Enhancing Effect of IL-17 on IL-1-Induced IL-6 and Leukemia Inhibitory Factor Production by Rheumatoid Arthritis Synoviocytes and Its Regulation by Th2 Cytokines¹

Rheumatoid arthritis (RA)³ is characterized by the chronic inflammation of the synovium with a hyperplasia of synovial lining cells that interact with blood-derived mononuclear cells (1). Monocyte/macrophage-derived cytokines such as IL-1β and TNF-α appear to play a pivotal role in local activation leading to joint destruction (2). They interact with synoviocytes to produce mediators of inflammation such as PGE₂, degrading enzymes such as collagenase, and other cytokines including granulocyte/macrophage (GM)-CSF, IL-6, and leukemia inhibitory factor (LIF) (3, 4, 5, 6, 7, 8).

Whereas T cells represent a large proportion of the inflammatory cells invading the synovial tissue, T cell-derived cytokines are less abundant in the joint than cytokines produced by the other cell types described above (9, 10). In particular, the reduced production of IL-4 may contribute to uncontrolled inflammation (11). A similar conclusion was reached for IL-10 and IL-13 (12). These cytokines, also defined as Th2 cytokines (13), have thus been classified as anti-inflammatory on the basis of their inhibitory effect on the production of IL-1, TNF-α, IL-6, and IL-8 by monocytes and synovium samples (11, 12, 14, 15, 16, 17).

Other T cell-derived cytokines such as IFN-γ, defined as Th1 cytokines (13), act in an opposite way, contributing to an inflammatory pattern (17). For newly defined cytokines, such classification is still pending. This is the case for IL-17, characterized recently as the human counterpart of mouse CTLA-8. This CD4 T cell-derived cytokine was found to directly activate fibroblasts and synoviocytes, leading to the production of IL-6, IL-8, PGE₂, and granulocyte-CSF (18). In addition, IL-17 was shown to sustain the maturation CD34-hemopoietic progenitors into neutrophils when cultured with fibroblasts (19). In keeping with its effect during inflammatory conditions, IL-17 was shown to have an enhancing effect on nitric oxide production by cartilage (20).

As this cytokine is produced only by activated T cells, we investigated the consequences of monocyte and T cell interaction on synoviocytes by looking at the effect of the T cell-derived cytokine IL-17 and the monocyte-derived cytokine IL-1 on cytokine production by synoviocytes. This was further studied in the presence of the regulatory cytokines IL-4, IL-13, and IL-10.

Purified Escherichia coli human rIL-4 (10⁷ U/mg), rIL-10 (10⁶ U/mg), rIL-13 (9.5 10⁶ U/mg), and rIL-17 were from Schering-Plough Research Institute (Kenilworth, NJ). Human rIL-1β (2 × 10⁸ U/mg) was purchased from Sigma (St. Louis, MO).

Rheumatoid synovium samples were obtained according to the revised criteria of the American College of Rheumatology (21) from patients with RA who were undergoing knee or wrist synovectomy or joint replacement. Synovium piece cultures were performed as already described (11). Briefly, fat and fibrous tissues were removed, and the synovium was cut into small pieces with a volume of approximately 5 mm³. Pieces of synovium were cultured in triplicate in complete medium made of α-MEM medium (Life Technologies, Grand Island, NY) with 2 mM l-glutamine, 100 U/ml penicillin, 50 mg/ml gentamicin, 20 mM HEPES buffer, and 10% FCS. Cultures were performed at 37°C in a 5% CO₂/95%-air humidified environment.

To isolate synoviocytes, synovium pieces were finely minced and digested with 4 mg/ml collagenase (Worthington, Freehold, NJ) in PBS-DMEM (Life Technologies) for 2 to 3 h at 37°C (7). After centrifugation, cells were suspended in complete medium and cultured in 100-mm culture petri dishes. After 48 h, nonadherent cells were removed. Adherent cells were cultured in complete medium, and at confluence were trypsinized and passaged in 150-cm² culture flasks. Synoviocytes were used between passage 3 and 8. At this time, they were a homogenous population of fibroblast-like cells, negative for the expression of CD1, CD3, CD19, CD14, and HLA-DR, and positive for the expression of CD10, CD44, and CD54, as determined by FACS analysis, using FITC-conjugated mAbs from Becton Dickinson (Mountain View, CA).

To obtain culture supernatants, synoviocytes were plated in 96-well dishes at 10⁴ cells/well in 200 μl of complete medium. The cytokines to be tested were added at the onset of the culture. Supernatants were collected after 7 days of culture.

IL-6 levels were measured by two-site sandwich ELISA as previously described (22). Briefly, supernatants or serial dilutions of IL-6 standards (Schering-Plough Research Institute) were incubated for 60 min at 37°C in 96-well microtiter plates (Nunc, Roskilde, Denmark), coated overnight at 4°C with mouse 39C3 anti-IL-6 mAb (1 μg/ml), and saturated for 90 min at 20°C with PBS 5% BSA. After washing, a biotinylated mouse anti-IL-6 mAb (1 μg/ml) was added and incubated for 90 min at 20°C. After subsequent incubation with peroxidase-coupled streptavidin and revelation with orthophenylenediamine (Sigma), the plates were read at 492 nm.

LIF levels were also measured by a two-site sandwich ELISA using two monoclonal anti-LIF Abs (23). Mouse 1F10 anti-human LIF mAb was used for coating and biotinylated mouse 7D2 anti-human LIF mAb for detection. After subsequent incubation with a streptavidin-peroxidase complex and revelation with OPD, the ODs were measured at 492 nm.

Synoviocytes (10⁴ cells/well) were incubated in 96-well plates in a final volume of 200 μl of their respective complete medium. Samples were preincubated at 37°C for 30 min with 1 μg/ml of the anti-IL-17 mAb5 (Schering-Plough Research Institute). mAb5 is a mouse IgG1 obtained after immunization with human rIL-17. 1 μg/ml of mAb5 was able to completely inhibit the IL-6 production induced by 50 ng/ml of IL-17, whereas the irrelevant mAb MX1 had no effect. This Ab had no effect on the action of IL-1, TNF-α, GM-CSF, or M-CSF (19). Following the preincubation step, IL-17 (50 ng/ml) or a 1:10 dilution supernatant (with or without anti-IL-17) was added at the onset of the culture for a 12-h incubation. Plates were washed before addition of fresh medium. Supernatants were collected after 48 h and stored at −20°C until cytokine assays.

Results were expressed as mean ± SEM of n separate experiments. Differences between IL-4- or IL-13- or IL-10-treated groups and the control group were compared with nonparametric Wilcoxon paired t test.

We investigated whether IL-6 production by synoviocytes could be modulated by the monocyte-derived cytokine IL-1β and by the T cell-derived cytokine IL-17. Synovial fibroblasts were incubated with different concentrations of IL-1 and IL-17. After 7 days of culture, supernatants were collected and assayed for IL-6 production by ELISA. When IL-17 at concentrations ranging from 0.01 to 1000 ng/ml and IL-1 at concentrations ranging from 0.01 to 1000 pg/ml were used alone, IL-6 production was increased in a dose-dependent manner (Fig. 1,A). IL-1β (100 pg/ml) induced a 33-fold increase of IL-6 production, while IL-17 (10 ng/ml) showed a lesser effect (17-fold increase) (Fig. 1 A).

When IL-1 and IL-17 were combined, there was an enhancing effect over that seen with IL-1 alone. At high concentrations of IL-17 (>10 ng/ml) and IL-1 (>10 pg/ml), an additive effect was observed until saturation was reached. At lower concentrations of IL-17 (1–10 ng/ml) and IL-1 (1–10 pg/ml) in combination, synoviocytes produced more IL-6 than the sum of IL-6 production by synoviocytes incubated with each cytokine, indicating a synergistic effect. This is further demonstrated in Figure 1 B where synovial fibroblasts were incubated with combinations of low concentrations of IL-1 (1 and 10 pg/ml) and IL-17 (1, 5, and 10 ng/ml). In particular, synergy was best observed with 1 ng/ml of IL-17 and 1 pg/ml of IL-1, which had no effect on IL-6 production when used alone. In time course studies, this synergistic effect was seen as early as 24 h after culture initiation (data not shown).

Similarly, we investigated how LIF production by synoviocytes could be modulated by IL-1β and IL-17 (Fig. 2). When IL-17 at concentrations ranging from 0.01 to 1000 ng/ml and IL-1 at concentrations ranging from 0.01 to 1000 pg/ml were used alone, LIF production was increased in a dose-dependent manner (Fig. 2 A). IL-1β (100 pg/ml) induced a 10-fold increase of LIF production, while IL-17 (10 ng/ml) showed a lesser effect (4-fold increase).

When IL-1 and IL-17 were combined, an additive effect was observed at both low and high concentrations of IL-1 and IL-17 until saturation was reached (Fig. 2,A). This is further demonstrated in Figure 2 B where synoviocytes incubated with low concentrations of IL-1 (1–10 pg/ml) and IL-17 (1–10 ng/ml) in combination produced more LIF than the synoviocytes incubated with each cytokine separately, indicating an additive effect. In time course studies, this additive effect was observed as early as 12 to 24 h after initiation of culture (data not shown).

Combined together, these results indicate that cytokines produced by monocytes (IL-1) and T cells (IL-17) can act together on synoviocytes at the low concentrations that can be expected to be present in vivo.

IL-4, IL-13, and IL-10 were tested for their effect on the secretion of IL-6 by synoviocytes stimulated with optimal concentrations of IL-1β and IL-17, alone and in combination. Levels measured after 7 days of culture are presented in Table I. Addition of IL-4 and IL-13 used alone at the onset of culture increased the spontaneous secretion of IL-6 by 4.5-fold. This effect was much lower than that of IL-1 as well as that of IL-17. However, with stimulated cells, a modest 30% enhancing effect of IL-4 and IL-13 was found on IL-6 production by IL-17-, IL-1β-, or IL-1β plus IL-17-stimulated synoviocytes. In contrast, IL-10 had rather a modest inhibitory effect on activated synoviocytes. This effect was less when IL-1β and IL-17 were combined.

As shown in Table II, and as opposed to the effect on IL-6, IL-4 and IL-13 reduced LIF production by unstimulated synoviocytes with a mean inhibition of 60 and 77%, respectively. Similar inhibitory effect was found with IL-1β- or IL-17-stimulated synoviocytes. This effect was less important when IL-1β and IL-17 were both added at the onset of culture. In contrast, IL-10 had no effect on LIF production.

To return to the in vivo situation, similar experiments were performed in the presence of low concentrations of IL-1β (1 pg/ml) and IL-17 (1 ng/ml). These concentrations of IL-17 and IL-1β alone had no effect on IL-6 production, whereas their combination induced the production of 30.3 ng/ml of IL-6, confirming the synergistic effect described above (Fig. 3). Addition of IL-4 and IL-13, but not IL-10, increased the spontaneous secretion of IL-6. The enhancing effect of IL-4 or IL-13 was more important on IL-17 or IL-17 plus IL-1β stimulation. For each of these conditions, IL-4 was in general more potent than IL-13. In contrast, IL-10 had no effect on IL-6 production by synoviocytes, except for an inhibitory effect on IL-17 plus IL-1β-stimulated synoviocytes.

Results obtained for LIF production with low concentrations of IL-1 and IL-17 are shown in Figure 4. Similar inhibitory effects of IL-4 and IL-13 were found with IL-17- or IL-1β-stimulated or unstimulated synoviocytes. This inhibition was less important when the two cytokines were combined. In contrast, IL-10 had no effect on LIF production by stimulated or unstimulated synoviocytes.

It was important to extend these results to natural IL-17 produced by RA synovium. To test its effect on synoviocytes, the biologic activity of IL-17 present in supernatants from RA synovium pieces was measured (Table III). Supernatants collected after 1 wk of culture of RA synovium pieces were preincubated with and without blocking anti-IL-17 mAb before being tested on synoviocytes. As a positive control, the IL-6-inducing effect of rIL-17, but not IL-1, was completely blocked. A 50% reduction of the induction of IL-6 and LIF production by RA synovium supernatants could be specifically blocked by anti-IL-17 mAb 5, but not by an isotype-matched irrelevant mAb, MX1.

The RA synovitis is characterized by cell interactions between bone marrow-derived cells such as monocytes, T cells, and B cells and resident mesenchymal cells, namely synoviocytes. The contribution of T cells in the proinflammatory cytokine release has been a matter of debate (24). It is clear that monocyte-derived cytokines such as TNF-α and IL-1 interact with synoviocytes, leading to the production of other cytokines such as IL-6, GM-CSF, and LIF (2). Contrasting with the abundance of monocyte and synoviocyte-derived proinflammatory cytokines, T cell-derived cytokines are difficult to detect in RA synovium (9, 10).

The objective of the present report was to examine the effect of IL-17, a specific T cell-derived cytokine, on synoviocytes in the presence of proinflammatory and anti-inflammatory cytokines. This allowed the study of the contribution of soluble factors in the interaction between T cells and synoviocytes.

IL-17 has been shown to activate the transcription nuclear factor NF-κB and to induce the production/expression of IL-6, IL-8, PGE₂, G-CSF, and ICAM-1 in fibroblasts and the production of nitric oxide by cartilage (18, 19, 20). Such effects are shared with proinflammatory cytokines such as TNF-α and IL-1. In the context of inflammation, IL-17 is of interest because it acts directly on various mesenchymal cells including chondrocytes and synoviocytes. Indeed, in RA synovium, T cells are in close contact with monocytes and fibroblasts. Thus it can be expected that factors produced by these cell types act in concert on the resident cells. Indeed, molecules involved in monocyte-synoviocyte interactions leading to enhanced cytokine production have been clarified (25). Here, we show a spontaneous production of biologically active IL-17 by cultures of RA synovium pieces. Results obtained with a blocking anti-IL-17 Ab indicate that IL-17 contributes to ∼50% of the IL-6- or LIF-inducing activity present in these supernatants. As a control, supernatants of cultures of osteoarthritis synovium, which contain a reduced T cell infiltrate, did not contain such IL-17-related activity (data not shown).

The results indicate that the combination of the most potent monocyte-derived cytokine on synoviocytes, namely IL-1, and IL-17 strongly increases their cytokine production. When used alone, both IL-1β and IL-17 induced IL-6 and LIF production by synovial fibroblasts in a dose-dependent manner. However, the effect of IL-17 was always less than that of IL-1.

These results were first obtained with optimal concentrations, which may not represent the situation found in vivo. Accordingly, stimulation was performed using low concentrations of IL-17 (1 ng/ml) and IL-1β (1 pg/ml) in combination. In such cultures, a synergistic effect of IL-1 and IL-17 was observed on the production of IL-6, whereas an additive effect was observed for LIF production. Similarly, whereas neither IL-17 nor TNF-α alone had any effect on the secretion of GM-CSF, the combination of the two cytokines induced its production by synoviocytes (19). In keeping with the in vivo relevance, these results indicate that cytokines produced by monocytes (IL-1 or TNF-α) and T cells (IL-17) can act together on synoviocytes at the low levels that can most probably be achieved in vivo. Such additive or even synergistic effects explain the important reduction of the IL-6- and LIF-inducing activity present in RA synovium supernatants with a blocking anti-IL-17 Ab (Table III).

The mode of action of IL-17 on IL-1 effect remains to be clarified. A differential effect on the induction of cytokine receptors may lead to synergy. Indeed, the combination of IFN-γ and TNF-α was found to induce TNF-Rα expression, leading to increased response to TNF-α (26). Similar findings were obtained for the effect of IL-10 on IL-2R expression by B cells (27). Such additive or synergistic effects could be expected because the respective IL-1 and IL-17 receptors appear to belong to different families (28). A complete characterization of the components of the multichain IL-17R is in progress. Further studies will thus be required to clarify the molecular basis of the interaction between IL-1 and IL-17.

The proinflammatory effects of IL-17 suggest its classification as a Th1 cytokine. This includes its effects on synoviocytes and the maturation of neutrophils. Most of the T cell clones derived from RA synovium have been classified as Th1 (29). Our recent studies suggest that some IFN-γ-producing RA synovium T cell clones also produce biologically active IL-17 (unpublished data). Such findings, combined with the demonstration of IL-17 production by RA synovium, are in line with the classification of IL-17 as a Th1 cytokine. This subset of T cells can then act directly on mesenchymal cells, leading to an increased proinflammatory pattern and mediating matrix destruction (20). It could be expected that such a concept also applies to other Th1-mediated chronic diseases such as multiple sclerosis, psoriasis, and diabetes (17). The exact proportion of IL-17-producing T cells and their function remain to be determined. Conversely, control of such an effect may lead to a therapeutic anti-inflammatory property.

The Th2 cytokines IL-4, IL-13, and IL-10 have been classified as anti-inflammatory on the basis of their inhibitory effect on the production of IL-1, TNF-α, IL-6, and IL-8 by monocytes and also by synovium targets (11, 12, 14, 15, 16, 17). We studied the effects of these cytokines on the production of IL-6 and LIF by RA synoviocytes stimulated by IL-1 and IL-17. Both IL-4 and IL-13 enhanced by a modest 30% the IL-1- and IL-17-induced production of IL-6. In contrast, they inhibited the production of LIF by 60 to 70%. IL-13 was in general less potent than IL-4. The enhancing effect of IL-4 or IL-13 was more potent with low concentrations of IL-17 than with IL-1β.

Surprisingly, the effects of IL-10 on IL-6 and LIF production were limited, in line with previous reports (7). IL-10 had no effect on LIF production, and a modest inhibitory effect of IL-6 production was found in response to IL-17 and/or IL-1. This is in contrast with its potent inhibitory effect on the secretion of proinflammatory cytokines by monocytes (15). Endogenous IL-10 has been shown to be produced by RA synovium acting as an anti-inflammatory cytokine (12). In addition, T cell clones from RA synovium also produced IL-10 (29). Because of its high endogenous production, addition of IL-10 had limited effect on cytokine production by RA synovium (30). On isolated synoviocytes, contrasting destructive and protective effects have been described. On one hand, IL-10 was found to down-regulate collagen I expression by fibroblasts while enhancing the production of collagenase and stromelysin (31). On the other hand, enhancement of tissue inhibitor of metalloproteinases (TIMP) production would limit destruction (32). In vivo studies will clarify the net effect of IL-10 on these targets.

The opposite effects of IL-4 and IL-13 on IL-6 and LIF production further demonstrate the complexity of the cytokine network. Both IL-6 and LIF use the common gp130 receptor chain to activate target cells (33). However, they are differently regulated by IL-4 and IL-13 when combined with IL-17 and IL-1, findings that extend our previous results with IL-1 (7). This effect on synoviocytes contrast with the common inhibitory effect of IL-4 and IL-13 observed with monocytes and whole synovium (7, 14, 16, 17). In addition, it should be underlined that the classification of IL-6 as a proinflammatory cytokine remains a matter of debate. Through the induction of acute phase proteins, IL-6 also acts as a anti-inflammatory cytokine (34). Such an effect appears to be further enhanced by IL-4 and IL-13 when cells are first exposed to in vivo proinflammatory signals such as IL-17 and IL-1 combined at low concentrations. This would further enhance their anti-inflammatory properties at the site of inflammation.

In conclusion, IL-17 appears to be a T cell cytokine produced by the RA synovium, which acts directly on synoviocytes, leading to an enhanced proinflammatory secretion profile when combined with monocyte-derived signals. Accordingly, control of the production and action of IL-17 may represent a therapeutic target for reducing the enhancing effect of monocyte-derived cytokines. Among the available molecules of interest, some Th2 cytokines may control part of the consequences of such interaction.