Hospital Universitario de Canarias, Universidad de La Laguna, Laguna, Spain, and Instituto Canario de Investigación del Cáncer, Instituto de Bio-Orgánica “Antonio González,” Universidad de La Laguna, Laguna, Spain
To explore the potential involvement of the chemokine system in synoviocyte-mediated tissue destruction in rheumatoid arthritis (RA), we studied the expression profile of chemokine receptors and their function in the migration, proliferation, and matrix metalloproteinase (MMP) production of cultured fibroblast-like synoviocytes (FLS) from RA patients.
The presence of CC and CXC chemokine receptors on cultured FLS was studied at the messenger RNA (mRNA) level by reverse transcriptase–polymerase chain reaction and at the cell surface expression level by flow cytometry. Variations in cytosolic calcium influx induced by chemokine stimulation were assessed by flow cytometry on Fura Red–preloaded FLS. Two-compartment transwell chambers were used for FLS chemotaxis assays. Cell growth was measured by a fluorescence-based proliferation assay. Gelatinase and collagenase activities were determined by a fibril degradation assay and zymography.
FLS constitutively expressed the receptors CCR2, CCR5, CXCR3, and CXCR4, both at the cell surface and mRNA levels, but failed to express CCR3 and CCR6. Significant intracytosolic calcium influx was observed on FLS challenged with monocyte chemotactic protein 1 (MCP-1), stromal cell–derived factor 1α (SDF-1α), and interferon-inducible protein 10 (IP-10). Stimulation with MCP-1, SDF-1α, IP-10, and monokine induced by interferon-γ enhanced the migration and proliferation of FLS. These chemokines, in addition to RANTES, increased in a dose- and time-dependent manner the gelatinase and collagenase activities in cell-free supernatants of cultured FLS. Interestingly, the chemokine-mediated up-regulation of MMP activities was significantly abrogated by the presence of anti–interleukin-1β, but not anti–tumor necrosis factor α, blocking antibodies.
These data suggest that through modulation of the migration, proliferation, and MMP production by FLS, the chemokine system may play a more direct role in the destructive phase of RA than is currently suspected, and thus emphasize the relevance of chemokines and their receptors as potential therapeutic targets in this disease.
Rheumatoid arthritis (RA) is a chronic disease that causes synovial inflammation and progressive destruction of the cartilage and bone of diarthrodial joints. Although the pathogenesis of RA is complex and not well understood, the basic mechanisms of the disease process are widely accepted. They include proliferation of synovial and endothelial cells, recruitment and activation of circulating proinflammatory cells, and secretion of cytokines and chemokines from macrophages and fibroblast-like synoviocytes (FLS) (1). All these changes result in hyperplasia of the synovial membrane, which is transformed into a granulated tissue (pannus) that invades and destroys the adjoining cartilage by the action of locally produced matrix metalloproteinases (MMPs), which are zinc-dependent matrix-degrading enzymes. In the current concept of RA pathogenesis, FLS play an essential role as effector cells in joint destruction through the production of MMPs, mainly collagenases and gelatinases (2, 3).
The chemokine system in humans comprises >50 small heparin-binding proteins and 20 transmembrane G protein–coupled receptors that were originally identified by their chemotactic activity on bone marrow–derived cells (4, 5). According to the position and spacing of N-terminal cysteine residues, chemokines can be divided into C, CC, CXC, and CX3C families, and consequently, their receptors have been designated as CR, CCR, CXCR, and CX3CR (6). As stated above, chemokines have been associated with the regulation of leukocyte trafficking to normal and inflamed tissues (4, 5). However, in addition to leukocytes, other nonhematopoietic cell types, including endothelial cells, smooth muscle cells, neurons, epithelial cells, fibroblasts, and several tumor lineage cells, also express chemokine receptors. This unrestricted cell type expression, added to the fact that chemokines can couple to distinct signaling pathways (7), suggests that the functions of chemokine receptors are not limited to cell traffic regulation. In this regard, gene transcription and cell proliferation, among others, are described as effects mediated by the chemokine system (7).
Synovial tissue and synovial fluid from RA patients contain increased concentrations of several chemokines, including interferon-inducible protein 10 (IP-10)/CXCL10, monokine induced by interferon-γ (Mig)/CXCL9, monocyte chemotactic protein 1 (MCP-1)/CCL2, macrophage inflammatory protein 1α (MIP-1α)/CCL3, interleukin-8 (IL-8)/CXCL8, and stromal cell–derived factor 1 (SDF-1)/CXCL12 (8, 9). These chemokines are produced by resident synoviocytes, as well as by infiltrating cells (10–13). It also has been documented that infiltrating cells in the rheumatoid synovium express CXCR3, CCR5, CCR3, CCR2, and CXCR2 (for review, see ref. 13). Based on these data, the chemokine system is considered to be implicated in RA pathogenesis via the recruitment and retention of monocytes and T lymphocytes into the joints (8, 14). However, although the chemokine system has been shown to be involved in many functions in several cell types, the actual contribution of the chemokine system to the mechanisms involved in cartilage and bone destruction beyond the regulation of cell trafficking remains to be established in RA.
In this study, we investigated this by evaluating the effect of activation of the CC and CXC chemokine systems on the migration, proliferation, and MMP production by FLS from RA patients. Our results show that FLS bear functionally active CXCR3, CCR2, CXCR4, and CCR5. Hence, MCP-1, IP-10, SDF-1α, RANTES, and Mig induce the migration and proliferation of FLS and increase the collagenase and gelatinase activities of this cell type. These findings suggest that the chemokine system may play a more direct role in the destructive phase of RA than is currently suspected, providing additional evidence for the potential benefit of chemokine targeting in RA.
MATERIALS AND METHODS
Samples of synovial tissue were obtained from 8 patients with RA, defined according to the criteria of the American College of Rheumatology (formerly, the American Rheumatism Association) (15). These patients were undergoing therapeutic joint surgery on the knee or hip.
Antibodies and recombinant products.
Monoclonal antibodies (mAb) to CCR2 (clones MCP-1R02 and MCP-1R05) (16) were gifts from Dr. Carlos Martínez (Centro Nacional de Biotecnología, Madrid, Spain). Monoclonal antibodies to CCR3 (clone 61828), CCR5 (clone 45531), CCR6 (clone 53103), and CXCR3 (clone 49801) were purchased from R&D Systems (Minneapolis, MN). Anti-CXCR4 mAb (clone 12G5) was purchased from PharMingen (San Diego, CA). Recombinant MCP-1 (CCL2), RANTES (CCL5), SDF-1α (CXCL12), MIP-1α (CCL3), IP-10 (CXCL10), and Mig (CXCL9) were purchased from R&D Systems. Functional blocking mAb against SDF-1α (clone 79014), MCP-1 (clone 24822), IP-10 (clone 33036), CCR5 (clone 45531), and CXCR4 (clone 44716) were purchased from R&D Systems. Recombinant human tumor necrosis factor α (TNFα) was obtained from BioSource International (Camarillo, CA), and recombinant human IL-1β and interferon-γ (IFNγ) were purchased from Sigma (St. Louis, MO). Recombinant human granulocyte–macrophage colony-stimulating factor (GM-CSF) was obtained from PeproTech (London, UK). Neutralizing monoclonal anti-human IL-1β (clone I3642) and anti-human TNFα (clone T6817) were both purchased from Sigma.
Cell isolation and culture.
Human FLS were isolated by enzymatic dispersion of synovial tissues obtained from RA patients. Briefly, aseptically collected tissue samples were washed several times with Dulbecco's modified Eagle's medium (DMEM) containing 4.5 gm/liter glucose (Life Technologies, Gaithersburg, MD) supplemented with 10 IU/ml penicillin, 0.1 μg/ml streptomycin, and were then dissected free of capsular tissues. With continuous stirring, the minced synovial fragments were incubated in serum-free DMEM with 0.1% bacterial collagenase (Roche Diagnostics, Mannheim, Germany) for 2 hours at 37°C. After the enzymatic digestion, synovial cells were pelleted by centrifugation, washed twice in phosphate buffered saline (PBS), plated on T75 tissue-culture plates (Corning, Corning, NY) in DMEM supplemented with 10% fetal calf serum (FCS) (BioWhittaker, Verviers, Belgium), penicillin (50 IU/ml), streptomycin (50 μg/ml), and L-glutamine (2 mM) (all from Biochrom, Berlin, Germany) (hereinafter referred to as cDMEM), and maintained in a humidified atmosphere of 5% CO2 at 37°C. After overnight incubation, nonadherent cells were removed and adherent cells were cultivated in cDMEM under the conditions described above. Upon reaching confluence, they were detached with trypsin/EDTA (Biochrom), split at a 1:3 ratio, and recultured in cDMEM under the same conditions. FLS at passages 3–6, when they were a homogeneous population (<1% CD11b and <1% Fcγ receptor II positive) (17), were used for the experiments.
Jurkat cells were obtained from American Type Culture Collection (Rockville, MD) and maintained in RPMI 1640 (Biochrom) supplemented with 10% FCS and 2 × 10−5M 2-mercaptoethanol (Sigma) at 37°C with 5% CO2.
Flow cytometry analysis.
RA FLS were incubated under adherent conditions in medium alone or in the presence of recombinant human TNFα (20 ng/ml), IL-1β (5 ng/ml), or IFNγ (100 units/ml) for 24 hours at 37°C, when indicated. Then, cells were detached with a nonenzymatic cell dissociation solution (Sigma), washed once with PBS, and incubated with the different mAb at 4°C for 30 minutes. After washing in PBS, cells were labeled with fluorescein isothiocyanate–labeled goat anti-mouse Ig (Dako, Glostrup, Denmark) at 4°C for 30 minutes. At least 5 × 104 cells from each sample were analyzed in a FACScan flow cytometer (BD Biosciences, San Jose, CA), and the data were collected (logarithmic scale) and expressed as the percentage of positive cells. The fluorescence gain was adapted so that <3% of the cells that stained with the negative control mAb showed fluorescence >10 on the log scale. The fluorescence produced by isotype-matched control mAb was considered background. When indicated, cell viability was determined in the FACScan by the spontaneous uptake of propidium iodide (10 μg/ml final concentration) (Sigma).
Using RNAwiz (Ambion, Austin, TX), total RNA was extracted from RA FLS that had been incubated overnight at 37°C in medium alone or in the presence of recombinant human TNFα (20 ng/ml) or IL-1β (5 ng/ml). Complementary DNA was obtained from 4 μg of total RNA by reverse transcription (Applied Biosystems, Warrington, UK) and amplified with primers as follows: for CXCR4, sense 5′-AGAACCAGCGGTTACCATGGA-3′ and antisense 5′-GAGTGTGACAGCTTGGAGATG-3′ (702 bp); for CCR5, sense 5′-CGCATCAAGTGTCAAGTCCAATC-3′ and antisense 5′-TGTAAACTGAGCTTGCTCGCT-3′ (1,014 bp); for CXCR3, sense 5′-GGAGCTGCTCAGAGTAAATCAC-3′ and antisense 5′-GCACGAGTCACTCTCGTTTTC-3′ (250 bp); for CCR2B, sense 5′-ATGCTGTCCACATCTCGTTCTCG-3′ and antisense 5′-TTATAAACCAGCCGAGACTTCCTGC-3′ (1,083 bp); and for CCR6, sense 5′-AGTGGATCCGAGGTCAGGCAGTTCTCCAG-3′ and antisense 5′-GTAGAATTCGCTGCCTTGGGTGTTGTATT-3′ (486 bp). The amplification was conducted for 35 cycles with an annealing temperature of 58°C for 40 seconds for CXCR4, 35 cycles with an annealing temperature of 60°C for 40 seconds for CCR5, 40 cycles with an annealing temperature of 60°C for 60 seconds for CXCR3, 30 cycles with an annealing temperature of 62°C for CCR2B, and 35 cycles with an annealing temperature of 57°C for 60 seconds for CCR6. Total RNA from Jurkat cells was used as a positive control for the CCR6 RT-PCR. PCR products were separated by electrophoresis through 1.5% agarose.
Confluent FLS from passages 3 and 4 were cultured for 4 hours in serum-free DMEM (hereinafter referred to as medium), then trypsinized and resuspended in Hanks' balanced salt solution (HBSS; BioWhittaker). Cells (3 × 106/ml) were suspended in a 5 μM solution of DMSO/Fura Red AM ester in Pluronic F-127 (1:1 final dilution; Molecular Probes, Leiden, The Netherlands), and the cell suspension was incubated at room temperature for 45 minutes. After washing twice in PBS, loaded cells were resuspended in HBSS at a final concentration of 1 × 106/ml. After stimulation with each chemokine (100 ng/ml), cells were screened continuously in a FACScan for 6 minutes to assess the reduction in mean fluorescence intensity induced by each chemokine and were related to the fluorescence obtained by calibration of the unstimulated cells. To assess the specificity of chemokine-mediated Ca++ mobilization, cells were preincubated with 0.75 μg/ml pertussis toxin (Calbiochem, Bad Soden, Germany) for 10 minutes before adding the different chemokines. The effect of the Ca++ ionophore A23187 (0.5 μmole/liter; Sigma) was used as a positive control for Ca++ influx. The variation in the intracellular Ca++ concentration induced by chemokines in FLS was normalized in each experiment, considering that cell fluorescence in the presence of ionophore A23187 represents the maximum intracellular Ca++ concentration (100%) and the fluorescence of cells in medium alone represents the minimum (0%). The percentage variation of the intracellular Ca++ concentration was calculated according to the following equation: ([FIx − FImedium]/[FIA23187 − FImedium]) × 100, where FIx is the fluorescence intensity of each condition, FImedium is the fluorescence intensity of cells in the absence of chemokines, and FIA23187 is the fluorescence intensity of cells treated with Ca++ ionophore A23187.
Transwell chemotaxis assays were performed in 24-well transwells (6.5-mm diameter, 8-μm pore size; Costar, Cambridge, MA), as previously described (18). FLS cultured as described above were detached with trypsin/EDTA and then washed with PBS. FLS (5 × 104) in 0.1 ml of medium were added to the upper chamber of 2-compartment transwells. Chemokines (50 ng/ml) or GM-CSF (25 ng/ml) were added to the medium in the lower compartment. The lower surfaces of the transwells had been coated with type I collagen (40 μg/ml for 1 hour at 37°C). The set was placed in a humidified 5% CO2 incubator at 37°C. After overnight incubation, both the cells in suspension in the lower compartment and the cells attached to the lower filter surface recovered by collagenase treatment (1 mg/ml for 10 minutes at 37°C) were pooled. All cell suspensions were washed once in PBS, resuspended in 0.5 ml PBS, and quantified in the FACScan using a known concentration of polystyrene fluorospheres as a control. When indicated, 0.75 μg/ml pertussis toxin was added to the cell suspension in the upper compartment. In all experiments, samples of FLS were labeled with propidium iodide to assess cell viability by flow cytometry. To normalize data from different experiments, the migration of FLS in response to medium alone was considered 100, and results are presented as the mean ± SEM of chemokine-mediated enhancement of transmigration.
Assays for gelatinase and collagenase activity.
Gelatinase and collagenase activity was determined in cell-free supernatants by a fibril degradation assay as well as zymography. FLS (6 × 104) were planted in 24-well plates (Corning) in cDMEM. After overnight incubation, wells were washed once with PBS, filled with 1 ml of medium, and incubated for 4 hours at 37°C. Then, the supernatant was removed and 250 μl of fresh medium, either alone or in the presence of different chemokines, was added. Time-course experiments were performed for 2–24 hours of incubation in the presence of 100 ng/ml of the different chemokines at 37°C. When dose-response studies were performed, FLS were incubated for 24 hours at 37°C with the different compounds at doses ranging from 0.01 to 200 ng/ml. In experiments with functional blocking mAb, FLS were preincubated with an excess of anti-human IL-1β (15 ng/ml; Sigma), anti-human TNFα (50 ng/ml; Sigma), or both together for 5 minutes, before adding the different chemokines. After these experiments, cell-free supernatants were collected and the MMP activity was determined.
We used the EnzChek gelatinase/collagenase assay kit (Molecular Probes), a fibril degradation assay that utilizes self-quenched fluorescein-conjugated gelatin and type I collagen, to determine the collagenase and gelatinase activity in cell-free supernatants. When these substrates are digested, highly fluorescent peptides are released and can be detected by fluorometry. The assay was performed according to the manufacturer's recommendations, and the fluorescence intensity was measured with a spectrofluorometer (Genios; Tecan, Männedorf, Switzerland). Data were expressed as MMP activity either in units produced by 1 × 106 FLS in 1 ml (units/ml/1 × 106 cells) or as a percentage of cells maintained in medium (considered 100%).
In zymography, 10 μl of supernatant from unstimulated FLS incubated in medium for 24 hours was mixed with 20 μl of buffer (500 mM Tris HCl, pH 6.8, 25% glycerol, 10% sodium dodecyl sulfate [SDS], and 1% bromphenol blue) and then resolved under nonreducing conditions by 9% SDS–polyacrylamide gel electrophoresis, embedded with 1 mg/ml gelatin (Calbiochem) or type II collagen (Sigma). Gels were rinsed 3 times in 2.5% Triton X-100 for 30 minutes at room temperature and then incubated in 50 mM Tris HCl, pH 7.5, 10 mM CaCl2, and 200 mM NaCl for 12 hours at 37°C. Then, gels were stained with Coomassie blue. Areas of gelatinolytic or collagenolytic activity produced transparent bands, which were negatively visualized using Adobe Photoshop 5 (Adobe Systems, San Jose, CA).
Enzyme-linked immunosorbent assays (ELISAs).
The specific ELISAs for human IL-1β and TNFα used in these experiments were purchased from ImmunoTools (Friesoythe, Germany). FLS were incubated in 24-well plates in 250 ml of medium (∼6 × 104 cell/well) in the presence or absence of 100 ng/ml of SDF-1α, MCP-1, and Mig. After 24 hours, supernatants were collected and the concentrations of soluble IL-1β and TNFα were assayed in duplicate wells following the manufacturer's recommendations. The sensitivities of ELISAs for IL-1β and TNFα in our laboratory were 3 and 10 pg/ml, respectively. Data were expressed as the mean ± SD cytokine concentration (in pg/ml).
Cell proliferation assay.
FLS proliferation was determined with the CyQuant cell proliferation assay, a fluorescence-based proliferation assay kit (Molecular Probes). Subconfluent cultured FLS (3 × 103/well) were cultured for 72 hours in 96-well plastic plates containing 5% FCS in DMEM, either alone or in the presence of 5 ng/ml IL-1β or 100 ng/ml of chemokines. Time and dose were selected according to previous dose- and time-dependent assays. When indicated, blocking mAb against MCP-1, SDF-1α, and IP-10 (R&D Systems) were added in excess prior to addition of stimuli, either independently or pooled and maintained for the whole incubation period. The density of cells in culture was determined according to the manufacturer's instructions (Molecular Probes). Briefly, after the incubation period, cells were frozen, then thawed and lysed by the addition of a buffer containing the green fluorescence CyQuant dye, which exhibits strong fluorescence enhancement when bound to cellular nucleic acids. Fluorescence was then measured directly using a fluorescence microtiter plate reader (Tecan) with an excitation wavelength of 485 nm and emission detection at 530 nm. In each experiment, a linear range was obtained from fluorescence measured in different known quantities of cells cultured in medium alone. Each sample was analyzed in triplicate. Data were expressed as a percentage of enhanced proliferation with respect to cells maintained in medium alone (considered 100%).
Differences between groups were examined for statistical significance using Wilcoxon's signed rank test. P values less than 0.05 were considered significant.
Expression of functionally active chemokine receptors on FLS from RA patients.
As a first step, we used RT-PCR to examine the presence of CCR2, CCR5, CCR6, CXCR3, and CXCR4 mRNA in cultured, unstimulated FLS. FLS from 3 RA patients constitutively expressed detectable amounts of mRNA for CCR2, CCR5, CXCR3, and CXCR4. However, this technique did not demonstrate the presence of CCR6 transcripts. Representative results of an RT-PCR are shown in Figure 1A. The control RT-PCR with mRNA from Jurkat cells was positive for CCR6 (results not shown), as previously described (19). Surface expression of chemokine receptors was then analyzed by indirect immunofluorescence and flow cytometry. Unstimulated FLS from 6 RA patients expressed CCR2, CCR5, CXCR3, and CXCR4, but not CCR3 and CCR6 (Figure 1B and Table 1). FLS stimulation for 24 hours with IL-1β (5 ng/ml), TNFα (20 ng/ml), or IFNγ (100 units/ml) did not significantly modify basal surface expression of any of the chemokine receptors studied (data not shown). These data represent the first demonstration of the presence of CXCR3 on FLS from RA patients at both the cell surface protein and mRNA levels.
Table 1. Chemokine receptor surface expression on FLS*
Clone monoclonal antibody
% positive cells, mean ± SEM
Values are the mean ± SEM of 8 independent experiments. FLS = fibroblast-like synoviocytes; MCP = monocyte chemotactic protein; TARC = thymus and activation–regulated chemokine; MIP = macrophage inflammatory protein; IP-10 = interferon-inducible protein 10; Mig = monokine induced by interferon-γ; SDF-1α = stromal cell–derived factor 1α.
= P < 0.05 versus the fluorescence of a control isotype-matched monoclonal antibody, by Wilcoxon's signed rank test.
Chemokine binding to chemokine receptors induces several intracellular changes, including mobilization of intracellular calcium (20). Calcium mobilization studies were performed to determine whether the chemokine receptors expressed by FLS from RA patients were functionally active. FLS were challenged with 100 ng/ml MCP-1, MIP-1β, IP-10, SDF-1α, and RANTES. The stimulation of FLS with SDF-1α, IP-10, and MCP-1 caused a significant increase in intracellular calcium (Figure 2). As expected, because chemokine receptors signal through a receptor coupled to Gi (21), preincubation with 0.75 μg/ml pertussis toxin (an inhibitor of the Gαi subunit of G proteins) abolished the chemokine-induced calcium flux in FLS. RANTES induced variable responses with significant intracellular calcium influx in 2 of 6 experiments, whereas MIP-1β did not induce a detectable effect in any experiment. These data demonstrate that FLS consistently express functionally active CCR2, CXCR3, and CXCR4 chemokine receptors.
Transmigration of FLS in chemokine gradient.
To determine whether chemokines could induce directed migration in FLS, transmigration assays were performed using a 2-compartment transwell system as described in Materials and Methods. IP-10, MCP-1, and SDF-1α (50 ng/ml) triggered a significant chemotactic response by FLS, increasing 2–2.5-fold compared with the basal transmigration. In contrast, neither RANTES nor MIP-1β induced a significant effect (Figure 3). Enhanced migration (2.5-fold) was attained by 25 ng/ml GM-CSF (data not shown), which was used as a positive control (18). Again, the presence of pertussis toxin (0.75 ng/ml) completely neutralized the effect of chemokine receptor agonists without a detectable reduction in cell viability, as assessed by propidium iodide incorporation after overnight incubation (data not shown). These results prove that some chemokines from the CC and CXC families are able to attract FLS, which may result in relevant pathogenic implications in RA.
Chemokine-induced increase in MMP activities of FLS.
MMPs are considered to be the primary proteolytic enzymes responsible for the degradation of cartilage in RA (2, 3). Resting FLS in culture release, mainly as inactive proenzymes, low levels of gelatinases (MMP-2 and MMP-9) and collagenases (MMP-1 and MMP-13) (22, 23). In order to validate our experimental conditions, we examined the gelatinase and collagenase activity of supernatants from unstimulated FLS in culture. Results of zymography in gelatin revealed a pair of bands at ∼65 kd (major band, proMMP-2) and 50 kd (MMP-2) and 2 minor bands at ∼90 kd (proMMP-9) and 85 kd (MMP-9) (Figure 4A). When type II collagen was used as a substrate (Figure 4B), zymography of FLS supernatants exhibited a major band at ∼60 kd (likely a mixture of proMMP-1 and proMMP-13) and a minor band at 45 kd (MMP-1), which was consistent with the results of previous studies (23). Then, we studied the potential role of chemokines in regulating MMP production by cultured FLS. Kinetic experiments in a fibril degradation assay showed that MCP-1, SDF-1α, Mig, IP-10, and RANTES (100 ng/ml), in this order of potency, increased gelatinase and collagenase activity in FLS supernatants, with a detectable effect only after 8 hours of incubation, reaching the maximum at 24 hours (Figures 4C and D). The induction of gelatinase and collagenase activity in FLS by chemokines was dose-dependent, with the optimal concentration between 100 and 200 ng/ml (Figures 4E and F). These effects were reversed by preincubation of cells with pertussis toxin (0.75 μg/ml) (Figure 5). Therefore, ligand binding to chemokine receptors expressed by FLS is able to induce gelatinase and collagenase activities.
Requirement of endogenous IL-1β for chemokine stimulation of MMP activity in FLS.
The finding that the induction of MMP activities by chemokines on FLS required more than 8 hours of incubation raised the possibility of an indirect effect. IL-1β and TNFα are potent stimulators of MMP production in FLS (24, 25). This, in addition to the fact that chemokines are able to modulate the production of several cytokines, including IL-1β and TNFα (26, 27), prompted us to investigate the potential effect of such cytokines on the induction of MMP activities by chemokine-treated FLS. The concentration of soluble IL-1β and TNFα was measured by ELISA in the supernatant of FLS maintained for 24 hours in medium alone or in the presence of different chemokines. We found that the basal production of IL-1β (mean ± SD 90 ± 30 pg/ml) by FLS was significantly increased by the presence of SDF-1α (190 ± 12 pg/ml), MCP-1 (160 ± 28 pg/ml), and Mig (148 ± 30 pg/ml) (4 experiments). However, TNFα was not detected in any of those FLS supernatants. The presence of blocking mAb against IL-1β or TNFα was able to inhibit 50–60% of the induction of collagenase (Figure 6) and gelatinase (data not shown) activity in FLS treated with IL-1β and TNFα, respectively. Interestingly, when FLS were preincubated with a blocking anti–IL-1β antibody, the collagenase and gelatinase activity induced by 24-hour stimulation with SDF-1α, MCP-1, or Mig was reduced 70–80% (Figure 6, data not shown for gelatinase activity). In contrast, the presence of a neutralizing anti-TNFα antibody had no effect on MMP production by chemokine-stimulated FLS (Figure 6, data not shown for gelatinase activity). These results demonstrate that endogenous IL-1β, but not TNFα, is necessary for chemokine-mediated induction of MMP activities in FLS.
Chemokine-induced FLS proliferation.
In RA, in situ proliferation of FLS contributes to synovial lining hyperplasia (28). A fluorescence-based assay was used to determine the effect of several chemokine receptor agonists on FLS proliferation. MCP-1, SDF-1α, IP-10, and Mig (100 ng/ml optimal concentration) were able to induce significant cell proliferation after 72 hours of culture in a range similar to that of IL-1β (5 ng/ml) (Figure 7A). RANTES also enhanced cell growth, but due to variability among experiments, the differences did not reach statistical significance (Figure 7A). Since 0.75 μg/ml pertussis toxin had a toxic effect on FLS after 72 hours of culture, the implication of the chemokine system in FLS proliferation was determined using blocking mAb against SDF-1α, MCP-1, and IP-10. These antibodies significantly inhibited the proliferation induced by their ligands (Figure 7B). A blocking mAb against CXCR4 also abrogated SDF-1α–mediated proliferation (data not shown). In the experiments in which cells responded to RANTES, significant inhibition of cell proliferation was achieved with a mAb against CCR5 (data not shown). These results demonstrate that CXCR3, CXCR4, CCR2, and, to a lesser extent, CCR5 were able to induce signals that resulted in FLS overgrowth.
This is the first report to describe the expression of express functionally active CXCR3 by RA FLS and to confirm the presence of CCR2, CCR5, and CXCR4. We also found that natural agonists to these receptors are able to induce migration, proliferation, and MMP activities by FLS, and that MMP activities induced by chemokines in FLS require endogenous IL-1β. These findings indicate that the chemokine system may be involved in the pathogenesis of RA beyond the regulation of cell trafficking and may provide additional evidence for the potential therapeutic benefit of targeting chemokines or their receptors in RA.
Current evidence shows that FLS act as the main effector cells in the joint destruction of RA through their ability to invade and degrade the connective tissue of cartilage and tendon (29). They can also stimulate the differentiation and activation of osteoclasts, resulting in bone erosion (30). Recent research has provided much information about signals that can target the FLS, such as mediators of inflammation, cytokines, and cell–cell and cell–extracellular matrix (ECM) interactions (31). However, despite the up-regulation of several chemokines, such as IP-10, Mig, MCP-1, SDF-1α, and RANTES, found in RA synovium (8, 10, 11, 32, 33), limited information is available about the expression profile and function of chemokine receptors in FLS. It has been shown that the production of IL-6 and IL-8 by FLS is enhanced by agonists of CCR2, CCR5, and CXCR4 (34) and that fractalkine increases the production of MMP-2 by FLS (35).
To our knowledge, this is the first study to demonstrate that RA FLS express CXCR3. We also confirmed previous reports of the expression of CCR2 and CCR5 (34). Additionally, our findings showing functional CXCR4 on FLS shed light on conflicting data concerning the presence of this receptor on RA synoviocytes (9, 34). The expression of all these receptors was demonstrated at both the cell surface protein and the mRNA levels. Several studies in the last decade have focused on the changes that cytokines induce in the basal expression of chemokine receptors in different cell types. These studies have shown that variations in the expression of chemokine receptors in response to cytokines are cell-type specific (36–40). In this regard, our data show that the profile of chemokine receptors expressed by RA FLS is very stable. Baseline surface expression of CCR2, CCR5, CXCR3, and CXCR4 was not significantly modified even after 24 hours of stimulation with TNFα, IL-1β, and IFNγ. Similar behavior has been described for CC receptors in TNFα-stimulated chondrocytes from osteoarthritis (OA) patients (41). Interestingly, natural agonists of CCR2, CXCR3, and CXCR4 induced a significant increase in intracellular calcium influx mediated by G protein–coupled complex, indicating that those receptors are able to originate intracellular signals and as a consequence be functionally active in FLS. The high variability in calcium mobilization obtained with RANTES in different FLS samples might be indicative of different receptor expression profiles, expression levels, or affinity states for this chemokine in different subsets of FLS. In this regard, the substantial inhibition of proliferation caused by a blocking anti-CCR5 mAb indicates that at least in some RA patients, FLS express functionally active CCR5.
The chemokine receptors described herein can be essential mediators of the pathologic mechanism of FLS in RA. FLS are thought to migrate over the cartilage and erode into the subchondral bone, ultimately leading to the formation of erosions. One major challenge for rheumatologists has been the identification of factors that direct proliferative synovium to destroy cartilage and bone in RA. Adhesion molecules and exposed fragments of enzymatically degraded ECM proteins can chemotactically guide cells with the appropriate receptors (42). In the same way, chemokines locally produced by chondrocytes (43) or by cells in the cartilage–pannus junction (13), immobilized by exposed proteoglycans, might lead to the migration of FLS that express an adequate repertoire of chemokine receptors. Our data show that the activation of the corresponding receptors by MCP-1, SDF-1α, and IP-10 induces specific transmigration of FLS. These findings suggest that chemokine receptors on FLS may act as a major driving force in the pannus invasion of cartilage in RA.
Unlike the synovial lining of healthy joints, which is only a few cells thick, RA synovium is characterized by synoviocyte hyperplasia. Proliferation of FLS is of pivotal importance for the formation of pannus (28). Key regulators of this phenomenon include the recently recognized macrophage migration inhibitory factor (44) and proinflammatory cytokines such as TNFα and IL-1β (45). Although the chemokine system has been implicated in the regulation of the proliferation of several cell types (46–48), its ability to regulate FLS growth has not been studied. We have found that MCP-1 and SDF-1α, along with IP-10 and Mig, which share CXCR3, may drive FLS proliferation to a similar extent as IL-1β. RANTES also induced cell overgrowth in the majority of samples, although a wide range of variability among experiments likely explains the lack of statistically significant differences.
FLS have been identified as the major source of matrix-degrading enzymes in RA. The ability of FLS to destroy cartilage ECM seems to be associated with the level of expression of MMP (49). In this study, we have firmly established that chemokines can induce proteolytic activity in FLS. Proenzymes of the 2 most important collagenases involved in this process, interstitial collagenase (MMP-1) and collagenase 3 (MMP-13) (50), were detected in zymography assays with FLS, in addition to active and inactive forms of gelatinolytic enzymes such as gelatinase A (MMP-2) and B (MMP-9). Previous reports have suggested the relevance of chemokines in the cartilaginous breakdown in OA through the release of stromelysin 1 (MMP-3) by normal or OA chondrocytes (9, 41, 51). Nevertheless, to our knowledge, no data have shown collagenase release in FLS induced by chemokines, and just the effect of fractalkine on gelatinase (MMP-2) production has been reported in FLS (35).
Our results demonstrate that MCP-1, SDF-1α, Mig, IP-10, and RANTES enhance collagenase and gelatinase activities of FLS in a dose- and time-dependent manner. The fact that the significant increase in MMP activity in the supernatants of chemokine-treated FLS required more than 8 hours of incubation before it could be detected raised the possibility of an indirect effect. Interestingly, our results using functional blocking antibodies showed that the effect of MCP-1, SDF-1α, Mig, IP-10, and RANTES is predominantly mediated by endogenous IL-1β, but not by TNFα. Therefore, chemokine stimulation up-regulates IL-1β expression by FLS which, through an autocrine/paracrine effect, would be responsible for the enhancement of MMP production. This is not surprising, since it has been shown that in vitro, some chemokines are able to modulate the production of several cytokines, such as IL-1β and TNFα (26, 27), which are well-known and potent stimulators of MMP production by FLS (25, 45).
Evidence from animal models of arthritis has shown that the blockade of chemokines or their receptors has beneficial effects not only on inflammation, but also on joint destruction, even in advanced phases of the disease (52–56). The therapeutic efficacy afforded by chemokine-targeting strategies in these models has been largely attributed to the inhibition of the recruitment of inflammatory cells. However, what portion of these benefits might be attributed to suppressing signals that lead to production of other cytokines, MMPs, or modulation of processes such as synovial cell proliferation and migration has not been clarified. In this regard, a recent study of a model of adjuvant-induced arthritis has demonstrated that CCR1, CCR2, and CCR5 activation in synovium is coincident with both peak inflammation and the initiation of different signaling pathways in affected rats compared with controls (57). Although the efficacy of modulation of chemokine signaling has not been established in humans, a recently reported phase Ib, double-blind, placebo-controlled clinical trial has shown that short-term (2 weeks) treatment with an oral CCR1 antagonist (58) improves signs and symptoms in RA patients and strikingly decreases cellularity in synovial biopsy samples. Long-term clinical studies to establish the safety and benefits of this approach in structural damage in the joint are warranted.
In summary, the findings of the current study suggest a broad involvement of the chemokine system in RA pathogenesis, which is not restricted to inflammatory cell influx and retention. We have shown that cultured RA FLS express a wide repertoire of chemokine receptors that can signal to drive migration, proliferation, and matrix degradation. These data, in addition to the aforementioned studies in animal models of arthritis, provide support for antichemokine therapies in RA to alleviate synovial inflammation and prevent joint destruction.
We thank Dr. C. Martinez for generously supplying the anti-CCR2 monoclonal antibodies. We also are indebted to Drs. F. Sánchez-Madrid and Manuel Feria for critical reading of the manuscript and to Drs. A. Laffón and Tomás González for their continuous support.