IL-10-Secreting Regulatory T Cells Do Not Express Foxp3 but Have Comparable Regulatory Function to Naturally Occurring CD4⁺CD25⁺ Regulatory T Cells ¹

The adaptive immune response evolved to enable the host to eradicate pathogens, but mechanisms also developed to prevent host autoreactivity and damage (1). Peripheral tolerance to self Ags, in addition to deletion and anergy mechanisms, involves active suppression mediated by regulatory T cell (T_Reg) ³ populations (2, 3, 4, 5). T_Reg populations described to date include the naturally occurring CD4⁺CD25⁺ T_Reg subset (CD4⁺CD25⁺ T_Reg) (2, 3, 4, 5), and Ag-driven IL-10-producing T_Reg (IL-10-T_Reg) (6, 7, 8, 9, 10, 11, 12) and TGF-β-T_Reg (13), which have been isolated under particular regimens of antigenic stimulation both in vitro and in vivo. The CD4⁺CD25⁺ T_Reg subset represents 5–10% of the adult peripheral CD4⁺ T cell compartment and is involved in preventing immune pathologies, including wasting disease, autoimmune thyroiditis, gastritis, oophoritis, and orchitis (14). CD4⁺CD25⁺ T_Reg protect lymphopenic mice from inflammatory bowel disease (IBD), autoimmune gastritis (15), wasting disease (4), and diabetes (16), and prevent transplant rejection (17) and immune response to tumors (18, 19). IL-10 produced by CD4⁺CD25⁺ T_Reg is involved in protection from induced colitis (20) and allograft rejection (21), in the control of naive T cell proliferation in vivo in immunocompromised mice (22) and in the response to the pathogen Leishmania major (23). IL-10 produced by T_Reg derived under defined regimens of antigenic stimulation, is involved in the control of autoimmune pathologies including experimental autoimmune encephalomyelitis (EAE) (11, 12, 24). Suppression of other pathologies such as autoimmune gastritis by CD4⁺CD25⁺ T_Reg (3, 15, 25) is IL-10 independent (25), as is the inhibition of T cell proliferation in vitro, which is mediated by cell-cell contact (26).

Most of the IL-10-producing cells described to date were heterogeneous or clones (6, 7, 8, 9, 10, 11, 12, 24, 27), and the molecular mechanisms for their derivation and full effector function are not clearly defined. For this reason, we previously devised a strategy of antigenic stimulation of naive CD4⁺ T cells using immunosuppressive drugs that reproducibly gave rise to a homogeneous population of IL-10-T_Reg (11). These IL-10-T_Reg could be derived from Ag-specific TCR transgenic mice on a recombination-activating gene knockout background that lack CD4⁺CD25⁺ T_Reg, showing that they can arise from naive CD4⁺ T cells not containing CD4⁺CD25⁺ T_Reg (11). These in vitro-derived IL-10-T_Reg blocked EAE, and their development and function was IL-10 dependent (11).

Recently, the forkhead/winged helix transcription factor Foxp3 was shown to be specifically expressed in CD4⁺CD25⁺ T_Reg and to control their development (28, 29, 30, 31). Absence or mutations in Foxp3 resulted in abrogation of their development and led to lymphoproliferative disorders, wasting disease, and IBD in mice (Refs. 29 and 30 ; reviewed in Ref.31), and lymphoproliferative disorders and autoimmune syndromes in humans (Refs.32, 33, 34, 35 ; reviewed in Refs. 31 and 36). Ectopic expression of Foxp3 in mouse CD4⁺CD25⁻ T cells prevented the induction of wasting disease and IBD in lymphopenic animals by these cells, and instead rendered them suppressive both in vitro and in vivo (28, 29, 30). CD4⁺CD25⁻ T cells ectopically expressing Foxp3 acquired the function to suppress induced IBD and the proliferation of CD4⁺CD25⁻ T cells in vitro (28, 29, 30).

Although CD4⁺CD25⁺ T_Reg can produce IL-10 (20, 23), the question remains as to the relationship between these T_Reg and Ag-driven IL-10-T_Reg (7, 9, 10, 11, 12, 24, 37, 38). In this study, we addressed this issue and show for the first time that homogeneous populations of IL-10-T_Reg, which can arise in the absence of naturally occurring T_Reg (11), do not express Foxp3 and yet inhibit the proliferation of T cells with comparable efficiency to ex vivo CD4⁺CD25⁺ T_Reg. This study demonstrates that Foxp3 is not essential for the function of T_Reg to inhibit T cell proliferation. The ability to derive IL-10-T_Reg with Ag and immunosuppressive drugs in the absence of Foxp3 may provide useful strategies for therapeutic intervention in inflammatory diseases.

BALB/c mice were used as a source of T cells and APC for in vitro studies. In some experiments, BALB/c DO11.10 mice transgenic for an OVA_323–339-specific TCRαβ crossed back with recombination-activating gene knockout mice were used as a source of naive T cells as described (11). The Tg4 transgenic mouse expressing TCRα and -β chains derived from the 1934.4 T cell hybridoma specific for the N-terminal, acetylated CD4-T cell epitope of myelin basic protein (MBP) (Ac1–9) (39) was used for in vivo tolerization protocols. Mice were bred and maintained under specific pathogen-free conditions at the National Institute for Medical Research or the University of Bristol School of Medical Sciences, and used between 8 and 12 wk of age.

For the derivation of IL-10-T_Reg in vitro with immunosuppressive drugs (11), and the medium control population, CD4⁺ subsets were enriched from spleen cell suspensions and purified as naive CD4⁺CD62L^highCD25⁻ T cells (>99% by MoFlo flow cytometer; Cytomation, Fort Collins, CO) (11). However, in this study, naive T cells were cultured using stimulation with immobilized anti-CD3 mAbs (10 μg/ml; clone 145.2C11; BD PharMingen, San Diego, CA) and soluble anti-CD28 mAbs (2 μg/ml; clone 37.51; BD PharMingen) in the presence of a combination of 4 × 10⁻⁸ M 1α,25-dihydroxyvitamin D3 (VitD3; BIOMOL Research Labs, Plymouth Meeting, PA) and of 5 × 10⁻⁸ M dexamethasone (Dex; Sigma-Aldrich, St. Louis, MO), and in the absence of anti-cytokine mAbs (medium controls used the same protocol in the absence of the drugs). Please note that the concentration of Dex has been increased in this study from that used in the previous study (11), to ensure that IL-4 and IL-2 are not produced by IL-10-T_Reg, in the absence of anti-cytokine mAbs, which is a prerequisite for their regulatory function in vitro (data not shown). CD4⁺CD25⁺ T_Reg and CD4⁺CD25⁻ control T cells were obtained from spleen cell suspensionsdepleted of B220, CD11b, and CD8-positive cells, and purified as CD4⁺-intermediate/CD25^bright staining cells or CD4^highCD25⁻ (>96 or >99% pure, respectively) as described (14, 26). Cells were stimulated with PMA (50 ng/ml; Sigma-Aldrich) and ionomycin (500 ng/ml; Sigma-Aldrich) in the presence of brefeldin A (10 μg/ml; Sigma-Aldrich) as described (11). Cells were stained with anti-IL-2-FITC (clone JES6-5H4), anti-IL-4-PE (clone 11B11), anti-IL-10-allophycocyanin (clone JES5-16E3), and isotype controls (all BD PharMingen) to assess cytokine production at the single-cell level as described (11). Labeled cells were analyzed by FACS and the obtained data were analyzed with CellQuest software (BD Biosciences, Mountain View, CA).

The acetylated N-terminal peptide of murine MBP (Ac1–9, AcASQKRPSQR) and the high MHC-affinity analog with a tyrosine substituting the wild-type lysine in position 4 (Ac1–9[4Y]) were prepared as described (12). Priming, tolerization, and challenge protocols were performed as described (12). Briefly, Tg4 transgenic mice were either given a single intranasal (i.n.) dose of PBS or primed in vivo for 6 h by a single i.n. dose of 100 μg of Ac1–9[4Y]. Tolerance was induced by giving nine doses of Ac1–9[4Y] at regular intervals over a period of 5 wk. Six hours before sacrifice, tolerized mice were challenged with a 10th i.n. dose of Ac1–9[4Y]. Purified CD4⁺ T cells (>95% CD4⁺ as determined by FACS) were obtained by positive selection of Tg4 splenocytes as described (12). Total splenocytes from naive, primed, or tolerized Tg4 mice were activated with Ac1–9[4K] (100 μg/ml), the cognate Ag, and supernatants were collected for cytokine immunoassay after 12–60 h (12). Alternatively, cells were incubated with cognate Ag for 20 h, and brefeldin A (Sigma-Aldrich) was added for the last 6 h; then cells were permeabilized and subsequently stained for cell surface CD4 and intracellularly with anti-IL-10-PE or the isotype control-PE (all BD PharMingen). Labeled cells were analyzed by FACS.

Freshly isolated T cells, primed or tolerized T cells, or in vitro-cultured T cells were stimulated for the indicated time points with immobilized anti-CD3 (2 μg/ml) and soluble anti-CD28 (2 μg/ml), or PMA (40 ng/ml) and ionomycin (100 ng/ml). RNA from the different T cell populations was extracted using RNeasy kit (Qiagen, Hilden, Germany) or by the TRIzol method for the in vivo populations, DNase treated (Roche, East Sussex, U.K.), and reverse-transcribed as described (11). cDNA was analyzed for the expression of Foxp3, IL-2, IL-10, and ubiquitin by real-time PCR assay using an ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, CA). Target gene mRNA expression was quantified using SYBR green (Applied Biosystems) and normalized to the ubiquitin mRNA levels.

CD4⁺CD25⁻ T cells (50 × 10³ cells) and the indicated numbers of IL-10-T_Reg or CD4⁺CD25⁺ T_Reg were cultured either alone or in combination (see Figs. 3 and 4). Cultures were performed in triplicate in 96-well round-bottom plates (Corning Costar, Cambridge, MA) with indicated numbers of T cell-depleted (anti-CD3; clone 17A2; BD PharMingen) and gamma-irradiated (3000 rad) spleen cells as an APC source and 0.5 μg/ml soluble anti-CD3, as described (26, 40). In some experiments, exogenous IL-2 (5 ng/ml), anti-IL10R mAb (clone 1B1.3a; 10 μg/ml; a kind gift from DNAX Research Institute (Palo Alto, CA); Ref.41), or isotype-matched control mAb were added to the cultures. In vitro proliferation was assessed by measuring incorporation of [³H]TdR (Amersham, Arlington Heights, IL) by liquid scintillation spectroscopy after pulsing with 18.5 kBq (0.5 μCi) for the last 6 h of a 72-h culture. Transwell experiments were performed in 24-well plates (Corning Costar) as described (26). Briefly, CD4⁺CD25⁻ T cells (0.5 × 10⁶ cells) were cultured with APC (0.5 × 10⁶ cells) and anti-CD3 (0.5 μg/ml). In some cultures, T_Reg cells (0.25 × 10⁶ cells) were added and separated by a semipermeable membrane (Corning Costar). Either anti-IL10R mAb (clone 1B1.3a; 10 μg/ml) or isotype-matched control mAb were added to the cultures. Alternatively, CD4⁺CD25⁻ T cells (see Fig. 4 Bi), IL-10-T_Reg, or CD4⁺CD25⁺ T_Reg (Bii) were labeled with CFSE (Molecular Probes, Eugene, OR) together with Thy1.2 mAbs (BD PharMingen) before the culture and analyzed for progressive halving of the CFSE label by FACS after 72 h.

To investigate the possible relationship between CD4⁺CD25⁺ T_Reg and Ag-driven IL-10-T_Reg, we analyzed the expression of Foxp3 in both populations. Homogeneous populations of IL-10-T_Reg were obtained in vitro by stimulating naive CD4⁺ T cells in the presence of a combination of the anti-inflammatory drugs VitD3 and Dex, as we have previously described (11). These cultures resulted consistently in >75% IL-10-producing T cells, upon restimulation for 4 h, which produced no detectable IL-2, IL-4 (Fig. 1,A), or IFN-γ (data not shown; Ref.11), whereas cells cultured in the absence of the drugs contained low numbers of IL-10-, IL-4-, IL-2- (Fig. 1 A), and IFN-γ-producing cells (data not shown; Ref.11). In contrast, freshly isolated CD4⁺CD25⁻ T cells or CD4⁺CD25⁺ T_Reg stimulated for 4 h in vitro produced very little IL-2 or IL-10 protein (data not shown).

Foxp3 mRNA was virtually undetectable in homogeneous populations of in vitro-derived IL-10-T_Reg as assessed by real-time quantitative RT-PCR (Fig. 1,B), in unstimulated cells or after subsequent restimulation through CD3 and CD28 for 6 h. In sharp contrast, freshly isolated CD4⁺CD25⁺ T_Reg specifically expressed high levels Foxp3 before or after stimulation, as previously demonstrated (28, 29, 30). Upon subsequent restimulation, IL-10-T_Reg that had developed in VitD3/Dex, expressed high levels of IL-10 and decreased IL-2 mRNA (Fig. 1,B), as compared with Ag-driven T cells derived in the absence of the drugs. As opposed to the IL-10 expression patterns of in vitro-generated IL-10-T_Reg, IL-10 mRNA was expressed at relatively lower levels in CD4⁺CD25⁺ T_Reg immediately after isolation and upon in vitro stimulation, and was also very low in CD4⁺CD25⁻ T cells (Fig. 1 B). Upon activation with anti-CD3/anti-CD28 mAb, CD4⁺CD25⁺ T_Reg did not express IL-2 mRNA, which was expressed albeit at low levels by freshly isolated CD4⁺CD25⁻ T cells upon activation.

To address whether other regimens for inducing IL-10-T_Reg such as peptide Ag stimulation in vivo (12) resulted in increased Foxp3 expression, we used a regimen of repetitive i.n. administration of the peptide Ac1–9[4Y] of MBP in the Tg4 transgenic mouse model to obtain in vivo Ag-driven IL-10-T_Reg as we have previously described (12, 24). Ex vivo-stimulated total CD4⁺ Tg4 T cells from i.n. tolerized mice contained ≈10% of IL-10-producing cells (Fig. 2,A). In contrast, total CD4⁺ Tg4 T cells from nontolerized mice contained <1% of IL-10-producing cells (data not shown). Differential production of either IL-10 or IL-2 by CD4⁺ Tg4 T cells from naive, primed, and i.n. tolerized mice was investigated in more depth in a kinetic immunoassay (Fig. 2,B). CD4⁺ T cells from PBS-treated or primed mice produced high IL-2 levels upon restimulation in vitro with the cognate Ag and APC with a peak between 48 and 72 h, decreased by 96 h, reflecting IL-2 consumption/regulation. CD4⁺ T cells from tolerized mice produced little to no IL-2 (Fig. 2 B). IL-10 production was maximal in CD4⁺ T cells from tolerized mice upon restimulation in vitro and maintained until 96 h, but low to undetectable in CD4⁺ T cells from primed or naive mice, respectively.

Ex vivo-purified total CD4⁺ T cells from unimmunized or primed Tg4 mice expressed very low Foxp3 levels and i.n. tolerization did not lead to increased Foxp3 expression in these total CD4⁺ Tg4 T cells (Fig. 2,C) as compared with the high levels expressed by CD4⁺CD25⁺ T_Reg. The low Foxp3 levels were further decreased upon in vitro stimulation of CD4⁺ T cells from PBS-treated, primed, or tolerized mice. Ex vivo-isolated CD4⁺ T cells from PBS-treated Tg4 mice did not express IL-10 and IL-2 mRNA, but in vivo priming of these mice with Ac1–9[4Y] 6 h before harvesting of the cells yielded minor IL-10 and considerable IL-2 mRNA levels (Fig. 2 C, upper panels) in keeping with the immunoassay data. Ex vivo-isolated CD4⁺ T cells from tolerized Tg4 mice expressed considerable IL-10 but reduced IL-2 mRNA levels. Upon in vitro restimulation of CD4⁺ T cells from Tg4-tolerized mice, IL-10 expression levels were further increased, whereas in vitro restimulation of CD4⁺ T cells from Tg4 primed mice did not up-regulate IL-10 expression, but showed a very significant increase in IL-2 expression.

In vivo-derived IL-10-T_Reg inhibit naive T cell proliferation in vitro in an IL-10-independent fashion (10, 12); however, these populations were heterogeneous and may have contained naturally occurring T_Reg. We addressed this issue using homogeneous IL-10-T_Reg derived in vitro with immunosuppressive drugs, which produce large amounts of IL-10, do not express Foxp3, and arise independently of naturally occurring T_Reg (Fig. 1; Ref.11). IL-10-T_Reg inhibited the proliferation of CD4⁺CD25⁻ T cells in response to anti-CD3 mAb, in the context of irradiated T-depleted spleen APC, at a ratio of one regulator to two responders, similarly to that previously reported for CD4⁺CD25⁺ T_Reg (Fig. 3,A). CD4⁺ T cells generated in vitro in the absence of the drugs produced IL-2 (Fig. 1,A) and failed to inhibit CD4⁺CD25⁻ T cell proliferation (data not shown). Whereas CD4⁺CD25⁺ T_Reg failed to proliferate when cultured with APC and anti-CD3 in the absence of CD4⁺CD25⁻ T cells (5, 26, 40) (Fig. 3,A), IL-10-T_Reg displayed a low basal proliferation level (Fig. 3 A). This seemingly raised the background proliferation levels of proliferation in the mixed cultures masking the full extent of their inhibition of CD4⁺CD25⁻ T cell proliferation.

Blocking IL-10R-mediated signaling during coculture of homogeneous populations of IL-10-T_Reg with CD4⁺CD25⁻ T cells at a ratio of 1:8 (and other ratios; data not shown) did not affect their inhibition of T cell proliferation in vitro (Fig. 3,B), similarly to cultures using CD4⁺CD25⁺ T_Reg whose function in vitro is IL-10 independent (B; Refs. 3 and 26). Neither anti-TGF-β (together with anti-IL-10R mAbs or alone) nor anti-CTLA-4 mAbs showed any effect on T cell suppression by either IL-10-T_Reg (data not shown) or by CD4⁺CD25⁺ T_Reg as previously shown in vitro (26), despite the enhanced levels of CTLA-4 by the IL-10-T_Reg (data not shown; Ref.12). Moreover, because IL-10-T_Reg expressed enhanced levels of glucocorticoid-induced TNFR family-related gene (GITR) (data not shown) and comparable to those of activated CD4⁺CD25⁺ T_Reg (42, 43), we tested whether GITR cross-linking would play a role in this in vitro assay. Anti-GITR mAbs overcame the suppression, but in this case, the enhancement of activated CD4⁺CD25⁻ T cell proliferation accounted for this effect (data not shown). Exogenous IL-2 completely abolished the inhibition of CD4⁺CD25⁻ T cell proliferation by either IL-10-T_Reg (Fig. 3,B) or CD4⁺CD25⁺ T_Reg (Fig. 3 B; Ref.26) and significantly raised the basal proliferation level of the CD4⁺CD25⁻ T cells and both T_Reg populations (B).

To address the issue whether cell-cell contact would mediate the suppression of naive T cell proliferation by IL-10-T_Reg, Transwell culture chambers were used. Separation of naive T cells from either T_Reg population resulted in the abrogation of suppression of T cell proliferation (Fig. 3 C) in response to anti-CD3 mAb, in the context of irradiated T-depleted spleen APC, as opposed to cultures in which both naive T cells and T_Reg were present in the lower chamber of the culture. Furthermore, blocking IL-10R-mediated signaling did not affect this abrogation of suppression.

We subsequently addressed whether the inhibitory potential of these IL-10-T_Reg is comparable to that of CD4⁺CD25⁺ T_Reg. IL-10-T_Reg and CD4⁺CD25⁺ T_Reg were titrated in vitro to compare their efficiency to inhibit CD4⁺CD25⁻ T cell proliferation in response to APC and anti-CD3 mAb. Strikingly, IL-10-T_Reg were as efficient as CD4⁺CD25⁺ T_Reg in suppressing CD4⁺CD25⁻ T cell proliferation at regulator-to-responder ratios ranging from 1:1 to 1:8 (Fig. 4,A, upper panels, •). However, the level of CD4⁺CD25⁻ T cell proliferation was reduced to <500 cpm with CD4⁺CD25⁺ T_Reg (background levels) and to only 1000 cpm with IL-10-T_Reg (Fig. 4,A, upper panels, •). This can be accounted for by the basal proliferation level of IL-10-T_Reg alone in contrast to the totally unresponsive CD4⁺CD25⁺ T_Reg (Fig. 4,A, top panel, □). Both T_Reg lost suppressive activity at T_Reg-to-T cell ratios of <1:4 when the activation threshold was altered by raising the amount of APC 5-fold in the system (Fig. 4,A, lower panels), although IL-10-T_Reg were much less efficient than CD4⁺CD25⁺ T_Reg (A, lower panels, •). Again, this reflected the basal proliferation of IL-10-T_Reg themselves vs CD4⁺CD25⁺ T_Reg (Fig. 4,A, lower panels, □). IL-10-T_Reg also seemed to control their own proliferation at high cell numbers (Fig. 4 A, left lower panel, □).

To clarify the ability of IL-10-T_Reg to inhibit T cell proliferation independently of their own proliferation, and to obtain an accurate assessment of their efficiency as compared with CD4⁺CD25⁺ T_Reg, we used CFSE-labeled cells. CFSE experiments demonstrated that, despite their basal proliferation level in response to APC and anti-CD3, IL-10-T_Reg were as potent as CD4⁺CD25⁺ T_Reg in controlling the proliferation of CD4⁺CD25⁻ T cells at a ratio of 1:2 T_Reg to CD4⁺CD25⁻ T cells (Fig. 4,Bi, left-hand panels) at low activation thresholds (50 × 10³ APC, which are the numbers used in the original studies for CD4⁺CD25⁺ T_Reg (26)). By raising 5-fold the amount of APC in the cultures, CD4⁺CD25⁺ T_Reg retained their inhibitory potential, whereas IL-10-T_Reg became slightly less efficient (Fig. 4,Bi, right-hand panel). Thus, the efficiency of IL-10-T_Reg vs CD4⁺CD25⁺ T_Reg was similar, although this was masked in cocultures analyzed by [³H]TdR incorporation by the different basal proliferation levels by each T_Reg population. In keeping with the [³H]TdR incorporation data (Fig. 3,B and data not shown), IL-10-T_Reg exerted their inhibitory function in vitro via an IL-10-independent mechanism as observed by CFSE decay (Fig. 4,Bi). In a number of experiments with very high APC numbers (250 × 10³) and varying IL-10-T_Reg-to-T cell ratios, a minor but nonreproducible effect of anti-IL-10R mAbs was occasionally observed, which in part was explained by a minor enhancement of proliferation of the IL-10-T_Reg themselves. However, in total, CFSE experiments also showed that IL-10-T_Reg themselves proliferate to a low level, in contrast to naturally occurring CD4⁺CD25⁺ T_Reg, which never proliferate (Fig. 4 Bii), and that blocking IL-10R signaling did not significantly alter their basal level of proliferation.

IL-10 produced by T_Reg controls inflammatory responses to self and nonself Ags (Refs.11 ,12 , and 20, 21, 22, 23 ; reviewed in Refs. 4 and 44). The relationship between Ag-driven IL-10-T_Reg and naturally occurring CD4⁺CD25⁺ T_Reg is poorly understood. We show in this study that IL-10-T_Reg obtained using either in vitro or in vivo regimens of antigenic stimulation do not express the transcription factor Foxp3 in contrast to CD4⁺CD25⁺ T_Reg. Despite their lack of Foxp3 expression, homogeneous populations of IL-10-T_Reg inhibited CD4⁺CD25⁻ T cell proliferation in vitro similarly to CD4⁺CD25⁺ T_Reg, and both T_Reg produced little to no IL-2. Like CD4⁺CD25⁺ T_Reg, the regulatory capacity of IL-10-T_Reg in vitro was independent of their intrinsic production of IL-10, was cell contact dependent, and was overcome by exogenous IL-2.

Mutations or absence of Foxp3 in mice and humans results in a lymphoproliferative disease characterized by cachexia and multiorgan lymphocytic infiltrates (Refs.29, 30, 31, 32, 33, 34, 35 ; reviewed in Refs. 31 and 36). Foxp3-expressing CD4⁺CD25⁺ T_Reg inhibited the development of this disease when administered to neonatal Scurfy mutant mice (30). Furthermore, ectopic expression of Foxp3 in CD4⁺CD25⁻ T cells prevented induction of lymphoproliferative and wasting diseases, and IBD by these cells in lymphopenic hosts, and rendered them suppressive both in vitro and in vivo (28, 29, 30). We show that Foxp3 expression is not necessary for T_Reg function to inhibit T cell proliferation in vitro, but may engender cells with this capacity via the induction of as-yet-unknown genes. IL-10-T_Reg that develop from naive T cells in the absence of CD4⁺CD25⁺ T_Reg (11), inhibited T cell proliferation in vitro with a similar efficiency to that of CD4⁺CD25⁺ T_Reg and may acquire this property postnatally upon certain conditions of Ag encounter by CD4⁺ T cells in the periphery.

Despite the differences in Foxp3 expression in CD4⁺CD25⁺ T_Reg and IL-10-T_Reg, both populations inhibited T cell proliferation in vitro, and expressed little to no IL-2 (Fig. 1 B), which may be a requisite for their regulatory function. Foxp3 was suggested to act as a negative modulator of IL-2 transcription (32). This can be clearly achieved in IL-10-T_Reg by different mechanisms including antigenic stimulation in the presence of immunosuppressive drugs, or by peptide-antigenic regimens leading to anergy (10, 11, 12, 27, 45). Both CD4⁺CD25⁺ and IL-10-T_Reg express enhanced levels of the IL-2R α-chain CD25 (data not shown), a requirement for their suppressive function if IL-2 consumption is a component of this in vitro suppression as suggested (T. Barthlott and B. Stockinger, unpublished observations). To date, the mechanism(s) whereby inhibition of proliferation and expression of IL-2 mRNA in T cells by CD4⁺CD25⁺ T_Reg (3, 26) and IL-10-T_Reg is achieved is unclear.

We show that homogeneous populations of IL-10-T_Reg, despite production of large amounts of IL-10, inhibit T cell proliferation in vitro in an IL-10-independent and cell contact-dependent fashion. In vitro assays were performed identically to assays described for CD4⁺CD25⁺ T_Reg (3, 26, 40) and included different ratios of T_Reg to CD4⁺ T cells, and tested different numbers of APC. Other T_Reg, including in vivo Ag-driven IL-10-secreting T_Reg (10, 12), which produce low levels of IL-10 and are heterogeneous, also inhibit T cell proliferation in vitro apparently via an IL-10-independent mechanism. However, one study suggested that IL-10 may partially mediate the inhibitory effect of the so-called Tr1 IL-10-T_Reg, although it is of note that neutralization of IL-10 never completely reversed the suppressive effect on T cell proliferation (7). These different results may reflect heterogeneity of T_Reg at various levels, including IL-10 production or the number of regulators vs T cells used in the assays. We now rule out these possibilities, confirming that even IL-10-T_Reg producing high IL-10 levels significantly inhibit T cell proliferation in the absence of IL-10 activity. The source of “naive” CD4⁺ T cells and possible contamination with effectors, the APC type and number used and their activation status may explain such discrepancies (7). We occasionally observed a minimal but nonreproducible effect of anti-IL-10R mAbs in relieving suppression of T cells by IL-10 T_Reg, at high APC number, and at certain T cell vs T_Reg ratios (data not shown), and possibly due to a minimal enhancement of proliferation of IL-10-T_Reg themselves in the presence of anti-IL-10R mAbs. Our study now shows definitively with the use of CFSE-labeled cells that the major suppression by IL-10-T_Reg of T cell proliferation in vitro under the conditions described by Thornton and Shevach (26) for CD4⁺CD25⁺ T_Reg is independent of IL-10.

In contrast to the lack of a role for IL-10 in the in vitro suppression of T cell proliferation by IL-10-T_Reg or CD4⁺CD25⁺ T_Reg, IL-10 is involved in the inhibition of IBD (20, 46) and wasting disease (20, 22), in the control of in vivo expansion of naive CD4⁺ T cells in lymphopenic hosts (22), and in the inhibition of EAE (11). These effects may be achieved via the inhibition of APC function by IL-10 (44, 46). In vitro assays used for inhibition of naive T cell proliferation by T_Reg may not reveal a role for IL-10 as a result of the duration of the assay and the APC type, and may not reflect the extent of APC activation that may occur in vivo. Toll-like receptor ligation of bone marrow-derived dendritic cells overcomes inhibition of naive T cell proliferation by CD4⁺CD25⁺ T_Reg partially through IL-6 production (47). Because IL-10’s major effects are to inhibit APC function, including production of proinflammatory cytokines (44, 48, 49), it is likely that IL-10 plays a role in T_Reg function upon activation of the innate immune response (44, 46, 50, 51). The in vitro assay showing robust inhibition of naive T cell proliferation by CD4⁺CD25⁺ T_Reg and IL-10-T_Reg independently of IL-10 may thus reflect the first layer of regulation in vivo before activation of the innate immune response. These findings are not incompatible with the hypothesis that control of cell numbers in vivo in lymphopenic mice and possibly intact mice, may first be subject to other mechanisms of regulation before a participation of IL-10 is evoked in suppression (3, 5, 36, 52).

IL-10-secreting cells can be induced by both in vitro and in vivo regimens of antigenic stimulation (6, 7, 8, 9, 10, 11, 12, 24, 27). However, the mechanisms underlying the induction of IL-10 expression in T_Reg and the possible relationship between IL-10 and Foxp3 expression is unclear (28, 29, 30, 31, 36). Studies addressing this issue in CD4⁺CD25⁻ T cells ectopically expressing Foxp3 vs CD4⁺CD25⁺ T_Reg yielded conflicting results (28, 29, 30). We now resolve this issue by showing that homogeneous populations of IL-10-T_Reg do not express Foxp3. It is of note that enhanced IL-10 expression in IL-10-T_Reg is always mirrored by a significant decrease in IL-2 expression (Figs. 1 and 2; Refs. 11 and 12) and that IL-2 expression is also absent in Foxp3-expressing CD4⁺CD25⁺ T_Reg, suggesting that IL-2 is extinguished by different mechanisms in both T_Reg types.

Finally, we demonstrate in this study that IL-10-T_Reg, which we have previously shown to have regulatory capacity in vivo (11, 12), do not express Foxp3 and yet inhibit naive T cell proliferation in vitro similarly to CD4⁺CD25⁺ T_Reg. These observations and the fact that we can derive these IL-10-T_Reg with immunosuppressive drugs from naive CD4⁺ T cells by antigenic stimulation in the absence of other naturally occurring T_Reg (11), and in the absence of Foxp3, suggests their potential use in early development in immunotherapy of inflammatory diseases. Withdrawal of immunosuppressive drugs later in development would ensure that the host could respond appropriately to environmental pathogens but be tolerant to self Ags encountered by T cells upon exit from the thymus.

We thank Drs. Douglas Robinson, Alexandre Potocnik, Margarida Saraiva, George Kassiotis, Benedict Seddon, and Paulo Vieira for helpful discussions and suggestions; Dr. Shimon Sakaguchi for sharing anti-GITR reagents; and Chris Atkins and Graham Preece for excellent assistance in obtaining sorted T cell populations.