Nitric Oxide–Induced Regulatory T Cells Inhibit Th17 but Not Th1 Cell Differentiation and Function

Nitric oxide mediates a variety of biological functions, including vascular relaxation, platelet aggregation, neurotransmission, as well as microbicidal, tumoricidal, and immune regulation (1–5). As such, NO is associated with some of the most important immunopathologies such as rheumatoid arthritis, diabetes, systemic lupus erythematosus, multiple sclerosis, and septic shock (6–10). NO is derived from the guanidino nitrogen atom(s) and molecular oxygen in a reaction catalyzed by the enzyme NO synthase. There are three forms of NO synthase: the neuronal form, the inducible form, and the endothelial-derived form. Neuronal NO synthase and endothelial-derived NO synthase produce amounts of NO required for physiological functions. Inducible NO synthase normally exists at a low level, is activated by a number of immunological stimuli such as IFN-γ, TNF-α, and LPS, and catalyzes the high output of NO, which can be cytotoxic and kill intracellular pathogens.

We have previously shown that a low dose of NO selectively enhances Th1 cell differentiation and expansion under Th1 polarizing conditions. This was mediated by enhancing the expression of IL-12Rβ2 through a cGMP-dependent pathway (11). We have also reported that NO can induce a subset of CD4⁺CD25⁺Foxp3⁻ regulatory T cells (NO-Tregs) from CD4⁺CD25⁻ T cells via p53, IL-2, and OX40 in a cGMP-independent manner (12). NO-Tregs suppress the proliferation and expansion of CD4⁺CD25⁻ effector cells in vitro and markedly attenuate the effector cell–mediated colitis and collagen-induced arthritis in the mouse in an IL-10–dependent manner. More recently, we reported that NO can also directly suppress the differentiation and function of polarized human and mouse Th17 cells via the downregulation of the aryl hydrocarbon receptor leading to attenuation of experimental autoimmune encephalomyelitis (EAE) (13). These results therefore indicate that NO is a key player in the modulation of inflammatory disease. To further explore the role of NO in immune regulation, we investigated the potential role of NO-Tregs in affecting the differentiation and function of different subsets of Th cells, particularly Th17 and Th1 cells.

Th17 cells are now defined as a distinct subset of Th cells that produce IL-17 with a signature transcription factor retinoid acid–related orphan receptor (ROR)γt. Th17 cells are associated with pathogenesis of human autoimmune diseases, including multiple sclerosis, rheumatoid arthritis, psoriasis, and inflammatory bowel disease (14–17). Thus, there likely exist rigorous endogenous control mechanisms to limit Th17 proliferation and function.

We report in this study that NO-Tregs effectively suppress the differentiation and expansion of Th17 but not Th1 cells. Unexpectedly, we found that natural Tregs (nTregs), although effective in suppressing Th1, are relatively ineffective in suppression of Th17. NO-Tregs suppress Th17 via a cell contact–dependent mechanism and also partially via IL-10. In vivo, NO-Tregs markedly attenuate EAE in association with the reduction of Th17 but not Th1. Taken together, these results support a key role of NO in the regulation of the immune system and further indicate compartmentalization of the regulatory mechanism in inflammatory disease.

Wild-type C57BL/6 and BALB/c mice were purchased from Harlan Olac. Il10^−/− mice of the C57BL/6 background were obtained from the National Institute of Medical Research (London, U.K.). Males and females mice were used at the age of 6–10 wk. All animal experiments were performed according to the University of Glasgow guidelines and the United Kingdom Home Office regulations. (Z)-1-[2-(2-aminoethyl)-N-(2-ammononioethyl)amino]diazen-1-ium-1,2-diolate (NOC-18, also known as DETA-NONOate) was purchased from Enzo Life Sciences (Exeter, U.K.). Anti-CTLA4 was purchased from eBioscience.

CD4⁺, CD4⁺CD25⁻, and CD4⁺CD25⁺ T cells were purified from the pool of spleens and lymph nodes of naive mice by negative selection (routine purity >98%) using an autoMACS (Miltenyi Biotec). Culture medium was RPMI 1640 supplemented with 10% (v/v) FCS (Lonza), 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 0.05 M 2-ME. To obtain NO-Tregs, CD4⁺CD25⁻ T cells were cultured in medium at 2.5 × 10⁴ cells/200 μl in 96-well round-bottom plates (Nunc) with soluble anti-CD3 (1 μg/ml; R&D Systems) and equal numbers of APCs (mitomycin C–treated spleen cells) for 5–6 d in the presence of 200 μM NOC-18 added at the beginning of culture. The cells were then harvested, washed, and added at graded numbers to other cell cultures as indicated. For Th17 cell differentiation, CD4⁺ T cells (5 × 10⁵ cells/ml) were cultured in round-bottom 96-well plates with equal numbers of APCs and soluble anti-CD3 (1 μg/ml), TGF-β (1 ng/ml), IL-6 (10 ng/ml), IL-1β (10 ng/ml), anti–IFN-γ (10 μg/ml), and anti–IL-4 (10 μg/ml) for up to 5 d. For Th1 cell differentiation, CD4⁺ T cells were cultured in round-bottom 96-well plates with equal numbers of APCs in the presence of soluble anti-CD3 (1 μg/ml), IL-12 (10 ng/ml), and anti–IL-4 (10 μg/ml) for 4 d. For Th2 cell differentiation, CD4⁺ T cells (5 × 10⁵ cells/ml) were cultured in 96-well round-bottom plates with equal numbers of APCs, anti-CD3 (1 μg/ml), IL-4 (10 ng/ml), and anti–IFN-γ (10 μg/ml) for 4 d. For Th9 cell differentiation, CD4⁺ T cells (5 × 10⁵ cells/ml) were cultured for up to 6 d in flat-bottom 96-well plates with plate-bound anti-CD3 (3 μg/ml) and soluble anti-CD28 (1.5 μg/ml), TGF-β (5 ng/ml), and IL-4 (10 ng/ml). For suppression assay, NO-Tregs or nTregs were added at the beginning of the polarization culture or established T cells at various Treg/effector cell ratios. The mixed cell populations were cultured for up to 6 d in 96-well round-bottom plates. In some experiments, NO-Tregs were treated with mitomycin C before being added to the suppression culture. In other experiments, NO-Tregs were placed in the upper well and the effector cells in the bottom well of a Transwell (Corning) culture. All cytokines and Abs were obtained from R&D Systems. Supernatants were collected and cytokine concentrations were determined by ELISA, and the cells were harvested for FACS analysis. Cell division was determined by labeling CD4⁺ T cells with CFSE (Molecular Probes), and fluorescence density was determined by FACS.

C57BL/6 mice were immunized s.c. on the back with 100 μg myelin oligodendrocyte glycoprotein (MOG)_35–55 peptide (Sigma-Genosys) in 50 μl PBS emulsified with an equal volume of CFA (total 125 μg Mycobacterium tuberculosis, strain H37RA; Difco, Detroit MI). Each mouse also received i.p. 100 ng/200 μl pertussis toxin (Sigma-Aldrich, Poole, U.K.) in PBS on days 0 and 2 after immunization. EAE was scored according to a 0–5 scale as follows: 0, no clinical sign; 1, complete loss of tail tone; 2, hind limb weakness; 3, hind limb paralysis; 4, forelimb involvement; 5, moribund.

Mice were euthanized in a CO₂ chamber and their spinal cords flushed out with PBS by hydrostatic pressure using a syringe attached to an 18-gauge needle; the brains were also harvested. For histology analysis, the tissues were then placed in 10% buffered formalin fixative overnight before being transferred to 75% ethanol. Tissues were then embedded in paraffin wax. Spinal cord and brain sections were stained with standard H&E. For FACS analysis, brains and spinal cords were cut into small pieces and digested with collagenase A (2 mg/ml; Roche Diagnostics) and DNase (1 mg/ml; Sigma-Aldrich) at 37°C for 40 min. Mononuclear cells were isolated by passing the tissue through a cell strainer (100 μM), followed by a Percoll gradient (70/30%) centrifugation. Mononuclear cells were removed from the interphase, washed, and resuspended in culture medium for further analysis. Draining lymph nodes (DLNs) were also harvested, passed through a cell strainer, and resuspended in culture medium for cellular analysis ex vivo by FACS.

For intracellular cytokine staining, the cells were restimulated for 4 h with 50 ng/ml PMA and 500 ng/ml ionomycin (both from Sigma-Aldrich) in the presence of GolgiStop (BD Bioscience). The cells were first stained for CD4, then permeabilized with Perm/Fix solution (eBioscience), and finally stained with anti–IL-17A, anti–IFN-γ, anti–IL-4, anti–IL-9, anti-RORγt, anti–T-bet, or anti-GATA3 (all from eBioscience). Data were acquired using a FACSCalibur (BD Biosciences) and analyzed using FlowJo software (Tree Star). Isotype-matched Abs (directly conjugated) were used as controls.

ELISAs were carried out with paired Abs for murine IL-9, IL-10, and IL-17A and IFN-γ (all from BD Biosciences) according to the manufacturer’s instructions. Sensitivity of the assays was <20 pg/ml.

Statistical analyses were performed using GraphPad Prism 5.0. Comparisons between two groups were performed using a two-tailed unpaired Student t test. Multiple groups were compared using a two-way ANOVA followed by a Bonferroni posttest. Values for all measurements are expressed as means ± SEM. A p value <0.05 was considered statistically significant. All experiments were performed at least twice.

We first investigated the possibility that NO-Tregs may inhibit the induction of Th17 cells. CD4⁺CD25⁻ T cells were purified from the lymph node and spleen of BALB/c mice. The cells were polarized to NO-Tregs with soluble anti-CD3 Ab and mitomycin C–treated spleen cells (as APCs) in the presence of 200 μM NOC-18 (a stable NO donor; 100 μM NOC-18 constantly release 200 nM NO with t_1/2 of 20 h) (18) as described previously (12). The cells were harvested on day 5 or 6 and checked for the NO-Treg phenotypes, which were CD4⁺CD25⁺Foxp3⁻, producing IL-10 but not IFN-γ, IL-4, IL-17, or IL-2 (Ref. 12 and data not shown). NO-Tregs were washed and added at the beginning of the culture of polarization to Th17 cells. Naive CD4⁺ T cells were labeled with CFSE and cultured with APCs, anti-CD3/CD28 plus TGF-β, IL-6, IL-1β, IL-23, anti–IFN-γ, and anti–IL-4 with graded CD4⁺/NO-Treg ratios. The replication of Th17 cells and the production of IL-17 were determined on day 4 by FACS and ELISA, respectively. NO-Tregs inhibited the polarization of Th17 cells in a cell ratio–dependent manner (Fig. 1A, 1B). The inhibition of Th17 polarization was confirmed by the reduction in the concentration of IL-17A in the culture supernatant (Fig. 1C). Furthermore, NO-Tregs also inhibited the expression of RORγt, the signature transcription factor of Th17 (19) (Fig. 1D). These results indicate that NO-Tregs may be potent Tregs for Th17 development.

We next compared the suppressive function of NO-Tregs with nTregs. NO-Tregs were prepared as above and nTregs (CD4⁺CD25⁺) were purified from the spleen and LN of naive BALB/c mice. CD4⁺ T cells were labeled with CFSE and polarized to Th17 as above in the presence of NO-Tregs or nTregs for 4 d. Whereas NO-Tregs strongly suppressed the differentiation of Th17 cells, nTregs had only a modest effect (Fig. 2A). Furthermore, whereas NO-Tregs markedly inhibited the number of IL-17⁺RORγt⁺ T cells, nTregs did not (Fig. 2B). We then compared the suppressive effect of NO-Tregs and nTregs on Th1 cell differentiation. Naive CD4⁺ T cells were polarized to Th1 cells in cultures containing anti-CD3/CD28 plus IL-12 and anti–IL-4 in the presence of NO-Tregs or nTregs. As expected, nTregs potently suppressed the differentiation of IFN-γ⁺ T cells (Fig. 2C) and the expression of T-bet on these cells, the signature transcription factor of Th1. In contrast, NO-Tregs failed to affect the differentiation of Th1 cells and T-bet expression (Fig. 2D). These results therefore demonstrate a differential effect of the suppressive functions of NO-Tregs and nTregs.

We next investigated the relative suppressive effect of NO-Tregs and nTregs on established Th17 and Th1 cells. Th17 and Th1 cells were cultured with two rounds of polarization as above, harvested on day 4, washed, and recultured with NO-Tregs or nTregs at various Treg/T cell (Th17 and Th1) ratios as described in 2Materials and Methods. Cells were harvested on day 4 and were analyzed by FACS (for cytokine and transcription factor expression) and by ELISA (for cytokines in the culture supernatants). Cell proliferation was assessed on day 3 by [³H]thymidine uptake. NO-Tregs potently decreased the percentage of IL-17⁺RORγt⁺ cells, whereas nTregs registered only a modest and not significant inhibition of Th17 cells (Fig. 3A). This differential suppressive effect was evident across a range of Treg/ Th17 ratio (Fig. 3B) and was supported by the results of the relative concentrations of IL-17A in the culture supernatants (Fig. 3C) and cellular proliferation (Fig. 3D). Under the present culture conditions, NO-Tregs are ∼16-fold more effective (on cell–for-cell basis) than nTregs in suppressing Th17 cells. In contrast, whereas nTregs potently decreased the percentage of established Th1 (IFNγ⁺/T-bet⁺) cells, NO-Tregs did not (Fig. 3E). Again, these differential suppressive effects of NO-Tregs and nTregs were observed over a range of Treg/Th1 ratios (Fig. 3F) and confirmed by the concentrations of IFN-γ in the culture supernatants (Fig. 3G) and by cellular proliferation (Fig. 3H). Under these culture conditions, nTregs are ∼8-fold more effective than NO-Tregs in suppressing Th1. We also examined the relative suppressive effects of NO-Tregs and nTregs on established Th2 cells and Th9 cells. NO-Tregs and nTregs were equally suppressive against Th2 (Supplemental Fig. 1) and Th9 cells (Supplemental Fig. 2). Taken together, these results corroborate the finding that NO-Tregs and nTregs have differential suppressive effects on the proinflammatory Th17 and Th1 cells.

We next investigated the mechanism by which NO-Tregs inhibit the differentiation of Th17 cells. We first tested the potential role of CTLA4, which has been reported to be associated with the suppressive activity of nTregs (20). NO-Tregs were added to the polarization culture of Th17 cells at 1:1 ratio in the presence of a neutralizing anti-CTLA4 Ab. The suppressive effect of NO-Tregs was not affected by anti-CTLA4 (Fig. 4A). We also tested the requirement of metabolically active NO-Tregs in the suppressive function. NO-Tregs were pretreated with mitomycin C, washed, and then added to the polarizing culture of Th17 cells. The mitomycin C treatment did not affect the suppressive activity of NO-Tregs (Fig. 4A). These results were confirmed by the relative concentrations of IL-17A in the culture supernatants (Fig. 4B). Because mitomycin C did not affect the function of NO-Tregs, we tested the possibility that NO-Tregs assert their suppressive activity via a cell contact–dependent mechanism. Th17 cells and NO-Tregs were cultured together or separated by a semipermeable membrane in a Transwell culture under Th17 polarizing conditions. The suppressive effect of NO-Tregs was partially reversed when cultured in a Transwell system (Fig. 4C). This result was also evident in the level of IL-17A in the culture supernatant (Fig. 4D). IL-10 has been shown to be important in nTreg-mediated suppression of effector cells (21). We therefore ascertained the potential role of IL-10 in NO-Treg–mediated suppression of Th17 cells. NO-Tregs were derived from WT or IL-10^−/− mice of the BALB/c background and cultured with CD4⁺ T cells under Th17 polarizing conditions. NO-Tregs from IL-10^−/− mice were significantly less suppressive than those from the WT mice (Fig. 4E, 4F). These results therefore indicate that the suppression of Th17 cells by NO-Tregs requires both cell contact and the presence of IL-10.

Finally, we investigated the role of NO-Tregs in vivo by using the EAE model, a disease in mice closely associated with Th17 (15, 22–24). EAE was induced in C57BL/6 mice by priming with MOG_35–55 peptide. The EAE mice were injected i.p. with 1 × 10⁶ NO-Tregs on day 10 after immunization, when the disease became apparent. Mice given NO-Tregs developed significantly less severe EAE compared with the control mice injected with PBS alone (Fig. 5A). The disease attenuation by NO-Tregs was confirmed by histological examination, which showed a marked reduction of leukocyte infiltration in the submeningeal area of the CNS of the mice receiving NO-Tregs compared with the control mice (Fig. 5B). The DLN cells were harvested on day 20 and analyzed by FACS ex vivo. The LN of the NO-Tregs recipients showed a marked reduction in the frequency of IL-17⁺ T cells compared with the PBS control mice. There was, however, no significant difference in the frequency of IFN-γ⁺ or IL-17⁺/IFN-γ⁺ T cells between the two groups of mice (Fig. 5C–E). We also analyzed the CNS cells of these mice by FACS. There was a significant reduction of the frequency of CD4⁺IL-17⁺ (Th17) cells in the NO-Treg recipients compared with controls (Fig. 5F–H). There was, however, no difference in the frequency of CD4⁺IFN-γ⁺ (Th1) and IL-17⁺IFN-γ⁺ double-positive cells in the CNS of the two groups (data not shown). Serum from NO-Treg recipients also produced less IL-17A but more IL-10 compared with the cells from the control mice (Fig. 5I). These results therefore show that NO-Tregs are able to attenuate EAE and that this effect is associated with the suppression of Th17 cells but not Th1 cells.

The existence and function of Tregs are arguably one of the most satisfying explanations of the maintenance of homeostasis of the immune system. Their key roles in immune regulation demand the possibility that Tregs exist in different forms to meet the challenges of distinct subsets of effector cells, mediating over exuberance of immune stimulations. Thus, in addition to nTregs, several other forms of Tregs have been described. Tregs are now known to be induced from CD4⁺CD25⁻ T cells by specific Ags in humans and mice (25–27). Tregs with a Th1-like phenotype have also been reported (28), whereas Tregs induced by IL-10 (Tr1) mediated their suppressive function in the absence of Foxp3 (29). A subset of TGF-β–induced Tregs (Th3) has also been described (30). NO-Tregs are distinct from these subsets of Tregs in several important aspects. NO-Tregs are CD4⁺CD25⁺, Foxp3⁻, GITR⁺, and CD27⁺, with a Th2-like phenotype, and they ameliorate inflammatory diseases in an IL-10–dependent manner (12). NO-Tregs differ from nTregs in that NO-Tregs do not express Foxp3. NO-Tregs are also different from Th3 in that NO-Tregs do not produce TGF-β. NO-Tregs are distinct from Th2 because Th2 are not suppressive against CD4⁺CD25⁻ effector cells. NO-Tregs are also different from Tr1 in that the induction of NO-Tregs is IL-10–independent. Importantly, note that IL-10 is the only detectable cytokine produced by NO-Tregs (12). This finding excludes the possibility that NO-Tregs may suppress Th17 via the production of IL-2, which has been shown to be a key inhibitor of Th17 differentiation (31).

nTregs have been reported to suppress Th17 (32). However, this suppression required a relatively high nTreg/Th17 ratio. In our hands, on a cell-for-cell basis, NO-Tregs are at least 16-fold more powerful than nTregs in suppressing Th17. In sharp contrast, nTregs are ∼8-fold more effective than NO-Tregs in suppressing Th1. The exact molecular mechanism for this reciprocal opposite effect is not clear, but the suppressions parallel the effect on the respective signature transcription factors of Th17 (RORγt) and Th1 (T-bet). Also note that despite intensive research for more than a decade, the molecular mechanism for nTreg-mediated suppression of effector cells remains elusive and controversial. NO-Tregs suppress Th17 in a cell contact–dependent manner and are partially dependent on IL-10. We have explored other possibilities by blocking the usual suspects (ICAM-1, LFA-1, OX40) with negative results (data not shown). It may be that IL-10 needs to act in a short range, effective only when NO-Tregs and Th17 are in close contact.

The 200–400 nM dose of NO used in our study is likely to occur in vivo in sites of acute infection and inflammation (18) and has been used routinely in experiments in vitro (4, 13). Although human macrophages produced only a low level of NO, humans do produce substantial amounts of NO comparable to those of rodents in vivo, and the role of NO and inducible NO synthase in humans is not in dispute. It has been recently reported that NO is produced in large amounts in the mouse by nonhematopoietic stroma cells such as fibroblastic reticular cells and lymphatic endothelial cells (33). The production of NO by these cell types in humans remains to be explored. Thus, it is likely that the findings reported in this study are equally applicable to humans.

We have demonstrated that NO-Tregs are effective in attenuating ongoing EAE. We have elected to test NO-Tregs in EAE because both Th17 and Th1 have been implicated in this disease model (22–24, 34). NO-Treg–mediated EAE attenuation is accompanied by a marked reduction in the frequency of Th17 cells in the DLN. In contrast, the frequency of Th1 cells remains unchanged. These results are consistent with the notion that NO-Tregs selectively suppress Th17 but not Th1 in vivo. It has been well documented that nTregs can suppress EAE probably through the suppression of Th1 cells (reviewed in Refs. 23, 24). We have demonstrated earlier that Th17 and Th1 cell infiltration in the DLN and spleen reflect the infiltration of these cells in the CNS (35). Also note that the disease attenuation by NO-Tregs was incomplete, compatible with a role for other effector cells such as Th1 in EAE. For this matter, we did not test the effect of NO-Tregs from IL-10^−/− mice, as the EAE model would not be sensitive enough to measure this partial effect.

Data reported in this study provide important functional characterization of NO-Tregs, which selectively suppress Th17 over Th1 cells. Additionally, we demonstrate the contrasting function of nTregs, which selectively suppress Th1 over Th17. The molecular mechanism involved is likely the selective inhibition of RORγt by NO-Tregs, and the reduction of T-bet expression by nTregs. nTregs exist under physiological conditions for the maintenance of immune balance. A high dose of NO is induced during infections, which also activate Th17 cells that can lead to a range of inflammatory autoimmune disorders. The selective suppression of Th17 by NO-Tregs would ensure the rapid and economic local suppression of excessive Th17 to avoid collateral damage, while leaving the general surveillance nature of Th1 intact. Our finding of the compartmentalization of Treg activities demonstrates an important biological function that may be of relevance to the evolution of the immune system.

The authors have no financial conflicts of interest.