Modulation of Fas-Dependent Apoptosis: A Dynamic Process Controlling Both the Persistence and Death of CD4 Regulatory T Cells and Effector T Cells¹

Abnormal control of immune responses can lead to autoimmune and inflammatory diseases. Three mechanisms control autoreactive and activated T cells, namely clonal deletion, clonal anergy, and active suppression by regulatory CD4⁺ T cell subsets (1, 2, 3, 4, 5). CD4⁺ T cells, which constitutively express the IL-2R α-chain (IL-2Rα or CD25), participate in this regulation. Indeed, elimination or inactivation of these regulatory T cells is associated with the onset of autoimmune and inflammatory diseases and with abnormal peripheral T cell homeostasis (3, 4, 5, 6, 7).

Regulatory CD25⁺CD4⁺ T cells differ functionally from CD25⁻CD4⁺ T cells, although the latter can acquire CD25 upon activation (8, 9). In particular, CD25⁺CD4⁺ T cells differ from their CD25⁻ counterparts by their cytokine profile. CD25⁺CD4⁺ T cells produce IL-10, an anti-inflammatory cytokine (10), but do not produce IL-2, making them dependent on exogenous IL-2 for expansion in vivo and in vitro (8, 9, 11, 12, 13). Conversely, CD25⁻CD4⁺ T cells produce IL-2 but not IL-10. Regulatory CD25⁺CD4⁺ and CD25⁻CD4⁺ T cells also differ by their susceptibility to clonal deletion. We have previously shown that CD25⁺CD4⁺ T cells are resistant to clonal deletion induced by viral superantigen (vSAG)³ in vivo (11). Although both CD25⁺CD4⁺ and CD25⁻CD4⁺ superantigen (SAG)-specific T cells were activated by vSAG, clonal deletion occurred at the expense of the CD25⁻CD4⁺ subset while the CD25⁺CD4⁺ T cell subset was spared.

CD25⁺CD4⁺ T cell resistance to clonal deletion may be important in maintaining a permanent population of regulatory T cells that can control and eventually suppress autoreactive T cells or harmful T cells activated during normal immune responses. Several experimental studies have shown that this regulatory population is needed to avoid autoimmune and inflammatory diseases. Mice rendered deficient in CD25⁺CD4⁺ T cells by neonatal thymectomy and nude mice reconstituted with CD25⁺-depleted T cells develop autoimmune diseases. Transfer of CD25⁺CD4⁺ T cells from normal mice can avoid the onset of autoimmune disease in these models (14, 15). CD4⁺ T cells expressing the CD45RB^low phenotype (which overlap with CD25⁺CD4⁺ T cells) control the inflammatory diseases that occur in mice transferred with CD45RB^highCD4⁺ T cells (16). IL-2Rα knockout (KO) and IL-2 KO mice, which cannot support CD25⁺CD4⁺ T cell expansion (12), develop inflammatory bowel disease and lymphadenopathy, suggesting that regulatory CD25⁺CD4⁺ T cells may also participate in the control of peripheral T cell homeostasis (17, 18). It was recently shown that regulatory CD45RB^lowCD4⁺ T cells can control the expansion of CD45RB^highCD4⁺ T cells in vivo (19). The mechanisms by which CD25⁺CD4⁺ T cells control immune responses are unclear. They may differ according to the model and the functions analyzed, and also according to the CD25⁺CD4⁺ T cell subset engaged. Various cytokines are also involved, of which IL-10 and TGF-β seem to be the most efficient (2, 20, 21, 22, 23). However, cytokines may participate to different degrees in different models and may also modulate different functions. IL-10 KO mice raised in a conventional environment develop inflammatory bowel disease but not lymphadenopathy, suggesting that IL-10 is involved not in T cell homeostasis but in the control of inflammatory reactions. The role of IL-10 in the suppression of T cell proliferation in vitro has not been demonstrated, but cell-cell interactions are known to be required (8, 13), and these may involve local cytokine production.

If CD4⁺ regulatory T cells are mandatory to avoid autoimmune and inflammatory diseases, and are also necessary to control T cell homeostasis, their resistance to apoptosis may underlie their persistence during ontogenesis and immune responses. By using polyclonal activation of purified regulatory CD25⁺CD4⁺ and CD25⁻CD4⁺ T cells, we found that the CD25⁺ subset is more resistant than the CD25⁻ subset to Fas-induced apoptosis. This is in keeping with our previous observation that CD25⁺CD4⁺ vSAG-specific T cells are resistant to clonal deletion in vivo. We then investigated whether CD25⁺CD4⁺ T cells could be sensitized to Fas-mediated apoptosis by contact with IL-2-producing CD25⁻CD4⁺ T cells in vitro. We found that sensitivity to apoptosis could be modulated in both subsets. This was associated with modulation of cytokine production. However, whereas CD25⁻CD4⁺ T cell apoptosis is highly dependent on the IL-2, the response of CD25⁺CD4⁺ T cells to Fas-induced apoptosis in vitro is IL-2 independent. The results suggest that regulation of apoptosis is a dynamic process that depends on the balance between CD25⁺CD4⁺ and CD25⁻CD4⁺ T cell numbers and is required for regulatory and effector T cell homeostasis.

C57BL/6 (H-2^b, Thy 1.2) mice were from CERJ (Le Genest, St. Isles, France), IL-2 KO (H-2^b, Thy 1.2) (24) and IL-10 KO (25) mice were from Transgenic Alliance (L’arbresle, France), and C57BL/Ba (H2^b, Thy 1.1) mice were from our own facilities. All mice were 6–10 wk old when used.

The following Abs were used for flow cytometry and cell culture. Purified anti-CD8 (clone 53-6.7), purified anti-CD3 (clone 145-2C11), biotinylated anti-CD25 (clone PC61), biotinylated anti-CD4 (clone RM4-5),and FITC anti-TCR (clone H57-597) were prepared and coupled in our laboratory. Rabbit anti-hamster IgG was from Jackson ImmunoResearch Laboratories (West Grove, PA). PerCP-anti-CD4 (clone RM4-5), PE-anti-CD69 (clone H1.2F3), PE-anti-Fas (clone JO₂), PE-anti-Thy1.2 (clone 53-2-1), purified anti-Fas, and purified hamster IgG were from BD PharMingen (San Diego, CA). Flow cytometry was performed using aFACSCalibur device (BD Biosciences, Mountain View, CA) and data were analyzed with CellQuest software (BD Biosciences).

Lymphocyte suspensions were prepared from peripheral (inguinal, axillary, cervical, and popliteal) lymph nodes. CD4⁺ T cells were prepared by negative selection. Briefly, lymph node cells were incubated with purified anti-CD8 and then with goat anti-rat Ig-coated magnetic beads (Dynabeads; Dynal Biotech, Oslo, Norway). B cells were depleted with goat anti-mouse Ig-coated Dynabeads. CD8⁺ T cells and B cells coated with beads were removed with a magnet. CD25⁺CD4⁺ T cells were positively selected. Briefly, the enriched CD4⁺ T cells were coated with biotinylated anti-CD25 and mixed with MACS Streptavidin Microbeads (Miltenyi Biotec, Paris, France). Cells coated with microbeads were selected by the magnetic field of a MACS separator (Miltenyi Biotec). Unlabeled (CD25⁻) cells were negatively selected and CD25⁺ cells were then eluted from the column after removing the magnetic field. The CD25⁺CD4⁺ and CD25⁻CD4⁺ populations were 90 and 95% pure, respectively.

Purified CD25⁺CD4⁺ and CD25⁻CD4⁺ lymph node T cells (2 × 10⁵ cells per well) were cultured for 70 h at 37°C in 5% CO₂-air in anti-CD3-coated 96-well culture plates, with or without 50 U/ml mouse rIL-2 (BD Biosciences). Plates were coated overnight with anti-CD3 (10 μg/ml; 50 μl per well). Plate-bound anti-CD3 was used for activation in the absence of APC, as APC was shown not to be mandatory for regulatory T cell function (26). All cells were grown in 200 μl of RPMI 1640 medium supplemented with 10% FCS (Life Technologies, Grand Island, NY), 100 U/ml penicillin, 100 μg/ml streptomycin, 1% sodium pyruvate (Biomedia, Foster City, CA), 5 mM HEPES (Life Technologies), and 10⁻⁵ M 2-ME. At the end of the culture period, the absolute number of living cells per well was measured with the trypan blue exclusion test. In some experiments, indicated numbers of CD25⁺CD4⁺ T cells from C57BL/6 (B6) or C57BL/Ba (Ba) mice were cultured alone or with 2 × 10⁵ purified CD25⁻CD4⁺ T cells from Ba or IL-2 KO mice. The two subsets were distinguished by their different Thy-1 markers. Pooled supernatants from several wells in each group were stored at −20°C for cytokine assay.

T cells (1 × 10⁶) activated in vitro were incubated for 30 min at 37°C in 5% CO₂-air with 1 μg/ml hamster anti-Fas Ab or hamster IgG as control. Cells were then washed and cultured for 1, 2, 4, or 20 h in 96-well plates (2 × 10⁵ cells per well) with 1 μg/ml goat anti-hamster IgG. At each time point, anti-Fas-induced apoptosis was measured with the annexin V test (Boehringer Ingelheim, Gagny, France). Cells were counted, washed in PBS/10% FCS, and stained with biotinylated anti-CD4, PE-anti-CD25, and streptavidin-allophycocyanin. Cells were then washed twice and stained with FITC-labeled annexin V according to the manufacturer’s instructions. The rate of apoptosis was determined by flow cytometry. Dead cells were excluded after labeling with 25 μg/ml propidium iodide, and annexin-positive cells were detected among CD25⁺CD4⁺ and CD25⁻CD4⁺ T cells. In mixed cultures of CD25⁻CD4⁺ and CD25⁺CD4⁺ T cells, the two subsets were distinguished by labeling with PE-conjugated anti-Thy1.2. In some experiments, hypodiploid nuclei (sub-G₁ peak) were detected by propidium iodide staining (27). Briefly, 2 × 10⁵ cells were fixed in 70% cold ethanol for 30 min, washed, and incubated with 10 μg/ml propidium iodide and subjected to flow cytometry analysis.

IL-2, IFN-γ, TNF-α, and IL-10 were assayed in culture supernatants with ELISA kits (R&D Systems, Abingdon, U.K.) according to the manufacturer’s recommendations. Results are expressed as picograms per milliliter of supernatant.

Data are expressed as means ± SEM. Comparisons were made using Student’s t test and differences were considered significant if p < 0.05.

Purified CD25⁺CD4⁺ and CD25⁻CD4⁺ lymph node T cells were activated with anti-CD3 for 70 h in vitro, then washed and counted before incubation (2 × 10⁵ living cells per well) with hamster anti-Fas or an isotype control Ab for 30 min. The cells were incubated for 1–20 h with anti-hamster IgG to determine the time of maximum apoptosis. The percentage of apoptotic cells was measured by two different techniques: the FITC-labeled annexin V method, which detects cells at an early stage of the apoptotic process, and the intracellular propidium iodide staining, which measures hypodiploid nuclei (sub-G₁ peak). As shown in Fig. 1,A, the highest percentage of annexin V-positive apoptotic CD25⁻CD4⁺ T cells was observed 1 h after Fas cross-linking (29% of CD25⁻CD4⁺ cells were annexin V positive and PI negative). When the Fas cross-linking period was increased from 1 to 20 h, the percentage of apoptotic CD25⁻CD4⁺ T cells fell gradually, possibly owing to gradual deletion of Fas-sensitive cells. In contrast, only 9% of CD25⁺CD4⁺ T cells were apoptotic 1 h after Fas cross-linking. No apoptotic cells were detected in this population when Fas cross-linking was extended up to 20 h. Results were confirmed using a second apoptosis test (Fig. 1 B). The percentage of apoptotic CD25⁻CD4⁺ T cells with hypodiploid nuclei was clearly higher (37%) than the percentage of apoptotic CD25⁺CD4⁺ T cells (7%) 1 h after Fas cross-linking. The percentage of CD25⁻CD4⁺ T cells with hypodiploid nuclei remained high up to 6 h after Fas cross-linking (data not shown). In subsequent experiments, Fas-dependent apoptosis was tested 1 h after Fas cross-linking with the annexin V method.

Regulatory CD25⁺CD4⁺ T cells proliferate in response to anti-CD3 in vitro, but only if IL-2 is added (8, 9, 12, 13). Exogenous IL-2 is also crucial for programmed cell death, enhancing sensitivity to Fas-dependent activation-induced cell death (AICD) (17, 28, 29, 30, 31, 32). Furthermore, apoptosis is linked to cell proliferation (33). However, the response to Fas cross-linking was unaffected when purified CD25⁺CD4⁺ and CD25⁻CD4⁺ lymph node T cells were first activated for 70 h with exogenous IL-2 (Fig. 1 A).

Because the kinetics of Fas-dependent apoptosis could differ in the two subsets with the duration of the activation period, we studied susceptibility to apoptosis earlier during activation by anti-CD3, in the presence and absence of IL-2 (Fig. 2). When the activation period was increased from 24 to 70 h, the percentage of apoptosis increased among CD25⁻CD4⁺ cells 1 h after Fas cross-linking. In contrast, CD25⁺CD4⁺ T cells remained resistant to apoptosis. Similar results were obtained when IL-2 was added during the activation period.

Thus, most CD25⁺CD4⁺ T cells were resistant to Fas-dependent apoptosis, and exogenous IL-2, added during the activation period, did not overcome this resistance.

Resistance to Fas-dependent apoptosis could be linked to a lack of activation or Fas expression following activation. Purified CD25⁺CD4⁺ and CD25⁻CD4⁺ lymph node T cells were incubated with anti-CD3 for 70 h in vitro, with and without IL-2, before measuring cell size and CD69 and Fas expression. Before activation, the cell size of the two populations was identical. In contrast, more CD25⁺CD4⁺ T cells than CD25⁻CD4⁺ T cells expressed the CD69 activation marker (26 vs 12%, respectively; Fig. 3,A). CD25⁺CD4⁺ T cells also expressed a slightly higher level of Fas protein on their surface. Cell size and CD69 expression increased in both subsets after 70 h of incubation with anti-CD3, although slightly less in CD25⁺ than in CD25⁻ cells (Fig. 3,B). When IL-2 was added during the activation period, CD25⁺CD4⁺ T cells proliferated (12) and cell recovery increased (Fig. 3,C) to the level observed with CD25⁻CD4⁺ T cells, which do not require exogenous IL-2 to proliferate. CD69 expression was identical when CD25⁻CD4⁺ T cells were activated with or without IL-2. CD69 expression was lower on CD25⁺CD4⁺ T cells activated with IL-2 than on CD25⁺CD4⁺ T cells activated without IL-2. However, the peak of CD69 expression on the CD25⁺ subset was maximum 48 h after activation with IL-2 (data not shown) and was already down-regulated after 70 h. Fas expression increased on both CD25⁺CD4⁺ and CD25⁻CD4⁺ T cells after anti-CD3 stimulation, in both the absence and presence of IL-2, although to a lesser degree on the CD25⁺ subset (Fig. 3 B).

Thus, despite the different sensitivity of CD25⁺CD4⁺ and CD25⁻CD4⁺ T cells to Fas-dependent apoptosis, both populations acquired an activated phenotype and expressed Fas on their surface after anti-CD3 stimulation.

CD25⁺CD4⁺ T cells constitutively express IL-10 mRNA (8, 11). It has been reported that IL-10 can prevent T cell apoptosis in vitro (34) and also protect T cells against AICD in vivo (35). Therefore, we investigated the possible involvement of IL-10 production by CD25⁺CD4⁺ cells in their resistance to Fas-induced apoptosis in vitro. Purified CD25⁺CD4⁺ and CD25⁻CD4⁺ T cells from B6 mice and IL-10-deficient mice were activated by anti-CD3 for 70 h before testing Fas-dependent apoptosis after 1 h of Fas cross-linking. As shown in Fig. 4, CD25⁺CD4⁺ T cells from both B6 and IL-10 KO mice were resistant to apoptosis, contrary to CD25⁻CD4⁺ T cells. Thus, IL-10 production by CD25⁺CD4⁺ T cells was not responsible for their resistance to Fas-dependent apoptosis in vitro.

Cell-cell contacts are necessary for the suppressor activity of regulatory CD25⁺CD4⁺ T cells in vitro (8, 13). Thus, we investigated whether the sensitivity of the two populations to Fas-dependent apoptosis was modified by mixing CD25⁺ and CD25⁻ cells in vitro. CD25⁺CD4⁺ and CD25⁻CD4⁺ T cells were cultured alone or together at ratios of 1:10 (the physiological steady-state ratio in vivo) and 1:1. Cells were incubated with anti-CD3 for 70 h in vitro. Following activation, cells were collected, washed, counted, and submitted for 1 h to Fas cross-linking. Cells initially selected as CD25⁺CD4⁺ or CD25⁻CD4⁺ T cells in Thy1.2⁺ and Thy1.2⁻ mice were detected in coculture by their respective expression of Thy1 (Fig. 5,A). The percentage of apoptotic CD25⁻CD4⁺ cells was high (20%), whether the latter were cultured alone or together with CD25⁺CD4⁺ T cells at a ratio of 10 CD25⁻ cells to 1 CD25⁺ cell (Fig. 5,B). In contrast, at a ratio of 1:1, most CD25⁻CD4⁺ T cells became resistant to Fas-induced apoptosis (7% of apoptotic cells). Thus, CD25⁺CD4⁺ T cells modulated the sensitivity of CD25⁻CD4⁺ T cells to apoptosis. At a 1:1 ratio, the percentage of apoptotic CD25⁺CD4⁺ T cells was not markedly different from that observed when CD25⁺CD4⁺ T cells were cultured alone (9 and 6%, respectively). However, at a ratio of one CD25⁺ cell to 10 CD25⁻ cells, the proportion of apoptotic CD25⁺CD4⁺ T cells increased to 23%. These results suggested that induction of Fas-dependent apoptosis could be modulated by interactions between the CD25⁺CD4⁺ and CD25⁻CD4⁺ T cell subsets, depending on the ratio between the two populations. To examine whether interactions between CD25⁻ and CD25⁺ cells during the activation phase were required to modulate Fas-induced apoptosis, each cell population was activated separately and then combined for 1 h during the apoptosis assay. Modulation of Fas-dependent apoptosis was no longer observed (Fig. 5 C). At a 1:1 ratio, CD25⁻CD4⁺ T cells remained sensitive to apoptosis, whereas, at a 1:10 ratio, CD25⁺CD4⁺ T cells remained resistant, as when cultured and tested separately. These results showed that the two cell subsets had to be combined during the activation phase to modulate their sensitivity to Fas-dependent apoptosis.

The ratio of CD25⁻CD4⁺ and CD25⁺CD4⁺ T cells during the activation phase might influence cytokine production and thus affect Fas-dependent apoptosis in vitro. Production of IL-2, which sensitizes CD4 T cells to Fas-dependent apoptosis (17, 28, 29, 30, 31, 32), is inhibited in a cell dose-dependent manner by interactions between CD25⁻CD4⁺ and CD25⁺CD4⁺ T cells (8, 13). Thus, we measured cytokine production during the activation phase (Fig. 6). Production of TNF-α in vitro was not significantly modified in the different culture conditions, while IFN-γ and IL-2 were decreased by cocultures in a cell dose-dependent manner, as previously shown (8, 36), compared with CD25⁻CD4⁺ T cells cultured alone. IL-10 production in the supernatants was increased by increasing the CD25⁺:CD25⁻ T cell ratio. However, we could not demonstrate the mandatory role of IL-10 in this system by using CD25⁺CD4⁺ T cells from IL-10 KO (data not shown).

To test the role of IL-2 in the increased sensitivity of CD25⁺CD4⁺ T cells to apoptosis, CD25⁻CD4⁺ T cells from IL-2 KO mice were cultured alone or with CD25⁺CD4⁺ T cells from normal Ba mice at the 10:1 ratio, then activated and submitted to Fas cross-linking. CD25⁻CD4⁺ cells from IL-2 KO mice were poorly sensitive to Fas-dependent apoptosis when cultured alone, as previously shown (28, 32). We then analyzed Fas-induced apoptosis in both subsets after mixed culture. As shown in Fig. 7 A, 24% of CD25⁺CD4⁺ T cells in mixed culture were apoptotic, suggesting that CD25⁻CD4⁺ cells could induce Fas-dependent apoptosis in CD25⁺CD4⁺ T cells through an IL-2-independent mechanism. Interestingly, sensitivity of the CD25⁻CD4⁺ subset from IL-2 KO mice to Fas-dependent apoptosis was also enhanced by this interaction, increasing from 5 to 15%.

Cytokine production in these culture conditions was measured after activation (Fig. 7 B). CD25⁻CD4⁺ T cells from IL-2 KO mice produced no measurable IL-2, significant amounts of TNF-α, and little IL-10 and IFN-γ. CD25⁺CD4⁺ T cells (0.2 × 10⁵) produced very low levels of each cytokine. The level of TNF-α was increased when CD25⁻CD4⁺ T cells from IL-2 KO mice were activated with normal CD25⁺CD4⁺ T cells at the 10:1 ratio (p < 0.02, activated together vs activated alone).

In conclusion, although modulation of apoptosis is associated with modulation of cytokine production, we cannot demonstrate the role of IL-10 or IL-2 in the modulation of the apoptosis sensitivity of CD25⁺CD4⁺ T cells.

Infectious mouse mammary tumor viruses (MMTV) encode a vSAG that activates vSAG-reactive T cells and then induces their clonal deletion. We have previously shown (37, 38) that the absolute number of vSAG-specific Vβ6⁺CD25⁺CD4⁺ T cells remains constant during infection by MMTV (SW), while the number of Vβ6⁺CD25⁻CD4⁺ T cells gradually falls. We concluded that CD25⁺CD4⁺ T cells were resistant to clonal deletion induced by vSAG activation in vivo (11). Other reports show the differential susceptibility of T cell subsets to AICD. The Th1 and Th2 T cell subsets are functionally distinct and are defined on the basis of their cytokine profiles. In vitro, Th1 effectors, contrary to Th2 effectors, undergo rapid Fas/Fas ligand-mediated AICD upon reactivation with Ag (39, 40). Resistance of Th2 clones to AICD is not due to a lack of Fas expression. The susceptibility of memory T cells to AICD also differs from that of naive T cells (41, 42). Primed T cells are more resistant than naive T cells to Fas-mediated AICD in vitro and in vivo. In the present study, we found that isolated CD25⁺CD4⁺ T cells were more resistant than their CD25⁻CD4⁺ counterparts to apoptosis through polyclonal activation by anti-CD3 and triggering of the Fas pathway. This resistance of isolated CD25⁺CD4⁺ T cells was not due to an inability to be activated or to express Fas. Indeed, anti-CD3 cross-linking in the absence of IL-2 up-regulated CD69 and Fas expression and increased the cell size of both populations. Fas expression was similar in the two populations.

CD25⁺CD4⁺ T cells do not produce IL-2, even after activation (8, 11, 13), and do not proliferate in response to anti-CD3 in the absence of IL-2 (8, 9, 12). Furthermore, IL-2 is involved in cell death by enhancing sensitivity to Fas-dependent AICD (29). T cells activated in vivo are not susceptible to Fas-dependent apoptosis when IL-2/IL-2R signaling is disrupted by a lack of IL-2 or its receptor, although Fas is expressed on their surface (17, 28, 30, 31, 32). Additionally, there is a relationship between proliferation and apoptosis: no apoptosis is detected among nonproliferating T cells following Ag stimulation (33). However, while addition of IL-2 to cultured CD25⁺CD4⁺ T cells induces their proliferation (8, 9, 12) and enhances their numbers in vitro (this study), it did not modify resistance to Fas-dependent apoptosis. Isolated activated CD25⁺CD4⁺ T cells were thus resistant to Fas-induced apoptosis independently of IL-2.

IL-2 also promotes T cell survival (43). CD25⁺CD4⁺ T cells show poor proliferation (8, 9, 12, 13) and poor survival when incubated in the absence of IL-2. Like human CD4⁺ T cells (44), these regulatory T cells are prone to nonspecific death in vitro, despite being resistant to specific Fas-mediated apoptosis. This susceptibility to cytokine privation in vitro may be explained by down-regulation of Bcl-2 or other members of the Bcl protein family (45, 46).

Resistance of CD25⁺CD4⁺ T cells to clonal deletion induced by vSAG in vivo (11) and to Fas-induced apoptosis after activation in vitro might also be dependent on IL-10, which is constitutively produced by these cells (8, 12). IL-10 protects T cells from apoptosis both in vitro, as shown for human T cells infected by EBV (34) and, in vivo, IL-10 KO mice injected with the bacterial SAG Staphylococcus aureus enterotoxin B display enhanced SAG-specific T cell clonal deletion in the thymus (35). However, we detected no difference between IL-10 KO and normal mice as regards Fas-induced apoptosis of isolated peripheral CD25⁺CD4⁺ and CD25⁻CD4⁺ T cells, suggesting that IL-10 is not involved in resistance, at least in vitro. It is possible that peripheral T cells compensate for their IL-10 deficiency by using another cytokine for Fas-induced apoptosis in vitro. TGF-β expressed by CD25⁺CD4⁺ T cells may be involved in protection from apoptosis, as both CD4⁺ and CD8⁺ T cell apoptosis is increased in mice lacking TGF-β (47, 48).

The suppressive function of regulatory CD25⁺CD4⁺ T cells is mediated in vitro by cell-cell contacts that lead to down-regulation of IL-2 production by the CD25⁻CD4⁺ subset (8, 13). Induction of Fas-mediated apoptosis might thus be regulated by such contacts. Interestingly, both the sensitivity of CD25⁻CD4⁺ T cells and the resistance of CD25⁺CD4⁺ T cells to Fas-dependent apoptosis were modified when the two populations were cocultured. The presence of both cell subsets during the activation phase is mandatory for this phenomenon, as mixing preactivated cells during the apoptotic assay alone is ineffective. Modulation depended on the ratio of the two cell subsets. At a ratio of 1:1, CD25⁻CD4⁺ T cells became resistant to apoptosis. CD25⁺CD4⁺ T cells regulated the activation and proliferation of CD25⁻ cells during the activation phase (data not shown) and thereby reduced the sensitivity of the latter to AICD. This is in keeping with previous reports showing that CD25⁺CD4⁺ T cells suppress IL-2 production and thus the proliferation of polyclonally activated CD25⁻CD4⁺ T cells in vitro (8, 13). To elucidate the mechanism involved in the modulation of apoptosis, we tested the role of IL-10 and/or IL-2 in this modulation. We found that coculture of CD25⁻CD4⁺ and CD25⁺CD4⁺ T cells at a 1:1 ratio totally inhibited both IL-2 and IFN-γ production and enhanced IL-10 production. However, we found that IL-10 privation did not modify the sensitivity of isolated CD25⁻CD4⁺ or CD25⁺CD4⁺ T cells to Fas-mediated apoptosis. An attempt to demonstrate the mandatory role of IL-10 in apoptosis regulation by using CD25⁺CD4⁺ T cells from IL-10 KO mice was inconclusive (data not shown). IL-10 is not the only suppressive factor used by CD25⁺CD4⁺ T cells for regulation (6). Alternative mechanisms may be used in the absence of IL-10 to counteract the effect of proapoptotic factors. Engagement of membrane-linked cytokines may be one of these mechanisms. Indeed, it has been suggested that membrane-bound TGF-β, expressed by CD25⁺CD4⁺ T cells, is involved in their regulatory function (47). TGF-β is also involved in apoptosis regulation (48), which may explain why cell to cell interaction is needed for both regulation and modulation of apoptosis (8, 13, 47).

Interestingly, these results are in keeping with our previous in vivo data showing that vSAG-induced clonal deletion spares not only the vSAG-specific CD25⁺CD4⁺ T cell subset but also a residual CD25⁻CD4⁺ T cell subset (11). The ratio of the two populations at the end of the clonal deletion process is 1:1. Resistance to apoptosis and its modulation may protect not only the regulatory suppressive population but also a small specific effector population, which may have a role in subsequent immune responses.

The question of why, when the ratio of the two subsets in coculture was 10 CD25⁻CD4⁺ T cells to 1 CD25⁺CD4⁺ T cell, the former remained sensitive and the latter became sensitive to apoptosis is intriguing. This inducible sensitivity of the regulatory T cell subset to apoptosis during immune responses may also be relevant to the general mechanisms controlling autoreactive T cells. Apoptosis of part of the regulatory population at some point of the immune response may be required to limit their control on the CD25⁻CD4⁺ T cell subset. Indeed, control of CD25⁻CD4⁺ clonal deletion by the CD25⁺CD4⁺ subpopulation has previously been shown in vivo (49). In mice infected at birth by MMTV (SW), depletion of CD25⁺CD4⁺ T cells by anti-CD25 Ab from birth onwards leads to more rapid and extensive death of vSAG-specific CD25⁻CD4⁺ T cells. This suggests that CD25⁺CD4⁺ T cells may hinder CD25⁻CD4⁺ T cell death in vivo. Deletion of autoreactive T cells might be able to proceed more rapidly when the number of CD25⁺CD4⁺ regulatory T cells is limited.

These results suggest that the sensitivity of both CD25⁺CD4⁺ and CD25⁻CD4⁺ T cells to Fas-dependent apoptosis can be modified during the course of an immune response, with the probable aim of maintaining an equilibrium between regulatory and effector T cells.

The mechanism by which CD25⁺CD4⁺ regulatory T cells become sensitive to apoptosis is not clear. CD25⁻CD4⁺ T cell IL-2 production did not explain the change in CD25⁺CD4⁺ T cell sensitivity to Fas-mediated apoptosis in coculture at a ratio of one CD25⁺ cell to 10 CD25⁻ cells, as CD25⁻CD4⁺ T cells from IL-2 KO mice had the same effect. Interestingly, CD25⁺CD4⁺ T cells correct the Fas-dependent cell death defect described in IL-2 KO mice (28, 32). The CD25⁺CD4⁺ T cell subset, which requires IL-2 to expand, is lacking in IL-2 KO mice (12). Thus, the abnormal AICD in these mice may be linked not only to the absence of IL-2 but also to the absence of IL-2-dependent CD25⁺CD4⁺ T cells. Interestingly, Wolf et al. (50) have shown that both the CD25⁺CD4⁺ and the CD25⁻CD4⁺ subsets control T cell activation and accumulation in IL-2-deficient mice when transferred in vivo. They suggested that the CD25⁺ subset might act on the initial activation step, whereas the CD25⁻ subset might act on the clonal contraction step mediated by AICD. Our results confirm that the two subsets interact to control AICD and reveal the existence of an additional, IL-2-independent mechanism.

Thus, CD25⁺CD4⁺ T cell resistance to Fas-dependent apoptosis induced by vSAG in vivo (11, 12) and to polyclonal activation in vitro can be modulated by environmental factors. Experimental interaction of regulatory CD25⁺CD4⁺ T cells and effector CD25⁻CD4⁺ T cells in vitro showed that sensitivity to Fas-dependent apoptosis depends on the ratio of these two populations during activation. During induction of tolerance to vSAG in vivo, the number of vSAG-specific CD25⁺CD4⁺ T cells remains identical to that in naive mice (12). This suggests that the modulation of sensitivity to apoptosis in vivo favors the persistence of the regulatory CD25⁺CD4⁺ T cell subset. Given the important role of CD25⁺CD4⁺ T cells in the control of immune responses, their resistance to apoptosis in vivo is of major interest.

We thank B. Rocha and C. Penit for critical reading of the manuscript, H. Veiga-Fernandes for help with propidium iodide staining, A. M. Joret for technical assistance, and S. Hamon for secretarial assistance.