Effector mechanisms of interleukin-17 in collagen-induced arthritis in the absence of interferon-γ and counteraction by interferon-γ

Recombinant mouse IFN-γ was derived from the supernatant fluid of Mick cells, a Chinese hamster ovary (CHO) cell line developed in our laboratory [18]. IFN-γ was purified by affinity chromatography to a specific activity of 10^8.5units/mg as described [19].

Anti-IL-17A antibody MM17F3 (IgG1-K) was derived from C57Bl/6 mice immunized with IL-17A-ovalbumin complexes as described [20]. Murine IgG1 monoclonal antibody (9E10), that recognizes part of the human c-myc protein, and is produced by the MYC1-9E10.2 (ATCC CRL 1729) hybridoma, was used as control antibody [21].

Generation and basic characteristics of the mutant mouse strain (129/Sv/Ev) with a disruption in the gene encoding for the α-chain of the IFN-γ R (IFN-γR KO) have been described [22]. These IFN-γR KO mice were backcrossed with wild-type DBA/1 mice for 10 generations to obtain IFN-γR KO DBA/1 mice. Homozygous IFN-γR KO mice were identified by PCR as described [23]. Wild-type and IFN-γR KO DBA/1 mice were bred in the Experimental Animal Centre of the Rega Institute for Medical Research at Leuven.

The generation and basic characterization of IFN-γ-deficient mice of the 129 × BALB/c strain have been described [24]. These mice were backcrossed for eight generations to the parental BALB/c strain. The signal transducer and activator of transcription (STAT)-1^-/- and interferon regulatory factor (IRF)-1^-/- mice, on a C57BL/6 background, were from Dr. David E. Levy of the New York University School of Medicine (New York, USA) and from Dr. Tak Mak of the Ontario Cancer Institute (Ontario, Canada), respectively. The generation and characteristics of these mice have been described [25, 26]. C57BL/6 mice (Harlan, Zeist, the Netherlands) were used as wild-type controls.

Chicken collagen type II (CII; Sigma-Aldrich, St Louis, MO, USA) was dissolved at 2 mg/ml in PBS containing 0.1 M acetic acid by stirring overnight at 6°C and emulsified in an equal volume of CFA (Difco Laboratories, Detroit, MI, USA) with added heat-killed Mycobacterium butyricum (1.5 mg/ml). Mice were sensitized with a single intradermal injection at the base of the tail with 100 μl of the emulsion on day 0, and treated with 0.2 mg of neutralizing anti-IL-17 or control antibody (in 250 μl PBS) once a week. Clinical and histological severity of arthritis were recorded following a scoring system as described [27, 28].

All animal experiments were approved by the local ethical committee (University of Leuven).

Blood samples were taken from the orbital sinus and were allowed to clot at room temperature for one hour and at 4°C overnight. Individual sera were tested for antibodies directed to chicken CII by ELISA as described [27]. For the determination of CII-specific IgG1, IgG2a and IgG2b antibody, plates were incubated for two hours with biotinylated rat antibody to mouse IgG1, IgG2a or IgG2b (Zymed Laboratories, San Francisco, CA, USA), followed by a one-hour incubation with streptavidin-conjugated peroxidase. For measurement of total IgG antibody, plates were coated with goat anti-mouse IgG (Jackson Immunoresearch Laboratories, West Grove, PA, USA; 10 μg/ml; 100 μl/well), followed by the same procedure as described in [27].

For evaluation of delayed-type hypersensitivity (DTH) reactivity, CII/CFA-immunized mice were subcutaneously injected with 20 μg of CII/20 μl PBS in the left ear and with 20 μl PBS in the right ear. DTH response was calculated as the percentage swelling (the difference between the increase of thickness of the left ear and the right ear, divided by the thickness of the right ear, multiplied by 100).

Spleens were harvested, gently cut into small pieces and passed through cell strainers (Becton Dickinson Labware, Franklin Lakes, NJ, USA). Red blood cells were lysed by two consecutive incubations (5 and 3 minutes at 37°C) of the suspension in NH₄Cl (0.83% in 0.01 M Tris-HCl, pH 7.2). Remaining cells were washed and counted.

To exclude any interference from endogenous IFN-γ, MEF were from IFN-γ KO BALB/c origin (unless otherwise mentioned). MEF were isolated from mouse embryos between 16 and 18 days of gestation, as described [28]. Two times 10⁵ MEF in a total volume of 300 μl Eagle's minimal essential medium (EMEM) containing 2% heat-inactivated FCS were seeded in chamber slides (LAB-TEK Brand Products, Nalge Nunc International, Naperville, IL, USA). After an incubation of 48 hours, cells were stimulated with IL-17 (20 ng/ml) (R&D systems, Abingdon UK) and/or TNF-α (20 ng/ml) (R&D systems, Abingdon UK) in the presence or absence of IFN-γ (100 units/ml) for 48 hours. Supernatants were collected and cells were harvested using a cell scraper (LAB-TEK Brand Products, Nalge Nunc International, Naperville, IL, USA). Cells were washed and pellets were used for PCR.

Single-cell splenocyte suspensions (0.5 × 10⁶ cells) were incubated for 15 minutes with the Fc-receptor-blocking antibodies anti-CD16/anti-CD32 (BD Biosciences Pharmingen, San Diego, CA, USA). Cells were washed with PBS (2% FCS) and stained with the indicated fluorescein isothiocyanate (FITC)-conjugated antibodies (0.5 μg) for 30 minutes, washed and incubated with the indicated phycoerythrin (PE)-conjugated antibodies for 30 minutes. FITC-conjugated anti-CD25 (7D4) and FITC-conjugated anti-B220 were purchased from BD Biosciences Pharmingen, San Diego, CA, USA). FITC-conjugated anti-CD11b (M1/70), FITC-conjugated anti-CD8 (53-6.7), PE conjugated anti-CD4 (L3T4), and PE-conjugated anti-Gr-1 (RB6-8C5) were from eBioscience (ImmunoSource, Halle-Zoersel, Belgium). Cells were washed, fixed with 0.37% formaldehyde in PBS and flowcytometric analysis was performed on a FACScan flow cytometer with Cell Quest^® software (Becton Dickinson, San Jose, CA, USA).

MEF were obtained as described above. RNA was extracted using the Micro-to-Midi Total RNA Purification System (Invitrogen Life Technologies, Carlsbad, CA, USA) in accordance with the manufacturer's instructions. cDNA was obtained by reverse transcription using Superscript II Reverse Transcriptase and random primers (Invitrogen Life Technologies, Carlsbad, CA, USA), in accordance with the manufacturer's instructions. For real-time PCR we used a TaqMan^® Assays-on-Demand™ Gene expression Product from Applied Biosystems (Foster City, CA, USA). Expression levels of granulocyte chemotactic protein-2 (GCP-2) (assay ID Mm00436451_g1, Applied Biosystems, Foster City, CA, USA), IL-6 (assay ID Mm00446190_m1, Applied Biosystems, Foster City, CA, USA), RANKL (assay ID Mm00441908_m1, Applied Biosystems, Foster City, CA, USA), IP-10 (assay ID Mm99999072_m1) and granulocyte macrophage colony-stimulating factor (GM-CSF; assay ID Mm99999059_m1) were normalized for 18S RNA (Cat. No. 4319413E, Applied Biosystems, Foster City, CA, USA) expression. Analysis was performed in an ABI Prism 7000 apparatus (Applied Biosystems, Foster City, CA, USA) under the following conditions: inactivation of possible contaminating amplicons by AmpErase uracil-N-glycosylase at 50°C for two minutes, initial denaturation at 95°C for 10 minutes, followed by 40 thermal cycles of 15 seconds at 95°C and 90 seconds at 60°C. The relative gene expression was assessed using the 2^-ΔΔCT method [29].

Expression of cytokines (i.e. IL-1β, IL-2, IL-4, IL-5, IL-6, IL-10, GM-CSF, IFN-γ and TNF-α) was determined by the Bio-Plex 200 system, Bio-plex Mouse Cytokine 8-plex assay and Bio-plex Mouse IL-6 Assay (Bio-Rad, Hercules, CA). IL-17, IP-10 and GM-CSF levels were measured by ELISA (R&D systems, Abingdon UK). GCP-2 was detected by an ELISA developed in our laboratory, as described [28].

IFN-γR KO mice develop CIA more readily than wild-type controls: symptoms of arthritis usually appear in IFN-γR KO mice from day 16 onwards compared with day 30 in wild-type animals [30]. In a first experiment, we confirmed the inhibitory activity of IFN-γ on the production of IL-17. Thus, IFN-γR KO and wild-type mice were immunized with CII in CFA. At day 21, a time point when the differences in disease symptoms between both groups are most pronounced, mice were injected with anti-CD3 and sera were collected. As expected, levels of IL-17 were more than four times higher in the sera of immunized IFN-γR KO mice as compared with those of wild-type counterparts (Figure 1a, in vivo). Similarly, anti-CD3-stimulated lymphocytes of immunized IFN-γR KO mice produced levels of IL-17 that were more than 10 times higher than those of wild-type mice (Figure 1a, in vitro). Therefore, to study the effector functions of IL-17 in arthritis, we chose IFN-γR KO mice for the induction of arthritis.

CII/CFA-immunized IFN-γR KO mice were injected with neutralizing anti-IL-17 or control antibody once a week starting from day 0 (day of immunization). In control-treated IFN-γR KO mice, symptoms of arthritis appeared from day 14 and reached a cumulative incidence of approximately 92%. In contrast, mice treated with anti-IL-17 antibody developed a significantly less severe form of arthritis with a lower incidence (25%) than control mice (Figures 1b, c). This protective effect was confirmed in two independent experiments. When the analysis was restricted to only arthritic mice, severity of arthritis was still lower in anti-IL-17-treated mice compared with control-treated mice (Figure 1d). In one such experiment, mice were sacrificed on day 25 for histological examination of the joints. As evident from data in Figure 2a, the reduced severity of arthritis in anti-IL-17-treated mice was associated with inhibition of infiltration of mono- and polymorphonuclear cells, hyperplasia and pannus formation (measured as the fraction of synovial inflammatory tissue, which has invaded bone tissue and forms bone erosion). In the control-treated group, several multinucleated osteoclast-like cells were detected at sites of bone erosion. Osteoclast-like cells and polymorphonuclear cells were completely absent in sections of anti-IL-17-treated mice. Figures 2b and 2c show light microscopy on hematoxylin-stained sections of both groups of mice.

Pathogenesis of CIA is generally considered to depend in part on both humoral and cellular immune responses against CII. We wanted to define whether the protection against CIA by anti-IL-17 antibodies results from modulation from either of these. Total IgG and anti-CII IgG, IgG1, IgG2a and IgG2b were determined in the sera of anti-IL-17- and control-treated IFN-γR KO mice on day 22 post-immunization. No differences in total IgG serum content were detected in the sera of anti-IL-17- and control-treated animals (Figure 3a). Titers of anti-CII antibodies were found to be reduced in the sera of anti-IL-17-treated mice, although values did not reach statistical significance (Figure 3b). As to subtypes of anti-CII IgG, no differences in anti-CII IgG1 and IgG2b could be detected. However, levels of anti-CII IgG2a were significantly lower in sera of anti-IL-17-treated IFN-γR KO mice as compared with control-treated mice (Figure 3c). DTH was tested on day 25 after immunization by injecting 20 μg of CII in the left ears and vehicle in the right ears. Bars in Figure 3d represent the percentages of swelling in the CII-challenged ears, normalized to the swelling of the PBS-challenged ears. A significantly lower DTH response to CII was observed in the anti-IL-17-treated mice as compared with that in control mice. Thus, the reduced severity of arthritis in anti-IL-17-treated mice appeared to be associated with reduced cellular immune responsiveness to CII.

The more severe form of arthritis in IFN-γR KO mice as compared with wild-type mice is accompanied with an extramedullary hemopoiesis and expansion of the CD11b⁺ cell population [31]. These expanding CD11b⁺ splenocytes, containing mostly immature mononuclear phagocytes and neutrophils, can act as a source of osteoclasts and may thus indirectly account for bone destruction in CIA [23] or may contribute to neutrophil inflammation in the joints. IL-17 is known to stimulate granulopoiesis in vivo [32] and to induce the production of hematopoietic cytokines such as IL-6, IL-8 and G-CSF in vitro [5], so we analysed the effect of IL-17 neutralization on the expansion of the CD11b⁺ cell expansion. Therefore, spleens were isolated from anti-IL-17-treated and control-treated IFN-γR KO mice on day 21 after immunization. The mean number of splenocytes was significantly lower in mice treated with anti-IL-17 as compared with control antibody (Figure 4a). To characterize these splenocytes, flowcytometric analysis was performed. Figure 4b shows that the anti-IL-17 antibodies inhibit the expansion of the CD11b⁺ cell population. In fact, treatment with anti-IL-17 antibodies resulted in significantly lower net numbers of CD11b⁺Gr-1^high neutrophils in the spleen. Numbers of CD4⁺ and CD8⁺ T cells were also significantly lower in anti-IL-17-treated mice, although to a lesser degree. B220⁺ cell numbers were comparable in both groups of mice. These data were confirmed by cytospin splenocyte preparations (Figure 4c). Significantly lower numbers of immature and mature neutrophils were observed in the spleen of anti-IL-17-treated as compared with control-treated mice. Although to a lesser degree, a significant reduction in the number of macrophages in the splenocyte population of anti-IL-17-treated mice was also seen. Treatment with anti-IL-17 antibodies did not significantly affect the number of lymphocytes in the splenocyte preparations.

The reduced expansion of the CD11b⁺ cell population following treatment with anti-IL-17 antibodies might result from inhibition of the production of hemopoietic cytokines. Therefore, sera of anti-CD3-challenged IFN-γR KO mice treated with anti-IL-17 or control antibodies were collected on day 21 and tested for the presence of cytokines. Levels of GM-CSF, IL-6 and IL-12, known to stimulate hemopoiesis, were significantly reduced in the sera of mice treated with anti-IL-17 antibodies as compared with control-treated mice (Figure 4d), possibly providing an explanation for the reduced expansion of CD11b⁺ cells in the spleen of anti-IL-17-treated mice. Expression of IFN-γ, known to inhibit the CFA-induced expansion of CD11b⁺ cells, was also found to be significantly reduced upon treatment with anti-IL-17. However, as IFN-γR KO mice were used, levels of IFN-γ have no effect in our model. With regard to the expression of other cytokines, levels of the pro-inflammatory cytokine TNF-α were slightly reduced upon treatment with anti-IL-17 antibodies (1133.3 ± 658.6 and 721.9 ± 562.3 pg/ml for the control- and anti-IL-17-treated mice, respectively), and expression of all other tested cytokines (IL-2, IL-4, IL-5 and IL-10) were comparable in the sera of both groups of mice (data not shown).

In a next set of experiments, we verified whether the absence of bone destruction and reduced influx of neutrophils in the joints, as well as the reduced production of hematopoietic cytokines such as IL-6 upon anti-IL-17-antibody treatment, is indirectly due to the reduced inflammation seen in these mice or may directly result from neutralization of IL-17. As to the inhibition of bone destruction, we compared the osteoclastogenic capacity of splenocyte populations of immunized anti-IL-17-treated and control-treated mice. Thus, splenocytes from both groups of mice were stimulated with macrophage colony-stimulating factor (M-CSF) and RANKL for induction of osteoclastogenesis. We found similar numbers and activity of osteoclasts in both groups of mice (data not shown), indicating that osteoclast precursors are present in both splenocyte populations. Possibly, the migration of these osteoclast precursors to the joints or the production of osteoclast stimulating factors such as RANKL in the joints is disturbed in anti-IL-17-treated mice. As fibroblasts are an important source of RANKL, IL-6 and the neutrophil-specific chemokine GCP-2 ([28]and our unpublished results), and because of the limited source of mouse synovial fibroblasts, we chose to use MEF for in vitro stimulation. Cells were stimulated with IL-17 in the absence or presence of TNF-α for 48 hours, and mRNA levels were measured using quantitative PCR. In addition, we tested the effect of IFN-γ on the induction of RANKL, IL-6 and GCP-2. As shown in Figure 5a, IL-17 induced the expression of GCP-2 mRNA in MEF. Moreover, a synergy between IL-17 and TNF-α in the induction of GCP-2 mRNA could be observed. These findings were confirmed at the protein level, using a GCP-2-specific ELISA (Figure 5b). IL-17 and TNF-α were also found to synergistically induce the expression of IL-6 mRNA (Figure 5c), IL-6 protein (Figure 5d) and RANKL mRNA (Figure 5e). These data indicate that IL-17 is able to induce the production of neutrophil-specific chemokines such as GCP-2. Through induction of IL-6 and RANKL, IL-17 can stimulate hemopoiesis and bone destruction. Importantly, expression of GCP-2, IL-6 and RANKL was counteracted by IFN-γ (Figures 5a to 5e). Thus, aside from inhibition of the production of IL-17, IFN-γ can inhibit the effector function of IL-17. To exclude that IFN-γ non-specifically inhibits cytokine production, we also tested the effect of IFN-γ on the production of keratinocyte-derived chemokine (KC), macrophage inflammatory protein (MIP)-2, IP-10, Regulated upon Activation, Normal T-cell Expressed, and Secreted (RANTES), monocyte chemotactic protein (MCP)-1, interferon-inducible T cell alpha chemoattractant (ITAC), and monokine induced by gamma interferon (MIG). IFN-γ was found to exhibit stimulatory effects on the production of MIP-2, IP-10, RANTES, ITAC and MIG induced by IL-17 and/or TNF-α (Figure 5f and data not shown). No significant effect of IFN-γ was observed on the production of M-CSF, KC and MCP-1 (Figure 5f and data not shown).

STAT-1 is a major mediator of cell activation by IFN-γ [33]. However, IFN-γ also activates STAT-1-independent pathways, and it has been proposed that these STAT-1-independent pathways mediate suppressive activities of IFN-γ. To investigate whether suppression of GCP-2, IL-6 and RANKL expression by IFN-γ was STAT-1-dependent, MEF from wild-type and STAT-1^-/- mice of the C57BL/6 strain were prepared. At the same time, we investigated whether the inhibition by IFN-γ was dependent on IRF-1, a transcription factor acting immediately downstream of STAT-1, by preparing MEF from IRF-1^-/- C57BL/6 mice. In each of the MEF, IL-17 and TNF-α synergistically induced the expression of GCP-2, RANKL and IL-6 mRNA (data not shown). To evaluate the importance of STAT-1 and IRF-1 in the inhibitory action of IFN-γ, the percentage of inhibition of the expression of GCP-2, RANKL and IL-6 mRNA induced by synergistic action of IL-17 and TNF-α was calculated and compared. As shown in Figure 6a, the inhibition by IFN-γ was significantly reduced in STAT-1^-/-, but unaffected in IRF-1^-/- MEF, as compared with wild-type MEF. For the inhibition of the expression of GCP-2, these results were confirmed at the protein level (Figure 6b). These data indicate that the inhibition by IFN-γ is STAT-1-, but not IRF-1-dependent.