IL-18 Enhances Collagen-Induced Arthritis by Recruiting Neutrophils Via TNF-α and Leukotriene B₄¹

Interleukin-18 is a member of the IL-1 cytokine family (1). Pro-IL-18 is cleaved by IL-1β-converting enzyme (ICE,⁴ caspase 1) to yield an active 18-kDa gp (2) that recognizes a heterodimeric receptor, consisting of unique α (IL-1R-related protein) and nonbinding β (accessory protein-like) signaling chains (3, 4), that is widely expressed on cells that mediate both innate and adoptive immunity. IL-18 is expressed by various cell types, including macrophages, dendritic cells, keratinocytes, osteoblasts, pituitary gland cells, adrenal cortical cells, intestinal epithelial cells, skin cells, and brain cells (5, 6, 7, 8, 9). IL-18 promotes proliferation and IFN-γ production by Th1, CD8⁺, and NK cells in mice and in humans (5). It shares some biological activities with IL-12, but lacks significant structural homology and serves as a costimulatory factor in the activation of Th1 cells (10). It promotes Th1 cell development via induction of IL-12R expression (11), and thereby synergizes with IL-12 for IFN-γ production (12). In the absence of IL-12, IL-18-mediated effects on T cells may extend beyond Th1 differentiation to include type 2 cytokine production (13, 14, 15), even in the absence of IL-4.

Recent reports indicate a role for IL-18 in the pathogenesis of several inflammatory diseases. In humans, IL-18 expression has been reported in psoriasis, inflammatory bowel disease, and sarcoidosis (16, 17, 18). We have demonstrated that IL-18 is present in significant levels in the rheumatoid arthritis (RA) synovium (19) where it induces and sustains articular Th1 cell responses and independently promotes TNF-α production. IL-18-deficient mice developed significantly reduced incidence and severity of collagen-induced arthritis (CIA), compared with wild-type mice, and are associated with suppressed TNF-α production and Th1 immune responses ex vivo. This reduction in disease and immune response was completely reversed by the administration of rIL-18 (20). Furthermore, anti-IL-18 Ab suppresses streptococcal cell wall-induced arthritis through an IFN-γ-independent mechanism (21). These data demonstrate clearly that IL-18 is of importance during developing and sustained inflammatory diseases. However, the precise mechanism by which IL-18 mediates the pathologic state remains to be explored.

We recently reported that IL-18 administration promoted neutrophil accumulation in vivo, whereas IL-18 neutralization suppressed the severity of footpad inflammation following carrageenan injection (22). Such findings suggest that IL-18 can induce inflammation by activating neutrophils, the infiltration and accumulation of which is a feature of many autoimmune lesions (23, 24, 25). However, the precise mechanisms by which this is accomplished were unclear. In this report we demonstrate that IL-18 activates and attracts neutrophils by inducing the production of TNF-α, which in turn induces the synthesis of leukotriene B₄ (LTB₄), a well-known chemoattractant of neutrophils (26, 27). This finding is supported by the additional observation that inhibition of LTB₄ synthesis attenuated IL-18-induced CIA. Furthermore, IL-18 strongly induced LTB₄ synthesis by human peripheral blood (PB) neutrophils. Therefore, data reported in this study significantly advance our understanding of the mode of action of IL-18 in inflammation and provide further rationale for IL-18 targeting as a novel strategy to treat inflammatory disease.

Male BALB/c, C57BL/6, and TNFR p55^−/− (28) mice were bred and maintained in the animal housing facility of the Department of Pharmacology, University of São Paulo (São Paulo, Brazil) as previously described (29). The p55^−/− mice had been bred into the BALB/c background for >10 generations. Male DBA/1 mice obtained from Harlan Olac (Bicester, U.K.) were used at 8–10 wk old and maintained at the Joint Animal Facilities, University of Glasgow (Glasgow, U.K.). All animal experiments conducted in this study were cared for in accordance with the National Institutes of Health or with the Home Office U.K. animal research guidelines. Recombinant murine (rm) IL-18 was obtained from PeproTech (London, U.K.). rmTNF-α (lot 99/532, specific activity 2 × 10⁵ IU/g) and sheep anti-mouse TNF-α antiserum (H92/B5) were gifts from Dr. S. Poole (National Institute for Biological Standards and Control, London, U.K.). LTB₄ receptor antagonist CP-105,696 (CP) was provided by Dr. M. Teixeira (Federal University of Minas Gerais, Minas Gerais, Brazil). LTB₄ synthesis inhibitor MK-886 (MK) was purchased from Calbiochem (San Diego, CA). All reagents were free of LPS contamination as measured by the Limulus amebocyte test (<0.01 ng/mg E-toxate; Sigma-Aldrich, Poole, U.K.).

Murine neutrophils were isolated from pooled venous blood samples via cardiac puncture and purified using Ficol-Hypaque modified medium (Cardinal Associates, Santa Fe, NM) according to the manufacturer’s instructions with viability >95% as determined by trypan blue exclusion. Neutrophils were then washed three times with Hank’s medium and resuspended in RPMI 1640 medium containing 0.01% BSA (Sigma-Aldrich). Human PB neutrophils from normal donors and RA patients were isolated as previously described (22).

Leukocyte migration was initiated in BALB/c, C57BL/6, and p55^−/− mice as described previously (29). Briefly, mice were injected i.p. with rmIL-18 (20 ng/animal), rmTNF-α (40 ng/animal), or PBS. At an indicated time point after injection, mice were sacrificed, and the peritoneal cavity cells were harvested by washing the cavity with 5 ml of PBS containing 1 mM of EDTA. Total counts were performed in a cell counter (COULTER A^CT; Coulter, Miami, FL), and differential cell counts were enumerated on cytocentrifuge (Thermo-Shandon, Pittsburgh, PA) slides stained with Rosenfeld. The differential count (>200 cells) was performed under a light microscope, and the results were presented as the number of neutrophils per cavity. To establish the role of IL-18 in neutrophil recruitment, mice were pretreated with either CP (3 mg/kg) 30 min before, or dexamethasone (1 mg/kg, Sigma-Aldrich), or MK (1 mg/kg), both at 1 h before rmIL-18 injection. In some experiments, either sheep anti-mouse TNF-α (25 μl/mouse) or control serum was administered i.p. 15 min before rmIL-18 i.p. injection, and neutrophil recruitment was evaluated 4 h later.

Chemotaxis was studied in 48-well microchemotaxis Boyden chambers (NeuroProbe, Cabin John, MD) containing 5 μm pore size polyvinylpyrrolidone-free polycarbonate membranes. rmIL-18 diluted in RPMI 1640/0.01% BSA (560 pg in 28 μl final volume) or medium alone was placed in the bottom chamber. Neutrophil suspension (50 μl of 10⁶ cells/ml) was then added to the top chamber. In some experiments, neutrophils were preincubated with MK (1 μM) for 20 min before being assayed. The chambers were then incubated for 1 h at 37°C in 5% CO₂, after which the membranes were removed, fixed, and stained with Diff-Quick stain kit (American Scientific Products, McGraw Park, IL). The number of neutrophils migrated to the lower side of the filter membrane was counted by light microscopy (×100) in five random fields in triplicate. Migration of the neutrophils toward RPMI/BSA, which indicates random migration, served as the negative control. Results were expressed as the number of emigrating neutrophils per microscopic field.

CD4⁺ T cells were positively purified from peritoneal lavage fluids using Biomagnetic separation (Dynabeads; Dynal Biotech, Oslo, Norway) according to the manufacturer’s instructions. The cells were stimulated with IL-18 (20 ng/ml) for 2 h, and aliquots of the supernatants were stored at −70°C for TNF-α determination. The purity of CD4⁺ T cells (>98%) was determined by FACS analysis (FACSort; BD Biosciences, San Jose, CA) (data not shown).

CIA was elicited in mice as previously described (30). Briefly, mice were immunized by intradermal (i.d.) injection of 200 μg of acidified bovine type II collagen (CII) (Sigma-Aldrich) emulsified in IFA (Difco, Detroit, MI). Collagen (200 μg in PBS) was injected again i.p. on day 21. Mice were monitored for signs of arthritis as previously described (30). Scores were assigned based on erythema, swelling, or loss of function present in each paw on a scale of 0–3, giving a maximum score of 12 per mouse. Paw thickness was measured with a spring-loaded dial-caliper (Kroeplin, Munich, Germany). For histological assessment, mice were sacrificed, and the hind limbs were removed and fixed in 10% neutral-buffered formalin. Then, 5 μm sections were stained with H&E or toluidine blue (both Sigma-Aldrich). The quantification of arthritis was by “treatment-blind” observer, and a score was assigned to each joint based on the degree of inflammation, synovial hyperplasia, and erosion as described previously (31).

DBA/1 mice were primed i.d. on day 0 with CII in IFA and boosted on day 21 with CII in saline. Some of the mice were injected i.p. with rmIL-18 (100 ng/injection diluted in PBS supplemented with 0.1% BSA (Sigma-Aldrich)). Cytokine treatment began 1 day before the primary and secondary immunization with collagen and continued for 6 consecutive days (i.e., from day −1 to day 4 and then again from day 20 to day 24). MK or vehicle control (4% methylcellulose in PBS; Sigma-Aldrich) was administered orally (4 mg/kg daily) from day 25 for 14 days.

Draining lymph nodes were removed on day 40 after primary immunization. Single-cell suspensions were prepared and cultured in triplicate at 2 × 10⁶ cells/ml in RPMI 1640 supplemented with 100 IU/ml penicillin, 100 μg/ml streptomycin, 25 mM HEPES buffer, and 10% FCS (Invitrogen, Paisley, U.K.) at 37°C in 5% CO₂. Cells were stimulated with graded concentrations of type II collagen in flat-bottom 96-well plates (Nunc, Roskilde, Denmark). Supernatants were collected after 72 or 96 h and stored at −20°C until assayed for cytokine concentration. Proliferation assays were performed in parallel cultures in U-bottom 96-well plates (Nunc) for 96 h and were pulsed with [³H]thymidine (Amersham, Aylesbury, U.K.) during the last 6 h of culture. Plates were then harvested and measured for incorporation of radioactivity as previously described (20).

The LTB₄ concentration in peritoneal lavage fluid was determined by RIA (DuPont/NEN, Boston, MA) as described previously (29). LTB₄ produced by human PB neutrophils was measured by a solid phase competition enzyme immunoassay (R&D Systems, Oxon, U.K.) according to the manufacturer’s recommendations. All cytokines and anti-collagen Ab levels were detected by ELISA. The Ab pairs used were: TNF-α, IFN-γ, IL-5, IL-6, and IL-10 (all BD PharMingen, San Diego, CA), and assays were performed according to the manufacturer’s instructions. Detection limits were as follows: IL-5, IL-6, and TNF-α, all at 10 pg/ml; IL-10 and IFN-γ, both at 20 pg/ml. Serum anti-collagen II Ab titers of individual serum were detected with biotin-conjugated anti-mouse IgG1 or IgG2a (BD PharMingen) followed by conjugated avidin peroxidase (Sigma-Aldrich) and developed with tetramethylbenzidine substrate (Kirkegaard & Perry Laboratories, Gaithersburg, MD).

For the results of both in vivo and in vitro migration experiments, the means from different treatments were compared by ANOVA. When significant differences were identified, individual comparisons were subsequently made by the Student’s t test with a Bonferroni correction factor for unpaired values. Clinical and histological scores were analyzed with the nonparametric Mann-Whitney U test. Differences between cumulative incidences at a given time point were analyzed by the χ² contingency analysis. LTB₄, cytokine, and collagen-specific IgG levels were compared using the Student’s t test.

We have previously shown that IL-18 plays an important role in neutrophil activation and is a potent inducer of neutrophil migration in vivo (22). This finding defines a novel pathway whereby IL-18 can amplify acute inflammation. Therefore, we explored the mechanism by which IL-18 mediates neutrophil migration. Because LTB₄ is a well-characterized chemoattractant of neutrophils (26, 27), we sought direct evidence that LTB₄ is a mediator of IL-18-induced neutrophil migration. BALB/c mice were injected i.p. with 20 ng/ml rmIL-18 in the presence of MK, CP, or as positive control, the corticosteriod dexamethasone at previously optimized concentrations. Neutrophil accumulation in the peritoneal cavity was measured 4 h later. As expected, IL-18 induced significant neutrophil migration compared with PBS. The migration was completely abrogated by dexamethasone and MK, and markedly inhibited by CP (Fig. 1,A). Furthermore, analysis of peritoneal lavage fluid showed that IL-18 induced the production of high concentrations of LTB₄, which was completely inhibited by prior in vivo administration of MK (Fig. 1,B). Therefore, these results strongly suggest that IL-18-induced neutrophil migration is LTB₄-dependent. This was supported by our in vitro analysis using the Boyden chamber. Murine neutrophils migrated significantly across a membrane toward an IL-18 concentration gradient. This IL-18-induced chemotaxis was also markedly inhibited by MK (Fig. 1 C).

To investigate whether IL-18-induced LTB₄ production by murine neutrophils is also applicable to human neutrophils, we cultured purified human PB neutrophils from healthy donors and patients with RA, with recombinant human IL-18 in the presence or absence of MK. The production of LTB₄ in culture supernatants was determined by ELISA. Fig. 2 shows that neutrophils from both RA and healthy donors produced significant amounts of LTB₄ in response to IL-18. The level achieved was comparable to that induced by fMLP, a well-known potent LTB₄ inducer. The production of LTB₄ induced by IL-18 and fMLP was completely inhibited by MK.

We next explored the mechanism by which IL-18 induced LTB₄ synthesis. We have previously reported that peritoneal exudate cells that are stimulated with TNF-α release LTB₄ (29) and that IL-18-activated macrophages (19) and neutrophils (22) produce TNF-α. Therefore, we investigated the role of TNF-α in IL-18-induced LTB₄ synthesis. Peritoneal fluid from mice injected with IL-18 contained significantly higher concentrations of LTB₄ compared with controls injected with PBS. The elevated LTB₄ production was completely abolished by prior injection of a neutralizing anti-TNF-α Ab (Fig. 3,A), suggesting that the IL-18-induced LTB₄ production was TNF-α-dependent. We then investigated the cell types other than neutrophils and macrophages that might produce TNF-α in response to IL-18. Because Th1 cells preferentially express IL-18R (11), we examined whether CD4⁺ T cells could produce TNF-α in response to IL-18. Purified CD4⁺ T cells produced significant amounts of TNF-α compared with cells cultured with medium alone (Fig. 3 B). Thus, it appears that IL-18 activates CD4⁺ cells, macrophages (19), and neutrophils (22) to produce TNF-α, which in turn induces LTB₄ synthesis that mediates neutrophil migration.

To confirm the role of TNF-α in IL-18-induced cell migration, we used both neutralizing anti-TNF-α Ab and TNFR (p55) knockout mice. BALB/c mice were injected i.p. with rmIL-18 in the presence of anti-TNF-α Ab or an irrelevant control Ab. Fig. 4,A shows that IL-18-induced neutrophil migration into the peritoneal cavity was completely inhibited by anti-TNF-α Ab. Furthermore, the administration of IL-18 into the peritoneal cavity of TNFRp55^−/− was unable to induce neutrophil accumulation (data not shown). The crucial role of TNF-α was confirmed in the in vitro Boyden chamber assay. The migration of neutrophils in response to IL-18, but not to fMLP gradient, was completely abolished by anti-TNF-α Ab (Fig. 4,B). Moreover, IL-18 induced chemotaxis of neutrophils from the wild-type C57BL/6 mice, but not cells from the TNFR (p55) knockout mice (Fig. 4 C).

We next investigated the relevance of our findings in an inflammatory disease model, CIA. We have previously shown that DBA/1 mice immunized with CII in IFA and boosted with CII in saline developed only a modest degree of CIA. The disease was significantly enhanced when the immunized mice were injected with rIL-18, showing that IL-18 plays a significant role in CIA (30). To determine whether IL-18-induced CIA is LTB₄ dependent, we investigated the effect of MK on the development of CIA in this model. DBA/1 mice were injected i.d. with CII in IFA and boosted i.p. 21 days later with CII in saline. IL-18 was administered i.p. before and during CII priming and upon subsequent challenge. The mice were treated orally daily with MK or carrier for 14 days from day 25. Results presented in Fig. 5,A show clearly that IL-18-induced CIA was markedly reduced by treatment with MK. Furthermore, the mean number of arthritic paws in MK-treated animals was also reduced (p < 0.05) compared with control mice (Fig. 5,B), suggesting clinical progression was modified by the blockade of LTB₄ synthesis. As expected, no difference in the incidence of arthritis was seen between the two groups throughout the study, because the treatment was initiated after the onset of the disease (Fig. 5 C).

To examine whether MK administration prevented articular destruction, we evaluated cartilage and bone integrity histologically. The joints of the control mice given carrier alone showed extensive infiltration of inflammatory cells into the articular compartment, synovial hyperplasia, and bone and cartilage erosion (Fig. 6, A and C). In contrast, only low levels of synovial hyperplasia and erosion were observed in mice treated with MK (Fig. 6, B and D). Histologic scores are summarized in Fig. 6 E. Together, these data clearly indicate that MK potently suppressed the development of IL-18-induced CIA and prevented progression of articular damage.

We next investigated whether the reduced articular inflammation may be reflected in changes in the CII-specific immunological profile of the IL-18-injected mice treated with MK. On day 40, draining lymph node cells from MK-treated or control mice were cultured with CII, and T cell proliferation and cytokine production were determined. Cells from MK-treated mice produced significantly less IL-6 and TNF-α compared with control animals, whereas T cell proliferation, IFN-γ, IL-5, and IL-10 responses remained unchanged (Fig. 7). The immune suppression by MK in vivo appeared to be Ag-specific, as Con A-induced production of IFN-γ, TNF-α, IL-5, and IL-6 in parallel cultures was not affected (data not shown). Although not statistically significant, a reduction of both serum anti-CII IgG2a and IgG1 levels was observed in mice treated with MK compared with carrier control (data not shown). However, it should be noted that although in this model MK inhibited IL-18-amplified CIA and was therefore consistent with the notion that IL-18 enhanced CIA via LTB₄, the effect of MK on LTB₄ synthesis driven by other inflammatory cytokines could not be excluded.

Data presented in this study demonstrate that IL-18-induced neutrophil chemotaxis is LTB₄-dependent. Furthermore, TNF-α plays a critical role in this process and leads to markedly enhanced CIA. The sequence of events may be as follows: IL-18 activates the production of TNF-α, which induces the synthesis of LTB₄ that attracts neutrophils to the site of inflammation where they can contribute to acute and chronic inflammatory responses. A number of reports have documented that LTB₄ antagonists could attenuate CIA (32, 33, 34), but the events leading to the induction of LTB₄ synthesis in CIA were hitherto unclear. Thus, data presented in this study not only reveal a novel function of IL-18, but also extend our understanding of the proinflammatory interactive role of IL-18, TNF-α, and LTB₄ in synovitis.

The possible mechanism of the IL-18-induced neutrophil migration and the resulting CIA is schematically represented in Fig. 8. In response to IL-18, a number of cell types, including CD4⁺ T cells (Fig. 3,B and Ref.35), neutrophils (22), and macrophages (19), produce substantial amounts of TNF-α, which in turn activates several cell types to synthesize LTB₄ (36, 37), importantly the mast cells (38). LTB₄ is a well-established chemoattractant for neutrophils, leading to local inflammation (26, 27). A critical role for mast cells in inflammatory arthritis has been proposed (39, 40). It should also be noted that in our CIA model, treatment with MK significantly inhibited the production of TNF-α (Fig. 7), suggesting that LTB₄ is also an inducer of TNF-α synthesis. Thus, the production of LTB₄ could form part of a positive feedback self-amplification circuit, perpetuating inflammatory processes.

The role of IL-18 in clinical and experimental arthritis has been well documented (19, 20, 21, 41). The involvement of TNF-α in IL-18-induced arthritis was also recognized (19, 35). Whereas the recruitment of neutrophils in the peritoneal cavity is rapid (peak at 4 h) in acute inflammatory response, IL-18 may also induce and sustain chronic inflammation in the adaptive response. Our earlier studies (29) have shown that in OVA-induced inflammation CD4⁺ T cells presented in the peritoneal cell suspension accounted for the release of TNF-α involved in neutrophil recruitment. Depletion of CD4⁺ T cells prevented the release of TNF-α and the chemotactic factor after OVA stimulation. CD4⁺ T cells, particularly the Th1 subset, which preferentially express IL-18R compared with Th2 cells (11), play a dominant role in RA (42, 43, 44). Hence, it is conceivable that IL-18, in addition to synergizing with IL-12 (12, 13) in the induction of CII-specific Th1 cells, can stimulate CD4⁺ T cells to produce TNF-α and thereby drive subsequent LTB₄ synthesis and neutrophil migration. The relative and differential roles of IL-18 in the induction of acute vs chronic inflammation are clearly important and require further exploration.

Cytokines occupy a central position in the pathogenesis of RA (45). Therapeutic blockade of TNF-α and IL-1β using soluble receptors or mAb suppresses murine CIA and RA itself (45, 46). However, clinical effects are transient, suggesting that critical pathways that maintain synovitis persist. Factors that up-regulate TNF-α production are clearly critical in disease chronicity (47). Data presented in this study suggest that IL-18, at least via its induction of TNF-α, may play such a role. The sequential events initiated by IL-18 postulated in this study and its relevance to human cells and in in vivo models further strengthen this contention and reveal several novel targets for potential therapeutic intervention in inflammatory disease in general and RA in particular.

We thank Peter Kerr (Department of Pathology, Western Infirmary, Glasgow, U.K.) for providing valuable technical assistance with histology.