Th17 cells, a subset of T cells involved in autoimmunity and host defense against extracellular Gram-negative infection, express both IL-17A and IL-17F. Both IL-17A and IL-17F can signal via the IL-17RA; however, IL-17F does so at a 1- to 2-log higher concentration than IL-17A. In this study, we show that the IL-17F homodimer via IL-17RA is a negative regulator of IL-17 production in T cells and suggest a mechanism whereby IL-17RA on T cells serves as an autocrine/paracrine regulator of IL-17 synthesis in T cells.
Interleukin-17 is an emerging cytokine family that consists of six family members (termed IL-17A through IL-17F) encoded by separate genes (1). Among these, IL-17A and IL-17F share the greatest homology and IL-17F presumably arose as a gene duplication event as IL-17F resides just 50,000 bp from IL-17 (1). Furthermore, both genes show relatively restricted expression in CD4+ T cells and are regulated by IL-23 and implicated in autoimmunity as well in host defense against extracellular bacteria (2, 3, 4). Both IL-17 and IL-17F can bind IL-17RA; however, IL-17F does so at a 1-log lower affinity than IL-17 (1, 5). Nonetheless, Abs against IL-17RA block chemokine induction by both IL-17 and IL-17F, suggesting that IL-17RA is required for signaling for both ligands (6). In addition to IL-17RA, it has been shown that IL-17RC is a receptor for IL-17F and is necessary for IL-17F-induced CXC chemokine induction in epithelial cells (7). Despite divergent receptor binding, the types of genes induced by IL-17 and IL-17F in both fibroblasts and epithelial cells are quite similar and dominated by CXC chemokines and growth factors such as IL-6 (6) (1). Moreover, IL-17 and IL-17F only have additive effects in terms of chemokine induction, and there is no evidence to date that these two ligands synergize in terms of cell signaling (1, 6). Thus, the rationale for having two isoforms with similar effects on target cells has remained undefined. In this study, we show in the context of hepatitis induced by Con A that IL-17RA is required for the development of hepatitis and it is also required to negatively regulate IL-17 and IL-17F by CD4+ T cells. Among potential ligands that mediate negative regulation of IL-17 by T cells, the IL-17F homodimer had a 50-fold more potent suppressive activity in T cells on the induction of IL-17 and IL-17F mRNA compared with the IL-17 homodimer. This suppressive activity of rIL-17F was recapitulated by antagonizing IL-17F signaling with soluble IL-17RC but also required cell surface IL-17RA. as the suppressive activity of IL-17F on IL-17 expression was absent in IL-17RA−/− T cells. These data suggest that IL-17F can function as a negative regulator of IL-17 production by CD4+ T cells.
Materials and Methods
Specified pathogen-free 9- to 12-wk-old C57BL/6 mice (from The Jackson Laboratory) or IL-17RA−/− mice (8) were bred in Children’s Hospital of Pittsburgh Animal Care Facilities. The IL-17RA−/− were backcrossed to C57BL/6 mice for 10 generations. IL-17 knockout (KO)2 mice on a C57BL/6 background were kindly provided by Dr. Yoichiro Iwakura (Tokyo, Japan) (9). All mice were housed under specific pathogen-free conditions within the Animal Care Facility of Children’s Hospital of Pittsburgh under an Institutional Animal Care and Use Committee-approved protocol, provided with water and food ad libitum, and housed under 12-h light/dark cycles.
Con A was purchased from Sigma-Aldrich and a serum transaminase ALT (Alanine aminotransferase) determination kit was purchased from Biotron. rIL-17 and F were purchased from R&D Systems. FITC-conjugated anti-CD4 and CD8, and PE-conjugated anti-IL-17, Abs were purchased from BD Biosciences. Serum IL-6 was measured using a Bioplex assay (Bio-Rad). IL-23, IL-21, IL12p40, and active TGF-β were measured using a commercial ELISA (eBioscience for IL-23 and R&D Systems for IL-21, IL12p40, and TGF-β). Anti-CD3/anti-CD28 coated microbeads were from Dynal.
Con A-induced hepatitis
Mice were injected i.v. with Con A (15 mg/kg) in sterile saline, and serum samples were taken or mice were killed at indicated time points. The extent of liver damage was evaluated by serum ALT levels and liver tissue histology.
Formalin-fixed tissues were paraffin embedded, sectioned, and stained with H&E at Histo-Scientific Research Laboratories.
Flow cytometry/ELISPOT assays
Liver mononuclear cells were separated from parenchymal hepatocytes and cell debris by centrifugation using 33% Percoll as described before (10).
The cells were stained with anti-CD4 and CD8 Abs to detect their surface expression and with anti-IL-17 Ab (BD Pharmingen) for their intracellular expression, according to the manufacturer’s directions. Precursor frequency of IL-17-producing CD4+ T cells in liver was performed by selecting CD4+ T cells with magnetic beads (Miltenyi Biotec) and culturing them in IL-17 ELISPOT plates (R&D Systems) with or without stimulation with PMA/Ionomycin.
Splenocytes were obtained from C57BL/6 and IL-17RA−/− mice via organ passage through a 40-μm nylon mesh filter and centrifuged at 1100 rpm, and red cells were lysed using NH4Cl. The cells were washed twice with PBS and subject to CD4+ T cell purification with anti-CD4 Ab-bound magnetic beads (Miltenyi Biotec) according to the manufacturer’s protocols. The purified CD4+ T cells were plated onto 48-well plates in RPMI 1640 plus 10% FCS in a 37°C 5% CO2 incubator. After 3 and 24 h of incubation with Con A or CD3/CD28 beads, cells and media from splenocyte cultures were harvested for mRNA and protein assay, respectively.
Total RNA isolation and quantitative PCR
The isolated liver and cultured cells were dissolved in TRIzol (Invitrogen) readily after the harvest. The total RNA was extracted and subjected to RT-PCR on the iCycler thermocycler (Bio-Rad), according to the manufacturer’s directions. Quantitative real-time PCR was performed using the iCycler Sequence Detection System for quantification with FAM and VIC (Applied Biosystems). Gene-specific primers and dual-labeled probe sequences for IL-17 and IL-17F cDNA were designed as follows (primer, primer, probe): IL-17, 5′-GCTCCAGAAGGCCCTCAGA-3′, 5′-CTTTCCCTCCGCATTGACA-3′, 5′-ACCTCAACCGTTCCACGTCACCCTG-3′; and IL-17F, 5′-AGG GCA TTT CTG TCC CAC GTG AAT-3′, reverse 5′-GCA TTG ATG CAG CCT GAG TGT CT-3′, probe 5′-CAT GGG ATT ACA ACA TCA CTC GAG ACC C-3′. All samples were normalized to GAPDH RNA content. TaqMan Rodent GAPDH Control Reagents (Applied Biosystems) was used for normalization of the samples. All PCRs were performed in a total volume of 50 μl. Relative quantification of the PCR signals was performed by comparing the cycle threshold value (Ct), in duplicate or triplicate, of the gene of interest of each sample and the reference gene GAPDH.
All data are presented as the mean ± SEM. Statistical significance (∗, p < 0.05 or ∗∗, p < 0.01) was determined by t test or ANOVA.
Role of T cell subsets and IL-17 in ConA-induced hepatitis
Con A is a lectin and has the ability to induce mitogenic activity in T lymphocytes and, when administered i.v., has been used as a model of autoimmune hepatitis (11) (12). i.v. administration of Con A results in significant liver injury and elevation of serum ALT levels in mice (Fig. 1,A). Induction of liver injury requires CD4+ T cells, as depletion of these cells significantly reduces serum ALT levels (Fig. 1,A). Depletion of CD8+ T cells did not attenuate liver injury; however, depletion of both CD4+ and CD8+ T cell reduce liver injury to a greater extent than just depleting CD4+ T cells alone. Furthermore, Rag1 KO mice that are deficient in T and B cells also showed significant protection against Con A induced liver injury yet still had elevated ALT levels compared with vehicle control mice (Fig. 1,B). To investigate whether this was due to perforin-expressing cells, we examined hepatitis in Con A-treated Rag1/perforin double KO mice. These mice showed a complete absence of liver injury (Fig. 1 B), suggesting that residual perforin expressing cells mediated the mild degree of hepatitis observed in the Rag1 KO mice.
To examine whether IL-17RA signaling is involved in the induction of hepatitis in this model, we administered Con A to wild-type (WT) mice and mice with a homozygous deletion of the IL-17 receptor (Fig. 1,C) (8). Both IL-17RA KO and IL-17 KO mice showed significantly reduced liver injury compared with WT control mice (Fig. 1,C). Consistent with a known role of CD4+ T cells and to a lesser extent CD8+ T cells, depletion of these T cell subsets, which showed reduction of ALT levels (Fig. 1,A), also showed reduction of serum IL-17 levels at 3 h after Con A administration (Fig. 1,D). Consistent with these data is the fact that IL-17 was undetectable in the serum of Rag1KO mice administered Con A (Fig. 1 E). Liver histology performed 22 h after injection with Con A in WT mice showed massive apoptosis and necrosis in subcapusular and midzonal regions, with accompanying marked hemorrhage. Meanwhile, in IL-17RA KO mice, the liver tissue showed maintained histological structure and less apoptosis and necrosis.
IL-17RA regulates IL-17 production in vivo
We next examined the time course of induction of serum IL-17 levels in these two strains of mice and found detectable IL-17 levels by ELISA as early as 1 h after administration of Con A with a peak value at 3 h (Fig. 2,A). In contrast to WT mice, IL-17RA KO mice had IL-17 levels at 1 h that were nearly 100-fold greater (Fig. 2,B). Peak IL-17 responses in IL-17RA KO were also at 3 h, and these mice continued to have a decline in serum IL-17, suggesting that the IL-17RA was not required for clearance of IL-17 from serum (Fig. 2,B). These elevated levels of IL-17 could not be explained by a non T cells source of IL-17 in these mice, as depletion of CD4+ and CD8+ T cell subsets resulted in significant reductions in serum IL-17 (Fig. 2,C) as in WT mice (Fig. 1,D). Furthermore IL-17RA KO mice had significantly higher levels of IL-17 message in mononuclear cells obtained from the liver 90 min after injection of Con A (Fig. 2,D) compared with WT control mice. Moreover, these elevated serum IL-17 levels could not be attributed to differences in serum IL-6, TGF-β, IL-21, IL23, or IL-12p40 levels in vivo (supplementary Fig. 1).3 Analysis of liver T cells for intracellular IL-17 expression by flow cytometry showed that both CD4+ and CD8+ T cells produce IL-17 in the liver (Fig. 3, A and B). Moreover, there was not a higher frequency of IL-17-positive cells in the liver of IL-17RA KO mice. However, the mean channel fluorescence of IL-17 staining was significantly higher (Fig. 3, C and D), which was particularly evident in CD4+ T cells (Fig. 3,C). Consistent with these findings, we observed similar precursor frequency of IL-17-producing CD4+ T cells in both WT and IL-17RA KO mice (Fig. 3 E) measured by ELISPOT. However, after stimulation with PMA/ionomycin, there were significantly more IL-17-producing cells in IL-17RA KO mice, suggesting that the capacity to produce IL-17 is greater in CD4+ T cells from IL-17RA KO mice.
IL-17RA regulates CD4+ T cell production of IL-17 production in vitro
Based on the fact that CD4+ T cells from IL-17RA KO mice appeared to produce greater amounts of IL-17 in vivo, we examined Con A responses of purified splenic CD4+ T cells in vitro. Varying number of purified CD4+ T cells were stimulated with Con A and, at all cell concentrations tested, secreted IL-17 levels were 1–2 logs higher compared with WT mice (Fig. 4,A). Again, we observed no differences in IL-6, TGF-β, IL12p40, or IL-21 levels in vitro (supplementary Fig. 2) 16 h after Con A stimulation. As IL-17F is often coexpressed with IL-17, we examined the induction IL-17F transcripts in WT and IL-17RA KO CD4+ splenocytes (Fig. 4,B). After 3 h stimulation with Con A, cells from IL-17RA KO mice had significantly greater induction of IL-17F message compared with WT mice (Fig. 4 B). To exclude abnormal T cell development confounding these results we analyzed CD4+ splenocytes from 5-day-old mice. Con A stimulation resulted in a similar level of IL-17 protein and induction of IL-17F message compared with 6- 8-wk-old WT and IL-17RA KO mice (data not shown). Moreover, there was no difference in IL-23R expression in WT vs IL-17RA KO in 6- to 8-wk old mice (data not shown). These data suggested that the IL-17RA was directly controlling expression of IL-17 and IL-17F.
To examine this further, we preincubated CD4+ splenocytes cells with a soluble murine IL-17RA:Fc protein (8), which has a Ka of 5 × 107 M−1 and a Kd for IL-17 of ∼75 nM. Cell pretreated with IL-17RA:Fc had only modest increases in transcripts for IL-17 and IL-17F (Fig. 4, C and D) compared with IL-17RA KO cells (Fig. 4, A and B), despite this concentration of IL-17RA:Fc neutralizing IL-17 homodimers in a bio-assay (data not shown). As IL-17F is also synthesized, and it has been reported that IL-17F has poor affinity for IL-17RA (5) but can bind IL-17RC (13), we examined whether soluble IL-17RC effected T cell production of IL-17 and IL-17F. The addition of a soluble IL-17RC molecule significantly augmented the induction of transcripts for IL-17 (Fig. 5,A) as well as for IL-17F (Fig. 5 B) in a dose-dependent manner.
Based on the differences obtained with soluble IL-17RA and IL-17RC, we examined whether the ligands IL-17 or IL-17F had differential effects on regulating the induction of gene expression for IL-17 or IL-17F in CD4+ splenocytes. CD4+ T cells pretreated with IL-17 had only a minimal suppression in the induction of transcripts for IL-17 (Fig. 6,A) or IL-17F (Fig. 6,B). In contrast, pretreatment with IL-17F potently suppressed the induction of IL-17 transcripts (Fig. 6,C) as well as IL-17 protein in the cell supernatant (Fig. 6,D). Similar suppression was observed if cells were simulated with Con A or CD3/CD28 beads (supplementary Fig. 3). Moreover, pretreatment with IL-17F resulted in a dose-dependent suppression of transcripts for IL-17F (Fig. 6,E), and this suppression of IL-17 and IL-17F required the IL-17RA, as no suppression occurred in CD4+ T cells from IL-17RA KO mice (Fig. 6, F and G). Under these conditions, we did not observe suppression of IL-22 production or RORγT transcripts (supplementary Fig. 3), suggesting that IL-17F was not interfering with Th17 differentiation during the 16 h study. We also compared IL-17RA expression in in vitro-polarized D011.10 Th17 cells or Th2 cells as previously described. Compared with Th2 cells, Th17 cells had higher expression of IL-17RA and a trend toward higher IL-17RC expression although this latter trend was not statistically significant (supplementary Fig. 3).
It has recently been shown that a subset of T cells termed ThIL-17 cells express IL-17 and IL-17F (3, 14). Moreover, these cells develop independent of the Th1 transcription factor T-bet and the Th2 transcription factor GATA-3 as well as Stat 4 and Stat 6 (15, 16). Although IL-23 regulates these IL-17-producing cells, IL-23R is not expressed on naive T cells and thus it has recently been demonstrated that both TGF-β1 and IL-6 are critical in the early transition of naive T cells to become IL-23-responsive ThIL-17 effector cells (17, 18). IL-17 has been shown to be a critical effector molecule produced by these cells as neutralization of this molecule has shown protection against tissue inflammation in experimental models of arthritis (19), multiple sclerosis (2), and inflammatory bowel disease (20). However, much less is known regarding the role of IL-17F- in ThIL-17-induced inflammation, which is also produced by this subset of T cells. IL-17F like IL-17 stimulates G-CSF, CXCL1, and IL-6 in fibroblasts and epithelial cells, and, in both murine fibroblasts and in human epithelial cells, this activity requires IL-17RA (6). Both molecules show marked synergy in these activities with costimulation with TNF-α, another product of ThIL-17 cells (6, 21). However, IL-17 and IL-17F only show additive activity in terms of CXCL1 or G-CSF induction in epithelial cells, suggesting a similar signaling mechanism (6).
These data question how two isoforms of a gene 50,000 bp apart and that show coordinate regulation in T cells arose in both mouse and human T cells (1). Using a model of Con A hepatitis, we show that IL-17 and IL-17RA signaling is critical for full development of hepatitis. These data differ from a recent manuscript showing that Con A-induced hepatitis was not reduced in IL-17A KO mice (22); however, this study used a significantly lower dose of Con A (10 ugm/gm). Moreover, our data show that T cells from IL-17RA KO mice show dysregulated IL-17 and IL-17F production. This was not due to developmental aspects of these mice as T cells from mice as young as 5 days showed similar dysregulation. Among potential ligands or the IL-17RA, the IL-17F homodimer potently suppressed Con A-induced IL-17 and IL-17F. Moreover, antagonizing IL-17F with IL-17RC, which binds IL-17F in vitro, augmented IL-17 and IL-17F induction in WT T cells. In support of this, Smith et al. (23) have recently shown that IL-17RA KO mice have an expansion of IL-17-producing T cells when cultured on plate-absorbed anti-CD3 and soluble anti-CD28 in the absence or presence of IL-23, followed by a rest period of 3 days. In addition to the expansion of cells, our data support higher per cell production of IL-17. Smith et al. (23) observed suppression of transcripts for IL-17 by both IL-17A and IL-17F, whereas IL-17F was more effective in our shorter-term stimulation studies. Another difference is we purified CD4+ T cells from spleen and liver, whereas Smith et al. (23) focused on bulk splenocytes. Neither our study nor the study by Smith et al. (23) has explored whether the recently described IL-17A/F heterodimer (24) plays a similar antagonistic role on IL-17 production and IL-17-producing T cell expansion.
These data strongly suggest that IL-17F through the IL-17RA negatively regulates IL-17 production in T cells. Meanwhile, it is also possible that highly elevated IL-17 levels in IL-17RA KO are controlled by the upstream cytokines that control the development of IL-17-producing T cells (25, 26, 27). To assess this possibility, we measured IL-12p40, IL-6, TGF-β, and dIL-21 both in vivo as well as in vitro in response to Con A. These studies showed no differences in serum levels of these cytokines after Con A administration and in vitro studies of stimulated CD4+ T cells also revealed no differences in the cytokine response between WT and IL-17RA KO mice. These data suggest that IL-17 production has a unique regulation system independent from cytokines that control Th17 T cell dependent and are regulated by IL-17RA and IL-17F. What remains unclear from these data is whether cell surface IL-17RC is also required for this negative regulation of IL-17 production in T cells as well, as it has recently been described that human and mouse IL-17 can signal via heterodimeric receptor complex consisting of IL-17RA and IL-17RC (28). Th17 cells express higher IL-17RA than Th2 cells (supplementary data 3) but not IL-17RC transcripts (data not shown). These data suggest that Th17 cells are equipped to respond to IL-17RA ligands more so than Th2 cells. In conclusion, these data suggest that IL-17F and IL-17RA regulate the amount of IL-17 released in the tissue by ThIL-17 cells and may serve to limit the extent of IL-17-induced inflammation.
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.
Abbreviations used in this paper: KO, knockout; ALT, alanine aminotransferase; Ct, cycle threshold; WT, wild type.
The online version of this article contains supplemental material.