Abstract
Rheumatoid arthritis (RA) is a chronic inflammatory disease that is mediated, in part, by proinflammatory factors produced by RA synovial tissue (ST) fibroblasts and macrophages, resulting in monocyte migration from the blood to the ST. To characterize the potential role of IL-17 in monocyte migration, RA synovial fibroblasts and macrophages were activated with IL-17 and examined for the expression of monocyte chemokines. The two potentially important monocyte chemoattractants identified were CCL20/MIP-3α and CCL2/MCP-1, which were significantly induced in RA synovial fibroblasts and macrophages. However, in vivo, only CCL2/MCP-1 was detectable following adenovirus IL-17 injection. We found that IL-17 induction of CCL2/MCP-1 was mediated by the PI3K, ERK, and JNK pathways in RA ST fibroblasts and by the PI3K and ERK pathways in macrophages. Further, we show that neutralization of CCL2/MCP-1 significantly reduced IL-17–mediated monocyte recruitment into the peritoneal cavity. We demonstrate that local expression of IL-17 in ankle joints was associated with significantly increased monocyte migration and CCL2/MCP-1 levels. Interestingly, we show that RA synovial fluids immunoneutralized for IL-17 and CCL2/MCP-1 have similar monocyte chemotaxis activity as those immunoneutralized for each factor alone. In short, CCL2/MCP-1 produced from cell types present in the RA joint, as well as in experimental arthritis, may be responsible, in part, for IL-17–induced monocyte migration; hence, these results suggest that CCL2/MCP-1 is a downstream target of IL-17 that may be important in RA.
Rheumatoid arthritis (RA) is a chronic systemic disorder that is characterized by infiltration of inflammatory cells that cause synovial hyperplasia and progressive destruction of cartilage and bone. RA was initially considered a Th1 cell-associated inflammatory joint disease; however, recent evidence from experimental models of arthritis indicates that IL-17–producing T cells play a major role in the initiation and progression of disease (1–4). IL-17 is produced in the T cell-rich areas of RA synovial tissue (ST), and high levels of IL-17 are present in RA synovial fluid compared with osteoarthritis synovial fluid (5, 6). We recently demonstrated that the percentage of Th17 cells is higher in RA synovial fluids compared with normal or RA peripheral blood (7).
Several groups demonstrated that IL-1β, -6, and -23 and TGF-β are essential for driving human Th17 cell differentiation from naive CD4+ and CD45RA+ PBLs (8, 9). Although it was initially suggested that the differentiation of human Th17 cells is independent of TGF-β, recently published data demonstrate that the absence of TGF-β mediates a shift from a Th17 profile to a Th1-like profile (8, 9). IL-6 and -1 play an important role in human Th17 cell polarization and are induced by IL-17 in the RA joint, thereby resulting in a positive feedback loop that may lead to self-perpetuating chronic disease. Interestingly, IL-1β is central to human Th17 differentiation; however TNF-α does not seem to be critical for this process (8, 10). Other investigators showed that TGF-β and IL-21 uniquely promote the polarization of Th17 cells from human naive CD4+ T cells (11).
In RA ST, T cells are in close proximity to RA ST fibroblasts and macrophages; therefore, interaction between the different cell types may occur. Previous studies showed that IL-17 induces RA ST fibroblasts to produce neutrophil and T cell chemokines, including IL-8, CCL20/MIP-α, CXCL1/growth-related oncogene-α, and CXCL2/growth-related oncogene-β (12–15). IL-17 is also capable of activating macrophages to express IL-1, TNF-α, cyclooxygenase-2, PGE2, and matrix metalloproteinase-9 (16–18). However, the role of IL-17 in the induction of monocyte chemokines from macrophages or RA synovial fibroblasts is not known.
To characterize the mechanisms by which IL-17 might contribute to the pathogenesis of RA, we recently demonstrated that IL-17 is directly chemotactic for monocytes (19). In the current study, we examined the hypothesis that IL-17 may also mediate monocyte migration through the induction of monocyte chemokines. For this purpose, we screened IL-17–activated in vitro differentiated macrophages and RA synovial fibroblasts for monocyte chemokines. Although IL-17 did not induce CCL5/RANTES, CCL3/MIP-1α, or CX3CL1/fractalkine in the RA synovial fibroblasts or in macrophages, it was effective in inducing CCL2/MCP-1 and CCL20/MIP-3α from both cell types. Consistent with these observations, microarray studies performed using RA synovial fibroblasts obtained from six patients demonstrated that IL-17 enhances monocyte chemokines, including CCL2/MCP-1 and CCL20/MIP-3α.
In the current study, we also showed that IL-17 induces production of CCL2/MCP-1 but not CCL20 from the cells present in the peritoneal cavity. We further demonstrate that induction of CCL2/MCP-1 by IL-17 is regulated by the activation of PI3K, as well as ERK and/or JNK pathways in the different cell types present in the RA joint, and that this is independent of TNF-α and IL-1β induction. In RA ST fibroblasts, TNF-α and IL-1β induce significantly higher levels of CCL2/MCP-1 compared with IL-17; however, these proinflammatory cytokines have similar ability in induction of CCL2/MCP-1 from macrophages. In an in vivo chemotaxis model, CCL2/MCP-1 plays an important role in IL-17–induced monocyte recruitment, because neutralization of CCL2/MCP-1 significantly reduced IL-17–mediated monocyte chemotaxis into the peritoneal cavity. Further, CCL2/MCP-1 is present in ankles that locally express IL-17, and the immunoneutralization of IL-17 and CCL2/MCP-1 in RA synovial fluids demonstrated similar effects on monocyte chemotaxis as the immunoneutralization of each one alone, indicating that both factors may induce monocyte migration through the same pathway. These results suggest that IL-17–activated monocyte migration in the RA joint may be due, in part, to the induction of CCL2/MCP-1. Hence, therapies directed against IL-17 and its downstream molecules may be beneficial for RA treatment.
Materials and Methods
RA ST fibroblast and macrophage isolation and culture
The studies were approved by the Northwestern University Institutional Review Board, and all donors gave informed written consent. RA and normal ST fibroblasts were isolated from fresh STs by mincing and digesting in a solution of dispase, collagenase, and DNase (20, 21). Cells were used between passages three and nine. RA and normal ST fibroblasts were treated with IL-17 for 0–8 h for mRNA studies, and cell supernatants were harvested after 24 h for protein studies. Monocytes were separated from buffy coats (Lifesource, Northbrook, IL) obtained from healthy donors. Mononuclear cells, isolated by Histopaque (Sigma-Aldrich, St. Louis, MO) gradient centrifugation, were separated by countercurrent centrifugal elutriation. Monocytes were used for chemotaxis or allowed to differentiate to macrophages, as previously described (22, 23). Macrophages were treated for 0–8 h for mRNA studies and 24 h for protein studies.
Real-time RT-PCR
7, 24). Relative gene expression was determined by the ΔΔ cycle threshold method, and the results were expressed as the fold increase above the 0 time point.
IL-17 signaling pathways in RA ST fibroblasts and macrophages
RA ST fibroblasts and macrophages (2 × 106/ml) were treated or not with IL-17 (50 ng/ml) for 0–120 min. Cell lysates were examined by Western blot analysis, as previously described (25). Blots were probed with phospho (p)-ERK, p-JNK, p-p38 MAPK, and p-AKT (Cell Signaling Technology, Beverly, MA; 1:1000 dilution) overnight; after stripping, they were probed with ERK, JNK, p38, and AKT (Cell Signaling Technology; 1:3000 dilution) for 1 h.
Inhibition of the signaling pathways in RA ST fibroblasts and macrophages
To define which signaling pathways mediate IL-17–induced CCL2/MCP-1 secretion, RA synovial fibroblasts and macrophages were incubated with inhibitors to p38 (SB203580; 10 μM), JNK (SP600125; 10 μM), ERK (PD98059; 10 μM) and PI3K (LY294002; 10 μM) for 1 h in 10% RPMI 1640. Cells were subsequently activated with IL-17 (50 ng/ml) for 24 h, and the media were collected to quantify the levels of CCL2/MCP-1 using ELISA.
Silencing RNA oligonucleotide studies in RA ST fibroblasts and macrophages
To confirm that specific chemical inhibitors suppress IL-17–induced CCL2/MCP-1 production, human ERK, JNK (only in RA ST fibroblasts), p38 MAPK, and AKT specific and nonspecific control (nontargeting silencing RNA oligonucleotides [siRNAs] pool) siRNAs (Dharmacon, Lafayette, CO) were transfected into RA ST fibroblasts and macrophages (cultured in six-well plates) by 4 μl/well Lipofectamine 2000 (Invitrogen), following the manufacturer’s protocol with several modifications. Immediately prior to transfection, the media were replaced with 2 ml/well fresh RPMI 1640 (no additives). siRNAs for ERK, JNK, p38, and AKT or control siRNAs were transfected at a final concentration of 100 nM. The transfection reactions were supplemented with 250 μl FBS at 2 h and 2250 μl 10% FBS in RPMI 1640 after 5 h. After 48 h (RA ST fibroblasts) or 72 h (macrophages), the cells were treated with IL-17 for 0–2 h to validate the reduction of each specific signaling pathway. Cells were then probed for ERK, JNK, p38 MAPK, or AKT (1:3000 dilution) and reprobed for tubulin (1:3000 dilution). After 24 h of IL-17 stimulation, CCL2/MCP-1 production was determined by ELISA in RA ST fibroblasts and macrophages transfected with AKT, ERK, JNK (only for RA ST fibroblasts), or p38 MAPK specific and nonspecific control siRNAs.
Ankle homogenization
Ankles were homogenized in a 50-ml conical centrifuge tube containing 1 ml Complete Mini-Protease Inhibitor Mixture (Roche Molecular Biochemicals, Indianapolis, IN) homogenization buffer. Ankle homogenization was completed on ice using a motorized homogenizer, followed by 30 s of sonication. Homogenates were centrifuged at 2000 × g for 10 min, filtered through a 0.45-μm pore-size filter (Millipore, Bedford, MA), and stored at −80°C until use. The concentration of protein in each tissue lysate was determined using a bicinchoninic acid assay (Pierce, Rockford, IL) and BSA as the standard (26–28).
Cytokine quantification
In vivo study protocol
The animal studies were approved by the Northwestern University Institutional Animal Care and Use Committee. The experiments were performed to examine the effect of IL-17 on monocyte migration in vivo. A recombinant human serotype 5 adenovirus (Ad) expressing murine IL-17 (Ad–IL-17) was a kind gift of J.K. Kolls and was constructed as reported previously (29). Six- to 7-wk-old C57BL/6 mice were injected i.p. with 108 PFU murine Ad–IL-17, Ad-CMV control, or PBS (only on day 1). Mice were sacrificed on days 1, 2, 3, or 4 (n = 5) postinjection, and the peritoneal lavage was collected in 8 ml PBS for quantifying the total number of cells, number of monocyte/macrophages, and the expression of murine IL-17 and CCL2/MCP-1. In a different experiment, mice were treated with Ad–IL-17 or Ad-CMV control i.p. The Ad–IL-17 group was treated i.p. with rabbit serum (Sigma-Aldrich) (n = 5) or anti-CCL2/MCP-1 [provided by W.J. Karpus and prepared as reported previously (30)] (n = 5), and the Ad-CMV group was treated with rabbit serum (n = 5) 2 h subsequent to the initial treatment. Mice were sacrificed on days 1 through 3 postinjection, and the peritoneal lavage was collected for quantifying the total number of monocyte/macrophages.
For local expression of IL-17 in mouse ankle joints, 4–6-wk-old C57BL/6 mice were injected intra-articularly with 107 PFU Ad–IL-17 or Ad-CMV control. Ankles were harvested on days 4 and 10 after Ad–IL-17 injection for ELISA and histological studies. Levels of IL-17 were quantified by ELISA from ankles harvested on days 4 and 10 after Ad–IL-17 injection.
Abs and immunohistochemistry
Mouse ankles were decalcified, fixed in formalin, embedded in paraffin, and sectioned in the pathology core facility of Northwestern University. Mouse ankles were stained with immunoperoxidase using Vector Elite ABC Kits (Vector Laboratories, Burlingame, CA), with diaminobenzidine (Vector Laboratories) as a chromogen. Slides were deparaffinized in xylene for 20 min at room temperature, followed by rehydration by transfer through graded alcohols. Ags were unmasked by first incubating slides in boiling citrate buffer for 20 min. Endogenous peroxidase activity was blocked by incubation with 3% H2O2 for 5 min. Nonspecific binding of avidin and biotin was blocked using an avidin/biotin blocking kit (Vector Laboratories). Nonspecific binding of Abs to the tissues was blocked by pretreatment of tissues in DakoCytomation Protein Block (Dako, Glostrup, Denmark) for 30 min. Tissues were incubated with rat anti-mouse F480 (1/250 dilution; AbD Serotec, Raleigh, NC) or a rat IgG control Ab (eBioscience, San Diego, CA). Slides were counterstained with Harris hematoxylin. Each slide was evaluated by a masked observer (A.M.M.) (27, 28, 31, 32). For F480 immunostaining, ST was graded on a scale of 0–100%, in which 0% indicates no staining and 100% indicates that all cells were immunoreactive. Score data were pooled, and the mean ± SEM was calculated in each data group (n = 5).
Flow cytometry
These studies were performed to quantify the number of monocyte/macrophages induced by Ad–IL-17 injection in the peritoneum compared with the controls. Cells in the lavage were collected, washed (0.5% BSA in PBS), and blocked by Fc blocker (CD16/CD32) (BD Bioscience, San Jose, CA). Thereafter, cells were stained with PE-conjugated F480 (Serotec), FITC-labeled CD11b (Serotec), and PE-CY7–conjugated Gr1 (eBioscience) or isotype control Abs (eBioscience) for 30 min. Cells were subsequently washed twice and resuspended in 0.5% BSA in PBS. The total number of leukocytes recruited into the peritoneal lavage in each treatment group was also quantified. Final results are shown as the total number of monocytes/macrophages (% monocytes/100 × total number of cells), which was determined by the fraction of total cells that met the criteria of F480 and CD11b positivity and Gr1 negativity.
Monocyte chemotaxis
Monocyte chemotaxis was performed to examine the role of IL-17 and/or CCL2/MCP-1 in RA synovial fluid-induced monocyte migration. For this purpose, chemotaxis was performed in triplicate for 2 h in Boyden chambers (Neuro Probe, Gaithersburg, MD) using FMLP (100 nM; Sigma-Aldrich) as the positive control and PBS as the negative control (33). Chemotaxis induced by RA synovial fluids was examined following the incubation of fluids (n = 6 fluids) with control IgG, anti–IL-17, anti CCL2/MCP-1, or both Abs (10 μg/ml) for 1 h prior to performing the assay (19).
Statistical analysis
The data were analyzed using the two-tailed Student t test for paired and unpaired samples or the Fisher exact test where appropriate; p values < 0.05 were considered significant.
Results
IL-17 induces CCL20/MIP-3α and CCL2/MCP-1 from RA synovial fibroblasts and macrophages
Using in vitro and in vivo assays, we observed that IL-17 is chemotactic for monocytes (19). Because not all of the in vivo monocyte chemotaxis could be attributed to the direct effect of IL-17, we hypothesized that the indirect effect of IL-17 may be through the induction of monocyte chemokines from adjacent cells. Because synovial fibroblasts and macrophages are important in the pathogenesis of RA, synovial fibroblasts and macrophages differentiated in vitro from normal monocytes were used to determine which monocyte chemokines might be induced by IL-17. The mRNA expression for CCL5/RANTES, CCL3/MIP-1α, CX3CL1/fractalkine, CCL20/MIP-3α, and CCL2/MCP-1 was determined in IL-17–treated macrophages and RA synovial fibroblasts. Although IL-17 did not induce CCL5/RANTES, CCL3/MIP-1α, or CX3CL1/fractalkine in RA synovial fibroblasts or normal macrophages, CCL20/MIP-3α (Fig. 1A, 1B) and CCL2/MCP-1 (Fig. 1C, 1D) were induced in both cell types.
CCL20/MIP-3α mRNA is induced in IL-17–activated RA synovial fibroblasts and macrophages. RA synovial fibroblasts (A, C) and normal macrophages (B, D) were activated with IL-17 (50 ng/ml) for 0–8 h. Real-time RT-PCR was used to identify CCL20/MIP-3α (A, B) and CCL2/MCP-1 (C, D), which was normalized to GAPDH. The results are presented as fold increase compared with the 0 time point (untreated cells) and represent mean ± SE. *p < 0.05 (n = 3–5).
TNF-α and IL-1β are more potent in inducing CCL2/MCP-1 expression in RA fibroblasts compared with IL-17, and IL-17–mediated CCL2/MCP-1 expression is independent of TNF-α and IL-1β production
To compare the potency of IL-17 with TNF-α and IL-1β in the induction of CCL2/MCP-1 expression, RA fibroblasts and macrophages were activated with each of the cytokines, and the expression of CCL2/MCP-1 was examined by real-time RT-PCR. Results from these experiments demonstrated that although TNF-α– and IL-1β–induced CCL2/MCP-1 was significantly higher in RA fibroblasts than that induced by IL-17, similar levels of CCL2/MCP-1 were expressed in macrophages that were activated by all three cytokines (Fig. 2A, 2B). Next, we asked whether IL-17–induced TNF-α or IL-1β expression had any effect on CCL2/MCP-1 transcription in RA ST fibroblasts or normal macrophages. Studies using neutralizing Abs to TNF-α and IL-1β demonstrated similar levels of IL-17–induced CCL2/MCP-1 as the IgG control, suggesting that IL-17–mediated CCL2/MCP-1 is independent of TNF-α and IL-1β induction in both cell types (Fig. 2C, 2D).
CCL2/MCP-1 mRNA induced in IL-17–activated RA synovial fibroblasts and macrophages is independent of TNF-α and IL-1β production. RA synovial fibroblasts (A) and normal macrophages (B) were activated with IL-17 (50 ng/ml), TNF-α (10 ng/ml), or IL-1β (10 ng/ml) for 4 or 6 h, and CCL2/MCP-1 expression was determined by real-time RT-PCR (n = 5–8). RA synovial fibroblasts (C) or normal macrophages (D) were treated or not with IgG, anti–TNF-α (10 μg/ml), or anti–IL-1β (10 μg/ml) prior to being treated with IL-17 (50 ng/ml) for 4 or 6 h. CCL2/MCP-1 mRNA was determined by real-time RT-PCR and normalized to GAPDH. The results are presented as fold increase compared with 0 time point (untreated cells) and represent the mean ± SE. *p < 0.05 (n = 3).
To determine which of these chemokines might be relevant in vivo, IL-17 was expressed in the peritoneal spaces of mice (Fig. 3A), and the fluids generated were examined by ELISA for the presence of the monocyte chemokines CCL2/MCP-1 (Fig. 3B), CCL5/RANTES, CCL3/MIP-1α, or CCL20 (negative data not shown). CCL2/MCP-1 was the only monocyte chemokine de-tected in the peritoneal lavage following the injection of Ad–IL-17. Because CCL2/MCP-1 was present in vivo following the expression of IL-17 and is a more potent monocyte chemoattractant compared with CCL20/MIP-3α (33), this study focused on the role of CCL2/MCP-1 in IL-17 monocyte chemotaxis, as well as the mechanism by which IL-17 activates CCL2/MCP-1 production in cells present in the RA joint.
Ad–IL-17 induces CCL2/MCP-1 production from the cells present in the peritoneal cavity. Ad–IL-17 and/or CMV control (108 PFU) was injected into the peritoneum of mice. Peritoneal lavage was collected to quantify the levels of IL-17 (A) or CCL2/MCP-1 (B) by ELISA. The values represent the mean ± SE. *p < 0.05; **p < 0.01; ***p < 0.005 (n = 5 mice per each time point and treatment group).
IL-17 induces CCL2/MCP-1 production by RA ST fibroblasts and macrophages
Based on these results, we asked whether IL-17 induced the secretion of CCL2/MCP-1 in synovial fibroblasts. IL-17 enhanced the secretion of CCL2/MCP-1 from normal and RA synovial fibroblasts. The level of IL-17–induced CCL2/MCP-1 was significantly higher in RA ST fibroblasts at baseline and following IL-17 treatment compared with normal synovial fibroblasts (Fig. 4A).
IL-17 induces CCL2/MCP-1 production in RA ST fibroblasts and macrophages. Normal (NL) and RA ST fibroblasts (A) or macrophages (B) were treated with PBS or IL-17 for 24 h, and CCL2/MCP-1 levels were determined by ELISA using the conditioned media. Values are mean ± SE. *p < 0.05 (n = 3–5).
Next, we examined the ability of IL-17 to induce CCL2/MCP-1 in macrophages. IL-17 induced the secretion of CCL2/MCP-1 2–3-fold (p < 0.05) in in vitro–differentiated macrophages (Fig. 4B), similar to the findings observed with RA synovial fibroblasts. These results demonstrate that IL-17 mediates CCL2/MCP-1 production from synovial fibroblasts and macrophages, cell types that are important in the pathogenesis of RA.
IL-17 induction of CCL2/MCP-1 is mediated by the PI3K, ERK, and JNK pathways
To identify the signaling pathways mediating IL-17–induced CCL2/MCP-1, IL-17–activated pathways were examined in RA ST fibroblasts by measuring the phosphorylated form of each kinase. The IL-17–mediated activation of the AKT pathway (15 min) occurs prior to that of ERK and JNK (30 min) or p38 (60 min) (Fig. 5A–D). Although chemical inhibitors to ERK, JNK, and PI3K suppressed IL-17–induced CCL2/MCP-1 secretion by 50–90% (Fig. 5E; p < 0.05), inhibition of p38 did not reduce the levels of CCL2/MCP-1 in RA synovial fibroblasts. Consistent with these observations, the siRNA-mediated forced reduction of AKT-1, ERK, and JNK, but not p38, suppressed (p < 0.05) IL-17–induced CCL2/MCP-1 production in RA synovial fibroblasts (Fig. 5F–J).
In RA ST fibroblasts, IL-17–induced CCL2/MCP-1 production is modulated by the PI3K, ERK, and JNK pathways. To determine the mechanism of IL-17 activation in RA ST fibroblasts, cells were stimulated with IL-17 (50 ng/ml) for 0–120 min, and the cell lysates were probed for p-ERK (A), p-JNK (B), p-p38 (C), or p-AKT (D). To examine which of the signaling pathways were associated with IL-17–induced CCL2/MCP-1 production, RA ST fibroblasts were treated or not with inhibitors to ERK (PD98059; 10 μM), JNK (SP600125; 10 μM), p38 (SB203580; 10 μM), or PI3K (LY294002; 10 μM). E, Cells were subsequently activated with IL-17 (50 ng/ml) for 24 h, and conditioned media were collected to quantify the levels of CCL2/MCP-1 using ELISA. To determine that siRNAs for each signaling pathway could effectively reduce the IL-17 signaling, siRNAs for ERK, JNK, p38, AKT, or control siRNAs were transfected. After 48 h, the cells were treated with IL-17 for 0–2 h. Cells were then probed for ERK (F), JNK (G), p38 MAPK (H), or AKT (I) (1:3000 dilution) and reprobed for tubulin (1:3000 dilution). J, IL-17–induced CCL2/MCP-1 production was determined by ELISA in RA ST fibroblasts transfected with siRNAs against ERK, JNK, p38 MAPK, AKT, or specific or nonspecific control siRNAs. These results are representative of three experiments. *p < 0.05.
Similar studies were performed with macrophages to determine which IL-17–activated signaling pathways were responsible for CCL2/MCP-1 production. Unlike RA synovial fibroblasts, IL-17 activated p38 and PI3K pathways in macrophages as early as 5 min, whereas ERK was phosphorylated later, at 60 min (Fig. 6A–C). Using chemical inhibitors, suppression of the PI3K and ERK pathways significantly (p < 0.05) reduced IL-17–induced CCL2/MCP-1 production in macrophages. Consistent with these observations, the siRNA-mediated reduction of AKT-1 and ERK suppressed IL-17 activation by 70% and 50%, respectively (p < 0.05), whereas the reduction of p38 had no effect (Fig. 6E–H). Collectively, our results suggest that although PI3K and ERK pathways modulate IL-17–induced CCL2/MCP-1 production in macrophages and RA synovial fibroblasts, activation of JNK by IL-17 plays a role in CCL2/MCP-1 secretion in RA synovial fibroblasts.
PI3K and ERK activation regulates IL-17–induced CCL2/MCP-1 production in macrophages. To determine the mechanism of IL-17 in macrophages, cells were stimulated with IL-17 (50 ng/ml) for 0–120 min, and the cell lysates were probed for p-ERK (A), p-p38 (B), or p-AKT (C). To examine which of the signaling pathways were associated with IL-17–induced CCL2/MCP-1 production, macrophages were treated or not with inhibitors to ERK (PD98059; 10 μM), p38 (SB203580; 10 μM), or PI3K (LY294002; 10 μM). D, Cells were subsequently activated with IL-17 (50 ng/ml) for 24 h, and the conditioned media were collected to quantify the levels of CCL2/MCP-1 using ELISA. To determine that siRNAs for each signaling pathway could effectively reduce the IL-17 signaling, siRNAs for ERK, p38, or AKT or control siRNAs were transfected into macrophages. Cells were harvested after 72 h and then probed for ERK (E), p38 MAPK (F), or AKT (G) (1:3000 dilution) and reprobed for tubulin (1:3000 dilution). H, IL-17–induced CCL2/MCP-1 production was determined by ELISA in macrophages transfected with siRNAs against ERK, p38 MAPK, or AKT or specific or nonspecific control siRNAs. These results are representative of three experiments. *p < 0.05.
CCL2/MCP-1 production contributes to IL-17–mediated monocyte migration into the peritoneal cavity
We chose to examine the role of CCL2/MCP-1 in IL-17–induced chemotaxis in an in vivo chemotaxis model in the peritoneal cavity rather than the ankle joints because of the ability to use flow cytometry, which is a more quantitative technique compared with histology. For this purpose, mice were injected i.p. with the Ad–IL-17 or the Ad-CMV control vector. Injection of Ad–IL-17 significantly increased (p < 0.005) the total number of monocytes recruited into the peritoneum on days 1 through 4 after injection of Ad–IL-17 compared with the Ad-CMV and PBS controls (Fig. 7A). Of interest, the total number of monocytes migrating to the peritoneum was significantly greater on day 1 after Ad–IL-17 injection (p < 0.005) compared with days 2–4 (Fig. 7A). This observation was consistent with higher levels of peritoneal IL-17 expression on day 1 (5–6-fold increase; p < 0.001) compared with days 2–4 (Fig. 3A), suggesting that the pattern of monocyte migration is similar to expression levels of IL-17 in the peritoneal cavity.
Neutralization of CCL2/MCP-1 reduces IL-17–mediated peritoneal monocyte recruitment. PBS, Ad–IL-17, or CMV control (108 PFU) was injected into the peritoneum of mice. A, Peritoneal lavage was collected to quantify the total number of monocytes recruited on days 1, 2, 3, and 4 postinjection. B, Next, the Ad–IL-17 group was treated i.p. with rabbit serum or anti-CCL2/MCP-1, and the CMV group was treated with rabbit serum for 2h subsequent to the initial treatment. Mice were sacrificed on days 1, 2, or 3 post injection and the peritoneal lavage was collected for quantifying the total number monocytes/macrophages. The values represent the mean ± SE. *p < 0.05; **p < 0.01; *** p < 0.001 (n = 5 mice per each time point and treatment group).
Experiments were performed to determine the role of CCL2/MCP-1 in IL-17–mediated monocyte migration in vivo. I.p. injection of neutralizing Ab to CCL2/MCP-1 significantly reduced IL-17–mediated monocyte migration (p < 0.05) compared with control injection (Fig. 7B). These observations suggest that IL-17–mediated monocyte migration was mediated, in part, through the induction of CCL2/MCP-1.
Local expression of IL-17 in mouse ankles upregulates monocyte migration and CCL2/MCP-1 levels
Local expression of IL-17 using an adenoviral vector (107 PFU) resulted in increased inflammation, synovial lining thickness, and bone erosion in the ankles of C57/BL6 mice compared with Ad-CMV–infected controls (107 PFU) (data not shown). The group treated with Ad–IL-17 demonstrated significantly greater ankle circumference (data not shown) on days 4 and 10 postinjection compared with the control group. Further, the concentration of IL-17 in the ankles of the IL-17–induced arthritis model was 1200 pg/mg on day 4 and 400 pg/mg on day 10 post Ad injection. F480 staining of ankles harvested on day 10 postinjection showed that mice treated with Ad–IL-17 had significantly greater macrophage staining compared with the control group (Fig. 8A–C). Additionally, CCL2/MCP-1 levels were markedly increased in the Ad–IL-17 group on days 4 and 10 postinjection compared with the control treatment group (Fig. 8D). These results suggest that IL-17 may be important for monocyte migration directly and indirectly through induction of CCL2/MCP-1.
Local expression of IL-17 increases monocyte migration and CCL2/MCP-1 expression in mouse ankles. Ad–IL-17 or Ad-CMV control was injected intra-articularly into the ankle joints of 4–6-wk-old C57BL/6 mice. Ankles at day 10 post-Ad injection were harvested, embedded in paraffin, and decalcified. Macrophages in the ankle were identified using F480 staining. Control ankles (A) had significantly lower macrophage staining compared with ankles locally expressing IL-17 (B) (original magnification ×200). C, Quantification of F480 staining in the Ad-CMV control and Ad–IL-17 groups. D, CCL2/MCP-1 levels were quantified from ankle homogenates harvested from days 4 and 10 post Ad–IL-17 or Ad-CMV control treatment and normalized to protein concentration. Values express mean ± SE (n = 5). *p < 0.05; **p < 0.01; ***p < 0.005.
Neutralization of IL-17 does not enhance the ability of anti–CCL2/MCP-1 to reduce RA synovial fluid-mediated monocyte migration
Experiments were performed to determine whether anti–IL-17 would synergize with anti–CCL2/MCP-1 in reducing RA synovial fluid-mediated monocyte chemotaxis. The concentrations of IL-17 and CCL2/MCP-1 in the six RA synovial fluids used for monocyte chemotaxis were 149 ± 27 pg/ml and 1268 ± 314 pg/ml, respectively, and our data from these experiments showed that immunodepletion of RA synovial fluid for IL-17 and CCL2/MCP-1 had similar effects on monocyte chemotaxis (Fig. 9) as did the neutralization of each factor alone, suggesting that both of these proinflammatory mediators induce monocyte migration through the same signaling pathway.
IL-17 and/or CCL2/MCP-1–immunodepleted RA synovial fluid had similar effects on monocyte chemotaxis. RA synovial fluids from six patients (1:20 dilution) were incubated with Abs to IL-17 (10 μg/ml), CCL2/MCP-1 (10 μg/ml), or both, as well as isotype control or PBS or FMLP, for 1 h prior to evaluating monocyte chemotaxis in response to RA synovial fluids. The values represent the mean ± SE. *p < 0.05.
Discussion
Recently, we showed that IL-17 mediates monocyte migration in vivo into sponges implanted s.c. in SCID mice and that it is capable of directly mediating monocyte migration in vitro and in RA (19). We hypothesized that direct effects of IL-17 might not account for all of its chemotactic ability in vivo. In this study, we demonstrated that IL-17 induces CCL2/MCP-1 production from RA ST fibroblasts and control macrophages independently of TNF-α and IL-1β, supporting an indirect role of IL-17 in monocyte migration. In macrophages, all three cytokines induced CCL2/MCP-1 to a similar extent; however, TNF-α– and IL-1β–induced CCL2/MCP-1 was significantly greater than that induced by IL-17 using RA fibroblasts. In fibroblasts, activation of the PI3K, JNK, and ERK pathways is responsible for IL-17–induced CCL2/MCP-1 production, whereas in IL-17–activated macrophages, PI3K and ERK regulate CCL2/MCP-1 production. Additionally, in this study we demonstrated that CCL2/MCP-1 production contributes to IL-17–mediated monocyte migration into the peritoneal cavity, suggesting that CCL2/MCP-1 plays an important role in IL-17–mediated monocyte migration in vivo. We also showed that local expression of IL-17 in mouse ankle joints increases monocyte trafficking and levels of CCL2/MCP-1. Finally, we demonstrated that Abs against IL-17 and CCL2/MCP-1 reduce RA synovial fluid-mediated monocyte migration to similar levels as does each Ab alone.
IL-17 plays a crucial role in regulating neutrophil recruitment. Intra-articular injection of IL-17 enhances neutrophil migration into the joints of mice (3). In the rat airway, IL-17 mediates neutrophil recruitment via the induction of IL-8 (CXCL1) (34). Neutrophil chemotaxis caused by conditioned media from IL-17–stimulated gastric epithelial cells was inhibited by neutralizing Ab to IL-8 but not to IL-17 (35). Consistent with these findings, we found that neutrophil migration was increased in the peritoneal cavity following Ad–IL-17 injection on days 1 and 2 (data not shown).
In our in vivo chemotaxis model, Ad–IL-17 resulted in the upregulation of CCL2/MCP-1 on day 1 postinjection. In contrast, other potential monocyte chemoattractants, including CCL20/MIP-3α, IL-6, TNF-α, CCL3/MIP-1α, and CCL5/RANTES, were undetectable in the peritoneal lavage, revealing a key role of CCL2/MCP-1 in IL-17–mediated monocyte chemotaxis in vivo. Although CCL20/MIP-3α was undetectable in the peritoneal cavity following the local expression of IL-17, it is possible that this chemokine has an important role in monocyte migration to RA joints because CCL20 is induced by IL-17 in macrophages and RA fibroblasts.
Our data characterize the pathways mediating IL-17–induced CCL2/MCP-1 expression. In macrophages and RA synovial fibroblasts, IL-17 activates the ERK, p38, and AKT pathways, as well as the JNK pathway in RA fibroblasts only. Using chemical inhibitors or siRNAs, suppression of ERK or PI3K in both cell types, and JNK in fibroblasts only, was effective in reducing IL-17–mediated CCL2/MCP-1. Previous studies showed that PI3K is the major pathway involved in the IL-17 induction of proinflammatory mediators, such as IL-6 and -8 and CXCL12/SDF-1, in RA synovial fibroblasts; however, the activation of p38 was not implicated in the process (36, 37). Similarly, in RA fibroblasts, other proinflammatory cytokines, such as IL-18, modulate CCL2/MCP-1 production through the activation of PI3K and JNK but not p38 MAPK (38). Consistent with our observations in fibroblasts and macrophages, activation of ERK but not p38 MAPK was essential for Mycobacterium tuberculosis H37Rv-induced CCL2/MCP-1 secretion by monocytes (39). Collectively, the data suggest that although PI3K, ERK, and JNK contribute to IL-17–mediated CCL2/MCP-1 production, p38 does not play a role.
In the current study, we found that the total number of monocytes recruited into the peritoneal cavity was significantly increased by IL-17. However, the total number of monocytes recruited was significantly greater on day 1 compared with days 2–4; this may be due to higher levels of IL-17 and CCL2/MCP-1 in the peritoneal cavity on day 1. It was shown that i.p. injection of TLR4 ligand, LPS, or thioglycolate can also induce CCL2/MCP-1 production as early as 15–24 h (40, 41). Our data also suggest that CCL2/MCP-1 plays an important role in IL-17–mediated monocyte migration, because inhibition of CCL2/MCP-1 suppresses this effect. However, no other potential monocyte chemoattractants were detectable in the peritoneal lavage following infection with Ad–IL-17, suggesting that the effects of IL-17 on monocyte chemotaxis may also be due to a direct effect on monocytes, as recently described (19). Monocytes in the circulation may be attracted to the higher gradient of IL-17 present in the peritoneal cavity. Monocytes use LFA1 (αLβ2) and VLA4 (α4β1), which bind to intracellular adhesion molecule-1 and VCAM-1 on the endothelial cells, for adhesion and migration in vivo (42–44). We have shown that IL-17 can induce VCAM-1 from human microvascular endothelial cells (S.R. Pickens and S. Shahrara, unpublished data), and other investigators demonstrated that IL-17 activated the production of intracellular adhesion molecule-1 (45). Therefore, these adhesion molecules may be important for facilitating IL-17–induced monocyte migration in vivo.
Consistent with the data from the peritoneal cavity, forced expression of IL-17 in the mouse ankle joints upregulated CCL2/MCP-1 concentration; this indicates that CCL2/MCP-1 is an important downstream target of IL-17 because it was produced in vitro in cell types present in RA synovium, as well as in vivo in the mouse peritoneal cavity and ankle joints. Interestingly, the expression patterns of CCL2/MCP-1 in the peritoneal cavity and the joints were different, which may be due to local differences in cell types and the area for distribution of IL-17. Specifically, the concentration of CCL2/MCP-1 became undetectable in the peritoneal cavity on days 2–4, whereas it remained elevated on days 4 and 10 in the IL-17–induced arthritis model. In the IL-17–induced arthritis model, CCL2/MCP-1 may be produced by synovial fibroblasts and macrophages, allowing for more prolonged expression. However, the structure of the peritoneal cavity is quite different and lacks synovial fibroblasts. The second reason is that the movement of the cells and the area for diffusion of IL-17 are much greater in the mouse peritoneal cavity compared with mouse ankles, as shown by our data. IL-17 levels decreased from 2000 pg/ml (day 1) to 400 pg/ml (day 2) in a day; however, in the Ad–IL-17–induced arthritis model, the level of IL-17 decreased from 1200 pg/mg (day 4) to 400 pg/mg (day 10) over a 6-d period.
The Ad–IL-17–induced arthritis model was associated with increased joint macrophage levels, which may be due to the expression of IL-17 and/or CCL2/MCP-1 in the mouse ankles. Experiments performed to determine the importance of IL-17 and CCL2/MCP-1 in RA synovial fluid-mediated monocyte trafficking revealed that both factors were equally important, and there was no synergistic effect when both proinflammatory mediators were neutralized. Interestingly, the inhibition of multiple factors in RA synovial fluid does not have an additive or synergistic effect on suppressing monocyte chemotaxis; this may be due to the fact that several of the chemotactic mediators use the same pathways and inhibition of one can disturb the equilibrium or that inhibition of monocyte migration cannot be detected because it is within the bell curve. Similar to our previously reported results with IL-17 (19), CCL2/MCP-1–induced monocyte migration is through activation of the p38 pathway (46). A recent paper demonstrated that chemokines competing for similar receptors or using similar signaling pathways do not synergize in monocyte chemotaxis (47). This lack of synergy is totally consistent with other reports (48) that showed that neutralization of several factors in RA synovial fluid does not reduce RA synovial fluid-mediated migration beyond the effect noted with each factor alone.
In summary, the observations presented in this article support the role of CCL2/MCP-1 in promoting IL-17–mediated chemotaxis. Together with our recent observation that IL-17 is also directly chemotactic for monocytes (19), these data support monocyte migration as an important mechanism by which IL-17 contributes to chronic inflammation, as observed in patients with RA. Given the importance of monocyte-derived macrophages in chronic inflammation (49), these data further support the role of IL-17 as a potential therapeutic target in RA.
Acknowledgments
Disclosures The authors have no financial conflicts of interest.
Footnotes
This work was supported in part by awards from the National Institutes of Health (AR056099, AR055240, AR048269, and NS34510), Arthritis National Research Foundation, and grants from Within Our Reach from The American College of Rheumatology.
Abbreviations used in this paper:
- Ad
- adenovirus
- Ad–IL-17
- adenovirus expressing IL-17
- NL
- normal
- p
- phospho
- RA
- rheumatoid arthritis
- siRNA
- silencing RNA oligonucleotide
- ST
- synovial tissue.
- Received June 19, 2009.
- Accepted February 11, 2010.
- Copyright © 2010 by The American Association of Immunologists, Inc.