Th17 cells require IL-6 and TGFβ for lineage commitment and IL-23 for maintenance. Unexpectedly, naive IL-6−/− splenocytes stimulated with anti-CD3 and IL-23 produced normal amounts of IL-17 during the first 24 h of culture. These rapid IL-6-independent IL-17 producers were identified as predominantly DX5+ TCRβ+ NKT cells, and a comparable response could be found using the invariant NKT-specific ligand α-galactosylceramide. Human NKT cells also produced IL-17. NKT cells constitutively expressed IL-23R and RORγt. Ligation of either TCR or IL-23R triggered IL-17 production and both together had a synergistic effect, suggesting independent but convergent pathways. IL-17 production was not restricted to a particular subset of NKT cells but they were NK1.1 negative. Importantly, in vivo administration of α-galactosylceramide triggered a rapid IL-17 response in the spleen. These data suggest an important biological role for innate IL-17 production by NKT cells that is rapid and precedes the adaptive IL-17 response.
Interleukin 17A is a major proinflammatory cytokine produced by a newly described lineage of helper T cells, Th17 cells. These cells are reported to require IL-6 and TGFβ for lineage commitment and IL-23 for maintenance (1, 2) and have been implicated as major effector cells in such prototypic murine models of autoimmunity as experimental autoimmune encephalomyelitis and collagen-induced arthritis (3, 4).
NKT cells are unique cells of the innate immune system that are defined by coexpression of NK lineage markers, such as NK1.1 and CD49b/DX5 in mice and CD56 in humans, with a TCR (5) and are characterized by their ability to rapidly produce immunoregulatory cytokines, such as IL-4 and IFN-γ, upon engagement of their TCR (6). Although there is still some debate over what defines an NKT cell, they are currently divided into three categories based on their reactivity to the glycolipid α-galactosylceramide (α-GalCer),4 TCR α-chain diversity, and CD1d dependency. Type I classical “invariant” NKT cells (iNKT cells) have invariant Vα14-Jα18 TCR α-chains and react to α-GalCer in a CD1d-dependent manner. Type II nonclassical NKT cells do not react to α-GalCer and have diverse TCR α-chains but are also CD1d dependent. Type III NKT cells are CD1d independent, do not respond to α-GalCer, and possess diverse TCR α-chains.
In the present study we initially set out to examine the role of IL-6 in the production of IL-17 in the primary immune response by using anti-CD3 and IL-23 stimulation. In contrast to the previously reported critical role of IL-6 in the differentiation of Th17 cells, we found that during the first 24 h of stimulation IL-6-knockout (IL-6 KO) splenocytes produced as much IL-17 as wild-type (WT) splenocytes. Detailed characterization revealed that this early IL-6-independent IL-17 production originated largely from NKT cells. We show here for the first time that IL-6 is not required for innate IL-17 production by NKT cells. Consistent with this, NKT cells constitutively express IL-23R and RORγt, similarly to memory T cells and unlike NK and naive T cells. IL-17 production is triggered by ligation of either the TCR alone or the IL-23R alone, and engagement of both has a synergistic effect. We show for the first time that human NKT cells produce IL-17 upon stimulation and that the administration of an iNKT ligand, α-GalCer, to mice results in a rapid IL-17 response in the spleen, suggesting evolutionary conservation and in vivo relevance of this phenomenon.
Materials and Methods
C57BL/6 (WT) and IL-6 KO mice were purchased from The Jackson Laboratory. Jα18 KO mice, CD1d KO mice and IL-23 receptor KO mice were generated as previously described (7, 8, 9) and maintained in the animal facilities of the National Institutes of Health in Bethesda MD or the National Institutes of Health in Frederick, MD in full compliance with institutional guidelines.
Flow sorting and intracellular cytokine analysis
Pooled splenocytes from naive mice or human peripheral mononuclear cells from the buffy coat of healthy donors (National Institutes of Health blood bank, protocol no. 99-C-0168) were sorted into NK, NKT, and T cell populations based on expression of DX5, TCR β-chain, and CD62L using a BD FACSAria cell sorter. iNKT cells were isolated by the binding of CD1d tetramers (ProImmune) loaded with α-GalCer (Alexis Biochemicals/Axxora). Populations were enriched by autoMACS (Miltenyi Biotech) before sorting. Final purity was 85–99%. Cells were stained for extracellular and intracellular markers as described (BD Biosciences protocol; www.bdbiosciences.com/pdfs/manuals/00-6088-559311-B.pdf). All Abs were purchased from BD Biosciences and flow cytometry was analyzed using a FACSCalibur (BD Biosciences) and FlowJo software (Tree Star).
Mouse cell culture was performed in HL-1 medium supplemented with 1% fresh frozen mouse serum. Cells were stimulated either by adding 2.5 μg/ml anti-CD3 (clone 145-2C11) directly into cell culture or by using plates precoated with anti-CD3 at 10 μg/ml. Where indicated, 10 ng/ml recombinant murine IL-23 was added (R&D Systems). α-GalCer was used at 100 ng/ml, presented by CD11c-positive cells isolated by autoMACS (Miltenyi Biotech). Mouse IL-17A and IFN-γ protein were detected in supernatants by DuoSet ELISA-matched Ab pairs (R&D Systems) as per the manufacturer’s instructions.
Purified human NKT cells were stimulated for 24 h with anti-CD3 (10 μg/ml; BD Biosciences) cross-linked with anti-mouse-IgG (5 μg/ml; Sigma-Aldrich) and recombinant human IL-23 (100 ng/ml; eBioscience). IL-6 was neutralized using goat-anti-human IL-6 (10 μg/ml; R&D Systems) and effective neutralization was confirmed by ELISA. Cytokine levels in supernatants were measured by multiplex ELISA (SearchLight; Pierce Biotechnology).
mRNA expression analysis
RNA was isolated using TRIzol (Invitrogen Life Technologies) and an RNeasy mini kit (Qiagen) before reverse transcription using SuperScript III (Invitrogen Life Technologies). Quantitative real-time PCR was performed using a 7900HT sequence detection system (Applied Biosystems). TaqMan eukaryotic 18S ribosomal RNA endogenous control (VIC/MGB probe) and primer/probe sets for the murine IL-23 receptor, RORc, IL-17A, and IFN-γ (FAM/MGB probe) were from Applied Biosystems.
Statistical analysis and data presentation
All experiments were repeated at least twice with identical results. Figures show data from representative experiments. Graphs and data analysis were done using GraphPad Prism (GraphPad Software).
Results and Discussion
Rapid initial production of IL-17 upon primary stimulation is IL-6 independent and derives from NKT cells
Splenocytes from untreated IL-6 KO and WT mice were activated in vitro by stimulation with anti-CD3 and IL-23. Unexpectedly, there was significant IL-17 production observed by ELISA already within 24 h of stimulation, which was similar in IL-6 KO and their WT counterparts (Fig. 1,A). This was in sharp contrast to previous reports that IL-6 was required for IL-17 production (1, 2). Beyond 24 h, IL-17 production by IL-6 KO splenocytes reached a plateau whereas with WT cells it continued to increase, suggesting that the window for IL-6 independent IL-17 production is brief. The rapid kinetics and unexpected IL-6 independence of this early IL-17 production suggested a TCR-bearing cell population other than naive conventional T cells. To dissect this further, naive spleen cells were fractionated by flow sorting based on expression of TCRβ and the NK cell marker DX5 into NKT cells (DX5+TCRβ+), T cells (DX5−TCRβ+), and NK cells (DX5+TCRβ−) and were stimulated with IL-23 and plate-bound anti-CD3 for 16 h. Analysis of culture supernatants by ELISA (Fig. 1,B) and the cells by intracellular cytokine staining (Fig. 1,C) revealed substantial production of IL-17 by NKT cells. No IL-17 production was detectable in NK cells and only a trace level was found within T cells, which was most likely attributable to the memory T cells in this population (as shown in Fig. 2).
Further confirmation that the rapid early IL-17 producers were NKT cells was obtained by using the iNKT-specific ligand α-GalCer. Intracellular staining of sorted DX5+TCRβ+ cells after stimulation with α-GalCer and purified CD11c+ APC revealed a distinct population of IL-17 producers (Fig. 1 D).
Our findings in mice prompted us to investigate innate production of IL-17 in humans. NKT cells were purified from the buffy coat of healthy human donors and stimulated with human anti-CD3/IL-23 for 24 h. Analysis of the supernatants by ELISA for IL-17 content demonstrated that human NKT cells, like mouse NKT cells, produce IL-17 without the need for exogenous IL-6 (Fig. 1 E). To examine the IL-6 dependence of this response, we neutralized IL-6 in vitro and found no effect on IL-17 production (data not shown).
NKT cells constitutively express the IL-23 receptor and the transcription factor RORγt
A crucial role of IL-6 in the differentiation of Th17 cells is thought to be the up-regulation of the IL-23 receptor (IL-23R) and the Th17 lineage-specific transcription factor RORγt (10). The rapidity and IL-6 independence of the IL17 response of NKT cells suggested that they might constitutively express these molecules. Previous studies have identified constitutive expression of IL-23R on memory, but not naïve, T cells (11). We quantitated both IL-23R and RORγt mRNA expression by real-time PCR in naive T cells (DX5−TCRβ+CD62L+), memory T cells (DX5−TCRβ+CD62L−), NK cells (DX5+TCRβ−), and NKT cells (DX5+TCRβ+) sorted from C57BL/6 spleen. The data revealed that IL-23R and RORγt are constitutively expressed in NKT cells and T memory cells (Fig. 2,A). Although RORγt was constitutively expressed, no IL-17 production was evident within any population before stimulation (Fig. 2,C). This is consistent with the notion that expression of RORγt is necessary, but not sufficient, for IL-17 production and that a trigger for IL-17 transcription is downstream of RORγt. This is further supported by the finding that despite high levels of IL-17 protein found in culture supernatants within 16 h of stimulation, the increase in RORγt mRNA was quite modest (Fig. 2,B), suggesting that constitutive expression could explain much of the observed induction of IL-17. Thus, the differential requirement for IL-6 in IL-17 induction in naive T cells vs NKT cells can be explained at least in part by constitutive expression of IL-23R and RORγt observed in the latter. Notably, T cells with memory phenotype displayed a similar ability as that of NKT cells to rapidly produce IL-17 after TCR ligation in an IL-6-independent manner (Fig. 2 C). This suggests that, in a naive host, NKT cells may fill a parallel niche in host response to that filled by memory T cells in an Ag-experienced host.
Activation of NKT cells through the TCR and IL-23 receptor is independent but synergistic
We next analyzed the relative contributions of TCR ligation and IL-23 to NKT-derived IL-17 production by using C57BL/6 WT and IL-23R KO mice. Flow cytometric analysis revealed anti-CD3 alone and IL-23 alone induced approximately equal numbers of WT NKT cells to produce similar levels of IL-17, as detected by ELISA (Fig. 3). Although the combination of both had an additive effect on the frequency of IL-17-producing cells, a synergistic effect in the amount of IL-17 production was evident from both the mean fluorescence intensity of intracellular IL-17 staining and by an ELISA of secreted IL-17 protein. IL-17 production by TCR ligation was unaffected in IL-23R KO NKT cells although, as expected, these cells did not respond to IL-23. These observations suggest that IL-17 production via TCR activation is not dependent on concomitant IL-23R stimulation and vice versa. However, the additive/synergistic effect of both signals together would suggest that they combine quantitatively to control IL-17 levels in the cell.
IL-17-producing NKT cells are not restricted to a particular NKT subset and are NK1.1 negative
NKT cells are currently divided into three types based on their reactivity to α-GalCer, TCR α-chain diversity, and CD1d dependency. iNKT cells (type I), are easily identifiable with α-GalCer-loaded, labeled CD1d tetramers, but there are no specific reagents to identify type II (TCRα diverse and CD1d dependent) and type III (TCRα diverse and CD1d independent) NKT cells. Therefore, to examine IL-17 production by all NKT cell types, we used a combination of iNKT-specific techniques and genetically modified mice. NKT populations were sorted based on expression of DX5 and TCRβ from the spleens of WT C57BL/6 (containing all subsets of NKT cells), Jα18 KO (lacking iNKT cells), and CD1d KO (lacking both iNKT and type II NKT cells) mice. IL-17 production in response to α-GalCer was evident only in WT mice, confirming that iNKT cells are capable of IL-17 production (Fig. 4,A). Notably, NKT cells from both KO mice responded to TCR stimulation with IL-17 production, allowing us to conclude that in addition to iNKT cells, type II and type III NKTs are able to produce IL-17 (Fig. 4 B).
Following the recent report of Michel et al. (12) showing that NK1.1-negative iNKT cells from the liver and lungs of mice produce IL-17, we examined NK1.1 phenotype dependence. We sorted DX5+TCRβ+ NKT cells from spleen based on their expression of NK1.1 and stimulated them with anti-CD3/IL-23 (because NK1.1 is rapidly down-regulated following TCR ligation, it could not be used after stimulation). IL-17 production was considerably higher in NKT cells lacking NK1.1 expression (Fig. 4,C). IL-23R expression and RORγt expression were also substantially higher in the “IL-17-ready” NK1.1-negative than in the NKT1.1+ population (Fig. 4 C), supporting the interpretation that their constitutive expression underlies the rapid production of IL-17.
Visualization of intracellular IL-17 in iNKT cells poses similar technical issues to that of NK1.1, as the invariant TCR is rapidly down-regulated after activation, just as IL-17 production is occurring. Therefore, to examine the relationship between NK1.1 expression and IL-17 in the iNKT subset, iNKT cells were isolated from naive spleen by sorting with α-GalCer/CD1d tetramers and were separated into NK1.1-positive and -negative fractions. Stimulation of these populations with anti-CD3/IL-23 or with α-GalCer induced high levels of IL-17 production within 24 h in only the NK1.1− population (Fig. 4 D).
Administration of α-GalCer elicits a rapid IL-17 response in vivo
To examine whether an innate IL-17 response can be induced from NKT cells in vivo, we injected α-GalCer into C57BL/6 mice (10 μg/mouse) and assessed the relative expression of IL-17A, IFN-γ, and RORγt mRNA by real-time quantitative RT-PCR in the spleen. α-GalCer injection induced a rapid, transient expression of IL-17 mRNA that peaked 6 h after α-GalCer treatment and declined thereafter, returning to baseline at 24 h (Fig. 5,A). Interestingly, RORγt was expressed at baseline and was not up-regulated by α-GalCer administration (Fig. 5,C), consistent with our findings in vitro. Because it has recently been suggested that IL-21 may play a role similar to that of IL-6 in the induction of RORγt and subsequent IL-17 expression in Th17 cells (10, 13), we also examined the kinetics of IL-21 expression. We found that it lagged considerably behind the rapid, dramatic increase in IL-17 RNA (Fig. 5 D), making it unlikely that IL-21 is a prerequisite for innate IL-17 expression. Together, our data suggests that neither IL-21 nor IL-6 are required for the production of innate IL-17 by activated NKT cells.
We also examined a more physiologically relevant iNKT ligand, glycosphingolipid derived from the bacteria Sphingomonas wittichii (provided by Dr. P. Savage, Brigham Young University, Provo, UT) and found that a similar pattern of response was observed, albeit the fold increase of mRNA over background was lower than that observed with α-GalCer (not shown). This is consistent with findings showing that the activity of this compound is lower than that of α-GalCer in stimulating cytokine production by NKT cells (12).
In conclusion, our data show that innate IL-17 production by NKT cells is not restricted to a particular NKT subset and differs significantly from adaptive IL-17 produced by Th17 effector cells. Unlike the adaptive response, innate IL-17 production by NKT cells is rapid, IL-6 independent, and mediated by engagement of TCR and constitutively expressed IL-23R. In addition, RORγt is also constitutively expressed in NKT cells and de novo induction does not appear to be required for a dramatic increase in levels of IL-17. It follows that the role of IL-17 from NKT cells is likely to differ considerably from the proinflammatory role of IL-17 from Th17 cells, which are implicated to be the effector cells in various models of autoimmunity (3, 4). Previous data from our laboratory indicated that IL-4 and IFN-γ produced by NKT cells are associated with protection from autoimmunity and the significance of innate IL-17 production by NKT cells to autoimmune disease is currently under investigation.
We are grateful to Dr. Dan Cua of Schering-Plough, DNAX division, for providing the IL-23R−/− mice. We also thank Dr. Jun Tang, National Eye Institute, National Institutes of Health for his assistance with real-time PCR of RORγt expression.
The authors 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.
This work was supported by National Institutes of Health, National Eye Institute intramural funding.
Abbreviations used in this paper: α-GalCer, α-galactosylceramide; iNKT, invariant NKT; KO, knockout; WT, wild type.