The two best-characterized types of CD4+ regulatory T cells (Tregs) are Foxp3+ Tregs and Foxp3− type 1 regulatory (Tr1) cells. The ability of Foxp3+ Tregs and Tr1 cells to suppress adaptive immune responses is well known, but how these cells regulate innate immunity is less defined. We discovered that CD44hiFoxp3− T cells from unmanipulated mice are enriched in Tr1 cell precursors, enabling differentiation of cells that express IL-10, as well as Tr1-associated cell surface markers, CD49b and LAG-3, and transcription factors, cMaf, Blimp-1, and AhR. We compared the ability of Tr1 cells versus Foxp3+ Tregs to suppress IL-1β production from macrophages following LPS and ATP stimulation. Surprisingly, Tr1 cells, but not Foxp3+ Tregs, inhibited the transcription of pro–IL-1β mRNA, inflammasome-mediated activation of caspase-1, and secretion of mature IL-1β. Consistent with the role for IL-10 in Tr1 cell–mediated suppression, inhibition of inflammasome activation and IL-1β secretion was abrogated in IL-10R–deficient macrophages. Moreover, IL-1β production from macrophages derived from Nlrp3A350V knockin mice, which carry a mutation found in cryopyrin-associated periodic syndrome patients, was suppressed by Tr1 cells but not Foxp3+ Tregs. Using an adoptive transfer model, we found a direct correlation between Tr1 cell engraftment and protection from weight loss in mice expressing a gain-of-function NLRP3. Collectively, these data provide the first evidence for a differential role of Tr1 cells and Foxp3+ Tregs in regulating innate immune responses. Through their capacity to produce high amounts of IL-10, Tr1 cells may have unique therapeutic effects in disease-associated inflammasome activation.
Regulatory T cells (Tregs) constitute an essential part of the central and peripheral tolerance machinery that negatively regulates innate and adaptive immune responses. Within the CD4+ T cell compartment, the best understood subsets of Tregs are Foxp3+ Tregs (1, 2) and IL-10–producing type 1 regulatory (Tr1) cells (3). Foxp3+ Tregs have intrathymic and extrathymic origins and potently suppress immune responses via a number of mechanisms (4). Distinct from Foxp3+ Tregs, Tr1 cells have an extrathymic origin, lack Foxp3 expression, and characteristically produce high amounts of IL-10 (3). The therapeutic potential of both Foxp3+ Tregs and Tr1 cells is being explored in multiple diseases, but whether there is a therapeutic advantage for either cell type remains largely unknown.
The ability of Foxp3+ Tregs and Tr1 cells to control adaptive immune responses is relatively well defined, but the role of these cells in regulating innate immunity is less well characterized. Foxp3+ Tregs and Tr1 cells are known to downregulate APC maturation and Ag presentation by inhibiting CD80/86 expression in a CTLA-4– and/or LAG-3–dependent manner (3, 5). In addition, Tr1 cells and Foxp3+ Tregs can kill myeloid APCs by expressing CD226 and granzyme B (6, 7), granzymes A and B, and perforin, respectively (5, 8). However, the relative ability of Tregs or Tr1 cells to control inflammasome activation, an important part of the innate immune response during tissue injury, is unknown.
Inflammasomes are multiprotein complexes activated in response to danger signals. Upon activation, the complex enables the cleavage and maturation of key proinflammatory proteins, IL-1β and IL-18, initiates pyroptosis, and provides a link between innate and adaptive immunity (9, 10). Unregulated inflammasome activation can lead to a variety of immune cell–driven pathologies. For example, gain-of-function mutations in the NLRP3 inflammasome that lower its activation threshold result in inappropriate IL-1β production and cryopyrin-associated periodic syndromes (CAPS) (11, 12). In addition, hyperactivity of the NLRP3 inflammasome underlies inflammation in many common chronic diseases, including type 2 diabetes, neurodegenerative diseases, and inflammatory bowel disease (13–16). Notably, many of these common chronic inflammatory diseases are also associated with changes in immunoregulatory cells, leading us to hypothesize that Foxp3+ Tregs and/or Tr1 cells could have a role in modulating inflammasome activity. In this study, we investigated the ability of Foxp3+ Tregs and Tr1 cells to regulate inflammasome activity and report evidence that only the latter are able to control this pathway.
Materials and Methods
Mice, monosodium urate–induced peritonitis, and autoinflammation in Nlrp3L351P/+ CreT mice
Foxp3GFP reporter (generation F11), Foxp3RFP, IL-10GFP reporter, B6.129S2-Il10rb tm1Agt/J (referred to as Il10rb−/−), B6.Cg-Tg (cre/ESR1)5Amc/J (referred to as CreT), Nlrp3A350VneoR/+ (referred to as MWS), Nlrp3L351PneoR/+ (referred to as FCAS), C57BL/6, and BALB/c mice were purchased from The Jackson Laboratory. Foxp3RFPIL-10GFP reporter mice were generated by crossing Foxp3RFP reporter mice with IL-10GFP mice. Foxp3GFPThy1.1 reporter mice were generated by crossing Foxp3GFP reporter mice with Thy1.1 mice. MWS CreT mice and FCAS CreT mice were generated by crossing CreT tamoxifen-inducible mice to Nlrp3A350VneoR/+ or Nlrp3L351PneoR/+knockin mice, respectively (17). FCAS CreT mice were injected with 50 mg/kg tamoxifen-free base (MP Biomedicals) into the peritoneal cavity for four consecutive days to induce the expression of mutant NLRP3 protein (17). Tr1 cells were transferred via i.v. injection immediately after the first injection of tamoxifen. The weight of the mice was monitored daily, and they were euthanized on day 4. Blood was collected, and serum cytokines were analyzed by Th1/2/17 cytokine bead array and IL-1β and KC flex sets (all from BD Biosciences). The cells were analyzed by flow cytometry.
For the monosodium urate (MSU) peritonitis model, 100 μg MSU (InvivoGen) was injected via i.p. injection, and the lavage was collected after 5 h by washing with 5 ml PBS. The cells collected were analyzed by flow cytometry. All mice were bred in-house and maintained under specific pathogen-free conditions at the animal facility at Child and Family Research Institute. The University of British Columbia Animal Care and Use Committee and the University of California San Diego Institutional Animal Care and Use Committee approved the experiments described in the study.
Cell isolation, stimulation, coculture, and suppression assays
To generate bone marrow–derived macrophages (BMDM), bone marrow aspirates were cultured in RPMI 1640 medium with rGM-CSF (50 ng/ml) for 7 d (18). Expression of the Nlrp3A350V/+ CreT transgene was induced by adding 4-hydroxytamoxifen (0.4 μg/ml) 24 h before stimulation. CD4+ T cells were isolated from spleen and lymph nodes from Foxp3GFP or Foxp3RFPIL-10GFP reporter mice using an anti-CD4 enrichment kit (StemCell Technologies) and then sorted into Foxp3+ Tregs, CD44hiFoxp3− Tr1 cells, and CD44int/loFoxp3− naive T cells on a BD FACSAria (Supplemental Fig. 1). After sorting, the cells were plated at 1 × 106/ml and stimulated with immobilized anti-CD3 (10 μg/ml), soluble anti-CD28 (2 μg/ml), and recombinant human IL-2 (200 U/ml). For some experiments, IL-27 (50 ng/ml; eBioscience) was added to the CD44int/loFoxp3− naive T cells as indicated. After 3 d, the T cells were washed and replated at 1 × 106/ml with fresh media and cytokines. Supernatants were harvested for experiments on day 4 or 5 as indicated.
For cocultures, T cells or their conditioned media were added to BMDM for 30 min or 16 h, respectively, at the ratio or concentration indicated in the figure legends. In some cases, 10 μg/ml IL-10 blocking Ab (JES5-2A5; eBioscience) or isotype control (IgG1κ, eBRG1; eBioscience) were added 30 min before adding T cells. The cultures were then stimulated with LPS (10 ng/ml) for 5 h, with addition of ATP (5 mM) for the final 1 h. Supernatants were collected for analysis by ELISA, and cells were lysed for analysis of mRNA or protein expression by Western blot. To test suppression of T cell proliferation, CD25−CD4+ T responder cells were labeled with cell proliferation dye eFluor670 (eBioscience) and cultured alone or with the indicated ratio of Tregs or Tr1 cells in the presence of irradiated APC and 1 μg/ml anti-CD3. Proliferation was analyzed by flow cytometry after 3 d.
Flow cytometry and ELISA
The following Abs were used for this study: CD4 (RM4-5), CD11b (M1/70, BD), CD25 (PC61.5), CD39 (24DMS1), CD44 (IM7), CD45Rb (16A), CD49b (HMa2), CD73 (TY/11.8), CD90.1 (HIS51), CD90.2 (53-2.1), CTLA-4 (UC10-4F10-11; BD Biosciences), F4/80 (BM8), Foxp3 (FJK-16S), Gr-1 (Rb6-8C5), Helios (22F6; BD Biosciences), IFN-γ (XMG1.2), IL-17 (eBio17B7), LAG-3 (C9B7W), LAP (TW7-16B4), programmed death-1 (J43), TIM-3 (8B.2C12), and Nrp1 (BAF566; R&D Systems) Abs were obtained from eBioscience unless otherwise specified. Dead cells were excluded using Fixable Viability Dye efluor780 (eBioscience). For cytokine staining, the cells were stimulated with 50 ng/ml PMA (Calbiochem, Mississauga, ON, Canada) and 1 μg/ml ionomycin (Sigma-Aldrich) for 4 h, with the addition of 10 μg/ml brefeldin A (Sigma-Aldrich) after 1 h. Intracellular staining was performed using Foxp3 staining buffer set (eBioscience). Flow cytometry was performed on BD FACSCanto, LSR II, or Fortessa, and analysis was performed on FlowJo version X0.7 (Tree Star). IL-1β ELISA was purchased from BioLegend and performed following the manufacturer’s instructions.
Western blotting for caspase-1 and pro–IL-1β
Macrophages were lysed in Laemmli buffer and proteins were denatured and separated on either 12 or 15% SDS-PAGE for analysis of IL-1β or caspase-1, respectively. Separated proteins were transferred and probed with Abs against IL-1β (R&D Systems), caspase-1 (Santa Cruz Biotechnology), and β-actin (Cell Signaling Technology).
Real-time PCR was performed on Applied Biosystems 7500 Fast Real-Time PCR System. The primer sequences were as follows (5′ to 3′): il1b, 5′-TGTAATGAAAGACGGCACACC-3′ and 5′-TCTTCTTTGGGTATTGCTTGG-3′; maf, 5′-AGGATGGCTTCAGAACTGGC-3′ and 5′-GGTCTCCACCGGTTCCTTTT-3′; ahr, 5′-AGCCGGTGCAGAAAACAGTA-3′ and 5′-CCAGGCGGTCTAACTCTGTG-3′; prdm1, 5′-CTGTACAAGCTGCCCCCAAG-3′ and 5′-TAAGGATGCCTCGGCTTGAA-3′; tgfb1, 5′-GGTGGACCGCAACAACGCCAT-3′ and 5′-GGGGTTCGGGCACTGCTTCC-3′; and 18s, 5′-CAAGACGGACCAGAGCGAAA-3′ and 5′-GGCGGGTCATGGGAATAAC-3′. Data were normalized to 18s rRNA using the 2−ddCt method.
Statistical analysis was performed in GraphPad Prism (version 6) using one-way ANOVA followed by Tukey’s multiple comparison tests unless otherwise as indicated. Significant differences were set at p < 0.05.
CD44hiFoxp3− T cells are enriched in Tr1 cell precursors
A major limitation to the study of Tr1 cells has been the lack of an efficient system to isolate and/or differentiate these cells. Current methods to differentiate Tr1 cells in vitro involve TCR stimulation in the presence of cytokines such as IL-10, IL-27, IL-6, or IL-12 (19–22) or with costimulation of CD46 or ICOS-L, or with tolerogenic dendritic cells (3, 20). These methods, however, are inefficient, resulting in relatively small (i.e., <30%) proportions of IL-10–producing cells, which often coproduce high levels of IFN-γ.
Recent evidence suggests that exposure of effector memory T cells to high m.w. hyaluronan, the ligand for CD44, can also induce IL-10 (23). We therefore speculated that expression of CD44 might identify cells that are poised to differentiate into Tr1 cells. To investigate the ability of CD44hi cells to differentiate into Tr1 cells, CD44hiFoxp3− T cells (hereafter Tr1), CD44int/loFoxp3− naive T (Tn) cells, and Foxp3+ Tregs were sorted from spleens and lymph nodes of unmanipulated Foxp3RFP × Il10GFP reporter mice (Supplemental Fig. 1A, 1B). Sorted cells were activated with anti-CD3/28 mAbs and IL-2, but in the absence of any polarizing cytokines and the fold expansion, expression of IL-10 and Foxp3 was analyzed on day 4. Tr1 cells expanded to a similar degree compared with Tregs, and both were significantly less proliferative than Tn cells (Fig. 1A). Whereas neither Tregs nor Tn cells contained a significant proportion of IL-10+ cells, >80% of cells in the putative Tr1 cell population were IL-10+ (Fig. 1B).
To confirm these finding were not specific for C57BL/6 mice, we sorted CD4+ T cells from BALB/c mice into Tregs (CD25+CD45Rblo), Tn cells (CD25−CD44int/lo), and Tr1 cells (CD25−CD44hi) (Supplemental Fig. 1C). Similar to data from C57BL/6 mice, the CD25−CD44hi Tr1 cells produced high amounts of IL-10 while remaining largely Foxp3 negative (Supplemental Fig. 1D, 1E).
In accordance with previous observations (20), Tn cells stimulated with IL-27 contained only ∼30% IL-10+ cells, of which ∼37% were IFN-γ (Supplemental Fig. 2A, 2B), and only cultures of Tregs expressed Foxp3 (Fig. 1B). Time-course experiments showed that the induction of IL-10 in CD44hiFoxp3− T cells occurred as early as day 1 and peaked after 4 d of stimulation (Fig. 1C). High expression of IL-10 in Tr1 cell cultures was confirmed by measuring IL-10 in supernatants after 4 d of stimulation, substantiating the fidelity of the Il10GFP reporter (Fig. 1D).
Because Tr1 cells are characterized by an overall profile of cytokine production, the capacity of activated CD44hiFoxp3− T cells to make cytokines in addition to IL-10 was examined. Because Tr1 cells are known to produce some IFN-γ but at a much lower level compared with Th1 cells (3), we confirmed this using in vitro–polarized Th1 cells and found that activated CD44hiFoxp3− T cells produced significantly less IFN-γ than Th1 cells (Fig. 1E) or Tr1 cells differentiated with IL-27 (Supplemental Fig. 2B). Furthermore, CD44hiFoxp3− T cells did not produce amounts of IL-2, IL-4, IL-17, TNF-α, or IL-6 that were significantly different compared with Tregs or Tn cells (Fig. 1F). They did, however, express similar levels of tgfb1 mRNA compared with Tregs (Fig. 1G). CD44hiFoxp3− T cells also suppressed the proliferation of CD4+ T cells, albeit less potently than Foxp3+ Tregs (Fig. 1H).
Tr1 cells derived from CD44hiFoxp3− T cells express characteristic cell surface markers and transcription factors
We next sought to determine whether Tr1 cells differentiated from CD44hiFoxp3− T cells express Tr1 cell–associated surface markers (24) and/or transcription factors. Compared with Tn cells or Foxp3+ Tregs, a significantly higher proportion of Tr1 cells expressed LAG-3 and/or CD49b, from both C57BL/6 mice (Fig. 2A) and BALB/c mice (Fig. 2B). Moreover, we found higher expression of multiple additional Tr1-associated markers including CD39 (Fig. 2C) and ICOS (Supplemental Fig. 2C) compared with Tn cells (3). Because IL-10 has also been implicated in T cell exhaustion (25, 26), we examined the expression of programmed death-1 and TIM-3 and found that Tr1 cells derived from CD44hiFoxp3− T cells expressed high levels of both markers (Fig. 2C, Supplemental Fig. 2C). Compared with Foxp3+ Tregs, however, they did not express high levels of LAP, CD73, or Helios (Fig. 2C).
We also analyzed the expression of Tr1-associated transcription factors (3, 20, 27, 28) after 4 d of activation and found that, in comparison with Tn, they have significantly higher expression of maf and prdm1 (gene for Blimp-1) and a trend toward increased ahr (Fig. 2D). In comparison with Foxp3+ Tregs, only expression of maf was significantly higher in Tr1 cells. Interestingly, high expression of these transcription factors was already present in ex vivo CD44hiFoxp3− cells (Supplemental Fig. 2D), suggesting that CD44hiFoxp3− cells contain Tr1 cell transcription machinery that may be programmed to upregulate IL-10 without the requirement for exposure to polarizing cytokines. Collectively, these data show that CD44hiFoxp3− T cells isolated from unmanipulated mice are poised to rapidly differentiate into Tr1 cells, representing a new and efficient way to obtain large numbers of these cells for further study.
Tr1 cells, but not Foxp3+ Tregs, suppress inflammasome activation in macrophages via soluble factors
To compare the ability of Tr1 versus Foxp3+ Tregs to suppress the inflammasome, T cells and macrophages were cocultured overnight and then stimulated with LPS for 4 h, followed by ATP for 1 h. The viability of the macrophages was not decreased in the presence of T cells (Supplemental Fig. 3). Surprisingly, we found that the production of mature IL-1β was significantly suppressed by Tr1 cells but not by Foxp3+ Tregs (Fig. 3A, 3B) in a dose-dependent manner (Supplemental Fig. 4A). In addition to ATP, Tr1 cells also suppressed LPS and MSU-induced IL-1β production (Supplemental Fig. 4B). Similarly, we found that IL-27–differentiated Tr1 cells inhibited IL-1β release (Supplemental Fig. 4C). In contrast, coculture of macrophages with Tn cells resulted in a significant increase in inflammasome activation. Because neither LPS nor ATP alone was sufficient for the production of IL-1β in the presence of T cells (Fig. 3A), Tn cells do not produce factors that can bypass the requirement for either LPS or ATP. The requirement for caspase-1 in IL-1β release was confirmed by the absence of IL-1β secretion in the presence of the caspase 1 inhibitor YVAD-FMK (Fig. 3A).
Because Tr1 cells act via cytokine-mediated suppression, we asked whether suppression of IL-1β secretion was mediated by soluble factors. Indeed, addition of media conditioned by the T cell subsets recapitulated the effect of cell cocultures: Tr1 cell–conditioned media potently suppressed IL-1β; Tregs had little effect; and Tn cells enhanced IL-1β (Fig. 3C, 3D). Notably, addition of rIL-10 had a similar suppressive effect (Fig. 3D), suggesting that the Tr1 cell–mediated suppression may be IL-10 dependent. As seen with the cocultures, the suppressive effect of Tr1 cells was dose dependent (Supplemental Fig. 4E). In contrast, despite containing ∼26 ng/ml rIL-10 (Fig. 1C), which should result in ∼80% suppression (Supplemental Fig. 4D), addition of conditioned media from Tregs failed to inhibit IL-1β production. Therefore, Tr1 cells, but not Foxp3+ Tregs, suppress IL-1β production via a cell contact–independent mechanism.
Because Tr1 cells produce high amounts of IL-10, we then investigated whether the suppressive effects of Tr1 cells on IL-1β production were mediated by IL-10. We first tested the effects of Tr1 cells on macrophages derived from Il10rb-deficient mice and found that, in comparison with wild-type (WT) macrophages, these cells were completely resistant to the suppressive effects of Tr1 cells (Fig. 3E, 3F) or their conditioned media (Fig. 3G, 3H). Moreover, Ab-mediated neutralization of IL-10 also reversed the suppressive effect of Tr1 cells (Supplemental Fig. 4F, 4G), suggesting that the suppression is dependent on IL-10R signaling.
Tr1 cells suppress il1b transcription and caspase-1 activation via an IL-10R–dependent mechanism
We next explored the mechanism by which Tr1 cells regulate the production of IL-1β, a process known to be tightly regulated by a two-step activation process (9, 10). The first step involves cytokine- or TLR-mediated activation of NF-κB, leading to transcription and translation of pro–IL-1β. The second step requires inflammasome activation and then caspase-1 cleavage and activation, followed by caspase-1–mediated proteolytic activation of IL-1β. For these experiments, we used media conditioned by T cells so that mRNA, and protein signals were exclusively derived from macrophages. Exposure of macrophages to media conditioned by Tr1 cells or to rIL-10 inhibited transcription of il1b mRNA (Fig. 4A). Tr1 cell–mediated inhibition of pro–IL-1β protein expression was demonstrated by Western blotting (Fig. 4B), although neither IL-10 or nor Tr1 conditioned media completely inhibited the processing of pro–IL-1β. In addition to transcriptional regulation of il1b, Tr1 cell conditioned media also decreased the generation of mature, cleaved caspase-1, without affecting the amount of procaspase-1(Fig. 4C). It is worth noting that IL-10 needed to be added at least 1.5 h before ATP to inhibit the production of IL-1β (data not shown), suggesting the effect of IL-10 is primarily on the priming step of inflammasome activation. To further investigate whether IL-10R signaling was required for the suppression of Tr1 cells, we found that Tr1 cells also failed to suppress the transcription of il1b mRNA (Fig. 4D) or cleavage of caspase-1 (Fig. 4E) in Il10rb-deficient macrophages. Thus, IL-10 is necessary for Tr1-mediated suppression of mature IL-1β production.
Tr1 cells suppress IL-1β production from macrophages with an NLRP3 gain-of-function mutation
To ask whether Tr1 cells could also control inflammasome activation in diseases caused by hyper-activation of this pathway, we used macrophages from tamoxifen inducible Nlrp3A350V/+ knock-in mice (Nlrp3A350V/+ CreT, referred to as MWS CreT), which carry a mutation corresponding to human cryopyrin associated periodic syndrome patients (17). MWS CreT macrophages were cultured in 4-hydroxytamoxifen to induce the expression of the mutant NLRP3. As expected, these macrophages no longer required ATP stimulation for the production of mature IL-1β (Fig. 5A, 5B), confirming constitutive activity of caspase-1. Consistent with our observations in WT macrophages, Tr1 cells, but not Tn cells, potently suppressed IL-1β production in MWS CreT macrophages, regardless of whether they were stimulated with LPS alone or with LPS and ATP (Fig. 5A, 5B). A similar suppressive effect was observed with Tr1 conditioned media or rIL-10 (Fig. 5C, 5D).
To investigate whether Tr1 cells protected from the systemic inflammation caused by the NLRP3 gain-of-function in vivo, we set up an adoptive transfer model using Nlrp3L351P-CreT (FCAS CreT) mice that have more rapid onset and severe disease than MWS CreT mice (29). FCAS CreT mice were injected with tamoxifen for four consecutive days to induce the expression of mutant NLRP3. In some mice, Tr1 cells were injected immediately following the first tamoxifen injection. Although all mice lost significant weight (Fig. 5E), mice that received Tr1 cells were partly protected, with a direct correlation between the percentage of Tr1 cells in the blood and decreased weight loss (Fig. 5F). Furthermore, transfer of Tr1 cells resulted in decreased amount of KC (CXCL1) in the serum of FCAS CreT mice (Fig. 5G), although amounts of IL-6, TNF-α, and IL-1β were not changed (data not shown). These data support the notion that Tr1 cells and IL-10 are important for controlling the pathological release of IL-1β that occurs in autoinflammatory diseases such as CAPS.
IL-10R signaling regulates inflammasome activation in vivo
To further explore the relevance of IL-10 in controlling inflammasome activation in vivo, we used the model of MSU-induced peritonitis (30). In this model, injection of MSU leads to activation of the NLRP3 inflammasome in resident macrophages, resulting in production of IL-1β and chemokines that recruit neutrophils. This neutrophil recruitment is inflammasome and IL-1 dependent and does not occur in mice deficient for components of the inflammasome complexes or the IL-1R (30). We injected the peritoneal cavity of WT or Il10rb-deficient mice with MSU and measured the recruitment of neutrophils to the peritoneum after 5 h. In WT mice, injection of MSU resulted in a rapid increase in the absolute number of Gr1+CD11b+F4/80− neutrophils in the peritoneal lavage fluid (Fig. 6A). In contrast, Il10rb−/− mice were impaired in controlling this inflammasome-mediated response and had a significant increase in the number of infiltrating neutrophils (Fig. 6A, 6B). Thus, IL-10R signaling has a critical role for regulating inflammasome activation and IL-1β production in vivo.
Clinical trials with Tr1 cells or Foxp3+ Tregs are being actively pursued in the hope that these cell-based therapies will offer durable and specific treatments for a variety of immune-mediated diseases such as graft-versus-host disease, transplant rejection, and inflammatory bowel disease. Activation of the inflammasome contributes to the pathology in all of these diseases, yet the ability of Tr1 cells or Foxp3+ Tregs to control this process is unknown. In this study, we found that Tr1 cells, but not Foxp3+ Tregs, suppress IL-1β production by negatively regulating both transcription of il1b mRNA and activation of caspase-1. Furthermore, Tr1 cells suppressed IL-1β production caused by a clinically relevant gain-of-function mutation in NLRP3. Combined with evidence that IL-10R signaling regulates inflammasome activation in vivo, these data suggest that Tr1 cells may have unique therapeutic advantages that are not shared by Foxp3+ Tregs.
A major limitation for studying Tr1 cells has been the lack of an efficient system to generate relatively homogeneous populations of cells for study. In this paper, we report a major advance and show that in vitro activation of CD4+CD44hiFoxp3− T cells from unmanipulated mice, without addition of any polarizing cytokines, results in a population of cells that appear to be Tr1 cells as judged by multiple parameters. They are >80% positive for IL-10, have the expected profile of cytokine production (IL-10hiIFN-γloIL-2−IL-4−), and express all of the major Tr1 cell–associated cell surface biomarkers, including CD49b and LAG-3 (3). The Tr1 cells also expressed higher levels of transcription factors important for the production of IL-10, specifically cMaf, Blimp1, and Ahr. It is worth nothing, however, that none of these transcription factors are specific for the Tr1 lineage because they are expressed by many other subsets of CD4+ T cells (3). Tr1 cells also had significantly higher expression of CD39, ICOS, CTLA-4, and programmed death-1 compared with naive T cells even after 4 d of activation and/or after resting when the transient activation-induced expression of some of these markers is diminished.
Because CD44 is a marker for T cell memory, in laboratory-raised mice, CD44-expressing cells have distinct characteristics compared with true Ag-experienced CD44+ memory T cells (31), including rapid and cytokine-driven proliferation in vivo that is independent of Ag stimulation. These cells do not seem to rely on commensal bacterial Ags because germ-free mice have a similar frequency of CD4+CD44hiFoxp3− cells (31). However, the function of these cells remains largely unknown. Our data suggest that these CD4+CD44hiFoxp3− cells have a distinct cytokine expression pattern and are poised to mount immunoregulatory rather than immunostimulatory responses. We also investigated the potential utility of CD44 as a marker for human Tr1 cells but found that human peripheral blood CD4+ T cells uniformly express high levels of CD44 (data not shown).
Consistent with the dominant role of IL-10 in Tr1-mediated suppression, we found that Tr1 cell–mediated inhibition of IL-1β transcription and caspase-1 activation is completely dependent on IL-10R signaling because macrophages from Il10rb-deficient mice were resistant to Tr1 cell–mediated inhibition of inflammasome activation. This finding is consistent with some papers showing that IL-10 inhibits transcription of il1β in human monocytes, murine peritoneal–derived macrophages, or macrophage cell lines (32–35) but not with others that failed to demonstrate IL-10–mediated transcriptional control of il1β (36–38). This discrepancy may be due to various factors including the poor stability of IL-10, time of delivery, and different regulatory mechanisms in monocytes and macrophages. Although in our in vitro system rIL-10 is as potent as Tr1 cells at suppressing IL-1β, in vivo Tr1 cells would likely be superior because of their ability to deliver IL-10 continuously at the site of inflammation.
Recently, two studies found that mice lacking IL-10 have significantly elevated inflammasome activation and IL-1β production, resulting in severe colitis (37) or Ag-induced arthritis (39). However, IL-10 knockout mice also have increased production of other proinflammatory cytokines that could contribute to inflammasome activation as much as the lack of macrophage exposure to IL-10. Our data complement these studies and demonstrate that Tr1 cells primarily regulate the priming steps of inflammasome activation but can also decrease caspase-1 activation. In contrast to the effects of Tr1 cells, which are mediated by soluble factors, a previous study found that memory T cells suppressed the inflammasome via a contact-dependent mechanism, with indirect evidence for a possible role of TNF family ligands (40). Therefore, the ability of different T cells to produce IL-10, and/or express inhibitory cell surface molecules, may be harnessed therapeutically to dampen innate inflammation.
Why did Foxp3+ Tregs fail to suppress the inflammasome? The role of IL-10 in Foxp3+ Treg-mediated suppression is controversial, with a wealth of literature arguing either for or against a role for this cytokine in different in vitro assays and disease settings (5). Findings from mice with a specific deletion of il10 in Foxp3+ Tregs suggest that these discrepancies are likely due to an essential role of IL-10 in Foxp3+ Treg-mediated control of tolerance at mucosal surfaces but not for autoimmunity (41). In our system, although Foxp3+ Tregs produced significantly more IL-10 than Tn cells (average of 26.1 ± 6.3 versus 1.7 ± 0.05 ng/ml), this amount was on average 10-fold lower than the amount produced by Tr1 cells. Notably, as little as 6 ng/ml rIL-10 efficiently suppressed the inflammasome, so lack of sufficient IL-10 production by Foxp3+ Tregs is likely not the explanation for why these cells cannot prevent activation of this pathway. Because murine Tregs express high levels of CD39 and CD73 that can convert ATP to adenosine and adenosine can potentiate inflammasome activation (42), Tregs may counteract the effect of IL-10 by this mechanism. Further investigation into the mechanistic basis for why Foxp3+ Tregs do not control inflammasome activation is warranted.
The therapeutic potential of Tr1 cells has been explored in various disease models (3) and in humans with two proof-of-principle trials in patients who received a hematopoietic stem cell transplant (43) or who had Crohn’s disease (44). We now provide evidence extending the spectrum of diseases that could be regulated by Tr1 cells to include autoinflammatory disorders. Specifically, we show that IL-10 from Tr1 cells suppresses inflammasome activation driven by a gain-of-function mutation in NLRP3 corresponding to human CAPS in vitro, and sufficient numbers of Tr1 cells can protect from inflammasome-driven weight loss and serum KC expression in FCAS CreT mice. Furthermore, IL-10R signaling regulates MSU-induced neutrophil infiltration. Overall, these data suggest that Tr1 cells may be uniquely able to resolve inflammation caused by tissue injury and inflammasome activation.
This work was supported by Canadian Institute of Health Research Grant MOP-93793 (to M.K.L.) and Crohn’s Colitis Canada (to T.S.S. and M.K.L.). Y.Y. holds a Canadian Institutes for Health Research doctoral research award, and M.K.L. holds a salary award from the Child and Family Research Institute.
The online version of this article contains supplemental material.
The authors have no financial conflicts of interest.