thank U. Matiba for technical assistance; K. Hexel and S. Schmitt for cell sorting; C. Watzl and K. Kohl for help with the conjugate assay; F. T. Wieland for suggestions regarding DAG analyses; R. Arnold, K. Gülow, J. Hoffmann, M. Li-Weber, T. Mock, and H. Weyd for helpful discussions and critical reading of the manuscript; and J. Buer (Essen, Germany) and R. Geffers (Braunschweig, Germany) for performing gene arrays. Funding: This research was supported by contract research “Forschungsprogramm Allergologie II” of the Baden-Württemberg Stiftung (P-LS-AL-18/2); the Alliance for Immunotherapy of the Helmholtz Society; SFB 405; CellNetworks (to B.B.); the German Research Foundation SFB 638 and TRR83 (to B.B. and M.H.); and a Ph.D. fellowship of the Helmholtz International Graduate School for Cancer Research at the Deutsches Krebsforschungszentrum (to A.S.). Author contributions: N.O. and A.S. designed, performed, and analyzed coculture assays, Western blotting, enzyme-linked immunosorbent assays, and RNA and proliferation assays and wrote the paper; A.S. designed, performed, and analyzed calcium measurements and coculture experiments for inositol phosphate extraction; N.O. and E.S.-P. designed, performed, and analyzed experiments involving detection of intracellular phosphoproteins by flow cytometry and gene array experiments; E.-M.W. provided assistance in the design, performance, and analysis of experiments; D.V. provided technical assistance; S.F., R.B., and N.O. designed, performed, and analyzed phosphorylated p65 by Western blotting; C.S.F., A.S., and E.S.-P. designed, performed, and analyzed Luminex experiments; M.H. and B.B. designed, performed, and analyzed lipid mass spectrometry; H.L. and G.W.M. designed, performed, and analyzed the HPLC analysis of inositol phosphates; P.R., M.G., and A.S. designed, performed, and analyzed experiments to set up the conditions for Ca2+ measurements; and E.S.-P. and P.H.K. designed and supervised the research and wrote the paper. Competing interests: The authors declare that they have no financial conflicts of interest.
View Abstract

Navigate This Article