|Home | About | Journals | Submit | Contact Us | Français|
Interferon-γ (IFN-γ) promotes a population of T-bet+ CXCR3+ regulatory T (Treg) cells that limit T helper 1 (Th1) cell-mediated pathology. Our studies demonstrate that interleukin-27 (IL-27) also promoted expression of T-bet and CXCR3 in Treg cells. During infection with Toxoplasma gondii a similar population emerged which limited T cell responses and were dependent on IFN-γ in the periphery but IL-27 at mucosal sites. Transfer of Treg cells ameliorated the infection-induced pathology observed in Il27−/− mice and this was dependent on their ability to produce IL-10. Microarray analysis revealed that Treg cells exposed to either IFN-γ or IL-27 have distinct transcriptional profiles. Thus, IFN-γ and IL-27 have different roles in Treg cell biology and IL-27 is a key cytokine that promotes the development of Treg cells specialized to control Th1 cell-mediated immunity at local sites of inflammation.
IL-27 is a member of the IL-6 and IL-12 family of cytokines. Early studies described it as an inducer of the T helper 1 (Th1) cell associated transcription factor T-bet, which enhances Th1 differentiation (reviewed in Hall et al., 2012). However, IL-27 is also an antagonist of inflammation associated with Th1, T helper 2 (Th2) and T helper 17 (Th17) cell responses in multiple settings (Stumhofer and Hunter, 2008) and the regulatory properties of IL-27 can be explained in part by its ability to limit IL-2 production, antagonize Th2 and Th17 cell responses and promote T cell production of IL-10. However, questions remain about the mechanisms used by IL-27 to limit immune pathology associated with Th1 responses (reviewed in Stumhofer and Hunter, 2008; Yoshida and Miyazaki, 2008).
CD4+ T cells that express the transcription factor Foxp3 (Treg cells), are an important means of immune suppression. Recent studies have demonstrated that during inflammation, specialized populations of Treg cells emerge that express transcriptional profiles similar to their effector counterparts (Esposito et al., 2010; Fujimoto et al., 2010; Koch et al., 2009). It has been suggested that this heterogeneity allows for regulation of specific types of immunity. For example, Treg cell expression of STAT3 is critical for limiting Th17 cell responses (Chaudhry et al., 2009), while expression of IRF4 allows control of Th2 cells (Zheng et al., 2009). During infections dominated by Th1 cells, Treg cells express Tbx21 and Cxcr3, genes associated with the presence of IFN-γ and expression of T-bet is required for their survival and proliferation (Koch et al., 2009). Whereas IFN-γ has been implicated in the development of this specialized Treg cell population, whether other environmental cues influence this program are unclear.
A subset of natural Treg cells (nTreg) express high levels of the IL-27Rα (Villarino et al., 2005). Paradoxically, there are reports that IL-27 is a direct antagonist of Treg cell conversion (Cox et al., 2011; Huber et al., 2008; Neufert et al., 2007; Stumhofer et al., 2007). The data presented here reveal that IL-27 directly promotes Treg cell expression of T-bet and CXCR3. In mice challenged with T. gondii, or other intracellular pathogens, a population of Treg cells emerged that express T-bet, CXCR3 and IL-10, and limited T effector responses. In mice that lack IL-27, this population is reduced at primary sites of infection but not at peripheral sites, such as in the spleen, where IFN-γ had a more prominent role. However, transcriptional profiling highlighted that IL-27 appeared to have a more dominant impact than IFN-γ on Treg expression of immunosuppressive genes such as Il10. Together, these studies identify distinct roles for IL-27 and IFN-γ in driving the T-bet+ subset of Treg cells which are specialized to control regional pathology during Th1 cell responses.
Although previous studies have demonstrated that 30–40% of natural Treg cells (nTreg) express the IL-27Rα, it was unclear whether this receptor was functional in Foxp3+ CD4+ T cells. In this study, we define nTreg cells as Foxp3+ CD4+ T cells isolated from mice, and “inducible” Treg cells as Foxp3+ CD4+ Treg cells generated from Foxp3− CD25− precursors in vitro (iTreg). To address whether nTreg cells respond to IL-27, naïve CD25+ T cells or Foxp3GFP+ cells were incubated with IL-27. Whereas unstimulated cells had negligible amounts of pSTAT1 or pSTAT3, IL-27 induced pSTAT1 and pSTAT3 in 30–40% of nTreg cells, (Figure 1A). Similarly, IL-27 induced pSTAT1 and pSTAT3 in 50–70% of iTreg cells (Figure 1B). It is notable that in Treg cells, IFN-γ and IL-10 also activate STAT1 and STAT3 respectively, but this was less than with IL-27 (Figure S1A). It is also relevant to note that previous reports have suggested that IL-27 antagonizes iTreg cell development (Huber et al., 2008; Neufert et al., 2007; Stumhofer et al., 2007; Cox et al., 2011) and in our experiments the frequency of Treg cells were initially reduced in the presence of IL-27, but Treg cells were generated and their numbers increased over time (Figure S1B, C). Together, these data suggest that existing and emerging Treg cell responses can be influenced by IL-27 and that IL-27 can actually promote Treg cell expansion.
Because IL-27 induces the expression of T-bet in effector CD4+ T cells, studies were performed to determine if IL-27 had a similar effect on Treg cells. When naïve Foxp3− CD4+ T cells were used to generate iTreg cells, those cultured in the presence of IL-27 expressed elevated levels of T-bet (Figure S1D). When Treg cells were differentiated in the presence of IL-27 plus α-IL-4 and α-IFN-γ (neutral conditions), it still promoted Treg cell expression of T-bet (Figure 1C). Thus, independent of its ability to promote IFN-γ, IL-27 promotes Treg cell expression of T-bet. It is notable that long-term TCR signaling was associated with the eventual up-regulation of T-bet in cultures of Treg cells, although the highest amounts of T-bet were always observed in the presence of IL-27 (Figure S1E).
Previous studies established that activation of T-bet in Treg cells promotes CXCR3 expression (Koch et al., 2009), a chemokine receptor involved in lymphocyte migration during Th1 responses (Lord et al., 2005). When iTreg cells were generated with IL-27, there was a consistent 4–5-fold increase in CXCR3 levels and nTreg cells incubated with IL-27 for 24 to 48hr expressed high levels of T-bet and CXCR3 (Figure 1C). Similarly, while iTreg cells stimulated with PMA/ionomycin did not produce IL-10 or IFN-γ those generated in the presence of IL-27 expressed IFN-γ and IL-10 (Figure 1D). Together, these data demonstrate that IL-27 promotes cytokine production by Treg cells and influences their proliferation, survival, and chemokine receptor expression.
Because IL-27 activates STAT1 and STAT3, experiments were performed to determine which pathway contributed to the up-regulation of T-bet and CXCR3 in iTreg cells. As noted earlier, addition of IL-27 to the cultures led to the induction of T-bet in 15–30% of the iTreg cells, and a modest but consistent increase in T-bet was observed in the Stat3−/− CD4+ T cells (Figure S1F). However, in the Stat1−/− Treg cell cultures, IL-27 did not induce T-bet (Figure 1E). In addition to T-bet, Eomesodermin (Eomes) has been shown to promote CXCR3 expression (Intlekofer et al., 2008) and polyclonal expansion of naïve cells CD4+ T cells in the presence of IL-27 led to increased levels of Eomes and T-bet, but in iTreg cells IL-27 did not promote Eomes (Figure S1G). This finding is consistent with the idea that TGF-β1 (required for the iTreg cell cultures) suppresses Eomes but not T-bet (Narayanan et al., 2010). Moreover, Treg cells generated with IL-27 from Eomes−/− T cells showed up-regulation of CXCR3, similar to wild-type (WT) Treg cells (Figure 1F), whereas Tbx21−/− iTreg cells had impaired up-regulation of CXCR3 (Figure 1F). These data suggest a model in which the ability of IL-27 to activate STAT1 drives T-bet and CXCR3 expression in a subset of iTreg cells.
To assess the effects of IL-27 on the phenotype of Treg cells in vivo, hydrodynamic gene delivery was used to administer DNA plasmid “mini-circles” that express IL-27 (IL-27MC) (Chen et al., 2005). Mice that received IL-27MC or GFP vector-only controls were sacrificed after 4 weeks for analysis of their T cell populations. No pathology was noted in either experimental group (data not shown) and while there was increased cellularity associated with IL-27MC (Figure S2A), there was no decrease in Treg cell frequency (Figure S2B). While total splenic CD4+ T cells from mice given IL-27MC had higher amounts of T-bet (Figure S2C), further analysis revealed that, consistent with the ability of IL-27 to promote a population of T-bet+ CXCR3+ Treg cells in vitro, only the Foxp3+ Treg cell population had higher T-bet and CXCR3 expression (Figure 2A–D).
Challenge with T. gondii results in the development of CD4+ effector T cells that produce IFN-γ and IL-10 (Jankovic et al., 2007; Roers et al., 2004). Analysis of Treg cells from infected mice following re-stimulation with PMA/ionomycin, and/or directly ex-vivo from IL-10 Vert-X reporter mice revealed that these cells were also a source of IFN-γ and IL-10 (Figure 3A, B). Following re-stimulation, a portion of these T. gondii-induced IL-10+ Treg cells co-expressed IFN-γ as well as T-bet and CXCR3 (Figure 3C–E). Further analysis of BrdU incorporation and Ki-67 expression revealed that infection promoted proliferation of these cells (Figure S3A, B). The appearance of CXCR3+ T-bet+ Treg cells coincided with the emergence of a population of cells that produced IL-27p28. While conventional DCs (CD11chi Class-II+) produced IL-12p40 in the spleen and the lamina propria lymphocyte (LPL) population, IL-27p28 production was the highest in the LPL and was made by monocytes and macrophages (CD11b+ CD11cLo) (Figure 3F).
To determine if Treg cells contributed to the control of the Th1 cell response during toxoplasmosis, DEREG mice were utilized to deplete Treg cells, and analysis of effector cell proliferation and cytokine production was performed. This treatment resulted in approximately 70% loss of Treg cells (Figure S3C) and an increased frequency of Ki-67+ effector CD4+ and CD8+ T cells (Figure S3D). This effect was accompanied by increased T-bet and IFN-γ expression in CD4+ and CD8+ T cells in response to soluble Toxoplasma antigen (STAg) (Figure 3G, Figure S3E). Thus, the Treg cells present in infected mice limited the Th1 effector cell response to T. gondii.
To determine if IL-27 was involved in generating Treg cell heterogeneity during Th1 responses, Il27−/− mice were challenged orally with T. gondii and their Treg cell populations were analyzed. In these experiments the Foxp3− CD4+ effector (CD44hi) T cells from wild-type or Il27−/− mice expressed equivalent levels of T-bet or CXCR3 (Figure 4A) but several Treg cell populations from the Il27−/− mice had reduced expression of T-bet, CXCR3 and IL-10 (Figure 4B–D, Figure S3F). Despite the systemic levels of IFN-γ and IL-12 (Figure 4E), a striking deficiency in T-bet+ CXCR3+ Treg cells was observed in the gut associated lymphoid tissue (GALT). In contrast, at peripheral sites like the spleen, these populations were not affected, consistent with the increased expression of IL-27 observed in the gut versus the spleen (Figure 3F). To determine whether IL-27 had a similar role in other models, WT or Ebi3−/− mice were infected with Leishmania major or Salmonella typhimurium. These infections also induced T-bet+ CXCR3+ Treg cells and in the absence of IL-27 there was a defect in Treg cell expression of T-bet and CXCR3 (Figure S3G, H). These results indicate a dominant role for IL-27 in Treg cell polarization at sites of ongoing Th1 cells responses.
To test whether the infection-induced CD4+ T cell-mediated pathology in Il27−/− mice could be ameliorated by Treg cells, WT and Il27−/− mice were infected with T. gondii, and at days 4, 7 and 10 post-infection, mice were provided with iTreg cells. This regimen was chosen because Treg cell homeostasis is altered during T. gondii infection due to increased Treg cell death (Oldenhove et al., 2009), and repeated transfer of Treg cells alleviates inflammation in several models (Darrasse-Jeze et al., 2009; Grainger et al., 2010; Mor et al., 2007; Zheng et al., 2006). While infected Il27−/− mice developed immune pathology and succumbed to acute infection, those that received IL-27-conditioned Treg cells, or Treg cells generated under neutral conditions, were rescued (Figure 4F). This effect was associated with reduced numbers of effector T cells and decreased serum alanine aminotransferase (ALT) (Figure S4A–C), suggesting that the transferred Treg cells limit effector cell expansion and pathology. Compared with the endogenous Treg cells, the Treg cells transferred into infected Il27−/− mice expressed higher levels of IL-10, IFN-γ, T-bet, CXCR3, and CTLA-4 and had increased proliferation as measured by Ki-67 (Figure S4D–G). When Treg cells were transferred into naïve mice, Foxp3 expression was not retained whereas in infected mice transferred cells sustained Foxp3 expression at sites of inflammation (Figure S4G), suggesting that environmental cues maintain these cells. Moreover, when adoptive transfer experiments were conducted with Treg cells generated from Il10−/− mice, recipient animals succumbed rapidly to infection, indicating that IL-10 is required for Treg cell-mediated rescue of Il27−/− mice (Figure 4F). These data establish that Treg cells can ameliorate the pathology observed in the Il27−/− mice and are consistent with a model in which Il27−/− mice infected with T. gondii have an underlying Treg cell defect, which contributes to the development of immune pathology. Nevertheless, these data have to be interpreted cautiously. The ability of neutral Treg cells to rescue these mice may be attributed to their eventual acquisition of T-bet and CXCR3 noted following TCR stimulation (Figure S1E). However, because neutral Treg cells could promote the survival of Il27−/− mice, Il27ra−/− and Ifngr1−/− iTreg cells were generated and transferred into infected Il27−/− mice to determine if expression of these receptors was required for protection (Figure S5A). These studies revealed that both cytokine receptors were required to rescue Il27−/− mice and suggest that although IL-27 is required for Treg cell expression of T-bet and CXCR3 at the sites of inflammation, signals through IFN-γ R1 also contribute to the rescue of these mice.
Although our studies have focused on the effects of IL-27 on Treg cells, our own data (Figure S5A) indicate that IFN-γ also has a role in these events and Treg cells from the spleens (but not the gut) of infected mice have high expression of IFN-γ R1 and IFN-γR2 (Figure S5B). To directly address the role of IFN-γ in generating Treg cell diversity during toxoplasmosis, infected mice were treated with α-IFN-γ starting 3 days post-infection. Neutralization of IFN-γ did not alter basal expression of T-bet or CXCR3 in Treg cells from naïve mice but in infected mice there was a significant decrease in levels of T-bet and CXCR3 in the spleen (Figure 5A, B). In contrast, there was no effect of IFN-γ depletion on Treg cell expression of T-bet and CXCR3 in the GALT of infected mice (Figure 5A, B). Furthermore, following re-stimulation, Treg cell production of IL-10 and IFN-γ was not lower in the spleen or the LPL in the absence of IFN-γ suggesting that IFN-γ is not critical for Treg cells production of these cytokines (Figure 5C, D). These results suggest that the effects of IFN-γ are prominent at peripheral sites whereas the effects of IL-27 are most apparent at the local sites of inflammation.
To better understand the differential effects of IFN-γ and IL-27 on Treg cells, a series of studies were performed to directly compare their signaling and their effects on Treg cell differentiation and transcriptional responses. When Treg cells were stimulated with IFN-γ or IL-27, both cytokines induced pSTAT1 but only IL-27 activated STAT3 and STAT5 (Figure 6A, B). However, although the kinetics of pSTAT1 signaling was similar, IL-27 induced markedly elevated amounts of pSTAT1 (Figure 6A, B). Next, the effects of IFN-γ and IL-27 on Treg cells were compared in vitro. By day 5 of iTreg cell culture, IFN-γ and IL-27 treatment resulted in increased T-bet and CXCR3 expression but IL-27 had a more profound effect (Figure 6C–E). In nTreg cell cultures, treatment with IL-27 resulted in higher T-bet expression compared to IFN-γ treatment although the cytokine-induced expression of CXCR3 was similar (Figure 6G). Perhaps the most notable difference was that while both IL-27 and IFN-γ -conditioned Treg cells produced IFN-γ, only the IL-27-conditioned Treg cells made IL-10 (Figure S6A).
Finally, expression profiling of Treg cells generated during IL-27 and IFN-γ treatment was performed. To obtain sufficient cells for analysis and ensure a uniform starting population, iTreg cells were generated under neutral conditions and then cultured in neutral conditions, or with IFN-γ or IL-27 for 10 and 48 hours. Microarray analysis identified 185 genes that were differentially expressed compared to neutral controls, either early (10hr) or late (48hr) following treatment (p value ≤ 0.05). Hierarchical clustering identified groups of genes that showed similar expression patterns across the three treatments and two time-points, and revealed three distinct clusters that are regulated in different ways by IL-27 compared to IFN-γ (Figure 7, Figure S6B). Cluster 1, (23 genes) was enriched for metabolic functional categories, and is suppressed by IL-27 while remaining relatively unchanged in the presence of IFN-γ (Figure 7A, B). In contrast, Cluster 2 (17 genes) is induced by IL-27, but not IFN-γ (Figure 7C, D). This cluster contains two subgroups: the first includes 10 genes induced early by IL-27 but were either not induced by IFN-γ, or weakly induced at 10 hours and were not sustained to 48 hours (Figure 7D). This subgroup includes Tbx21, Il10, Lag3, Ccr5, Il12rb1 and Il12rb2 highlighting a group of genes involved in T cell activation that are co-coordinately regulated by IL-27. The second subgroup includes 7 genes that are not affected early by either cytokine but are induced at 48hr by IL-27 (Figure 7D). Finally, a third cluster of 45 genes was identified that is enhanced by IL-27 and IFN-γ, but for which IL-27 was a more potent inducer (Figure 7E, F). This cluster includes many canonical STAT1 target genes, further demonstrating the ability of IL-27 to act as an activator of STAT1-dependent gene transcription in Treg cells. Thus, while IFN-γ and IL-27 can give rise to a phenotypically similar population of Treg cells, these two cytokines have distinct transcriptional effects on Treg cells that are indicative of a more complex biology and may reflect the regional effects observed in vivo.
In the last decade, there has been a growing appreciation of the inhibitory properties of IL-27 in the setting of autoimmunity and inflammation (Stumhofer and Hunter, 2008; Yoshida and Miyazaki, 2008). Indeed, studies with Il27ra−/− mice infected with T. gondii have identified a role for IL-27 in limiting the production of IL-2, antagonizing Th17 cells and promoting effector cell production of IL-10 (reviewed in Hall et al., 2012). In early studies it did not appear that the enhanced inflammation observed in these mice was a consequence of a defective Treg cell response, as IL-27 did not alter Treg cell activity in suppressor assays, nor are there any overt differences in the frequency of Treg cell populations in Il27ra−/− mice (Villarino et al., 2003; and this study). The data presented here highlight that the ability of IL-27 to promote a specialized population of Treg cells contributes to its suppressive activities in multiple experimental models and indicate that this is one of the many pathways that limit overt T cell-mediated inflammation in this model (Aliberti et al., 2002; Bhadra et al., 2011; Buzoni-Gatel et al., 2001; Gazzinelli et al., 1996).
It has been suggested that Treg cells have a limited role during toxoplasmosis (Couper et al., 2009; Jankovic et al., 2007) but the Treg cell depletion studies presented here indicate that these cells are relevant. The finding that Treg cell transfers can reverse the infection-induced pathology observed in Il27−/− mice provides additional support for the idea that Treg cells are operational during toxoplasmosis. However, these transfer experiments have to be interpreted carefully, and illustrate how understanding the mechanism by which adoptively transferred Treg cells suppress in vivo remain elusive. Although our system clearly shows that Treg cells need to express Il27ra, Ifngr1 and Il10 to rescue Il27−/− mice, it is unclear how these signals are integrated. Since the transfer of activated Treg cells can rescue the Il27−/− mice, it implied that signals other than IL-27 (such as IFN-γ or TCR stimulus) could be important. However, given these data, if IFN-γ or TCR were sufficient to contribute to Treg cell function in vivo we would expect that it could contribute to protection upon adoptive transfer of Il27ra−/− Treg cells, which is not the case. We cannot rule out a role for IL-27Rα independently of IL-27 and/or gp130 as the IL-27Rα has the capacity to signal via JAK1-STAT1 independently of gp130 (Pradhan et al., 2007; Takeda et al., 2003). Moreover, the IL-27Rα is also a component of the receptors for cytokine-like factor-1 (CLF-1) and humanin, and there is an alternatively spliced form of IL-27Rα which may be involved in trans-signaling (reviewed in Hall et al., 2012). Thus, it remains to be determined if the IL-27-independent properties of IL-27Rα play a role in Treg cell function during infection. Alternatively, our mRNA expression data (Figure S6C) suggest the possibility that Treg cells may also produce IL-27 implying that autocrine signaling through IL-27Rα may be important for Treg cells function. Nevertheless, these observations raise fundamental questions about whether this Th1-like Treg cell population is derived from nTreg cells or iTreg cells, whether these cells are specific for Toxoplasma and how IFN-γ R1, IL-27Rα and IL-10 contribute to their function in vivo.
Although the emphasis of these studies has been on the role of IL-27 during infection, our findings imply that IL-27 and IFN-γ act in distinct sites during infection, perhaps a consequence of local differences in the cytokine environment. However, despite the systemic elevation of IFN-γ present in infected Il27−/− mice, there is still a defect in the generation of T-bet+ Treg cells at sites of inflammation, indicating a critical role for IL-27 in mediating Treg cell expression of T-bet and CXCR3. Additionally, the ability of IL-27, but not IFN-γ, to promote the expression of IL-10 in Treg cells, highlight distinct functions of these cytokines.
The observation that in naïve mice only a subset of Treg cells expresses the IL-27 and IFN-γ receptors raises the question of whether these Treg cells subsets are “hard-wired” to deal with specific types of inflammation. Indeed, little is known about the heterogeneity in Treg cell expression of cytokine receptors, and whether this predicts their capacity to influence distinct types of inflammation. In the context of trying to understand Treg cell heterogeneity, microarray analysis of Treg cells isolated from different anatomical sites revealed that distinct subsets of Treg cells exist with non-overlapping transcriptional profiles (Feuerer et al., 2009; Feuerer et al., 2010). Similarly, when we compared the transcriptional profiles of Treg cells treated with IFN-γ and IL-27, these data revealed at least three clusters of genes that are differentially regulated. The largest cluster consists of many known STAT1 target genes, reinforcing the notion that IL-27 is a potent inducer of STAT1-mediated transcription. It is unclear how the specific STAT1 genes that we have identified as IL-27 targets might contribute to Treg cell activity but these data will aid in the selection of candidates for future studies of Treg cell function. In addition, we have identified a subset of genes that are specifically induced by IL27 but not IFNγ, providing candidate mediators for IL-27-specific Treg cell function. For instance, we found that Ly6c1 is strongly induced by IL-27. Although its role in Treg cell biology has not been addressed, it has been implicated in the function and homing of effector T cells (Jaakkola et al., 2003; Marshall et al., 2011). These data also highlight that examining a select few phenotypic markers (such as T-bet and CXCR3), may over-simplify the complex heterogeneity that exists in Treg cells during infection.
Many studies have defined how factors such as IL-2, TGF-β and Foxp3 have a prominent role in the homeostasis and function of Treg cells (Apostolou et al., 2008; Bayer et al., 2007; Fontenot et al., 2005). While there is good evidence that an ongoing immune response can limit Treg cell function and differentiation (Caretto et al., 2010; Mantel et al., 2007; Pasare and Medzhitov, 2003; Wei et al., 2007), recent reports indicate that by utilizing the same transcription factors as their effector counterparts, Treg cells may become specialized to operate in distinct inflammatory environments (Chaudhry et al., 2009; Koch et al., 2009; Zheng et al., 2009; Lu et al., 2010). These findings have led to models in which environmental cues promote the development of specialized Treg cell subsets. The finding that IL-27 promotes a specialized subset of Treg cells delineates a unique pathway by which they are influenced by the inflammatory environment. These findings may be directly relevant to human disease and it is notable that a loss of Treg cells has been observed in human patients with inflammatory bowel disease (IBD) (Eastaff-Leung et al., 2010) and a recent report linked a polymorphism in the IL-27p28 loci, associated with reduced production of IL27 transcripts, to increased susceptibility to IBD (Imielinski et al., 2009). The finding that IL-27 can profoundly influence Treg cell populations in the gut may offer a partial explanation for the susceptibility of these particular patients to IBD.
Il27−/− mice were generated by Lexicon Pharmaceuticals, Inc. Ebi3−/− mice are described elsewhere (Yang et al., 2008). Wild-type C57BL/6J (WT), IL-10−/−, Swiss Webster and CBA/CaJ mice were purchased from Jackson laboratory. Tbx21−/− and CD4-Cre × Eomesfl/fl mice were provided by S. Reiner. Stat1−/− mice were purchased from Taconic labs. Il27ra−/− mice were provided by C. Saris (Amgen), Foxp3GFP reporter mice were obtained from V. Kuchroo, Stat3fl/fl mice were obtained from L. Heninghausen, and Vert-X IL-10 reporter mice were provided by C. Karp, and bred in our facility. All mice were housed in a specific-pathogen free environment at the University of Pennsylvania School of Veterinary Medicine in accordance with federal and institutional guidelines. The ME49 strain of T. gondii was maintained in Swiss Webster and CBA/CaJ, and used as a source of tissue cysts for oral (100 cysts) or i.p. (20 cysts) infections. For Treg cell depletion experiments, WT or DEREG mice (Lahl et al., 2007) were infected orally with 10 cysts of ME49. Two days post-infection, mice were injected with diphtheria toxin (DT) (Calbiochem) in endotoxin-free PBS. 1 µg of toxin was injected i.p. for 7 consecutive days. For the depletion of IFN-γ, WT mice were infected orally and starting at day 3 post-infection treated with 2mg anti-mouse IFN-γ (clone XMG1.2, BioXcell), or ratIgG control (Sigma) every 2 days.
For iTreg cells, CD4+ CD25−, or Foxp3GFP− cells were separated by FACS or MACs sorting (Miltenyi). Cells were rested for 30 min in complete RPMI (cRPMI) media (1% penicillin/streptomycin, 2mM L-glutamine, 10% fetal bovine serum, 0.1% beta-mercaptoethanol, 1% non-essential amino acids, 1mM sodium pyruvate and 20mM HEPES) (GIBCO). Cells were cultured at 1×106 cells/mL in αCD3-coated (1µg/mL; clone 145-2C11; eBioscience) 96-well U-bottom plates (Costar) in cRPMI containing αCD28 (1µg/mL; clone 37.51; eBioscience), recombinant human (rHu) TGF-β2 (5ng/mL, eBioscience), with or without (rHu IL-2; 100U/mL; Proleukin), anti-IFNγ (10µg/mL; clone XMG1.2) and anti-IL-4 (10 µg/mL; clone 11B11) blocking antibodies, recombinant mouse (rMu) IL-27 (50ng/mL; Amgen), and with or without rMu IFN-γ (50ng/mL; R&D). Media was added every 2 days with initial cytokines and neutralizing antibodies. For nTreg cells, sorted Foxp3+ cells were cultured on plate-bound αCD3 in media containing αCD28, 100U/mL rHu IL-2, 5ng/mL rHu TGF-β2, with or without neutralizing antibodies to IFN-γ or IL-4, with rMu IL-27, or rMu IFN-γ.
Single cell suspensions from the spleens, mesenteric lymph nodes (mLN) and Payer’s patches were prepared using standard methods. For the analysis of lamina propria lymphocytes (LPL), small intestines were collected in PBS at 4°C, cut longitudinally, and fecal contents were removed in PBS. Epithelial cells were stripped (5mM EDTA and 1mM DTT) in cRPMI, followed by digestion (0.16U/mL Liberase TL) (Roche) for 30 min at 37°C, and processed for lymph ocytes.
Cells were stained in FACS buffer (0.5% BSA, 2mM EDTA in PBS) with Fc block (2.4g2, BD) containing live/dead fixable Amcyan (Invitrogen), using the following surface antibodies: CD4 Percp-Cy5.5 (RM4-5, eBioscience), CD8a PE-Texas Red® (53-6.7, Abcam), CD44 PE-Cy7 (IM7, eBioscience), CD62L APC-eFluor780® (MEL-14, eBioscience), TCR-β Alexa Fluor700® (H57597, BioLegend), CD3 eFluor450® (17A2, eBioscience), CD25 APC-eFluor780® (PC61.5, eBioscience), CXCR3 PE and APC (220803, R&D) and PE-Cy7 (CXCR3-173, BioLegend). All intracellular staining was done using the Foxp3/transcription factor staining buffer set (eBioscience) for Foxp3 Alexa Fluor488® and eFluor450® (FJK-16s, eBioscience), T-bet eFluor660® and FITC (4B10, eBioscience and BioLegend respectively), Ki-67 Alexa Fluor488® and Alexa Fluor647® (B56, BD), CTLA-4 PE (UC10-4F10-11, BD), and EOMES PE (Dan11mag, eBioscience).
To detect cytokine production, isolated cells were cultured in cRPMI in six replicates at 1×106 cells/mL in a 96-well U-bottom plate, with PMA and ionomycin for 5hr with Brefeldin A (Sigma) and monensin (BD) golgi inhibitors. Cells were rinsed, stained for surface markers at 4°C, and fixed with 4% PFA in PBS for 10 min at RT. Intracellular cytokines were detected by staining in FACs buffer containing 0.5% saponin (Sigma), IL-10 APC (JES5-16E3 , eBioscience), IFNγ PE-Cy7 or Percp-Cy5.5 (XMG1.2, eBioscience) and Foxp3. Intracellular IL-10 was detected ex vivo using Vert-X reporters (Madan et al., 2009) and antibodies for GFP (polyclonal rabbit anti-GFP, eBioscience and FITC-conjugated rat anti-rabbit Jackson Immunoresearch) with staining for Foxp3.
NTreg cells were isolated ex vivo, and iTreg cells were generated in vitro. Before stimulation, cells were washed in 0.5% BSA RPMI and rested at 4°C for 20 min. WT and Il27ra−/− Treg cells were incubated with 50ng/mL of the following cytokines: rMu IL-27; rMu IL-10 (R&D); and rMu IFN-γ for 20 min or over various time points at 37°C in the presence of Amcyan fixable live/dead marker, followed by immediate fixation on ice in 4% PFA for 20 min. Following PBS rinse, cells were permeabilized in 90% methanol on ice for 1 hour or stored at −20°C overnight. Staining was performed in Fc block with BD PhosFlow antibodies to pSTAT1 (pY701) PE, pSTAT3 (pY705) Alexa Fluor488®, and pSTAT5 (pY694) Alexa Fluor647® with antibodies to T-bet, Foxp3, TCR-β, CD4 and CD25.
The p2øC31.RSV.hAAT.bpA plasmid was provided by Dr. Zhi-Ying Chen (Stanford University, Stanford, CA) and the vector modified to include unique 5’ PmeI and 3’ PacI restriction sites flanking hAAT for directional cloning of cDNA’s. PCR amplification was used to place 5’ PmeI and 3’ PacI cloning sites on the linked mIL-27 cDNAs which were ligated with the modified mini-circle plasmid. Mini-circle DNA was produced as described (Chen et al., 2005) with minor modifications for overnight cultures, Terrific broth containing 100ug/ml ampicillin was inoculated and incubated for 18hr shaking at 270 rpm. Endotoxin free Qiagen megaprep kits were used for DNA purification which was resuspended endotoxin-free Tris EDTA. Mini-circle DNA was dialyzed in Midi MWCO 3.5kDa tubes overnight against Tris EDTA and mini-circle DNA was verified by restriction digestion and sequencing. For a hydrodynamics-based transfection procedure (Liu et al., 1999), 20mg of mini-circle DNA in 2ml of Ringer’s solution was administered via tail vein injection within 5–8 seconds.
IL-27p28 (DuoSet, R&D), IL-12p40 (C17.8 and C15.6 biotin) and IFN-γ (AN18 and R4-6A2 biotin, eBioscience) were measured by ELISA. To detect intracellular IL-27p28 and IL-12p40 LPL dendritic cells (LpDCs) were enriched using the 1-Step 1.077/265 gradient (Accurate Chemical & Scientific Corp.) and incubated for 6–8hr at 37°C in the presence of Brefeldin A and Monensin an d then surface stained in Fc block including Amcyan live/dead dye, for MHC Class II I-A I-E Alexa Fluor 700® (M5/114.15.2, BioLegend), CD11c PE-Cy7 (N418, eBioscience), CD11b PerCP-Cy5.5 (M1/70, eBioscience), NK1.1 Pacific Blue (PK136, BioLegend), CD3 eFluor450®, and CD19 eFluor450® (1D3, eBioscience), and then fixed with 4% PFA. Cytokines were detected by staining for IL-12p40 PE (C15.6, BD) and IL-27p28 Alexa Fluor 647® (clone MM27-7B1, BioLegend) for 30 min in 0.5% Saponin.
WT or Il27−/− animals were infected i.p. with 20 cysts ME49 and monitored for morbidity. Neutral Treg cells were generated from naïve WT CD25− CD4+ T cells or Th1 Treg cells were generated from naïve WT or Il10−/− CD4+ CD25− T cells. At day 4 post-infection, mice received either i.v. PBS or 2–4×106 Treg cells followed by 2 injections of Treg cells on day 7 and 10, and were monitored for survival.
For whole genome expression microarray, iTreg cells were generated under neutral conditions for seven days to ensure a homogeneous starting population of Treg cells. Cells were harvested and then exposed to neutral, IL-27 or IFN-γ Treg culture conditions for 10 or 48 hours. RNA was isolated using RNeasy Plus (Qiagen) and quality was assessed by Bioanalyzer (Agilent). Biotin labeled complementary RNA (cRNA) was made using the Illumina TotalPrep RNA amplification kit. An Illumina MouseWG-6 version 2 expression beadchip was hybridized with cRNA from three biological replicates and scanned on a beadscan unit. Data were quantile normalized and differential expression analysis was carried out using GenomeStudio v1.8 software (Illumina). Genes were considered differentially regulated by IL-27 or IFN-γ if expression level changed ≥1.5 fold compared to neutral controls with a corresponding diffscore ≥13 or ≤-13 (equivalent to a p value of ≤ 0.05). Data was deposited on the Gene Expression Omnibus (GEO) database for public access (GSE38686). Hierarchical clustering (Eisen et al., 1998) and heat map tools available on GenePattern (Reich et al., 2006) were used to analyze and display microarray data. Gene ontology (GO) enrichment analysis was done using the Database for Visualization and Integrative Discovery (DAVID) (Dennis et al., 2003) with enrichment defined relative to entire microarray. Only GO terms shared by three or more genes and which had an enrichment p value of < 0.05 were considered for analysis.
Statistical significance was determined by a two-tailed unpaired Student’s t-test. Error bars indicate standard deviation of the mean. For p values: *, p <0.05; **, p < 0.01; ***, p < 0.001.
This work was supported by the State of Pennsylvania, NIH grants to CAH, BJ, DSR and AOH. CAH: AI 071302 AI084882. BJ: R21-AI090234-01. DSR: R37-AI28724. AOH: AI055428. We thank Kenneth A. Platt at Lexicon Pharmaceuticals, Inc. for the generation of the Il27−/− mice. We thank Igor Brodsky and Dan Campbell for helpful discussions.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.