The repeated injection of low doses of bacterial superantigens (SAg) is known to induce specific T cell unresponsiveness. We show in this study that the spleen of BALB/c mice receiving chronically, staphylococcal enterotoxin B (SEB) contains SEB-specific CD4+ TCRBV8+ T cells exerting an immune regulatory function on SEB-specific primary T cell responses. Suppression affects IL-2 and IFN-γ secretion as well as proliferation of T cells. However, the suppressor cells differ from the natural CD4+ T regulatory cells, described recently in human and mouse, because they do not express cell surface CD25. They are CD152 (CTLA-4)-negative and their regulatory activity is not associated with expression of the NF Foxp3. By contrast, after repeated SEB injection, CD4+CD25+ splenocytes were heterogenous and contained both effector as well as regulatory cells. In vivo, CD4+CD25− T regulatory cells prevented SEB-induced death independently of CD4+CD25+ T cells. Nevertheless, SEB-induced tolerance could not be achieved in thymectomized CD25+ cell-depleted mice because repeated injection of SEB did not avert lethal toxic shock in these animals. Collectively, these data demonstrate that, whereas CD4+CD25+ T regulatory cells are required for the induction of SAg-induced tolerance, CD4+CD25− T cells exert their regulatory activity at the maintenance stage of SAg-specific unresponsiveness.
Accumulating evidence indicates that CD4+CD25+ T regulatory cells are essential for the maintenance of self-tolerance. This notion was first substantiated by the observation that neonatal thymectomy of mice leads to the development of fatal autoimmune disorders that are prevented by the transfer of CD4+CD25+ T cells purified from normal mice (1). Likewise, the autoimmune diseases induced in T cell-deficient animals by transferred CD4+CD25− T cells can be abrogated by the coinjection of CD4+CD25+ T cells (2, 3). On the contrary, depletion of CD4+CD25+ T cells leads to the development of various autoimmune disorders in genetically susceptible strains of mice; reconstitution of the depleted population prevents autoimmunity (4, 5, 6). Elimination of CD4+CD25+ T cells can also enhance immune responses to non-self Ags such as infectious agents, tumor Ags, alloantigens and superantigens (SAg)3 (7, 8, 9, 10, 11). Moreover, expansion of CD4+CD25+ T cells or augmentation of their activity can suppress allograft rejection (12, 13, 14, 15). Thus, the activity of natural CD4+CD25+ T regulatory cells not only controls the occurrence of autoimmune disease but can also suppress the response of T cells to exogenous Ags.
Staphylococcal exotoxins, including toxic shock syndrome toxin-1 and staphylococcal enterotoxin serotypes A, B, C1, C2, C3, D, E, G, H, and I, cause toxic shock syndrome in both humans and animals (16). These microbial products are all members of a family of structurally related proteins called SAgs that bind to the MHC class II molecules on the APCs and to TCR bearing-specific Vβ fragments (17). This trimolecular interaction leads to massive proliferation of T cells and uncontrolled release of proinflammatory cytokines, including IL-1, IL-2, IFN-γ, and TNF-α, which causes life-threatening toxic shock syndrome (16). Others and our own studies have shown that SAg-dependent toxic shock can be prevented in mice by repeated exposure to low doses of SAg (1–10 μg per dose) (18, 19, 20, 21). One of the striking consequences of repeated SAg injection is that T cells from treated mice fail to secrete proinflammatory cytokines in response to subsequent SAg challenge. This state of unresponsiveness, however, does not effect the production of IL-10 nor type 2 cytokines because IL-4, IL-5, and IL-10 are readily detected in assays involving the stimulation of unresponsive T cells (19, 21). Remarkably, the inability to produce type 1 cytokines such as IL-2 and IFN-γ can be adoptively transferred to syngeneic recipients (21). This transfer appears to be dependent on the secretion of IL-10 and requires the presence of CD4+CD8− T cells in the transferred spleen cell population. These observations led us to postulate that repeated exposure to SAgs induces the development of CD4+ T regulatory cells capable of suppressing the primary response of SAg-specific T cells. This hypothesis was also supported by the observation that, in BALB/c mice made unresponsive to staphylococcal enterotoxin B (SEB), 10–15% of SEB-specific TCR Vβ8 (TCRBV8) bearing CD4+ T cells over-expressed CD152 (CTLA-4) (21), a molecule described recently as preferentially expressed by T regulatory cells (3, 10, 22). We undertook the present study to determine the capacity of CD152high CD4+ TCRBV8+ T cells from unresponsive animals to regulate primary T cell responses to SEB.
Materials and Methods
Animals and tolerizing protocol
Four- to 6-wk-old BALB/c female mice were obtained from Harlan Netherland (Horst, The Netherlands). Induction of unresponsiveness was performed as previously described, i.e., by three i.p. injections of 5–10 μg SEB (Toxin Technology, Sarasota, FL) on day 0, 2, and 4 (21).
Cells were immunostained with various combinations of fluorescence-conjugated Abs and analyzed by flow cytometry. Fluorochrome- and biotin-conjugated Abs were all from BD PharMingen (Erembodegem, Belgium). Abs were APC-conjugated rat anti-mouse CD4 (clone RM4-5), biotinylated rat anti-mouse CD25 (clone 7D4), FITC-conjugated mouse anti-mouse TCRBV8 (clone F23.1), FITC-conjugated rat anti-mouse TCRBV6 (clone RR4-7), and PE-conjugated hamster anti-mouse CD152 (clone UC10-4F10-11). Second-step reagents included CyChrome- or PE-conjugated streptavidin. For analysis of intracellular CD152, cells were fixed before with 2.0% paraformaldehyde and permeabilized using 0.5% saponin. Samples were analyzed on four color-fitted FACSCaliber cytometer using CellQuest software (BD Biosciences, Erembodegem, Belgium).
Purification of T cell subsets
Two to four days after the last SEB injection, spleens were removed and prepared into single-cell suspensions. CD4+ T cells were isolated using magnetic separation. Briefly, cells were incubated with FITC-conjugated rat anti-mouse CD8 (clone 53-6.7; BD PharMingen), FITC-conjugated rat anti-mouse B220 (clone RA3-6B2; BD PharMingen), and FITC-conjugated mouse anti-mouse Ia (clone 25-9-17; BD PharMingen) mAbs. Cells were then washed and incubated with anti-FITC microbeads (Miltenyi Biotec, Bergisch-Gladbach, Germany). Depletion of CD8+, Ia+, and B220+ cells was conducted by magnetic separation with MACS separation columns according to the suggested protocol (Miltenyi Biotec). The CD4+ T cell-enriched population always contained <0.3% CD8+, Ia+, and B220+ cells as assessed by flow cytometry. To separate TCRBV8+ from TCRBV8− CD4+ T cells, the enriched CD4+ T cells were incubated with FITC-conjugated mouse anti-mouse TCRBV8 mAbs (clone F23.1; BD PharMingen). The cells were then incubated with anti-FITC microbeads (Miltenyi Biotec) and magnetic separation was performed on MACS separation columns (Miltenyi Biotec). The retained cells were eluted from the column as TCRBV8+ CD4+ T cells. Purity of the cell preparation was >98.1% TCRBV8+ CD4+ T cells. The flow-through contained TCRBV8− CD4+ T cells and was <0.4% TCRBV8+ CD4+ T cells.
For some experiments, CD4+ T cell-enriched spleen cells were separated according to their expression of CD25. Cells were incubated with biotinylated rat anti-mouse CD25 mAbs (clone PC61; BD PharMingen). They were washed then incubated with streptavidin-conjugated microbeads, and magnetic separation was conducted on MACS column (Miltenyi Biotec). The CD4+CD25− T cells passed through the column. The retained cells were eluted as CD4+CD25+ T cells.
In vitro assays
Suppression was assayed on the decreased capacity of normal spleen cells to secrete IL-2 and IFN-γ and to proliferate in response to in vitro stimulation with SEB. Serial dilutions of spleen cells or purified subsets of T cells from unimmunized BALB/c mice or mice made unresponsive to SEB were mixed with normal BALB/c spleen cells (1 × 106 per well) in U-shaped 96-well culture plates (final volume 200 μl) and stimulated for 60 h with SEB (1 μg/ml). Proliferation for the last 16 h of culture was assessed by [3H]TdR incorporation assay. Culture supernatants were taken after 48 h of stimulation with SEB, and their contents of IL-2, IL-10, and IFN-γ were determined with ELISA kits (R&D Systems, Abingdon, U.K.) using manufacturer’s procedure.
Stimulation by immobilized anti-CD3 mAbs was conducted as follows: Hamster anti-CD3 mAbs (clone 145-2C11; 10 μg/ml) were incubated overnight at 4°C in 50 μl PBS per well of flat-bottom 96-well plates. After three washes with PBS, plates were seeded with cells (1 × 106 per well) and cultured for 48 h. Culture supernatants were harvested and levels of IFN-γ and IL-10 assessed by ELISA kits (R&D Systems) using manufacturer’s protocol.
In vivo assays
To assess the capacity of mice made tolerant to SEB to suppress in vivo the primary proliferative response of SEB-specific T cells, spleen cells from normal BALB/c mice were stained with CFSE (1 μM, 30 min, 37°C; Molecular Probes, Eugene, OR). The cells (50 × 106) were injected via the tail vein into normal or SEB-pretreated BALB/c mice. One day later, 10 μg of SEB were injected i.p. After two more days, spleen cells were labeled with CyChrome-conjugated anti-CD4 and PE-conjugated anti-TCRBV8 mAbs (BD PharMingen). CFSE-stained CD4+ TCRBV8+ cells were then analyzed for green fluorescence decay on a FACSCaliber (BD Biosciences).
For experiments involving the depletion of CD25+ cells, 4-wk-old mice were thymectomized according to Monaco et al. (23) protocol. Briefly, an elliptic incision of the skin was made on the anterior surface of the neck and the thymus was removed by suction through a glass tip inserted in the anterior mediastinum. Thymectomized animals were then depleted of CD25+ cells by three injections of 0.5 mg rat anti-mouse CD25 mAbs (clone PC61) that was kindly provided by Dr. O. Leo, Institut de Biologie et de Medecine Moleculaires, Université Libre de Bruxelles (Gosselies, Belgium), applied on every other day. Six days after the last injection, depletion was assessed by fluorescence cytometry by double staining with FITC-conjugated rat anti-mouse CD25 mAbs (clone 7D4; BD PharMingen) and APC-conjugated rat anti-mouse CD4 (BD PharMingen). CD25+ cell-depleted animals were then repeatedly injected with SEB and their survival was observed. In some experiments, undepleted or CD25+ cell-depleted spleen cells, or purified CD4+CD25+ T cells from normal or unresponsive animals were injected i.p. in CD25+ cell-depleted thymectomized mice one day before SEB injection. Depletion or purification of CD25+ cells was conducted by using magnetic bead separation technology as previously described.
Expression analysis by real-time RT-PCR
We synthesized an oligo-dT-primed first-strand cDNA using the ImProm-II Reverse Transcription System (Promega Benelux, Leiden, The Netherlands) to use as a template for real-time RT-PCR. Expression of Foxp3 was measured using the primers 5′-CCCAGGAAAGACAGCAACCTT-3′ and 5′-TTCTCACAACCAGGCCACTTG-3′, and the internal probe 5′-FAM-ATCCTACCCACTGCTGGCAAATGGAGTC-TAMRA-3′. β-actin was used as an endogenous reference and measured with specific primers 5′-CTAAGGCCAACCGTGAAAAG-3′ and 5′-AGCCTGGATGGCTACGTACAT-3′, and the probe 5′-FAM-TGACCCAGATCATGTTTGAGACCTTCA-TAMRA-3′. Other components of the reaction were from the LightCycler-FastStart DNA Master Hybridization probes (Roche, Mannheim, Germany). PCR cycling conditions were 95°C for 10 min and 45 cycles of 95°C for 10 s, 55°C for 10 s, and 72°C for 5 s.
Data were collected using LightCycler Data Analysis software (Roche). A standard curve was generated with a dilution series of a reference cDNA sample. The software determines by using the threshold cycle (CT) the relative quantity of each unknown sample. Data are expressed as normalized Foxp3 expression, which was obtained by dividing the relative quantity of Foxp3 for each sample by the relative quantity of β-actin for the same sample. The Foxp3 primers do not amplify genomic DNA.
Repeated injection of SEB induces SEB-specific T cells with immune regulatory activity
Repeated injection of SEB in BALB/c mice induced unresponsiveness, or anergy, as subsequent in vitro challenge of isolated spleen cells with SEB did not stimulate cell proliferation nor IL-2 secretion (Fig. 1,A, left panels). Unresponsiveness also extended to IFN-γ production (Fig. 1,A, left panels). This unresponsive state was not the direct result of deletion of responsive T cells because the spleen of repeatedly injected mice consistently contained more SEB-specific CD4+ TCRBV8+ T cells than that of normal BALB/c mice (7.1 ± 1.5 × 106 cells vs 5.7 ± 1.0 × 106 cells, respectively). Spleen cells from SEB-injected animals, however, significantly produced more IL-10 in response to SEB stimulation than primary T cells (Fig. 1,A, left panels). Mixed cultures of spleen cells isolated from untreated and repeatedly injected animals revealed the suppressive activity exerted by unresponsive spleen cells on primary T cell response (Fig. 1,A, right panels). This activity could be detected up to 20 days after the last injection of SEB (data not shown). Its intensity was cell number-dependent because increasing the number of unresponsive cells in the mixed cultures enhanced the suppressive effect on cell proliferation and IL-2 and IFN-γ secretion (Fig. 1,A, right panels). On the contrary, levels of IL-10 detected in mixed cultures stimulated with SEB increased dramatically as more suppressive cells were added to primary cultures and could be directly correlated to the intensity of the suppressive activity (Fig. 1 A, right panels). Remarkably, the amount of IL-10 produced in mixed cultures was much higher than that secreted in cultures of single-cell populations added together, as if IL-10 production was strongly enhanced by the presence of unresponsive spleen cells.
The suppressive activity of spleen cells from BALB/c mice repeatedly injected with SEB was also investigated in vivo. We previously reported that unresponsiveness toward SEB can be adoptively transferred by injection of unresponsive spleen cells into naive syngeneic recipient mice, and IL-2 and IFN-γ produced in primary response to SEB stimulation are not detected in the serum of adoptively transferred recipients (21). In the present study, we assessed the capacity of unresponsive spleen cells to suppress in vivo the capacity of normal SEB-specific CD4+ T cells to proliferate in response to primary SEB stimulation. Spleen cells from BALB/c mice were stained with CFSE and transferred into untreated or 3× SEB-treated mice. One day later, recipients were injected with 10 μg of SEB and after two more days, CFSE-positive SEB-specific CD4+ TCRBV8+ T cells were analyzed by flow cytometry for sign of proliferation. As seen in Fig. 1 B, SEB stimulation of BALB/c recipients not pretreated with SEB induced the proliferation of CD4+ TCRBV8+ T cells, and most of the cells (86.0 ± 10.2%) underwent at least 1 cycle of cell division. By contrast, about half of CD4+ TCRBV8+ T cells transferred in recipients pretreated with three injections of SEB did not proliferate following SEB stimulation; most CSFE-labeled cells that did proliferate in these mice went only through 1 or 2 cycles of cell division. The poor ability of SEB-specific T cells transferred into SEB-pretreated mice to enter cell division was also reflected by the lower percentage of TCRBV8+ T cells among CFSE+ CD4+ cells after SEB stimulation (23.5 ± 3.7% cells in 3× SEB-treated mice vs 30.4 ± 4.5% in untreated mice). Thus, the observation that SEB-specific T cells transferred into SEB-pretreated recipients have a lower capacity to proliferate in response to SEB is consistent with the idea that SEB-unresponsive spleen cells exert an immune regulatory function on SEB-specific primary T cell response.
We have shown that spleen cells from BALB/c mice rendered tolerant by chronic exposure to SEB do not suppress SEA-specific primary response (21), suggesting that suppression observed in this model is indeed promoted by SEB-specific T cells. To verify this hypothesis, CD4+ SEB-specific TCRBV8+ and SEB-nonspecific TCRBV8− T cells were purified from the spleen of SEB-treated BALB/c mice and tested in vitro for their suppressive activity. As depicted in Fig. 2 A, only CD4+ TCRBV8+ T cells suppressed the primary response of normal BALB/c spleen cells, whereas adding CD4+ TCRBV8− T cells to primary cultures did not inhibit responsiveness to SEB. Again, suppression affected cell proliferation, as well as the secretion of IL-2 and IFN-γ. On the contrary, IL-10 production was detected in cultures containing CD4+ TCRBV8+ T cells. Taken together, these results demonstrated that the suppressive activity of unresponsive spleen cells was mediated by SEB-specific CD4+ T cells.
The suppressive activity of SEB-unresponsive spleen cells is mediated by unconventional CD4+CD25−CD152low T regulatory cells
We have previously reported that the spleen of BALB/c mice repeatedly injected with SEB contains a subpopulation of SEB-specific CD4+ TCRBV8+ T cells that over-express CD152 (CTLA-4) (21). Because CD152 was described recently to be preferentially expressed by natural CD4+ T regulatory cells (3, 10, 22), we wished to determine whether CD152+CD4+ TCRBV8+ T cells present in the spleen of SEB-treated mice could mediate the suppression of primary response to SEB. However, because immunostaining for CD152 requires cell membrane permeabilization, sorting viable cells according to their expression of CD152 was not possible. Nevertheless, it is known that in addition to CD152, naturally occurring CD4+ T regulatory cells also express the α-chain of the IL-2 receptor (1, 2), namely CD25, and it is possible that CD152+CD4+ TCRBV8+ T cells present in 3×-SEB-treated mice could also express this molecule. Therefore, we analyzed the expression of CD25 and CD152 on the surface of CD4+ TCRBV8+ T regulatory cells. As shown in Fig. 2,B, all CD152+CD4+ TCRBV8+ T cells isolated from the spleen of mice repeatedly injected with SEB expressed CD25. The cell population could be seen up to 10 days after the last injection of SEB; CD152 over-expression was not detected thereafter (data not shown). By contrast, CD152+CD25+ cells were not present in CD4− TCRBV8+ T cells nor among SEB-nonreactive TCRBV6+ T cells from SEB-treated mice (Fig. 2,B). They were also absent from the spleen of untreated animals (Fig. 2 B).
To determine the exact phenotype of the suppressor cell population, CD4+ T cells purified from the spleen of SEB-treated BALB/c mice were sorted on their expression of CD25 and compared for their suppressive activity with cells isolated from unimmunized animals. As depicted in Fig. 3 and consistent with the function of CD4+CD25+ T regulatory cells, adding CD4+CD25+ T cells from unimmunized mice to cultures of SEB-stimulated BALB/c whole spleen cells inhibited the production of IL-2 in a dose-dependent manner. Cell proliferation was moderately suppressed. On the contrary, in agreement with their effector function, CD4+CD25− T cells from unimmunized mice increased cell proliferation as well as the amount of IL-2 detected in the mixed cultures. IFN-γ production was not modified by addition of CD4+CD25+ or CD25− T cells purified from unimmunized mice, confirming previous observations that CD4+ T cells are not the main source of IFN-γ in primary T cell response to SEB (24). To our surprise, CD4+CD25−, not CD25+, T cells purified from animals tolerant to SEB could suppress effectively in vitro primary response to SEB (Fig. 3, A and B). This was observed not only for cell proliferation, but also for cytokine secretion (IL-2 and IFN-γ). On the contrary, adding CD4+CD25+ T cells to primary T cell cultures dramatically increased both cell proliferation and production of IFN-γ following SEB stimulation (Fig. 3,B). The production of IL-2 in these cultures, however, was not modified (Fig. 3 B). Thus, these observations demonstrated the presence of unconventional CD4 T regulatory cells in the spleen of mice made unresponsive to SEB that could control proliferation, as well as cytokine production, of primary SEB-specific T cells. Unlike natural CD4+ T regulatory cells, these cells did not express CD25 and were CD152-negative.
It has been recently shown that the transcription factor Foxp3 is a crucial regulator of immune cell function and is required to generate CD4+CD25+ T regulatory cells (25, 26). To assess whether Foxp3 expression was associated with the induction of CD4+CD25− T regulatory cells, we compared the abundance of the Foxp3 transcript in CD4+CD25− and CD4+CD25+ T cells derived from untreated or SEB-unresponsive mice. Strong expression was detected in CD4+CD25+ T cells isolated from both types of mice (Fig. 4). By contrast, RNA from CD4+CD25− T cells isolated from normal or unresponsive animals did not contain Foxp3 transcripts. This confirmed that Foxp3 was specific to the CD4+CD25+ T cell lineage (25, 26) and was not induced in CD4+CD25− T cells. Thus, the generation and activity of SEB-induced CD4+CD25− T regulatory cells appears to be independent of Foxp3 expression.
One surprising observation was the incapacity of SEB-specific CD4+CD25+ T cells from unresponsive animals to suppress primary T cell response to SEB. In normal BALB/c mice, these cells exert a strong IL-10-dependent control of the burst of SEB-induced production of cytokine, including IL-2 (11). Thus, it appears that, in mice chronically injected with SEB, while SEB-specific CD4+CD25− T cells have gained immune regulatory function, SEB-specific CD4+CD25+ T cells might have lost, or modified, their capacity to regulate type 1 T cell response. This assumption was investigated in assays involving T cell stimulation by CD3 mobilization. As shown in Fig. 3,C, whereas CD4+CD25− T cells, consistent with their immune regulatory capacity, were good producers of IL-10, CD4+CD25+ T cells secreted large amounts of IFN-γ (>0.05 μg/ml) in response to anti-CD3 stimulation (Fig. 3,C). This situation contrasted with that observed when cells from unimmunized controls were stimulated with anti-CD3 and where CD4+CD25+ T cells did not produce IFN-γ and secreted more IL-10 than CD4+CD25− T cells (Fig. 3 C).
CD4+CD25+ T cells control the induction phase of SEB-specific unresponsiveness
Recently, CD4+CD25− T cells exerting immune regulatory function have been described in two human reports (27, 28). Their suppressive activity was shown to be cell contact-independent and partially mediated by IL-10 and soluble TGF-β (27, 28). Remarkably, others and we have reported that suppression of primary T cell response by SAg-specific T cells isolated from mice receiving repeated injection of SAg is mediated, although not entirely, by IL-10 and TGF-β (20, 21). These observations suggested that the CD25− T suppressor cells identified in our study might represent the in vivo mouse counterpart of the human CD4+CD25− T suppressor cells described recently in vitro. Human CD4+CD25− T cells have been shown in vitro to acquire their suppressive function from CD25+ regulatory T cells via cell-cell contact (27, 28). To analyze whether in vivo development of mouse CD4+CD25− T cells depended on the activity of CD4+CD25+ T cells, BALB/c mice, thymectomized and depleted in CD25+ cells by treatment with anti-CD25 Abs (Fig. 5,A), were submitted to tolerogenic SEB treatment (3× 10 μg of SEB). As depicted in Fig. 5,B, CD25+ cell-depleted mice suffered toxic shock syndrome and did not survive SEB treatment. On the contrary, undepleted thymectomized animals survived indefinitely (Fig. 5 B) and became tolerant to SEB as their spleen cells could suppress in vitro primary T cell response to SEB (data not shown). Thus, tolerance induction by chronic exposure to SEB requires the presence of CD4+CD25+ T cells in the periphery.
SAg-specific CD4+CD25− T regulatory cells mediate their suppressive function independently of CD4+CD25+ T cells
The next question we addressed was to determine whether CD4+CD25− T suppressor cells could mediate suppression of primary T cell response in the absence of CD4+CD25+ T cells. Thymectomized CD25+ cell-depleted BALB/c mice were reconstituted with normal or CD25+ cell-depleted spleen cells or with purified CD4+CD25− T cells isolated from untreated or tolerant donors. One day after reconstitution, the animals were given i.p. repeated injection of SEB. As seen in Fig. 5,C, splenocytes from normal mice did not protect from SEB-induced death, regardless of the presence of CD25+ cells in the cellular inoculum. On the contrary, animals receiving cells isolated from tolerant donors survived indefinitely to repeated SEB injection (Fig. 5,C). Transfer of unresponsiveness in this system was strictly independent of CD25+ T cells because CD25− spleen cells from unresponsive mice were as efficient as whole spleen cells to suppress SEB-induced toxic shock in CD25+ cell-depleted mice (Fig. 5,C). Collectively, these data strongly supported the concept that CD4+CD25− T regulatory cells mediate their suppressive function independently of CD4+CD25+ T cells. This was also demonstrated in experiments involving reconstitution of thymectomized CD25+ cell-depleted mice with CD4+CD25− T cells purified from the spleen of tolerant mice. As seen in Fig. 5 C, transfer of purified CD4+CD25− T cells indeed suppressed SEB-induced toxic shock in the absence of CD25+ cells.
The removal of recent thymic emigrants reveals the presence of CD152− T effector cells as well as CD152+ T regulatory cells in CD4+CD25+ TCRBV8+ T cells from mice made unresponsive to SEB
Our observation that CD4+CD25+ T cells isolated from tolerant animals failed to regulate T cell primary response to SEB (Fig. 3,B) raised the possibility that this cell population was heterogenous and could contain effector as well as regulatory cells. In T cell response to SEB, effector function, as defined by the capacity to proliferate and to secrete IFN-γ, mostly results from the stimulation of nonmemory cells (29). Thus, in our protocol of repetitive injection of SEB, it was anticipated that the suppressive activity of CD4+CD25+ T cells from unresponsive mice could be revealed only after the removal of recent thymic emigrants entering the spleen between two injections of SEB. To investigate this possibility, BALB/c mice were thymectomized and 2 wk later they were given three injections of SEB. In contrast with CD4+CD25+ TCRBV8+ T cells from euthymic BALB/c mice that received repeated injection of SEB (Fig. 2,B), the proportion of cells that expressed high levels of CD152 was much higher (>30%) in CD4+CD25+ TCRBV8+ T cells isolated from SEB-treated thymectomized mice (Fig. 6,A). Indeed, CD4+CD25+CD152− TCRBV8+ T cells were virtually absent in these animals (Fig. 6,A). Remarkably, in the absence of CD4+CD25+CD152− cells, CD4+CD25+CD152+ cells from tolerant thymectomized mice were as efficient as CD4+CD25− T cells for the suppression of primary response to SEB (Fig. 6 C). Taken together, these results demonstrated that CD4+CD25+ cell failure to control primary T cell response in nonthymectomized tolerant mice was due to the presence of CD4+CD25+CD152− effector cells in the suppressive cell population.
Interestingly, 1 wk after depleting CD25+ T regulatory cells by anti-CD25 mAb treatment (30), the spleen of thymectomized BALB/c mice contained a subpopulation of CD4+ T cells that expressed low levels of CD25 and were CD152-negative (Fig. 6,A). When assayed for suppressive function, these CD4+CD25+ T cells were unable to control the production of IL-2 and IFN-γ in primary response to SEB (Fig. 6,C). Moreover, unlike nondepleted thymectomized mice, repetitive injection of lower sublethal dose of SEB (5 μg per injection) induced a massive increased number of SEB-reactive CD4+ TCRBV8+ T cells in the spleen of CD25+ cell-depleted animals (Fig. 6,B). These expanded cells were anergic, as they exhibited reduced proliferation and IL-2 secretion following SEB challenge (data not shown), and did not express CD25 (Fig. 6,A). Like CD4+CD25− T cells issued from nondepleted thymectomized mice, they were capable of inhibiting the secretion of IL-2 in SEB-stimulated primary cultures (Fig. 6,C). However, the cells could only marginally control the production of IFN-γ (Fig. 6 C). Taken together, these results demonstrated that in the absence of CD25+ T regulatory cells, SEB-reactive CD4+ T cells could differentiate into CD25− T suppressor cells. However, our in vivo and in vitro data also showed that this differentiation was more effective in the presence of CD25+ inducer T regulatory cells.
Because SAg can bind to specific TCRBV fragments, SAg-specific T cells can easily be identified by immunocytometry. This property has been successfully used in the past to study mechanisms of tolerance induction. Indeed, thymic deletion events required for shaping a mature T cell repertoire were first described in models using SAg as deleting agent (31, 32). Moreover, T cell unresponsiveness induced by exposure to SAg has long been considered as the only in vivo model to study T cell anergy (33). Our model is the first non-TCR-transgenic in vivo system where T cells responsible for suppression can be precisely identified and studied.
We show in this study that, on the one hand, CD4+CD25− SAg-specific T cells isolated from unresponsive mice can exert a regulatory activity on primary T cell response to SEB. On the other hand, we observed that CD4+CD25+ T cells are required to induce an efficient SEB-specific unresponsiveness. Thus, whereas CD4+CD25+ T cells act as “inducer” T suppressor cells for the establishment of SEB-induced tolerance, CD4+CD25− “effector” T suppressor cells can only exert their regulatory activity during the maintenance phase of tolerance.
Our study is not the first one to report the regulatory activity of CD4+CD25− T cells and in many situations CD4+CD25− T cells are as effective as CD4+CD25+ T cells in controlling T cell-mediated diseases. The development of experimental autoimmune encephalomyelitis in TCR-transgenic mice can be efficiently controlled by CD4+ T cells depleted of CD25+ cells (6). In a TCR-transgenic model of spontaneous diabetes, disease is prevented by CD4+CD25+ or CD4+CD25− T cells in a similar dose-response protection effect (34). In another model where mice express simultaneously a TCR transgene and its agonist ligand, Apostolou et al. (35) also showed that both T cell subsets had identical capacity to block naive T cell proliferation in vitro and in vivo. In nontransgenic rats, Stephens and Mason (36) observed that both peripheral CD25+ and CD25−CD4+ T cells could prevent the induction of autoimmune diabetes. In transplantation, Graca and colleagues (15) showed that, although CD4+CD25+ T cells did have a suppressive role in transplantation tolerance, so did CD4+CD25− T cells.
Nevertheless, the vast majority of models studying regulatory T cells have clearly established the critical role played by CD4+CD25+ T cells in the control of T cell response (37). The reasons for these discrepancies regarding the segregation of regulatory activity with the CD25 marker are not clear. We propose that the different frequency of regulatory T cells in the various experimental systems may account for the different outcome. In the Stephens and Mason (36) model previously described, thymectomy eliminated the pathogenic CD25− recent emigrants, increasing the proportion of CD25− T cells with regulatory function. Moreover, most models acknowledging the suppressive function of CD4+CD25− T cells are TCR transgenic ones (6, 34, 35). Immunomodulation applied in these systems could induce regulatory activity in a large number of Ag-specific T CD4+CD25− cells. Similarly, modulation by repeated injection of SAg could also increase the number of SAg-specific regulatory CD25− T cells in the periphery. Alternatively, activation induced cell death previously reported to occur after stimulation by SAg (33) may selectively remove effector cells from the CD25− population, thereby unmasking regulatory cell activity. Our concept has one important implication, i.e., CD4+CD25− regulatory T cells can be derived from precursors by appropriate conditioning stimulation as demonstrated recently by Roncarolo and colleagues (38, 39). In these studies, human CD4+CD25− T cells could be rendered anergic by IL-10 and differentiated into T regulatory cells (Tr1) by repeated activation in the presence of IL-10 and IFN-α.
On the contrary, most studies on transplantation tolerance identified CD4+CD25+ T cells as the only cell subset with regulatory function (12, 13, 14). This might be explained by the low frequency of alloantigen-specific T regulatory cells within the CD25− cell compartment in animals carrying long-term surviving grafts. Supporting this concept is the observation made by Graca et al. (15) that CD4+CD25− T cells from tolerant recipients are required in much larger number than CD4+CD25+ T cells to transfer transplantation tolerance.
One intriguing observation made in our study was that CD4+CD25+ T cells purified from animals rendered tolerant to SEB were apparently unable to suppress primary T cell response to SEB in vitro. Moreover, adding purified CD4+CD25+ T cells to primary culture exacerbated IFN-γ production. This observation contradicts the commonly accepted view that active suppression results from the direct activity of CD4+CD25+ T regulatory cells. One possible explanation is that SEB-specific CD4+CD25+ T cells have lost or modified their regulatory capacity after repeated exposure to SEB. It is also possible that they are heterogeneous and contain effector as well as regulatory T cells, thereby masking the effect of regulatory populations within. This former possibility is supported by the recent work of Levings et al. (39) on the heterogeneity of human clones generated by repeated anti-CD3 stimulation of purified CD4+CD25+ T cells. In these conditions, about half of the clones (20 of 44) inhibited the proliferation of naive CD4+ T cells. FACS analysis also revealed that there was an absolute correlation between constitutive expression of CD152 (CTLA-4) and suppressive function. Interestingly, CD25+CD4+ TCRBV8+ T cells from BALB/c mice repeatedly injected with SEB contained both CD152+ and CD152− cells, at about one to one ratio (Fig. 2). Using thymectomized animals to remove CD4+CD25+CD152− T cells, we were able to unmask the regulatory function of CD4+CD25+CD152+ T cells present in the spleen of SEB-unresponsive mice (Fig. 6). Therefore, it appears that, in mice tolerant to SEB, CD4+CD25+CD152+ T regulatory cells are unable to control CD4+CD25+CD152− T effector cells. One possible explanation for this failure could be that the suppressive function of CD25+ T regulatory cells would largely depend on direct cell contact (40). This might not be sufficient enough when the activity of a large number of Ag-primed effector cells has to be controlled.
Several previous studies have associated the suppressive function of CD4+ T regulatory cells with soluble or membrane-bound regulatory cytokines, such as IL-10 and TGF-β (41, 42, 43). However, CD25+ T cells from IL-10−/− and TGF-β−/− mice were shown to be fully competent suppressors (44, 45). It remains possible that regulatory cytokines may not be produced by CD25+ T cells themselves, but are produced by other cells, including CD4+CD25− T cells, as a result of interaction with the suppressor cells. As previously mentioned, recent reports of in vitro human studies propose that one of the properties of CD4+CD25+ T regulatory cells (and it might be the only one) is to convey suppressor activity to CD4+CD25− Th cells (27, 28). Hence, it is tempting to speculate that the acquisition of regulatory function by SAg-specific CD4+CD25− T cells also depends on the activity of CD4+CD25+ T regulatory cells. Our study indeed provides direct in vivo evidence to support this concept because the depletion of CD4+CD25+ T cells prevented the development of SAg-specific unresponsiveness, and consequently, the generation of CD4+CD25− T regulatory cells after repeated injection of SEB. Induction of CD4+CD25− T regulatory cells by CD4+CD25+ T cells has been reported to depend on cell-cell contact, to be cytokine-independent and to induce high level of IL-10 in CD4+CD25− T cells (27, 28). The resulting IL-10-producing T cells are then able to suppress T cell proliferation in an IL-10-dependent fashion (27). The observations that 1) IL-10 is produced by SEB-specific CD4+CD25− T regulatory cells (present study) and 2) suppression of primary response toward SEB by adoptive transfer of unresponsive spleen cells can be abrogated by injection of neutralizing anti-IL-10 Abs (21) suggest that IL-10 might play a major role in CD25− T cell-mediated immunosuppression of SEB-specific response.
One of the cytokines that has been closely linked to tolerance induction is IL-2. IL-2 has been known for some time for its crucial role in activation-induced cell death of T lymphocytes (46). Its role in the generation or function of CD4+CD25+ T regulatory cells is now emerging (37). Recently, in a TCR-transgenic mouse model where full protection against spontaneous experimental autoimmune encephalomyelitis could be achieved by the transfer of wild-type CD4+CD25− T cells, Furtado et al. (47) showed that responsiveness to IL-2 was required for the suppressive function. This implied that CD25-negative regulatory cells could respond to IL-2 after induction of CD25 expression and through the transient formation of high-affinity IL-2R αβγ complexes. Whether responsiveness to IL-2 is required for the development or function of SAg-specific CD4+CD25− T regulatory cells is unknown. However, strong and ephemeral expression of CD25 is known to occur after stimulation of anergic SAg-specific T cells in mice repeatedly injected with SAg (48).
We believe that the understanding of the mechanisms leading to the emergence of T regulatory cell populations after repeated exposure to bacterial SAg could bring important information on how infectious agents modulate the immune system of their host. This could have important implications in the control of autoimmunity as well as for the induction of tolerance in transplantation.
During the review process of this study, Grundstrüm et al. (49) reported that both CD4+CD25+ and CD4+CD25− SAg-specific T cells isolated from TCR-transgenic mice treated with SEA three consecutive times could efficiently control primary response to SAg stimulation. Both types of regulatory cell were reported to express elevated levels of CD152, suggesting that both subsets could use the same effector mechanism of suppression. This was not observed in our study. These discrepancies might result from the use of different T cells in the experiments, i.e., normal T cells for our studies and transgenic T cells expressing high level of TCR in Grundstrüm et al. (49) work.
We are very thankful to Dr. Oberdan Leo for his critical review of our manuscript.
This work was funded by the Fonds National de la Recherche Scientifique of Belgium and the Commission of the European Union. P.F. is a Research Fellow of the Fonds National de la Recherche Scientifique of Belgium. L.P. is supported by the Fonds pour la Formation à la Recherche dans l’Industrie et dans l’Agriculture of Belgium.
Abbreviations used in this paper: SAg, superantigen; SEB, staphylococcal enterotoxin B.