The activity of regulatory T cells (Treg) is widely accepted to play a central role in preventing pathogenic immune responses against self-Ags. However, it is not clear why such regulation breaks down during the onset of autoimmunity. We have studied self-Ag-specific Treg during the induction of spontaneous diabetes. Our data reveal a shift in the balance between regulatory and pathogenic islet-reactive T cells in the pancreas-draining lymph nodes during disease onset. Treg function was not compromised during disease initiation, but instead conventional T cells showed reduced susceptibility to Treg-mediated suppression. Release from Treg suppression was associated with elevated levels of IL-21 in vivo, and provision of this cytokine abrogated Treg suppression in vitro and in vivo. These data suggest that immunological protection of a peripheral tissue by Treg can be subverted by IL-21, suggesting new strategies for intervention in autoimmunity.

Peripheral tissues are protected from autoimmune attack by a repertoire of regulatory T cells (Treg)3 that respond to self-Ags. The need for ongoing suppression is emphasized by the rampant autoimmunity that follows ablation of the Treg compartment in adult mice (1). Although Treg ablation can induce disease in experimental settings, it is not clear whether a decrease in this population is responsible for triggering spontaneous autoimmunity. Despite an early report that Treg proportions were decreased in insulin-dependent diabetes mellitus 1 patients (2), subsequent studies have failed to confirm this (3, 4). Such analyses are necessarily focused on peripheral blood as the most accessible lymphocyte source in humans, and the extent to which these results reflect Treg proportions at other sites is not clear. In particular, the homeostasis of self-Ag-specific Treg within inflamed tissues and their draining lymphoid organs may be quite different from what is observed in the systemic circulation. CD25+ T cells are known to be present at inflammatory sites including the pancreas in insulin-dependent diabetes mellitus 1 (5) and the synovium in rheumatoid arthritis (6), with an enrichment of these cells being reported in the latter case. However, a clear picture of how Treg numbers change during the onset of spontaneous autoimmunity is currently lacking.

An alternative possibility is that autoimmunity arises not because of altered Treg numbers but rather because of diminished Treg function. In support of this, differences in Treg suppressive function have been reported in the setting of diabetes (7), rheumatoid arthritis (8), multiple sclerosis (9), and autoimmune polyglandular syndrome II (10). However, such differences are not universally observed (3). There is also evidence that autoimmunity may be associated with resistance of T cells to Treg-mediated suppression (11), possibly due to altered TGF-β sensitivity (12).

To explore how regulatory mechanisms are overcome during the onset of autoimmunity, we have analyzed both Treg numbers and Treg function in a mouse model of spontaneous diabetes. In this model, OVA is expressed as a self-Ag in the pancreas under the control of the rat insulin promoter (RIP) and coexpression of the DO11.10 transgene permits the development of OVA-specific DO11 T cells (DO11 × RIP-mOVA mice). We have previously shown that both conventional T cells (Tconv) and Treg specific for OVA develop in these animals (13), the emergence of the latter reflecting activity of the transgenic insulin promoter in the thymus (14). In this report, we show that after an initial period of Treg-mediated disease control, DO11 × RIP-mOVA mice develop autoimmune diabetes with 100% penetrance. Thus, the Tconv that arise in these animals are capable of triggering the destruction of OVA-expressing pancreatic β cells. Our experiments reveal that progression to autoimmunity is associated with a decrease in the proportion of regulatory to Tconv specifically in the pancreatic lymph node, with no change being detectable in peripheral blood or in nondraining lymphoid tissues. Early in life CD25+ Treg are able to prevent diabetes; however, this protective mechanism ultimately breaks down and overt disease develops. Interestingly, adoptive transfer studies indicate that this breakdown does not reflect a decrease in the suppressive function of Treg but instead relates to Tconv acquiring resistance to Treg suppression. Analysis of cytokine expression revealed that resistance is associated with induction of IL-21, and we show that provision of IL-21 is highly effective at counteracting Treg suppression in vitro and in vivo.

DO11.10 TCR-transgenic mice and BALB/c mice were purchased from The Jackson Laboratory. RAG-2−/− mice were purchased from Taconic Farms. RIP-mOVA mice on a BALB/c background that express a membrane-bound form of OVA under the control of the RIP (from line 296-1B) were a gift from W. Heath (Walter and Eliza Hall Institute, Melbourne, Australia). DO11.10 mice and RIP-mOVA mice were crossed as previously described (13). All mice were housed in the University of Birmingham Biomedical Services Unit and used according to Home Office and institutional regulations.

Blood glucose levels were monitored by Glucometer Ascensia Elite XL (Bayer). Six hundred milligrams per deciliter is the maximum measurable glucose reading. Mice are typically considered diabetic following readings of >250 mg/dl. Where indicated, DO11 × RIP-mOVA mice were injected i.p. with 0.5 mg of anti-CD25 Ab (PC61) or control rat IgG at 3 and 4 wk of age or at 5 and 6 wk of age. Blood glucose levels were monitored weekly and mice were killed at 6 or 7wk.

Acetone-fixed frozen 5-μm tissue sections were stained with biotinylated KJ-126 mAb that was detected with streptavidin-ABComplex-alkaline phosphate (DakoCytomation) and anti-mouse IgD that was detected with HRP-conjugated anti-sheep IgG (both Binding Site). Alkaline phosphatase activity was detected using naphthol AS-MX phosphate and Fast Blue salt with levamisole. HRP was detected using diaminobenzidine tetrahydrochloride solution.

Lymphocytes isolated from the indicated tissues were stained for 10 min at 4°C using Ab concentrations previously determined by titration. Where indicated, pancreas samples were torn into small pieces in cold PBS containing 5% FBS, 56 mM glucose, 2 μg/ml aprotinin, and 50 μg/ml TLCK (both Roche). Following centrifugation, samples were resuspended in 2 ml of prewarmed PBS containing 15% FBS and Liberase CI (Roche). Samples were subsequently passed through a cell strainer and centrifuged with Lympholyte-M (Cedarlane Laboratories) and washed with PBS before Ab staining. All Ab were purchased from BD Biosciences, unless otherwise stated. Staining Ab used were KJ-126 (Caltag Laboratories) CD25 (PC61), CD4 (L3T4), and Foxp3 (FJK-16s; eBioscience). For intracellular Foxp3 staining, cells were fixed and permeabilized according to the manufacturer’s instructions. Gates were set using isotype-matched controls. Statistics were performed using an unpaired two-tailed t test with a 95% confidence interval.

Combined peripheral LN (axillary, inguinal, brachial, popliteal, and mesenteric) cells from DO11 × RIP-mOVA mice at the indicated age were stained for CD4, CD25, and KJ-126 and purified using high-speed cell sorting (Mo-Flo; DakoCytomation). In brief, 0.2–0.4 × 106 cells (CD4+KJ+CD25+ or CD4+KJ+CD25) were injected i.v. into mOVA/rag recipients. Cotransfer experiments used a 1:1 ratio of Tconv:Treg. For in vitro activation, DO11 T cells were incubated for 4 days with 1 μg/ml OVA323–339 peptide and BALB/c splenocytes in RPMI 1640 supplemented with 1 mM l-glutamine, penicillin, streptomycin, HEPES, 5 × 10−5 M 2-ME, and 10% FBS (Sigma-Aldrich). Blood glucose levels were monitored every 3–4 days. For analysis of proliferation in the pancreatic LN, LN cells from DO11+Thy1.1+ mice were injected such that 1.5 × 106 of DO11+ T cells were transferred. An equivalent number of Thy1.2+ purified OVA-specific Treg (KJ+CD25+) was transferred where indicated. IL-21 (PeproTech) or PBS was injected daily i.p. at a dose of 1 μg/mouse. Mice were killed at day 4.

RNA was isolated from whole LN using RNAzol B according to the manufacturer’s instructions (Biogenesis). cDNA was prepared using avian reverse transcriptase and oligo(dT) primers (Invitrogen Life Technologies). Primers and probes were purchased from Applied Biosystems. Gene expression was quantified using the Applied Biosystems Prism 7700 sequence detection system and normalized to β-actin levels.

Magnetic separation (Miltenyi Biotec) was used to purify CD4+CD25 and CD4+CD25+ cells from BALB/c LN. CD4+CD25 (2.5 × 104) cells were cultured per well with 0.1 μg/ml anti-CD3 and 1.5 × 105 CD19+ B cells in the presence of the indicated ratio of CD4+CD25+ cells. Fifty to 100 ng/ml IL-21, 250 ng/ml IL-4, 5 ng/ml IL-12, or 25 ng/ml TNF-α (all PeproTech) were added where indicated. Proliferation was assessed at day 3 by [3H]thymidine incorporation.

We have previously reported the development of OVA-specific Treg in mice coexpressing OVA-specific transgenic T cells (DO11) and a source of Ag (RIP-mOVA) (13). Treg differentiation in these animals reflects the activity of the RIP intrathymically (13, 15) such that DO11 T cells encounter OVA during their development in the thymus. Since our original analysis used CD25 as a Treg marker and this is now known to be a relatively poor marker of Treg in the thymus (16), we confirmed the presence of Ag-specific Treg in the thymus of DO11 × RIP-mOVA mice by Foxp3 staining. Fig. 1,A shows that OVA-specific single-positive CD4 cells in the thymus of DO11 × RIP-mOVA mice contain a substantial fraction of Foxp3+ cells. Development of this population was Ag dependent since few OVA-specific single- positive CD4 cells expressed Foxp3 in DO11 mice lacking OVA (Fig. 1,A). Thus, Foxp3 staining confirmed the development of Ag-specific Treg in the thymus of DO11 × RIP-mOVA mice. Conventional OVA-specific CD4 T cells (Tconv) also develop intrathymically in parallel with the Treg; while some of these are deleted, others survive to seed the periphery (13). To assess whether these OVA-specific T cells were capable of instigating autoimmune tissue damage, blood glucose levels were measured in DO11 × RIP-mOVA mice. This analysis revealed that DO11 × RIP-mOVA mice developed overt diabetes with 100% penetrance by 20 wk of age (Fig. 1,B). Fig. 1 C illustrates pancreatic islet infiltration by DO11 T cells and B cells in an 18-wk-old animal. Thus, despite the presence of an Ag-specific Treg population, DO11 × RIP-mOVA mice still develop spontaneous autoimmune diabetes.

FIGURE 1.

Development of spontaneous diabetes in DO11 × RIP-mOVA mice despite Ag-specific Treg. A, Thymus cell suspensions from 5-wk DO11 or DO11 × RIP-mOVA mice were stained for expression of CD4, CD8, DO11 TCR, and intracellular Foxp3. Plots are gated on single-positive CD4 cells expressing the DO11 TCR. B, Blood glucose levels in DO11 × RIP-mOVA mice measured at different weeks of age. Each symbol represents an individual mouse. C, Frozen pancreas sections from an 18-wk- old DO11 × RIP-mOVA mouse were stained for the DO11 TCR (brown) and B220 (blue).

FIGURE 1.

Development of spontaneous diabetes in DO11 × RIP-mOVA mice despite Ag-specific Treg. A, Thymus cell suspensions from 5-wk DO11 or DO11 × RIP-mOVA mice were stained for expression of CD4, CD8, DO11 TCR, and intracellular Foxp3. Plots are gated on single-positive CD4 cells expressing the DO11 TCR. B, Blood glucose levels in DO11 × RIP-mOVA mice measured at different weeks of age. Each symbol represents an individual mouse. C, Frozen pancreas sections from an 18-wk- old DO11 × RIP-mOVA mouse were stained for the DO11 TCR (brown) and B220 (blue).

Close modal

One possible explanation for the failure of Ag-specific Treg to control diabetes induction was that they were unable to migrate to Ag-bearing sites in the periphery. To address this issue, we measured the proportion of Foxp3+ cells in Ag-draining and nondraining LN. Representative staining for Foxp3 is shown in Fig. 2,A. In the nondraining LN (inguinal LN) and spleen of 6-wk-old mice, ∼20% of the CD4 compartment comprised Foxp3+Treg. However, in the pancreatic LN (draining the site of self-Ag expression (13)), the proportion of Treg was greatly enriched with up to 65% of the CD4+ population expressing this transcription factor (Fig. 2,B). This suggests that the OVA-specific Treg exported from the thymus were able to respond to peripherally expressed Ag and accumulate in the Ag-draining lymphoid tissue. Three-week-old mice did not exhibit significant accumulation of Treg in the pancreatic LN (Fig. 2,B), perhaps reflecting the lack of presentation of pancreas-derived Ags in the pancreatic LN of juvenile animals (17). We also assessed whether the inability of Treg to prevent diabetes reflected a failure of Treg to enter the pancreas itself. Using enzymatic digestion to recover lymphocytes that had infiltrated the pancreas, we showed that Treg were well represented in this population in 6-wk-old animals. In fact, the proportion of Foxp3+ Treg within the T cells infiltrating the pancreas was higher than that seen in nondraining LN at both 6 and 12 wk, although the enrichment was not as marked as in the pancreatic LN (Fig. 2 B). Thus, diabetes induction did not reflect a failure of Foxp3 Treg to respond to self-Ag in the periphery and accumulate at relevant sites.

FIGURE 2.

Enrichment of Treg in the pancreatic LN of DO11 × RIP-mOVA mice, and shift in ratio during disease progression. A, Representative intracellular staining for Foxp3 in lymphocyte populations isolated form the inguinal LN and pancreatic LN. Figures in brackets show the percentage of Foxp3+ cells within the CD4+ population. B, Percentages of Foxp3+ Treg in inguinal (Ing) LN, spleen (Spl), pancreatic LN (PanLN), and digested pancreas (Pan) of DO11 × RIP-mOVA mice at the indicated age. Red lines show mean value. C, Proportion of Foxp3+ cells within the CD4+ population in pancreatic LN and pancreas was correlated with blood glucose levels. DO11 × RIP-mOVA mice between the ages of 6 and 13wk were used. Value of p reflects comparison of normoglycemic (100–200 mg/dl blood glucose) and diabetic (≥600 mg/dl) animals (n ≥ 7 for each group). Cross shows mean value for the inguinal LN. Proportion of Foxp3+ cells in inguinal LN, spleen, and blood did not alter during disease progression (our unpublished data). Data are compiled from more than five independent experiments.

FIGURE 2.

Enrichment of Treg in the pancreatic LN of DO11 × RIP-mOVA mice, and shift in ratio during disease progression. A, Representative intracellular staining for Foxp3 in lymphocyte populations isolated form the inguinal LN and pancreatic LN. Figures in brackets show the percentage of Foxp3+ cells within the CD4+ population. B, Percentages of Foxp3+ Treg in inguinal (Ing) LN, spleen (Spl), pancreatic LN (PanLN), and digested pancreas (Pan) of DO11 × RIP-mOVA mice at the indicated age. Red lines show mean value. C, Proportion of Foxp3+ cells within the CD4+ population in pancreatic LN and pancreas was correlated with blood glucose levels. DO11 × RIP-mOVA mice between the ages of 6 and 13wk were used. Value of p reflects comparison of normoglycemic (100–200 mg/dl blood glucose) and diabetic (≥600 mg/dl) animals (n ≥ 7 for each group). Cross shows mean value for the inguinal LN. Proportion of Foxp3+ cells in inguinal LN, spleen, and blood did not alter during disease progression (our unpublished data). Data are compiled from more than five independent experiments.

Close modal

Diabetes induction and Treg accumulation suggested that both Tconv and Treg, respectively, were exposed to pancreas-expressed Ag in DO11 × RIP-mOVA mice. We therefore set out to assess whether the balance between these two populations shifted during progression to diabetes. The percentage of Foxp3+ cells within the CD4+ population provides a measure of the relative balance between Treg and Tconv. Foxp3+ cells were measured in DO11 × RIP-mOVA mice ages between 6 and 13 wk and these data were plotted against blood glucose levels. The proportion of Treg in the pancreatic LN decreased during disease onset such that the percentage in the most diabetic animals (600 mg/dl blood glucose) was lower than that in normoglycemic (100–200 mg/dl) animals (34.8 ± 3.6 vs 51.2 ± 5.6, respectively, Fig. 2,C). Despite the decrease associated with disease progression, Treg nevertheless remained enriched in the pancreatic LN compared with nondraining LN in diabetic animals (34.8 ± 3.6 in pancreatic LN vs 16.8 ± 2.4 in inguinal LN; see cross in Fig. 2,C). In contrast to the decline in Treg proportions in the pancreatic LN, the proportion of Treg within the pancreas remained remarkably constant throughout the course of disease (Fig. 2 C). This suggested that the ratio of Treg:Tconv in the pancreatic draining LN is uncoupled from that seen in the target tissue itself. The proportion of Treg in the nondraining LN, spleen, and blood also remained unchanged during progression to overt autoimmunity (L. E. Clough and L. S. K. Walker, unpublished data). Thus, the only site at which the ratio of Treg:Tconv was altered during disease progression was the pancreas-draining LN.

The relative decline in Treg in the pancreatic LN during progression to diabetes suggested the possibility that Treg were protecting from disease in younger animals. Alternatively, given that all DO11 × RIP-mOVA mice developed diabetes, it was conceivable that the Treg were not functioning at all in this model. In an attempt to distinguish between these possibilities, we administered anti-CD25 mAb to deplete CD25+ Treg in DO11 × RIP-mOVA mice. Because Treg accumulation in the pancreatic LN occurred between the ages of 3 and 6wk, we chose to test the requirement for endogenous Treg in this time frame. Anti-CD25 mAb was injected at 3 and 4wk of age and mice were examined at 6 wk of age. Recipients of anti-CD25 mAb showed a clear loss of CD4+CD25+ cells when stained with an Ab that binds a distinct epitope from that used for the depletion (Fig. 3,A). Analysis of CD4+Foxp3+ cells (Fig. 3,B) confirmed that this reflected Treg depletion rather than CD25 down-regulation as reported in the context of certain anti-CD25 mAb depletion protocols (18). Strikingly, depletion of CD25+ cells triggered diabetes in all treated animals by the age of 6 wk while mice treated with control Ab mice remained normoglycemic (Fig. 3 C). In NOD mice, administration of anti-CD25 mAb after 4 wk of age is far less efficient at promoting diabetes than the same treatment administered at 3 wk of age (19). We therefore tested the effect of CD25 depletion in 5-wk-old DO11 × RIP-mOVA mice. Animals that received anti-CD25 mAb at 5 and 6 wk of age were all diabetic by 7 wk of age, whereas recipients of control Ab remained healthy (average blood glucose of 504 ± 87.1 vs 133 ± 7.5, respectively, n = 4). Thus, the fact that 6-wk- old DO11 × RIP-mOVA mice remained healthy despite possessing numerous self-reactive T cells was attributable to the protective effect of endogenous Treg.

FIGURE 3.

Depletion of CD25+ cells causes diabetes in young DO11 × RIP-mOVA mice. DO11 × RIP-mOVA mice were injected i.p. with 500 μg of anti-CD25 (PC61) or control rat IgG (Ctr Ab) at 3 and 4 wk and analyzed at 6 wk of age. A, Peripheral blood samples from control Ab or anti-CD25-injected animals were stained for CD4 and CD25 (clone 7D4). B, Lymphocyte populations were isolated from different sites of control Ab or anti-CD25-injected DO11 × RIP-mOVA mice and stained for CD4 and intracellular Foxp3. Graph shows the percentage of Foxp3+ cells within the CD4+ population for pancreatic LN (PanLN) and pancreas (Pan) samples. C, Blood glucose readings in 6-wk-old control Ab or anti-CD25-injected DO11 × RIP-mOVA mice. Data are compiled from two separate experiments.

FIGURE 3.

Depletion of CD25+ cells causes diabetes in young DO11 × RIP-mOVA mice. DO11 × RIP-mOVA mice were injected i.p. with 500 μg of anti-CD25 (PC61) or control rat IgG (Ctr Ab) at 3 and 4 wk and analyzed at 6 wk of age. A, Peripheral blood samples from control Ab or anti-CD25-injected animals were stained for CD4 and CD25 (clone 7D4). B, Lymphocyte populations were isolated from different sites of control Ab or anti-CD25-injected DO11 × RIP-mOVA mice and stained for CD4 and intracellular Foxp3. Graph shows the percentage of Foxp3+ cells within the CD4+ population for pancreatic LN (PanLN) and pancreas (Pan) samples. C, Blood glucose readings in 6-wk-old control Ab or anti-CD25-injected DO11 × RIP-mOVA mice. Data are compiled from two separate experiments.

Close modal

CD25 depletion demonstrated that Treg in young DO11 × RIP-mOVA mice were functional and that their activity was required to prevent diabetes. However, it was possible that Treg function declined over time, resulting in the onset of disease. Consistent with this hypothesis, it has been suggested that CD25+ cells in NOD mice exhibit reduced suppression when taken from diabetic animals (12, 20, 21). To investigate this idea, we compared the suppressive capacity of Treg from 6-wk normoglycemic DO11 × RIP-mOVA mice with that of 12- to 13-wk-old diabetic mice of the same strain. For these experiments, we used an adoptive transfer model of diabetes based on injection of purified cell populations into RAG-deficient RIP-mOVA mice (mOVA/rag) (22). Sorted CD4+KJ+CD25+ cells from either 6-wk (healthy) or 12- to 13-wk (diabetic) DO11 × RIP-mOVA mice were adoptively transferred into mOVA/rag recipients. Their ability to suppress diabetes induced by cotransferred Tconv (CD4+KJ+CD25) sorted from either healthy or diabetic donor mice was assessed by measuring blood glucose levels (Fig. 4). Representative time-course data showed that CD25 cells from healthy 6-wk-old mice induced diabetes with similar kinetics to those from diabetic donors (Fig. 4,A; ○ vs •, respectively). Cotransfer of Treg was in most cases sufficient to prevent disease; however, there was an indication that Treg from 6-wk donors were not able to prevent diabetes induced by CD25 cells from diabetic donors (Fig. 4A; ▪). Fig. 4B shows pooled data from several experiments. When diabetes was induced by CD25 cells from 6-wk-old donors, Treg isolated from diabetic animals were clearly just as competent at preventing disease as Treg from healthy mice (Fig. 4,Bii). Remarkably, when disease was induced by CD25 cells from diabetic donors (Fig. 4,Biii), Treg from 6-wk-old donors were clearly no longer able to suppress the induction of diabetes. This suggested that CD25 cells from diabetic mice were less susceptible to Treg suppression than CD25 cells from healthy donors. There was no evidence from our experiments that Treg from diabetic animals showed decreased suppressive capacity: in fact, recipients of Treg from diabetic animals maintained lower blood glucose than recipients of Treg from 6-wk animals (Fig. 4,Biii; 253.3 ± 128.8 vs 503.3 ± 70.83, respectively). Treg isolated from the peripheral LN of diabetic or healthy donors showed equivalent expression of Foxp3, CD25, and CTLA-4 (data not shown). One potential explanation for the inability of CD25 cells from diabetic animals to be suppressed is that Treg might act to prevent T cell activation and therefore be ineffective against T cells that have already undergone Ag stimulation. To address this issue, DO11 T cells were activated in vitro before adoptive transfer and the ability of Treg to suppress their pathogenicity was measured. Importantly, Treg were highly effective at preventing disease induced by in vitro-activated DO11 cells (Fig. 4 Biv). Thus, T cell activation per se was not sufficient to confer resistance to Treg suppression: instead such resistance appeared to reflect the nature of T cell activation in mice experiencing autoimmunity.

FIGURE 4.

Treg from diabetic mice are functional in vivo but CD25 cells show resistance to suppression. In brief, 0.2–0.4 × 106 purified CD4+KJ+CD25 (25) were injected i.v. into RIP-mOVA/rag recipient mice. Where indicated, an equal number of CD4+KJ+CD25+ (25+) cells were cotransferred. Donor cells were derived from either healthy 6-wk-old or diabetic 12- to 13-wk-old (diab) DO11 × RIP-mOVA mice. Blood glucose levels were measured every 3–4 days after cell transfer. A, Time course of blood glucose levels in one representative experiment showing a single recipient of each cell transfer. Mice protected from diabetes were still normoglycemic at day 40 postinjection (data not shown). B, Collated data showing blood glucose readings 25–30 days postinjection of RIP-mOVA/rag mice with the indicated cell populations (n ≥ 4 per group). iv, DO11 T cells were activated for 4 days in vitro with 1 μg/ml OVA323–339 peptide and BALB/c splenocytes and coinjected Treg were from 6-wk DO11 × RIP-mOVA mice. Data in B are combined from four independent experiments.

FIGURE 4.

Treg from diabetic mice are functional in vivo but CD25 cells show resistance to suppression. In brief, 0.2–0.4 × 106 purified CD4+KJ+CD25 (25) were injected i.v. into RIP-mOVA/rag recipient mice. Where indicated, an equal number of CD4+KJ+CD25+ (25+) cells were cotransferred. Donor cells were derived from either healthy 6-wk-old or diabetic 12- to 13-wk-old (diab) DO11 × RIP-mOVA mice. Blood glucose levels were measured every 3–4 days after cell transfer. A, Time course of blood glucose levels in one representative experiment showing a single recipient of each cell transfer. Mice protected from diabetes were still normoglycemic at day 40 postinjection (data not shown). B, Collated data showing blood glucose readings 25–30 days postinjection of RIP-mOVA/rag mice with the indicated cell populations (n ≥ 4 per group). iv, DO11 T cells were activated for 4 days in vitro with 1 μg/ml OVA323–339 peptide and BALB/c splenocytes and coinjected Treg were from 6-wk DO11 × RIP-mOVA mice. Data in B are combined from four independent experiments.

Close modal

We reasoned that resistance to Treg suppression might be conferred by the local cytokine environment during the interaction of T cells with self-Ag. We therefore examined cytokine expression in pancreatic LN derived from either 6-wk (healthy) or 12- to 13-k (diabetic) DO11 × RIP-mOVA mice. mRNA was extracted from whole pancreatic LN and cytokine levels were measured by quantitative PCR. In the light of reports that TNF-α and IL-6 have the capacity to release T cells from Treg-mediated suppression (23, 24), we were particularly interested in whether an increase in expression of these cytokines was associated with loss of suppression and progression to diabetes in our model. However, only modest differences in the expression of these cytokines was observed (Fig. 5,A). In contrast, mRNA for IL-21, the newest member of the class I cytokine family (25), was below detection levels in healthy 6-wk- old mice but was clearly present in diabetic 12- to 13-wk-old animals (Fig. 5,A). Further analysis of pancreatic LN cells from diabetic animals indicated that CD4+CD25 cells, but not CD4+CD25+ cells, produced IL-21 (Fig. 5,B). Il-21 mRNA could also be detected in purified CD4+CD25 cells from the inguinal LN, although levels were more variable (mean value 57.13 ± 83.68 for CD4+CD25 cells, undetectable in CD4+CD25+ cells). These findings raised the possibility that IL-21 might play a role in conferring resistance to suppression, an idea that has been suggested by recent human studies (26). To test this hypothesis, we assessed the effect of adding IL-21 to an in vitro suppression assay using concentrations previously used by others to examine biological activity of this cytokine (26, 27, 28). Strikingly, provision of IL-21 was highly effective at abrogating Treg suppression, conferring significant release from suppression even at the highest ratios of Treg:T cell (Fig. 5,C). In most experiments, IL-21 did not enhance the proliferation of CD4 cells incubated in the absence of Treg (Fig. 5,C), although occasionally a modest increase in proliferation was observed (our unpublished data). To exclude the possibility that the increased [3H]thymidine incorporation in cocultures exposed to IL-21 reflected Treg proliferation, rather than abrogated suppression, duplicate wells were analyzed by flow cytometry. The percentage of Foxp3+ cells was markedly lower in the presence of IL-21, consistent with outgrowth of the Foxp3 population rather than proliferation of the Treg population (Fig. 5,C). In line with this, IL-21 did not markedly increase the proliferation of Treg stimulated with anti-CD3 and APC alone (our unpublished data). Thus, in the presence of IL-21, the proliferation of CD25 T cells was no longer inhibited by Treg. Since Treg derived from diabetic animals appeared to have enhanced regulatory capacity (Fig. 4B), we investigated the possibility that suppression by these cells might be resistant to the effects of IL-21. However, IL-21 was still able to abrogate suppression when Treg derived from diabetic animals were cocultured with T cells from normal BALB/c mice (84% suppression in the absence of cytokine reduced to 27% in the presence 100 ng/ml IL-21). A number of other cytokines have been reported to abrogate the effects of Treg suppression including IL-4 (29) and IL-12 (30). Although suppression was somewhat impaired in the presence of these cytokines, we found that IL-21 was far more effective at preventing suppression, particularly at low ratios of Treg:T cell (Fig. 5 D).

FIGURE 5.

IL-21 is overexpressed in diabetic animals and can counteract Treg suppression in vitro. A, Whole pancreatic LN from either healthy 6-wk-old or diabetic (Diab) 12- to 13-wk-old DO11 × RIP-mOVA mice were subjected to mRNA extraction, reverse transcription, and real-time PCR for the indicated cytokines. Symbols represent individual mice. B, Sorted CD4+KJ+CD25+ or CD4+KJ+CD25 cells from the pancreatic LN of 12- to 13-wk-old diabetic DO11 × RIP-mOVA mice were analyzed for IL-21 mRNA as above. C, BALB/c CD4+CD25 cells (2.5 × 104) were incubated with 1 μg/ml anti-CD3 and APC either alone or with differing ratios of CD4+CD25+ cells in the presence of the indicated concentration of IL-21. At day 3, wells were pulsed with [3H]thymidine and harvested for proliferation analysis or duplicate wells were stained for intracellular Foxp3 and analyzed by flow cytometry. Histograms indicate Foxp3 staining in gated CD4 cells in cocultures (1:1 ratio Treg:T cell) incubated in the presence or absence of 200 ng/ml IL-21. Data are representative of four independent experiments. D, A suppression assay was set up as in B in the presence of 50 ng/ml IL-21, 250 ng/ml IL-4, 5 ng/ml IL-12, or 25 ng/ml TNF-α where indicated. Triplicate values were averaged and expressed as a proportion of maximal proliferation. Data are representative of at least two experiments for each cytokine. ∗, p < 0.01 and ∗∗, p < 0.001.

FIGURE 5.

IL-21 is overexpressed in diabetic animals and can counteract Treg suppression in vitro. A, Whole pancreatic LN from either healthy 6-wk-old or diabetic (Diab) 12- to 13-wk-old DO11 × RIP-mOVA mice were subjected to mRNA extraction, reverse transcription, and real-time PCR for the indicated cytokines. Symbols represent individual mice. B, Sorted CD4+KJ+CD25+ or CD4+KJ+CD25 cells from the pancreatic LN of 12- to 13-wk-old diabetic DO11 × RIP-mOVA mice were analyzed for IL-21 mRNA as above. C, BALB/c CD4+CD25 cells (2.5 × 104) were incubated with 1 μg/ml anti-CD3 and APC either alone or with differing ratios of CD4+CD25+ cells in the presence of the indicated concentration of IL-21. At day 3, wells were pulsed with [3H]thymidine and harvested for proliferation analysis or duplicate wells were stained for intracellular Foxp3 and analyzed by flow cytometry. Histograms indicate Foxp3 staining in gated CD4 cells in cocultures (1:1 ratio Treg:T cell) incubated in the presence or absence of 200 ng/ml IL-21. Data are representative of four independent experiments. D, A suppression assay was set up as in B in the presence of 50 ng/ml IL-21, 250 ng/ml IL-4, 5 ng/ml IL-12, or 25 ng/ml TNF-α where indicated. Triplicate values were averaged and expressed as a proportion of maximal proliferation. Data are representative of at least two experiments for each cytokine. ∗, p < 0.01 and ∗∗, p < 0.001.

Close modal

To test whether IL-21 was able to counteract Treg suppression in vivo, we examined T cell responses to pancreas-expressed protein in an adoptive transfer system (Fig. 6). This assay takes advantage of the fact that OVA expressed in the pancreas is trafficked to the pancreatic LN where it can cause T cell proliferation (13). Thy1.1+ DO11 T cells injected into RIP-mOVA mice proliferated in the pancreatic LN but not the inguinal LN (Fig. 6). Cotransfer of Thy1.2+ OVA-specific Treg suppressed this proliferation. Injection of IL-21 abrogated the effects of Treg suppression, such that the proliferative response of the Thy1.1+ T cells was restored. These data provide compelling evidence that provision of IL-21 can overcome Treg suppression in the context of a naturally processed and presented, albeit transgenic, pancreatic Ag. Since IL-21 is up-regulated in DO11 × RIP-mOVA mice at the time of disease progression, and this cytokine interferes with Treg suppression both in vitro and in vivo, collectively these data suggest a novel mechanism for the release of autoreactive T cells from Treg control during induction of spontaneous autoimmunity.

FIGURE 6.

IL-21 counteracts Treg suppression in vivo. Rip-mOVA mice were injected with 1.5 × 106 Thy1.1+ DO11 T cells alone or with an equivalent number of Thy1.2+ DO11 Treg. Where indicated, mice received daily i.p. injections of 1 μg of IL-21. Four days later, recipient mice were killed and the proliferation of DO11+Thy1.1+ T cells was assessed by Ki67 staining. Injected OVA-specific cells were identified by KJ126+ staining then further subdivided on the basis of Thy1.1 staining to distinguish conventional (Thy1.1+) from regulatory (Thy1.1) cells. Each plot shows pooled lymphocytes from two recipient mice.

FIGURE 6.

IL-21 counteracts Treg suppression in vivo. Rip-mOVA mice were injected with 1.5 × 106 Thy1.1+ DO11 T cells alone or with an equivalent number of Thy1.2+ DO11 Treg. Where indicated, mice received daily i.p. injections of 1 μg of IL-21. Four days later, recipient mice were killed and the proliferation of DO11+Thy1.1+ T cells was assessed by Ki67 staining. Injected OVA-specific cells were identified by KJ126+ staining then further subdivided on the basis of Thy1.1 staining to distinguish conventional (Thy1.1+) from regulatory (Thy1.1) cells. Each plot shows pooled lymphocytes from two recipient mice.

Close modal

This study has used a mouse model of spontaneous autoimmune diabetes to address whether disease onset is associated with changes in Treg number or function. We show that over time the ability of Treg to control autoimmunity is undermined by the acquisition of resistance to suppression within the autoreactive T cell population. This is consistent with the intrinsic increase in pathogenic potential of CD25 cells in diabetic NOD mice reported by Adorini and colleagues (11). Our study suggests that this phenomenon is entirely attributable to reduced sensitivity to Treg suppression because the kinetics of disease induction by CD25 cells isolated from diabetic vs healthy donors, when transferred alone, was indistinguishable. Our data are also consistent with those of Chatenoud and colleagues (12) who reported decreased susceptibility of CD25 cells to suppression in diabetic NOD mice. Because Ag specificity was unrestricted in their experiments, the enhanced pathogenicity of CD25 cells following adoptive transfer to NOD/SCID could have reflected the presence of more islet-specific T cells in diabetic animals, rather than altered responses on a per cell basis. Our experiments show that even when T cell specificity is fixed, by sorting cells with a transgenic TCR, CD25 cells isolated from diabetic animals show increased resistance to Treg suppression.

We have found that escape from suppression in DO11 × RIP-mOVA mice is associated with elevated levels of IL-21 in vivo. Provision of exogenous IL-21 during the in vitro coculture of CD25 and CD25+ cells is highly effective at releasing T cells from Treg suppression. Furthermore, we show that suppression of a T cell response to pancreas-derived protein in vivo is also abrogated by administration of IL-21. Mice expressing IL-21 as a transgene exhibit inflammatory infiltrates in multiple tissues (31), and the role of this cytokine in promoting generation of IL-17-expressing lymphocytes has attracted recent attention (28, 32, 33). Our data, in conjunction with the human studies by Peluso et al. (26), suggest that an additional mechanism by which IL-21 might promote autoimmune pathology is by counteracting Treg suppression. Given that the major source of IL-21 is activated CD4 cells (25), this suggests that T cells have the capacity to orchestrate their own escape from Treg suppression. Further characterization of the IL-21-producing population in vivo will be an important step; however, this is currently limited by a lack of suitable reagents for intracellular IL-21 staining. Our data are consistent with the notion that IL-21 acts on Tconv rather than Treg, because Treg derived from IL-21-expressing diabetic animals retained the capacity to suppress. However, further studies are required to confirm this. What is clear from our study is that IL-21 can be induced during the onset of spontaneous autoimmunity and that the presence of this cytokine can override Treg suppression. Of note, elevated levels of IL-21 have been reported in the NOD mouse (34), raising the possibility that this may contribute to the reduced susceptibility to Treg suppression observed in these animals (11, 12).

Although IL-21 is known to influence NK cell development and function and to modulate B cell responses, some confusion exists over its role in T cells. Accordingly, IL-21 has been reported to inhibit T cell IFN-γ production under Th1-inducing conditions (35, 36) and IL-21R−/− T cells show enhanced IFN-γ (37), yet paradoxically IL-21 can also promote IFN-γ production (in synergy with IL-15) (38) and blocking IL-21 can reduce IFN-γ levels (39). Given the ability of regulatory T cells to inhibit IFN-γ production (22, 40, 41), it is plausible that some of these discrepancies can be explained by the ability of IL-21 to release T cells from Treg suppression.

Using our transgenic model, we have been able to analyze pathogenic and regulatory populations specific for a single self- Ag. Within a polyclonal population, the Foxp3+ Treg population constitutes only a minor fraction of lymphocytes; however, the relative numbers of pathogenic and Treg specific for any given self-Ag are not known. In this study, we show that when one focuses on cells specific for a single self-Ag, the regulatory population is numerically dominant in the draining LN at the time of disease regulation. This dominance wanes during progression to overt diabetes. Notably, Treg proportions in peripheral blood and spleen are a poor predictor of diabetes progression. In fact, the only location at which Treg demise correlated with disease onset was the pancreatic LN. This may explain the lack of correlation between Treg proportions in peripheral blood and disease state in humans (3, 4, 7). Interestingly, even after disease onset, the proportion of Treg remained higher in the pancreatic LN than in nondraining LN, suggesting that increased Treg numbers per se are not necessarily indicative of successful suppression. This is compatible with the augmentation of Treg numbers observed in the synovial fluid of rheumatoid arthritis patients (6) and in the CNS of mice (23) and humans (42) experiencing experimental autoimmune encephalomyelitis or multiple sclerosis respectively.

The observation that Treg proportions in the pancreas remained constant during disease progression was somewhat surprising given the differences observed in the pancreatic LN. Available data suggest that Treg may exert suppressive function both in the pancreatic LN (22, 41) and within the pancreas itself (43, 44); however, the mechanism of action at these two locations may differ. Whether suppression at a given site attenuates proliferation as well as cytokine production, and whether one Treg population invokes another (45), will alter the ratio of Treg observed. The efficiency with which activated T cells at different sites recruit or maintain Treg will also impact on this ratio. Accordingly, increased production of IL-2 in NOD Idd3 congenics augments Treg proportions in the pancreatic LN (46). Thus, it is conceivable that the mechanisms that determine the ratio of Treg:Tconv in the tissue and draining LN differ.

One recent study in the NOD mouse did not reveal a decrease in the percentage of Foxp3+ cells in the pancreatic LN during progression to diabetes (47). Two factors are likely to underlie this discrepancy. First, diabetes progression in DO11 × RIP-mOVA mice is likely to be more synchronous than in NOD colonies, where onset of disease is less predictable and frequently penetrance is not complete. Second, trends in Treg proportions may exist at the level of a single specificity, but may be masked when a mixture of different islet reactivities is analyzed during a polyclonal autoimmune response. It is likely that the changes in Treg proportions evident in this study are only seen because the Ag specificity of pathogenic and Treg has been fixed by a TCR transgene, augmenting the respective populations to a readily measurable size. This has allowed visualization of the interplay between pathogenic and regulatory populations specific for a single self-Ag. Notwithstanding the exaggerated precursor frequencies of the TCR-transgenic system, the concept that aggressive and protective T cells might share a common specificity draws support from the observation that pathogenic and Treg use overlapping sets of TCRs (48). Indeed, by tracking a single self-reactive TCR preferentially used by Foxp3+ cells, these authors were able to demonstrate this same TCR within the CD25 population, albeit at a very low frequency. Recent extensive analysis of TCR usage in Treg and non-Treg supports the idea that TCRs are shared but present at differing frequencies between these two subsets (49). An overlap between the TCR repertoires of CD25 and CD25high cells has also been documented in humans (50). Thus, a scenario in which Treg and pathogenic T cells recognize the same self-ligand is not inconceivable.

Evidence from the NOD model of diabetes has suggested that the functional capacity of Treg may be reduced in older diabetic animals (12, 20, 21). Our data do not support this and indeed point to increased Treg function during diabetes, mirroring the enhanced suppressive capacity of Treg isolated from the CNS of mice experiencing experimental autoimmune encephalomyelitis (51). The impaired function of Treg from diabetic NOD mice was reported to correlate with reduced surface TGF-β on Treg (21) or decreased TGF-β mRNA within Foxp3+ cells (20). However, Chatenoud and colleagues (12) reported no change in surface TGF-β expression on Treg between 6-wk healthy and older diabetic NOD animals. Moreover, it has been shown that Treg do not themselves have to make TGF-β to mediate TGF-β-dependent suppression (52), suggesting that the ability to synthesize TGF-β mRNA may not be the determinant factor for Treg suppression. An alternative possibility is that CD25+ populations from diabetic mice are more prone to contamination with activated Tconv than cell preparations from healthy mice, and that this leads to an apparent reduction in suppressive capacity. In this regard, it should be noted that CD25+ cells from diabetic NOD mice were shown to produce high amounts of IFN-γ upon in vitro stimulation (12); although it is not inconceivable that IFN-γ is produced by regulatory subsets (53), it is also possible that this reflects contamination with activated pathogenic cells.

IL-21 is known to inhibit de novo induction of Foxp3+ Treg in response to TGF-β (28, 32, 33). Our data indicate that in addition to thwarting differentiation of adaptive Treg, IL-21 also cripples the effectiveness of natural Treg. Because recent work shows that polymorphisms within the genes for IL-21 and its receptor are associated with type 1 diabetes in humans (54), these findings provide strong impetus for targeting IL-21 in this disease setting. The efficacy of IL-21 blockade in animal models of rheumatoid arthritis (55) and systemic lupus erythematosus (56) is encouraging in this respect.

We are grateful to Abul Abbas, Qizhi Tang, and David Sansom for critical reading of this manuscript and Nadia Sarween for immunohistology.

We have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by a Medical Research Council Career Development Fellowship (to L.S.K.W.). L.E.C. is funded by a Biotechnology and Biological Sciences Research Council/AstraZeneca Case Studentship.

3

Abbreviations used in this paper: Treg, regulatory T cell; RIP, rat insulin promoter; Tconv, conventional T cell; mOVA, mouse OVA; LN, lymph node.

1
Kim, J. M., J. P. Rasmussen, A. Y. Rudensky.
2007
. Regulatory T cells prevent catastrophic autoimmunity throughout the lifespan of mice.
Nat. Immunol.
8
:
191
-197.
2
Kukreja, A., G. Cost, J. Marker, C. Zhang, Z. Sun, K. Lin-Su, S. Ten, M. Sanz, M. Exley, B. Wilson, et al
2002
. Multiple immuno-regulatory defects in type-1 diabetes.
J. Clin. Invest.
109
:
131
-140.
3
Liu, W., A. L. Putnam, Z. Xu-Yu, G. L. Szot, M. R. Lee, S. Zhu, P. A. Gottlieb, P. Kapranov, T. R. Gingeras, B. Fazekas de St. Groth, et al
2006
. CD127 expression inversely correlates with FoxP3 and suppressive function of human CD4+ T reg cells.
J. Exp. Med.
203
:
1701
-1711.
4
Brusko, T., C. Wasserfall, K. McGrail, R. Schatz, H. L. Viener, D. Schatz, M. Haller, J. Rockell, P. Gottlieb, M. Clare-Salzler, M. Atkinson.
2007
. No alterations in the frequency of FOXP3+ regulatory T-cells in type 1 diabetes.
Diabetes
56
:
604
-612.
5
Bottazzo, G. F., B. M. Dean, J. M. McNally, E. H. MacKay, P. G. Swift, D. R. Gamble.
1985
. In situ characterization of autoimmune phenomena and expression of HLA molecules in the pancreas in diabetic insulitis.
N. Engl. J. Med.
313
:
353
-360.
6
van Amelsfort, J. M., K. M. Jacobs, J. W. Bijlsma, F. P. Lafeber, L. S. Taams.
2004
. CD4+CD25+ regulatory T cells in rheumatoid arthritis: differences in the presence, phenotype, and function between peripheral blood and synovial fluid.
Arthritis Rheum.
50
:
2775
-2785.
7
Lindley, S., C. M. Dayan, A. Bishop, B. O. Roep, M. Peakman, T. I. Tree.
2005
. Defective suppressor function in CD4+CD25+ T-cells from patients with type 1 diabetes.
Diabetes
54
:
92
-99.
8
Ehrenstein, M. R., J. G. Evans, A. Singh, S. Moore, G. Warnes, D. A. Isenberg, C. Mauri.
2004
. Compromised function of regulatory T cells in rheumatoid arthritis and reversal by anti-TNFα therapy.
J. Exp. Med.
200
:
277
-285.
9
Viglietta, V., C. Baecher-Allan, H. L. Weiner, D. A. Hafler.
2004
. Loss of functional suppression by CD4+CD25+ regulatory T cells in patients with multiple sclerosis.
J. Exp. Med.
199
:
971
-979.
10
Kriegel, M. A., T. Lohmann, C. Gabler, N. Blank, J. R. Kalden, H. M. Lorenz.
2004
. Defective suppressor function of human CD4+ CD25+ regulatory T cells in autoimmune polyglandular syndrome type II.
J. Exp. Med.
199
:
1285
-1291.
11
Gregori, S., N. Giarratana, S. Smiroldo, L. Adorini.
2003
. Dynamics of pathogenic and suppressor T cells in autoimmune diabetes development.
J. Immunol.
171
:
4040
-4047.
12
You, S., M. Belghith, S. Cobbold, M. A. Alyanakian, C. Gouarin, S. Barriot, C. Garcia, H. Waldmann, J. F. Bach, L. Chatenoud.
2005
. Autoimmune diabetes onset results from qualitative rather than quantitative age-dependent changes in pathogenic T-cells.
Diabetes
54
:
1415
-1422.
13
Walker, L. S., A. Chodos, M. Eggena, H. Dooms, A. K. Abbas.
2003
. Antigen-dependent proliferation of CD4+CD25+ regulatory T cells in vivo.
J. Exp. Med.
198
:
249
-258.
14
Anderson, M. S., E. S. Venanzi, Z. Chen, S. P. Berzins, C. Benoist, D. Mathis.
2005
. The cellular mechanism of Aire control of T cell tolerance.
Immunity
23
:
227
-239.
15
Kurts, C., W. R. Heath, F. R. Carbone, J. Allison, J. F. Miller, H. Kosaka.
1996
. Constitutive class I-restricted exogenous presentation of self antigens in vivo.
J. Exp. Med.
184
:
923
-930.
16
Fontenot, J. D., J. L. Dooley, A. G. Farr, A. Y. Rudensky.
2005
. Developmental regulation of Foxp3 expression during ontogeny.
J. Exp. Med.
202
:
901
-906.
17
Hoglund, P., J. Mintern, C. Waltzinger, W. Heath, C. Benoist, D. Mathis.
1999
. Initiation of autoimmune diabetes by developmentally regulated presentation of islet cell antigens in the pancreatic lymph nodes.
J. Exp. Med.
189
:
331
-339.
18
Kohm, A. P., J. S. McMahon, J. R. Podojil, W. S. Begolka, M. DeGutes, D. J. Kasprowicz, S. F. Ziegler, S. D. Miller.
2006
. Cutting edge: Anti-CD25 monoclonal antibody injection results in the functional inactivation, not depletion, of CD4+CD25+ T regulatory cells.
J. Immunol.
176
:
3301
-3305.
19
Billiard, F., E. Litvinova, D. Saadoun, F. Djelti, D. Klatzmann, J. L. Cohen, G. Marodon, B. L. Salomon.
2006
. Regulatory and effector T cell activation levels are prime determinants of in vivo immune regulation.
J. Immunol.
177
:
2167
-2174.
20
Pop, S. M., C. P. Wong, D. A. Culton, S. H. Clarke, R. Tisch.
2005
. Single cell analysis shows decreasing FoxP3 and TGFβ1 coexpressing CD4+CD25+ regulatory T cells during autoimmune diabetes.
J. Exp. Med.
201
:
1333
-1346.
21
Gregg, R. K., R. Jain, S. J. Schoenleber, R. Divekar, J. J. Bell, H. H. Lee, P. Yu, H. Zaghouani.
2004
. A sudden decline in active membrane-bound TGF-β impairs both T regulatory cell function and protection against autoimmune diabetes.
J. Immunol.
173
:
7308
-7316.
22
Sarween, N., A. Chodos, C. Raykundalia, M. Khan, A. K. Abbas, L. S. Walker.
2004
. CD4+CD25+ cells controlling a pathogenic CD4 response inhibit cytokine differentiation, CXCR-3 expression, and tissue invasion.
J. Immunol.
173
:
2942
-2951.
23
Korn, T., J. Reddy, W. Gao, E. Bettelli, A. Awasthi, T. R. Petersen, B. T. Backstrom, R. A. Sobel, K. W. Wucherpfennig, T. B. Strom, et al
2007
. Myelin-specific regulatory T cells accumulate in the CNS but fail to control autoimmune inflammation.
Nat. Med.
13
:
423
-431.
24
Valencia, X., G. Stephens, R. Goldbach-Mansky, M. Wilson, E. M. Shevach, P. E. Lipsky.
2006
. TNF downmodulates the function of human CD4+CD25high T-regulatory cells.
Blood
108
:
253
-261.
25
Parrish-Novak, J., S. R. Dillon, A. Nelson, A. Hammond, C. Sprecher, J. A. Gross, J. Johnston, K. Madden, W. Xu, J. West, et al
2000
. Interleukin 21 and its receptor are involved in NK cell expansion and regulation of lymphocyte function.
Nature
408
:
57
-63.
26
Peluso, I., M. C. Fantini, D. Fina, R. Caruso, M. Boirivant, T. T. MacDonald, F. Pallone, G. Monteleone.
2007
. IL-21 counteracts the regulatory T cell-mediated suppression of human CD4+ T lymphocytes.
J. Immunol.
178
:
732
-739.
27
Pelletier, M., A. Bouchard, D. Girard.
2004
. In vivo and in vitro roles of IL-21 in inflammation.
J. Immunol.
173
:
7521
-7530.
28
Korn, T., E. Bettelli, W. Gao, A. Awasthi, A. Jager, T. B. Strom, M. Oukka, V. K. Kuchroo.
2007
. IL-21 initiates an alternative pathway to induce proinflammatory TH17 cells.
Nature
448
:
484
-487.
29
Thornton, A. M., C. A. Piccirillo, E. M. Shevach.
2004
. Activation requirements for the induction of CD4+CD25+ T cell suppressor function.
Eur. J. Immunol.
34
:
366
-376.
30
King, I. L., B. M. Segal.
2005
. Cutting edge: IL-12 induces CD4+CD25 T cell activation in the presence of T regulatory cells.
J. Immunol.
175
:
641
-645.
31
Parrish-Novak, J., D. C. Foster, R. D. Holly, C. H. Clegg.
2002
. Interleukin-21 and the IL-21 receptor: novel effectors of NK and T cell responses.
J. Leukocyte Biol.
72
:
856
-863.
32
Nurieva, R., X. O. Yang, G. Martinez, Y. Zhang, A. D. Panopoulos, L. Ma, K. Schluns, Q. Tian, S. S. Watowich, A. M. Jetten, C. Dong.
2007
. Essential autocrine regulation by IL-21 in the generation of inflammatory T cells.
Nature
448
:
480
-483.
33
Zhou, L., I. I. Ivanov, R. Spolski, R. Min, K. Shenderov, T. Egawa, D. E. Levy, W. J. Leonard, D. R. Littman.
2007
. IL-6 programs TH-17 cell differentiation by promoting sequential engagement of the IL-21 and IL-23 pathways.
Nat. Immunol.
8
:
967
-974.
34
King, C., A. Ilic, K. Koelsch, N. Sarvetnick.
2004
. Homeostatic expansion of T cells during immune insufficiency generates autoimmunity.
Cell
117
:
265
-277.
35
Wurster, A. L., V. L. Rodgers, A. R. Satoskar, M. J. Whitters, D. A. Young, M. Collins, M. J. Grusby.
2002
. Interleukin 21 is a T helper (Th) cell 2 cytokine that specifically inhibits the differentiation of naive Th cells into interferon γ-producing Th1 cells.
J. Exp. Med.
196
:
969
-977.
36
Suto, A., A. L. Wurster, S. L. Reiner, M. J. Grusby.
2006
. IL-21 inhibits IFN-γ production in developing Th1 cells through the repression of eomesodermin expression.
J. Immunol.
177
:
3721
-3727.
37
Frohlich, A., B. J. Marsland, I. Sonderegger, M. Kurrer, M. R. Hodge, N. L. Harris, M. Kopf.
2007
. IL-21 receptor signaling is integral to the development of Th2 effector responses in vivo.
Blood
109
:
2023
-2031.
38
Strengell, M., S. Matikainen, J. Siren, A. Lehtonen, D. Foster, I. Julkunen, T. Sareneva.
2003
. IL-21 in synergy with IL-15 or IL-18 enhances IFN-γ production in human NK and T cells.
J. Immunol.
170
:
5464
-5469.
39
Monteleone, G., I. Monteleone, D. Fina, P. Vavassori, G. Del Vecchio Blanco, R. Caruso, R. Tersigni, L. Alessandroni, L. Biancone, G. C. Naccari, et al
2005
. Interleukin-21 enhances T-helper cell type I signaling and interferon-γ production in Crohn’s disease.
Gastroenterology
128
:
687
-694.
40
DiPaolo, R. J., D. D. Glass, K. E. Bijwaard, E. M. Shevach.
2005
. CD4+CD25+ T cells prevent the development of organ-specific autoimmune disease by inhibiting the differentiation of autoreactive effector T cells.
J. Immunol.
175
:
7135
-7142.
41
Tang, Q., J. Y. Adams, A. J. Tooley, M. Bi, B. T. Fife, P. Serra, P. Santamaria, R. M. Locksley, M. F. Krummel, J. A. Bluestone.
2006
. Visualizing regulatory T cell control of autoimmune responses in nonobese diabetic mice.
Nat. Immunol.
7
:
83
-92.
42
Feger, U., C. Luther, S. Poeschel, A. Melms, E. Tolosa, H. Wiendl.
2007
. Increased frequency of CD4+CD25+ regulatory T cells in the cerebrospinal fluid but not in the blood of multiple sclerosis patients.
Clin. Exp. Immunol.
147
:
412
-418.
43
Chen, Z., A. E. Herman, M. Matos, D. Mathis, C. Benoist.
2005
. Where CD4+CD25+ T reg cells impinge on autoimmune diabetes.
J. Exp. Med.
202
:
1387
-1397.
44
Brode, S., T. Raine, P. Zaccone, A. Cooke.
2006
. Cyclophosphamide-induced type-1 diabetes in the NOD mouse is associated with a reduction of CD4+CD25+Foxp3+ regulatory T cells.
J. Immunol.
177
:
6603
-6612.
45
Tarbell, K. V., L. Petit, X. Zuo, P. Toy, X. Luo, A. Mqadmi, H. Yang, M. Suthanthiran, S. Mojsov, R. M. Steinman.
2007
. Dendritic cell-expanded, islet-specific CD4+CD25+ CD62L+ regulatory T cells restore normoglycemia in diabetic NOD mice.
J. Exp. Med.
204
:
191
-201.
46
Yamanouchi, J., D. Rainbow, P. Serra, S. Howlett, K. Hunter, V. E. Garner, A. Gonzalez-Munoz, J. Clark, R. Veijola, R. Cubbon, et al
2007
. Interleukin-2 gene variation impairs regulatory T cell function and causes autoimmunity.
Nat. Genet.
39
:
329
-337.
47
Mellanby, R. J., D. Thomas, J. M. Phillips, A. Cooke.
2007
. Diabetes in non-obese diabetic mice is not associated with quantitative changes in CD4+CD25+Foxp3+ regulatory T cells.
Immunology
121
:
15
-28.
48
Hsieh, C. S., Y. Zheng, Y. Liang, J. D. Fontenot, A. Y. Rudensky.
2006
. An intersection between the self-reactive regulatory and nonregulatory T cell receptor repertoires.
Nat. Immunol.
7
:
401
-410.
49
Pacholczyk, R., J. Kern, N. Singh, M. Iwashima, P. Kraj, L. Ignatowicz.
2007
. Nonself-antigens are the cognate specificities of Foxp3+ regulatory T cells.
Immunity
27
:
493
-504.
50
Fazilleau, N., H. Bachelez, M. L. Gougeon, M. Viguier.
2007
. Cutting edge: size and diversity of CD4+CD25high Foxp3+ regulatory T cell repertoire in humans: evidence for similarities and partial overlapping with CD4+CD25 T cells.
J. Immunol.
179
:
3412
-3416.
51
McGeachy, M. J., L. A. Stephens, S. M. Anderton.
2005
. Natural recovery and protection from autoimmune encephalomyelitis: contribution of CD4+CD25+ regulatory cells within the central nervous system.
J. Immunol.
175
:
3025
-3032.
52
Fahlen, L., S. Read, L. Gorelik, S. D. Hurst, R. L. Coffman, R. A. Flavell, F. Powrie.
2005
. T cells that cannot respond to TGF-β escape control by CD4+CD25+ regulatory T cells.
J. Exp. Med.
201
:
737
-746.
53
Sawitzki, B., C. I. Kingsley, V. Oliveira, M. Karim, M. Herber, K. J. Wood.
2005
. IFN-γ production by alloantigen-reactive regulatory T cells is important for their regulatory function in vivo.
J. Exp. Med.
201
:
1925
-1935.
54
Asano, K., H. Ikegami, T. Fujisawa, M. Nishino, K. Nojima, Y. Kawabata, S. Noso, Y. Hiromine, A. Fukai, T. Ogihara.
2007
. Molecular scanning of interleukin-21 gene and genetic susceptibility to type 1 diabetes.
Hum. Immunol.
68
:
384
-391.
55
Young, D. A., M. Hegen, H. L. Ma, M. J. Whitters, L. M. Albert, L. Lowe, M. Senices, P. W. Wu, B. Sibley, Y. Leathurby, et al
2007
. Blockade of the interleukin-21/interleukin-21 receptor pathway ameliorates disease in animal models of rheumatoid arthritis.
Arthritis Rheum.
56
:
1152
-1163.
56
Herber, D., T. P. Brown, S. Liang, D. A. Young, M. Collins, K. Dunussi-Joannopoulos.
2007
. IL-21 has a pathogenic role in a lupus-prone mouse model and its blockade with IL-21R.Fc reduces disease progression.
J. Immunol.
178
:
3822
-3830.