IL-25 inhibits the Th17 response within the CNS

We have recently described an IL-23–dependent, IL-17– producing T cell population that is highly pathogenic in autoimmune inflammatory diseases (13). Here, we found the same Th17 cells as well as IFNγ-producing T cells in the CNS of il25^−/− mice at an earlier time point and in greater numbers than in the WT mice (Fig. 2). 2 d before expected disease onset, the CNS of il25^−/− mice contained fivefold more IL-17⁺ cells compared with the WT (Fig. 2 A). In contrast, the total numbers of IFNγ-producing T cells were similar in the il25^−/− and WT control mice. Also, the total number of inflammatory macrophages was comparable between il25^−/− and WT mice (Fig. S2, available at http://www.jem.org/cgi/content/full/jem.20061738/DC1). At the peak of disease, the il25^−/− mice continue to have greater numbers of IL-17–producing T cells as well as IFNγ and TNF producers (Fig. 2 B).

Open in a separate window

Figure 2.

Increased numbers of CNS-infiltrating inflammatory Th17 cells in il25^−/− mice. (A) Intracellular cytokine staining of FITC-gated CD4⁺ T cells isolated from the CNS of mice at day 7 after immunization (left, flow cytometric analysis; right, calculated number of cytokine-producing cells). (B) Number of cytokine-producing CD4⁺ cells in CNS taken at peak of disease. For A and B, cells from four mice were pooled for intracellular cytokine analysis. Data are representative of two experiments. (C) Disease score of il25^−/− mice immunized with MOG/CFA (n = 4 mice per group) and treated with antibodies against IL-17 or IFNγ or with rat IgG1 control antibody. Data shown are representative of two experiments.

To examine whether the disease exacerbation observed in the il25^−/− mice was mediated by Th17 and/or Th1 cells, we injected anti–IL-17 or anti-IFNγ antagonist mAbs into EAE-primed mice. Blockade of IL-17 dramatically inhibited EAE, indicating that Th17 cells play a major role in EAE pathogenesis in il25^−/− mice (Fig. 2 C). In contrast, anti-IFNγ treatment had no effect, consistent with studies demonstrating that Th1 cells only play a minor role in EAE (13). Thus, the absence of IL-25 leads to a more rapid and intense infiltration of pathogenic Th17 cells into the target organ, which drives an accelerated and exacerbated disease phenotype.

IL-25 is crucial for Th2 cytokine expression in the periphery

During EAE priming, myelin antigen–specific T cells are generated in lymph nodes draining the sites of immunization. In these lymphoid organs, T cell differentiation is dependent on the cytokines produced by antigen-presenting cells. In draining lymph node (DLN) cells from il25^−/− mice, we found a twofold higher expression of IL-23p19 mRNA than in WT controls at day 7 after immunization (Fig. 3 A). Interestingly, the expression of IL-12p35 was not regulated by IL-25. To determine whether IL-25 affected T cell priming and T cell subset differentiation, we cultured DLN cells at day 7 after immunization in the presence of MOG peptide to assess myelin antigen–specific T cell responses. The Th2 cytokines IL-4, -5, and -13 were significantly reduced in DLN cultures from il25^−/− mice. IL- 17 and IFNγ levels were slightly higher in il25^−/− DLN cultures; however, the difference did not reach statistical significance (Fig. 3 B). In addition to Th2 responses, regulatory T cells play a crucial role in limiting chronic inflammation. We found that il25^−/− mice generated Foxp3⁺ regulatory T cells in numbers similar to WT controls during the course of EAE (Fig. 3 C). These data suggest that IL-25 plays a role in driving IL-4, IL-5, and IL-13 production but does not affect regulatory T cell development.

Open in a separate window

Figure 3.

Reduced Th2 cytokine production but normal regulatory T cell response in il25^−/− DLN cells. (A) mRNA expression in MOG-stimulated DLN cells taken at day 7 after immunization. Expression levels are normalized to ubiquitin. n = 5. Obtained DLN samples were pooled in three different pools for mRNA processing. (B) In vitro cytokine recall response. DLN cells from MOG-immunized il25^−/− and WT mice were isolated at day 7 after immunization. Cells were stimulated with MOG peptide for 3 d and then restimulated with anti-CD3/anti-CD28 for 2 additional days. Supernatants were from 24- or 72-h cultures as indicated. Data are representative of three experiments. All cytokines were measured by cytometric bead array except for IL-13 and IL-17, which were detected by ELISA as described in Materials and methods. ND, not detectable. *, P < 0.05 (unpaired two-tailed t test). (C) Generation of regulatory T cells in il25^−/− and WT mice before (pooled sample of two mice per group) and 7 d after (pooled sample of four mice per group) immunization with MOG/CFA. DLN cells were stained for CD4, CD25, and Foxp3.

Exogenous IL-25 inhibits EAE development and progression

We next determined whether exogenous IL-25 can suppress acute as well as relapse-remitting EAE. We used replication-deficient adenoviral vectors (rAdVs) to induce local CNS or systemic IL-25 expression (33). SJL mice were used for these studies because they exhibit a multiple relapse-remitting pattern of clinical disease. Using rAdV gene transfer vectors, we found that intracerebral delivery of IL-25–rAdV 1 d before immunization fully protected SJL mice from EAE (Fig. 4 A). Mice treated with GFP-control-rAdV were not protected. An intracerebral injection of 3 × 10⁹ viral particles (Fig. 4 A) or i.v. injection of 10¹⁰ particles (not depicted) were equally effective. Injection of IL-25–rAdV 10 d after initial priming could also block EAE (Fig. 4 A). The histological analysis of spinal cords 18 d after immunization shows large numbers of parenchymal and perivascular infiltrates in the control group. IL-25–rAdV–treated mice, in contrast, show minimal to no infiltrate (Fig. 4 B and Table S1, available at http://www.jem.org/cgi/content/full/jem.20061738/DC1). We next examined whether IL-25 treatment can inhibit the effector phase of CNS inflammation using a passive transfer model of EAE. Recombinant adenoviral vector–mediated IL-25 expression prevented EAE development when the treatment was given at the time of cell transfer, or even 3 d after cell transfer (Fig. 4 C).

Open in a separate window

Figure 4.

IL-25 inhibits the effector phase of EAE and suppresses clinical disease relapse. (A) Clinical score of SJL mice (n = 5 per group) injected with 3 × 10⁹ particles of indicated rAdV by intracerebral ventricular injection 1 d before or 10 d after immunization. (B) Histopathological analysis of spinal cords from SJL mice at day 18 after immunization. Mice were treated with GFP-rAdV (left) or IL-25-rAdV (right) at day 1. (bottom) 10-fold magnification of boxed area in top panels. Bars, 100 μm. Results of histological analysis are summarized in Table S1 (available at http://www.jem.org/cgi/content/full/jem.20061738/DC1). (C) Clinical score of mice (n = 4–5 per group) that received 2 × 10⁷ activated PLP-specific lymphocytes. Recipient mice were treated with 3 × 10⁹ particles of indicated rAdV at day 0 (top) or day 3 (bottom) after PLP-cell adoptive transfer. (left) Kinetics of mean disease score; (right) maximum disease score per individual mouse during 16 d of observation. (D) Clinical score of EAE-primed mice (n = 7–10 per group) injected with 3 × 10⁹ particles of indicated rAdV by an intracerebral ventricular route during disease remission (20 d after immunization, arrow). Mice were observed for two disease-relapse cycles. Results are representative of at least two experiments.

In multiple sclerosis, patients usually suffer from episodes of clinical disease relapse separated by periods of remission. Because IL-25–rAdV could protect from EAE when administered at either the priming stage (day 1 treatment) or during the effector stage (day 10 treatment), we tested whether IL-25 could be used therapeutically to modify clinical disease. Remarkably, we found that IL-25–rAdV completely blocked EAE relapse when the treatment was given after the first episode of paralysis (Fig. 4 D). In contrast, control-rAdV–treated mice suffered two rounds of disease relapse over a period of 51 d. Collectively, these results suggest that besides its ability to prevent disease, IL-25 may regulate effector T cells and reduce their pathogenic functions.

IL-25 induced Th2 and suppressed Th17 responses

IL-25–deficient mice exhibited increased Th17 responses that promoted severe EAE pathology (Fig. 2). This result led us to examine the in vivo mechanisms of action of IL-25 treatment during disease induction. Daily s.c. injection with mouse IL (mIL)-25 protected the cytokine-treated mice from EAE (Fig. 5 A) as compared with saline-treated C57BL/6 mice as well as il25^−/− mice. Given that the severe disease observed in il25^−/− mice appeared to be driven by IL-17 rather than IFNγ (Fig. 2 C), we examined whether IL-25 could inhibit EAE in ifng^−/− mice. Indeed, IL-25 treatment protected ifng^−/− mice from EAE, suggesting that IL-25 regulation of the Th17 pathway is independent of IFNγ (Fig. 5 B).

Open in a separate window

Figure 5.

IL-25 induces Th2 and suppresses Th17 independent of IFNγ. (A) Clinical score of immunized WT and il25^−/− mice compared with immunized WT mice injected daily s.c. with mIL-25 protein from day 1 to the end of the experiment. Results are representative for at least four experiments with five to six mice per group. (B) Clinical score of ifng^−/− (GKO) mice (n = 5 per group) injected daily with PBS or mIL-25 (s.c.) from days 1–12. Results are representative of two experiments. (C) DLN cells from MOG-immunized WT or ifng^−/− mice treated with IL-25 or saline were analyzed for cytokine gene expression at day 7 after immunization (five mice per group; DLN cells were processed in three different pools for mRNA analysis). Cytokine gene expression was normalized to ubiquitin. (D) In vitro cytokine recall response. DLN cells from MOG-immunized WT mice treated with PBS or mIL-25 daily s.c. were isolated at day 7 after immunization. Cells were stimulated with MOG peptide for 3 d and restimulated with anti-CD3/anti-CD28 for 2 additional days. Cytokine levels in supernatants were measured by CBA or ELISA. Cells from individual animals were cultured separately, and data are mean cytokine levels ± SD of three to five mice per treatment group. For IFNγ detection, nine mice per treatment group are shown. Data are representative of at least two experiments. *, P < 0.05. (E) DLN cells from MOG-immunized WT and ifng^−/− (GKO) mice treated with IL-25 or saline (pool of four mice per group) were stimulated in vitro with MOG for 3 d followed by incubation with IL-2 for another 2 d. Frequencies (FACS plots) and absolute numbers (bar graph) of MOG-specific IL-17–producing CD4⁺ cells were assessed. Data shown are representative of two experiments.

To further characterize the immune responses regulated by IL-25, we profiled the ex vivo gene expression pattern of DLN cells 7 d after MOG/CFA priming. We observed a marked down-regulation of IL-17 in IL-25–treated C57BL/6 as well as ifng^−/− mice (Fig. 5 C). Both genotypes showed an increased expression of IL-13 after IL-25 treatment. Interestingly, IL-25 did not regulate the expression of total IFNγ mRNA levels in response to priming with MOG peptide and CFA. It is notable that most of the IFNγ response in the DLNs is directed against the mycobacterial component of the adjuvant (only 1–2% MOG-specific T cell response and up to 10–30% anti-mycobacterial response). Thus, it is important to determine the myelin-specific T cell response. The in vitro MOG-specific cytokine recall analysis showed that DLN cells from IL-25–treated mice produced less IL-17 and IFNγ as compared with saline-treated controls (Fig. 5 D). Type 2 cytokines such as IL-4, IL-5, and IL-13 were markedly increased in IL-25–treated animals. Intracellular cytokine staining of DLN cell cultures revealed lower relative and absolute numbers of IL-17–producing CD4⁺ T cells in IL-25–treated ifng^−/− as well as C57BL/6 mice (Fig. 5 E). These findings established that treatment with recombinant IL-25 can suppress autoimmune inflammation by inducing Th2 while suppressing Th1 and Th17 responses.

EAE induction.

For studies in gene-deficient mice, we used the C57BL/ 6-MOG_33-55 model of EAE, as il25^−/−, ifng^−/−, il4 ^−/−, il4ra ^−/−, and il13 ^−/− were bred on a C57BL/6 background. C57BL/6 mice were immunized with 100 μg MOG_33-55 peptide and CFA containing 1 mg H37Ra s.c. on day 0 and 100 ng pertussis toxin (PTX) i.v. on day 0 and again on day 2. For disease-prevention studies in relapse-remitting EAE, SJL mice were immunized with 50 μg PLP_139-151 plus CFA containing 200 μg H37Ra s.c. on day 0. SJL mice were given 100 ng PTX on day 0 only. For passive transfer EAE, donor SJL mice were immunized with 50 μg PLP_139-151 in CFA (no PTX adjuvant was used for the donor mice). 7–8 d after PLP priming, DLN cells were stimulated for 3–4 d with 20 μg/ml PLP_139-151, and 2 × 10⁷ viable cells were transferred into naive recipient SJL mice. Clinical scores were as follows: tail weakness, grade 1; hind limb weakness, grade 2; partial hind limb paralysis and weight loss, grade 3; paraplegia and incontinence, grade 4; quadriplegia and wasting, grade 5; and moribund, grade 6.

IL-25, IL-25–rAdV, and antibody treatment.

Recombinant mouse IL-25 protein was administered daily s.c., beginning at 1 d before MOG immunization. For administration into the CNS, rAdV gene transfer vectors were suspended in 10 μl PBS and injected into the right-lateral cerebral ventricle using a 28-gauge needle, as described previously (33). For peripheral delivery, rAdVs suspended in 100 μl PBS were injected into the tail vein. Antibodies for neutralization of IFNγ (clone XMG1.2, rIgG1; 200 μg per mouse) and IL-17 (clone 1D10, rIgG1; 1 mg per mouse) and the corresponding isotype control (rat IgG1; 1 mg per mouse) were administered s.c. once a week starting 1 d before immunization.

Isolation of CNS inflammatory cells.

C57BL/6 mice were immunized as described. At day 7 (before disease onset) and at peak disease (day 11 in the il25^−/− and day 14 in WT), mice were killed, and CNS-infiltrating T cells, macrophages, and resident microglia were isolated by digesting brain and spinal cord homogenate with collagenase and DNase followed by Percoll gradient centrifugation as described previously (57). We determined the number of CD4⁺CD45^hi T cells and CD4⁻CD11b⁺CD45^hi inflammatory macrophages in the CNS by multiplying the percentage of lineage marker–positive cells by the total number of mononuclear cells isolated from the CNS. For further characterization, intracellular cytokine staining was performed (see the following section). For analysis of IL-25 mRNA expression in the CNS, inflammatory macrophages and resident microglia were isolated from irradiation bone marrow chimeric mice produced by the injection of CD45.1 B6 bone marrow cells into irradiated CD45.2 C57BL/ 6 mice. The CD11b⁺CD45.1^hiCD45.2^neg inflammatory macrophages and CD11b⁺CD45.2^loCD45.1^neg resident microglia (radiation-resistant cells remaining with the host CD45 allotype) were purified by three-color flow cytometric sorting from >100 irradiation bone marrow chimeric mice (11). Purified and sorted cells from naive or diseased animals were pooled (microglia or inflammatory macrophages) and analyzed by quantitative real-time PCR for IL-25.

Intracellular staining.

CNS-infiltrating cells were obtained from MOG/CFA-immunized mice as described in the previous paragraph. Cells were stained with anti–CD4-FITC, fixed, and permeabilized with Cytofix/Cytoperm (BD Biosciences). Intracellular staining with antibodies against IL-17, IFNγ, and TNF (BD Biosciences) was performed and analyzed by flow cytometry. T cells were identified by gating on CD4. The number of cytokine-producing cells was calculated from the percentage of cytokine-positive CD4⁺ T cells. We determined the number of CD4⁺ T cells in the CNS by multiplying the percentage of lineage marker–positive cells by the total number of mononuclear cells isolated from the CNS. For intracellular cytokine staining of DLN cells, DLNs were taken at day 7 after immunization. After 3 d of culture with MOG, cells were incubated with IL-2 for 2 d. Cells were then incubated with PMA/ionomycin in the presence of Golgi-Plug (BD Biosciences) for 5 h, and intracellular staining was performed as described. Absolute cell numbers were obtained by multiplying relative numbers with total number in culture plate well. For intracellular Foxp3-staining, DLNs were taken at day 7 after immunization and processed immediately for CD4, CD25, and Foxp3 staining with the Mouse regulatory T cell staining kit (eBioscience).

Cytokine assays.

Mice were immunized as described. DLNs (axillary, brachial, and inguinal) were isolated at day 7, before disease onset. Single-cell suspensions were plated at 6 × 10⁶ cells/well in 24-well plates in the presence of 100 μg/ml MOG_33-55 peptide in cRPMI. Supernatants were collected after 3 d of culture. Cells were rested for 4 d in the presence of 1–2 ng/ml IL-2. After 7 d, cells were restimulated with 10 μg/ml αCD3 (plate bound) plus 5 μg/ml αCD28, and supernatants were collected at 24 and 72 h. Alternatively, supernatants were collected at 48 h after restimulation. Cytokine production was analyzed by sandwich ELISA (IL-13 and IL-17 kits; R&D Systems) or cytometric bead array (Mouse Th1/Th2 cytokine CBA and Inflammation CBA kits; BD Biosciences). For the CBA assay, 50 μl of supernatant was incubated with a mixture of beads coated with capture antibodies for IL-2, IL-4, IL-5, IFNγ, or TNF (for Th1/Th2 kit), or IL-6, IL-10, MCP-1, IFNγ, TNF, or IL-12p70 (for Inflammation kit). Addition of the PE-conjugated detection antibodies forms a sandwich complex. After 2 h of incubation and one wash, samples were analyzed by flow cytometry. Standard curves were generated from analysis of titrated cytokine standards (at least 10 serial dilutions) using the BD CBA analysis software, which runs on Excel (Microsoft). IL-13 and IL-17 in samples from in vitro Th17 assays were assessed using a multiplexing bead array (Linco) according to the manufacturer's recommendations. In brief, 25 μl of supernatants were incubated with assay buffer, and fluorescent beads were conjugated to capture antibodies for the cytokines of interest. After addition of the detection antibody mix and PE-conjugated streptavidin, plates were read on a LUMINEX 100 machine (Luminex). Data were analyzed using Masterplex QT (MiraiBio) and GraphPad Prism (GraphPad).

Histological analysis.

Spinal cords were taken at day 18 after immunization. Fixated tissue was embedded in paraffin wax, and 5-μm-thick longitudinal sections were stained with hematoxylin and eosin. Lesions were counted, and each scored based was on the size of the inflammatory infiltrate (score 1, <10 cells; score 3, 10–50 cells; score 5, >50 cells). The total number of lesions was assessed per section, and mean and SD were calculated per treatment group. Cumulative score per section was divided by total number of lesions to obtain a mean lesion score per section. Mean and SD were calculated per treatment group.

Quantitative real-time PCR.

DLN were taken at day 7 after immunization from five mice per indicated group and processed in three different pools. Single-cell suspensions were prepared by mashing lymph nodes through a 70-μm cell strainer (BD Biosciences). CNS leukocytes were prepared and purified as described. DC cultures were prepared as stated in the following section. Total RNA was prepared from frozen cell pellets using RNeasy columns (QIAGEN), and 5 μg were subjected to treatment with DNase (Ambion) and reverse transcribed using Superscript II (Invitrogen) with oligo (dT)15 primers (Roche) and random hexamers (Promega). Cytokine-specific mRNA was measured by real-time quantitative PCR (Applied Biosystems). Gene expression levels were normalized to ubiquitin.

In vitro cultures.

Spleens and lymph nodes were taken from naive DO11.10 × RAG^−/− BALB/c mice, which only produce OVA peptide–specific naive T cells. Cells were also isolated from naive C57BL/6 MOG_33-55 TCR Tg (clone 2D2) mice or from WT C57BL/6 mice (for DC isolations). CD4⁺ as well as CD11c⁺ cells were purified by magnetic bead separation (MACS; Miltenyi Biotec) after collagenase digestion of tissue. For Th17 polarization, CD11c⁺ cells (5 × 10⁵ cells per well in a 24-well plate) were activated with 0.5 ng/ml LPS and incubated with OVA peptide as well as neutralizing antibodies against IL-4 (BVD4-1D11; 10 μg/ml) and IFNγ (XMG1.2; 10 μg/ml) for 2 h before addition of CD4⁺ cells in a 1:1 ratio. 20 ng/ml IL-23 was added to expand Th17 cells. Depending on the experimental setup, 100 ng/ml IL-25, 100 ng/ml IL-3, or 10 μg/ml anti–IL-13 were added to the cultures. Cultures with no cytokine addition served as a control. After 72 h, cells were incubated with 5 ng/ml IL-2 for an additional 48 h. Supernatants were then sampled for cytokine analysis, and cells were stimulated with PMA/ionomycin for 5 h in the presence of Golgi-Plug (BD Biosciences). Intracellular staining for IL-17 was performed as described. For DC cultures, 10⁶ CD11c⁺ DCs were plated in a 48-well plate, rested overnight, and incubated with or without IL-13 1 h before addition of LPS. Cells were harvested after 24 h and prepared for gene expression analysis.

We thank Thomas Backstrom and Vijay Kuchroo for providing the MOG-TCR transgenic mice. We also thank Alissa Chackerian, Mandy McGeachy, and Brent McKenzie for helpful discussion.

SP Biopharma (formerly DNAX Research) is supported by the Schering-Plough Corporation. The authors have no other conflicting financial interests.