Mediation of the proinflammatory cytokine response in rheumatoid arthritis and spondylarthritis by interactions between fibroblast-like synoviocytes and natural killer cells.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 58, No. 3, March 2008, pp 707–717 DOI 10.1002/art.23264 © 2008, American College of Rheumatology Mediation of the Proinflammatory Cytokine Response in Rheumatoid Arthritis and Spondylarthritis by Interactions Between Fibroblast-like Synoviocytes and Natural Killer Cells Antoni Chan,1 Andrew Filer,2 Greg Parsonage,2 Simon Kollnberger,1 Roger Gundle,3 Christopher D. Buckley,2 and Paul Bowness4 IL-1␤, and IL-15, was increased in cocultures of NK cells and FLS, particularly in those from RA and SpA patients. Production of interferon-␥, RANTES, and matrix metalloproteinase 3 (MMP-3) by NK cell and FLS coculture was greatest in SpA patients. Surface expression of IL-15 on FLS was significantly increased in SpA and RA patients, but not OA patients. Blockade with an IL-15 monoclonal antibody resulted in increased apoptosis of NK cells. Conclusion. FLS promote the migration, activation, and survival of NK cells. The interaction of NK cells with FLS results in increased IL-15 expression by FLS and the production of proinflammatory chemokines, cytokines, and MMPs, which may contribute to joint inflammation. This response was much more marked in SpA and RA patients as compared with OA patients. Objective. Fibroblast-like synoviocytes (FLS) are potentially directly involved in the propagation of inflammation. We have previously shown evidence of an expanded activated population of natural killer (NK) cells in spondylarthritis (SpA) patients. In the present study, we sought to determine whether the interaction between NK cells and FLS from SpA patients results in a proinflammatory response. Methods. Autologous NK cells and FLS were obtained from 6 patients with SpA, 4 patients with rheumatoid arthritis (RA), and 8 patients with osteoarthritis (OA). Physical interactions between NK cells and FLS were studied by time-lapse phase-contrast microscopy. Fluorescence-activated cell sorting was used to study the activation, proliferation, and survival of NK cells in contact with FLS. Cytokine and stromal factor production were measured by a multiple cytokine bead assay. Results. NK cells both adhered to and migrated beneath the FLS monolayer (pseudoemperipolesis). FLS from SpA and RA patients supported increased pseudoemperipolesis, activation, cytokine production, and survival of NK cells. The production of proinflammatory cytokines, including interleukin-6 (IL-6), IL-8, The synovial membrane lining comprises 2 dominant cell populations. These consist of macrophage-like synoviocytes (type A synoviocytes) and fibroblast-like synoviocytes (FLS; type B synoviocytes), the latter distinguished by their surface marker expression profile (1). Two-thirds of native synoviocytes are FLS, which function to provide structural integrity and to maintain the articular surface in health (2). Recent evidence shows that FLS are not passive players in the immune response (3) and may regulate the switch from acute resolving inflammation to chronic persistent inflammation (4). Little is known of the role of FLS in spondylarthritis (SpA). In rheumatoid arthritis (RA), FLS are involved in the recruitment, activation, and survival of T cells in the joint (5). FLS participate actively in cell recruitment by producing chemokines, including monocyte chemotactic protein (MCP), interleukin-8 (IL-8), RANTES, and IL-16 (6,7). Retention of T cells within the joint is aided by the production of the chemokine Supported by the Arthritis Research Campaign and the Medical Research Council, UK. 1 Antoni Chan, PhD, MRCP, Simon Kollnberger, PhD: John Radcliffe Hospital, Oxford, UK; 2Andrew Filer, MBChB, PhD, Greg Parsonage, PhD, Christopher D. Buckley, MBBS, FRCP: University of Birmingham, Birmingham, UK; 3Roger Gundle, DPhil: Nuffield Orthopaedic Centre, Oxford, UK; 4Paul Bowness, MB BChir, DPhil: John Radcliffe Hospital, Oxford, and Nuffield Orthopaedic Centre, Oxford, UK. Address correspondence and reprint requests to Antoni Chan, PhD, MRCP, Medical Research Council Human Immunology Unit, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, Headley Way, Headington, Oxford OX3 9DS, UK. E-mail: firstname.lastname@example.org or email@example.com. Submitted for publication March 27, 2007; accepted in revised form November 26, 2007. 707 708 CHAN ET AL stromal cell–derived factor 1 (CXCL12) by FLS and by the expression of its receptor CXCR4 on infiltrating T cells (8,9). The prolonged survival of T cells in the rheumatoid joint is supported principally by type I interferons (IFNs) produced by FLS (10). FLS also produce IL-6, which may regulate the switch from innate to acquired immunity through differential control of leukocyte recruitment, activation, and apoptosis (11,12). Natural killer (NK) cells are effector cells of the innate immune system; they comprise 10–15% of peripheral blood lymphocytes (13). They express CD56, but not B cell or T cell markers such as CD3 (14). NK cells play an important role in the defense against virusinfected and tumor cells (15). NK cells bear a number of cytokine receptors, including those for IL-2, IL-12, IL15, IL-21, and IFN␣/␤. These cytokines can promote NK proliferation, cytotoxicity, and production of cytokines such as IFN␥, tumor necrosis factor ␣ (TNF␣), granulocyte–macrophage colony-stimulating factor (GM-CSF), IL-5, IL-13, macrophage inflammatory protein 1␣/␤ (MIP-1␣/␤), and RANTES (16–18). We have recently shown an expansion of NK cells bearing the killer cell immunoglobulin-like receptor 3DL2 (KIR3DL2) in the peripheral blood and synovial fluid of SpA patients (19). Ex vivo, KIR-3DL2⫹ NK cells from SpA patients had an activated phenotype and were protected from activation-induced cell death. We hypothesized that NK cells may interact with resident FLS in the joint through cell-to-cell contact as well as with cytokines and that this could result in a proinflammatory response. In this study, we found that ex vivo, isolated NK cells physically interact with FLS. This interaction results in mutual activation and production of proinflammatory cytokines and other factors that may contribute to joint damage, including IFN␥, IL-6, IL-8, GM-CSF, MCP-1, and matrix metalloproteinase 3 (MMP-3). Furthermore, we found evidence of a significant role of IL-15 in this interaction. PATIENTS AND METHODS Patients and samples. Synovial tissue samples were collected from the knee joints of 6 SpA, 4 RA, and 8 osteoarthritis (OA) patients who were undergoing arthroplasty or synovectomy. SpA was defined according to the European Spondylarthropathy Study Group criteria (20). Four SpA patients fulfilled the modified New York criteria for ankylosing spondylitis (AS) (21). Of the 4 AS patients, 3 were HLA–B27 positive (HLA–B*2702/2705) by DNA typing. The remaining 2 SpA patients, both of whom were HLA–B27 positive, had psoriatic arthritis. RA patients fulfilled the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) revised criteria (22). OA of the knee was defined according to the ACR criteria (23). The main demographic and clinical data on the study patients are shown in Table 1. This study had ethical approval from the Oxford Radcliffe Trust local (COREC C00.114) and the South Birmingham Local Ethics Committee (LREC 5735). Isolation of peripheral blood mononuclear cells (PBMCs). Peripheral blood was collected into 50-ml polypropylene tubes (BD Falcon, Bedford, MA) prepared with 100 l (5,000 units) of sodium heparin. PBMCs were isolated by centrifugation at 2,000 revolutions per minute for 20 minutes using a Ficoll-Hypaque gradient (Lymphoprep; Nycomed Pharma, Oslo, Norway). Cells at the interface of the 2 phases were collected and washed twice with RPMI 1640 medium (Gibco BRL, Grand Island, NY), and cells were resuspended in R10 (RPMI 1640 medium supplemented with 2 mM L-glutamine, 100 IU of penicillin/streptomycin, and 10% fetal calf serum [FCS]). Both fresh and cryopreserved cells were compared by repeated stainings, and consistent results were obtained, with a ⬍5% difference (19). FLS and NK cell culture. Fine strips of synovial tissue ⬍1 mm3 in volume were suspended in 10 ml of wash buffer (RPMI 1640 and 20 mM HEPES). After centrifugation at 200g for 10 minutes, the supernatant was discarded and the pellet resuspended in wash buffer. The wash step was repeated 2 more times. The pellet was resuspended in 10 ml of digestion buffer (RPMI 1640, 20 mM HEPES, and 0.2% collagenase type 1A) and incubated for 4–5 hours at 37°C, with vigorous shaking. The digestion mixture was then centrifuged at 200g for 10 minutes, the supernatant was discarded, and the pellet was resuspended in 10 ml of fresh medium (RPMI 1640, 20% FCS, 1% glutamine, 1% penicillin/streptomycin, 1% nonessential amino acids, and 1% sodium pyruvate; Sigma-Aldrich, St. Louis, MO) and cultured to confluence as previously described (24,25). FLS had characteristic morphology and expressed fibronectin and prolyl-4-hydroxylase, but not CD1, CD3, CD19, CD31, CD68, CD80, CD86, von Willebrand factor, or cytokeratin (26). FLS were used between passages 3 and 5. NK cells were enriched from PBMCs using a magneticactivated cell sorting NK-negative isolation kit (Miltenyi Biotech, Bergisch Gladbach, Germany). The purity of the cells was determined before they were used, and ⬎99% were found to be CD3–,CD56⫹ by flow cytometry. NK cells and FLS coculture conditions. FLS were passaged and added to 96-well flat-bottomed plates (BD Falcon) at a density of 3,000 cells/well. Experiments were performed with confluent FLS prepared 24 hours before contact. Once confluence was confirmed by microscopy, the fibroblasts were washed twice in phosphate buffered saline (PBS) and cocultured for 7 days with 1 ⫻ 105 NK cells in R10 (RPMI 1640 medium supplemented with 2 mM L-glutamine, 100 IU of penicillin/streptomycin, and 10% FCS). We have previously shown that NK cell differentiation and proliferation occur after 5 days of NK cell/FLS coculture, with many NK cells undergoing apoptosis after 10 days of coculture without the addition of cytokines (27). Therefore, analyses were performed on day 7. The control monolayers of FLS remained confluent at 7 days in the absence of NK cells. All experiments were performed in duplicate. A 24-well Transwell system (Corning Costar, Cambridge, MA) was used in some coculture experiments. This system consists of 2 compartments separated by a porous MUTUAL STIMULATION OF FLS AND NK CELLS IN INFLAMMATORY ARTHRITIS Table 1. 709 Demographic, laboratory, and clinical features of the study patients, by diagnostic group* Age, mean (range) years Sex, no. male/female No. (%) HLA–B27⫹ No. of HLA–B27⫹ SpA patients AS PsA BASDAI score, mean ⫾ SD No. (%) RF⫹ DAS28 score, mean ⫾ SD No. (%) taking DMARDs No. taking MTX No. taking SSZ No. (%) taking corticosteroids No. taking 5 mg/day of prednisolone Previous anti-TNF␣ therapy† No. who took infliximab No. who took etanercept No. who took adalimumab SpA patients (n ⫽ 6) RA patients (n ⫽ 4) OA patients (n ⫽ 8) 39.7 (28–53) 4/2 5 (83) 40.8 (28–59) 1/3 0 (0) 46.7 (31–77) 4/4 0 (0) 3 2 3.9 ⫾ 1.0 0 (0) – 2 (33) 1 1 0 (0) 0 0 (0) 0 0 0 – – – 3 (75) 4.1 ⫾ 1.1 4 (100) 4 2 1 (25) 1 1 (25) 1 0 0 – – – 0 (0) – 0 (0) 0 0 0 (0) 0 0 (0) 0 0 0 * SpA ⫽ spondylarthritis; RA ⫽ rheumatoid arthritis; OA ⫽ osteoarthritis; AS ⫽ ankylosing spondylitis; PsA ⫽ psoriatic arthritis; BASDAI ⫽ Bath Ankylosing Spondylitis Disease Activity Index; RF ⫽ rheumatoid factor; DAS28 ⫽ Disease Activity Score in 28 joints; DMARDs ⫽ disease-modifying antirheumatic drugs; MTX ⫽ methotrexate; SSZ ⫽ sulfasalazine. † None of the patients were receiving an anti–tumor necrosis factor ␣ (anti-TNF␣) agent at the time of the study. matrix (0.4 m), which allows the exchange of soluble factors while preventing direct contact. FLS were grown to confluence in the bottom well, and NK cells were either added to the same well (allowing contact) or to the top well (avoiding contact). Anti-human IL-6 monoclonal antibody (mAb) and anti-human IL-15 mAb (both from R&D Systems, Abingdon, UK) blocking antibodies and isotype control antibody (BD PharMingen, San Jose, CA) were added at a concentration of 10 g/ml to the FLS for 30 minutes at 4°C. Neutralizing polyclonal goat IgG anti–IL-15R␣ antibody (R&D Systems) or control goat IgG was used at 1 g/ml and was incubated with NK cells for 30 minutes at 4°C. NK cells were subsequently added to the FLS. Antibodies, fluorescence-activated cell sorter (FACS) analysis, and immunostaining. NK cells were washed in PBS in the presence of 1% heat-inactivated FCS and 0.1% sodium azide (weight/volume) and incubated for 30 minutes on ice with saturating amounts of the following directly conjugated anti-human mAb: CD3, CD16, CD94, and CD56 (Serotec, Oxford, UK). The following directly conjugated anti-human mAb were also used in this study: CD68, CD69, CD80, CD86, CD90, HLA–DR, vascular cell adhesion molecule 1, NKp46, NKp30, and IFN␥ (BD Biosciences, San Jose, CA) and IL-15 (R&D Systems). Cells were washed twice after incubation with mAb. Cells were fixed in 1% paraformaldehyde and analyzed with a FACSCalibur cell sorter using CellQuest software (Becton Dickinson, Mountain View, CA). Intracellular cytokine staining was used to detect IFN␥ release using a Cytofix/ Cytoperm kit (BD Biosciences). As positive controls for intracellular cytokine staining, phorbol myristate acetate (10 ng/ml) and ionomycin (1 ng/ml) were used (Sigma-Aldrich, Dorset, UK). Pseudoemperipolesis assays. Fibroblasts were seeded onto glass chamber slides at a density of 1 ⫻ 105/well and cultured for 3 days. Prior to coculture, the fibroblast layer was washed, and 400 l of R10 medium was added. A 100-l volume of autologous NK cells (2.5 ⫻ 105) was added to the well, and the coculture was gently mixed, then incubated under stasis for a period of 2 hours. Pseudoemperipolesis was assessed by counting phase-dark cells in 3 independent fields and was expressed as a percentage of the total input NK cells. Phase-dark cells were cells that had migrated into and under the fibroblast layer; phase-light cells were cells that either remained nonadherent or adhered to the surface of the fibroblast layer. A Zeiss Axiovert 200 microscope system with settings for phase contrast was used (Carl Zeiss Instruments, Welwyn Garden City, UK). Images were captured and combined into time-lapse movies using a Hamamatsu C4742-95 camera and Simple PCI software (Hamamatsu Photonics, Welwyn Garden City, UK). Survival assays. NK cells (1 ⫻ 105 cells) and FLS (3,000 cells) were cultured for 7 days in 200 l of R10 in flat-bottomed 96-well plates. Cells were harvested and stained for annexin V and 7-aminoactinomycin D (7-AAD) according to the manufacturer’s instructions (BD Biosciences) and then analyzed by FACS. Annexin V–negative, 7-AAD–negative cells were designated fully viable, annexin V–positive, 7-AAD– negative cells were designated as undergoing early apoptosis with membrane integrity present, and annexin V–positive, 7-AAD–positive cells were designated as end-stage apoptotic cells committed to death. Labeling and analysis with 5,6-carboxyfluorescein diacetate N-succinimidyl ester (CFSE-DA). NK cells were labeled with 5 M CFSE-DA (Molecular Probes, Eugene, OR) 710 CHAN ET AL RESULTS High levels of NK cell pseudoemperipolesis in RA and SpA FLS. We have previously shown that fibroblasts taken from different sites interact with T cells in a different manner, such that synovial fibroblasts, but not skin fibroblasts, can support T cell pseudoemperipolesis (9). However, no comparison has yet been made between fibroblasts derived from the same anatomic sites but representing different diseases. Time-lapse phase-contrast microscopy showed that NK cells exhibited pseudoemperipolesis when in contact with autologous FLS from individuals with OA, SpA, and RA (Figure 1). NK cells from SpA and RA patients exhib- Figure 1. High levels of natural killer (NK) cell pseudoemperipolesis in fibroblast-like synoviocytes (FLS) from rheumatoid arthritis (RA) and spondylarthritis (SpA) patients, but not osteoarthritis (OA) patients, as determined by time-lapse phase-contrast microscopy. Pseudoemperipolesis was assessed by counting phase-dark cells (those that migrated into and under the fibroblast layer) in 3 independent fields. Results were expressed as a percentage of the total input NK cells cocultured with autologous FLS. Values are the mean and SD of triplicate determinations of samples from 6 SpA, 4 RA, and 4 OA patients. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01. ns ⫽ not significant. at 37°C as described previously (28). After culture, the CFSE-DA content and surface phenotype were simultaneously analyzed by FACS. NK cells were stimulated with 100 IU/ml of IL-2 (PeproTech, London, UK) as a positive control. CFSE-DA was used to determine the number of divisions that NK cells had undergone during culture. Cell division in CFSE-DA–labeled NK cells was calculated based on the sequential halving of fluorescence intensity in daughter cells; thus, the cell numbers in each peak were divided by the expected progeny for those divisions (divided by 2 for 1 division, by 4 for 2 divisions, by 8 for 3 divisions, etc.) as described elsewhere (29). Multiple cytokine bead assay. NK cells were cocultured with FLS, and supernatants were harvested on day 7 and stored at –80°C before being assayed. Cytokine assays were performed using the Bio-Plex (multiplex) cytokine assay according to the manufacturer’s instructions (Bio-Rad, Hercules, CA) and as previously described (30). Cytokines in the supernatants were measured with the Bio-Plex suspension array system instrument and were analyzed using Bio-Plex Manager software (both from Bio-Rad). Statistical analysis. Data were analyzed using Student’s 2-tailed t-test. When comparing more than 2 groups, one-way analysis of variance was performed. The level of significance was set at P ⬍ 0.05. Figure 2. Coculture of natural killer (NK) cells and fibroblast-like synoviocytes (FLS) from patients with spondylarthritis (SpA), rheumatoid arthritis (RA), and osteoarthritis (OA) and NK cell activation, as determined by fluorescence-activated cell sorter analysis. A, Expression of CD69 on NK cells cultured in media only, with FLS, or with FLS in the presence of interleukin-6 (IL-6)–blocking monoclonal antibody (mAb). Results are from a representative patient of 6 SpA, 4 RA, and 8 OA patients evaluated. Shaded histograms show the percentages of CD69⫹ NK cells; open histograms show NK cells stained with isotype control antibody. B, Expression of HLA–DR on NK cells cultured in media only, with FLS, or with FLS in the presence of IL-6–blocking mAb. In all cases, autologous NK cells and FLS were cocultured. Values are the mean and SD of cells from 6 SpA, 4 RA, and 8 OA patients. ⴱ ⫽ P ⬍ 0.05. MUTUAL STIMULATION OF FLS AND NK CELLS IN INFLAMMATORY ARTHRITIS 711 Table 2. Cytokine and stromal factor levels in cell culture supernatants, by diagnostic group* Analyte concentration NK cell ⫹ media NK cell ⫹ FLS FLS ⫹ media Analyte† SpA RA OA SpA RA OA SpA RA OA IL-1␤ IL-2 IL-4 IL-5 IL-6 IL-7 IL-8 IL-10 IL-12 IL-13 IL-17 IL-15 TNF␣ IFN␥ GM-CSF MIP-1␤ MCP-1 G-CSF VEGF RANTES IP-10 bFGF PDGF EGF MMP-3 MMP-9 MMP-13 1⫹ – – – – – – – – – – 1⫹ – 1⫹ – 1⫹ – – 1⫹ 2⫹ – – – – – 1⫹ – 1⫹ – – – – – – – – – – – – 1⫹ – 1⫹ – – 1⫹ 2⫹ – – – – – 1⫹ – – – – – – – – – – – – – – – – – – – 1⫹ 1⫹ – – – – – 1⫹ – 1⫹ – – 1⫹ 5⫹ – 6⫹ – – 3⫹ – 3⫹ 1⫹ 3⫹ 5⫹ 2⫹ 5⫹ 2⫹ 3⫹ 4⫹ 1⫹ – – 1⫹ 5⫹ 2⫹ 1⫹ 1⫹ – – 1⫹ 5⫹ – 6⫹ – – 2⫹ – 3⫹ 1⫹ 2⫹ 5⫹ 2⫹ 5⫹ 1⫹ 3⫹ 3⫹ 1⫹ – – 1⫹ 4⫹ 2⫹ 1⫹ 1⫹ – – 1⫹ 5⫹ – 5⫹ – – 1⫹ – 1⫹ 1⫹ 1⫹ 4⫹ 1⫹ 4⫹ 1⫹ 2⫹ 1⫹ 2⫹ – – 1⫹ 2⫹ 2⫹ 1⫹ 1⫹ – – – 5⫹ – 1⫹ – – – – – – – 3⫹ – 2⫹ – 1⫹ – – – – 1⫹ 1⫹ – – 1⫹ – – – 4⫹ – 1⫹ – – – – – – – 3⫹ – 2⫹ – 1⫹ – – – – 1⫹ 1⫹ – – 1⫹ – – – 4⫹ – 1⫹ – – – – – – – 3⫹ – 2⫹ – 1⫹ – – – – 1⫹ 1⫹ – – * Supernatants obtained on day 7 from natural killer (NK) cells cultured in R10, NK cells cocultured with fibroblast-like synoviocytes (FLS) in R10, and FLS cultured in R10 were analyzed for cytokine and stromal factors by multiple cytokine bead assay. Duplicate wells were analyzed for each sample, and background levels in wells containing only R10 were subtracted. Values are the mean of 3 independent experiments (each performed in duplicate) in 6 spondylarthritis (SpA), 4 rheumatoid arthritis (RA), and 8 osteoarthritis (OA) patients. Results are defined as follows: – ⫽ undetectable or ⬍1 pg/ml, 1⫹ ⫽ 1–50 pg/ml, 2⫹ ⫽ 50.1–100 pg/ml, 3⫹ ⫽ 100.1–500 pg/ml, 4⫹ ⫽ 500.1–1,000 pg/ml, 5⫹ ⫽ 1,000.1–5,000 pg/ml, and 6⫹ ⫽ 5,000.1–10,000 pg/ml. † IL-1␤ ⫽ interleukin-1␤; TNF␣ ⫽ tumor necrosis factor ␣; IFN␥ ⫽ interferon-␥; GM-CSF ⫽ granulocyte–macrophage colony-stimulating factor; MIP-1␤ ⫽ macrophage inflammatory protein 1␤; MCP-1 ⫽ monocyte chemotactic protein 1; G-CSF ⫽ granulocyte colony-stimulating factor; VEGF ⫽ vascular endothelial growth factor; IP-10 ⫽ IFN␥-inducible 10-kd protein; bFGF ⫽ basic fibroblast growth factor; PDGF ⫽ platelet-derived growth factor; EGF ⫽ epidermal growth factor; MMP-3 ⫽ matrix metalloproteinase 3. ited significantly increased levels of pseudoemperipolesis compared with those from OA patients (P ⬍ 0.05). NK cell activation and interactions between NK cells and RA or SpA FLS. To determine whether the NK cell/FLS interaction could lead to activation of NK cells, highly purified NK cells (⬎99% CD3–,CD56⫹) were cocultured with autologous FLS for 24 hours. CD69 and HLA–DR were used as surrogate markers of NK cell activation (31). As shown in Figure 2, the expression of CD69 and HLA–DR was increased on NK cells after coculture with FLS. This was significantly higher in the SpA and RA groups as compared with the OA group. Thus, a mean ⫾ SD of 69.4 ⫾ 4.8% (n ⫽ 6) and 64.5 ⫾ 6.9% (n ⫽ 4) of NK cells from the SpA and RA groups, respectively, expressed CD69 when cocultured with FLS, as compared with 51.7 ⫾ 4.8% (n ⫽ 8) of NK cells from the OA group (P ⬍ 0.05 for both comparisons) (Figure 2A). Significant increases in NK cell HLA–DR expression were also observed upon coculture with FLS, and these were also more marked in cells from the SpA and RA groups. Thus, 22.3 ⫾ 3.5% (n ⫽ 6) and 23.4 ⫾ 3.1% (n ⫽ 4) of NK cells from the SpA and RA groups, respectively, expressed HLA–DR when cocultured with FLS, as compared with 14.9 ⫾ 2.3% (n ⫽ 4) of NK cells from the OA group (P ⬍ 0.05 for both comparisons) (Figure 2B). Given that IL-6 was one of the most abundant cytokines produced by FLS under our experimental conditions (see Table 2), we wished to determine 712 CHAN ET AL Figure 3. Fibroblast-like synoviocyte (FLS) promotion of contact-dependent natural killer (NK) cell production of interferon-␥ (IFN␥) and proliferation, as determined by fluorescence-activated cell sorter analysis. A, Permeabilized CD3–,CD56⫹ NK cells showing IFN␥ secretion versus CD56 expression in NK cells cultured in media only, with autologous FLS, with FLS in the presence of a Transwell after 24 hours, or with phorbol myristate acetate (PMA)/ionomycin (ION) (positive control) after 6 hours. Results are from a representative patient of 6 spondylarthritis (SpA), 4 rheumatoid arthritis (RA), and 8 osteoarthritis (OA) patients evaluated. Values are the percentage of cells in each compartment. B, Intensity of 5,6-carboxyfluorescein diacetate N-succinimidyl ester (CFSE-DA) labeling of NK cells (shaded histograms) cultured with media only, with FLS, with FLS in the presence of a Transwell, or with 100 IU/ml of interleukin-2 (IL-2; positive control) after 7 days. Open histograms show the autofluorescence of unlabeled cultured NK cells. Arrow shows major cell populations that have divided once. Results are representative of 3 experiments using cells from 6 SpA, 4 RA, and 8 OA patients. Values are the percentage of cells that have undergone division. MUTUAL STIMULATION OF FLS AND NK CELLS IN INFLAMMATORY ARTHRITIS 713 Figure 4. Natural killer (NK) cell survival promoted by fibroblast-like synoviocytes (FLS) and up-regulation of interleukin-15 (IL-15) expression on FLS by coculture with NK cells, as determined by fluorescence-activated cell sorter analysis. A, CD3–,CD56⫹ NK cells showing annexin V staining versus 7-aminoactinomycin D (7-AAD) staining of NK cells cultured for 7 days in media only, with IL-15–blocking monoclonal antibody (mAb), with autologous FLS, with FLS in the presence of IL-15–blocking mAb, or with FLS in the presence of isotype control antibody (Ab). Values are the percentage of cells in each compartment. B, FLS showing cell surface IL-15 expression (shaded histograms) after culture for 7 days in media only or with NK cells. Open histograms show isotype control IgG1 antibody (background staining). Cells were first gated on CD56–, NKp30–, and NKp46 mAb. Results are representative of 3 experiments using cells from 6 spondylarthritis (SpA), 4 rheumatoid arthritis (RA), and 8 osteoarthritis (OA) patients. whether IL-6 was involved in the activation of NK cells in contact with FLS. Addition of saturating concentrations of neutralizing mAb against IL-6 to the NK cell/FLS coculture did not reduce the expression of CD69 or HLA–DR on NK cells. FLS stimulation of NK cell IFN␥ production and proliferation. We next sought to establish whether coculture of NK cells with FLS is associated with increased NK cell effector function. IFN␥ secretion from NK cells was measured by intracellular cytokine staining and FACS analysis. Figure 3A shows the results of a representative experiment of 3 that were performed for each patient. NK cells derived from patients in each of the arthritis groups did not constitutively secrete significant quantities of IFN␥ when cultured in media alone (overall mean ⬍1%; mean ⫾ SD in SpA patients, 0.9 ⫾ 0.3%, 714 in RA patients, 0.9 ⫾ 0.3%, and in OA patients, 0.5 ⫾ 0.2%). In coculture with FLS, intracellular IFN␥ was increased after 24 hours as compared with baseline. This was higher in SpA (14.6 ⫾ 2.3%) and RA (14.0 ⫾ 1.5%) patients than in OA patients (4.2 ⫾ 3.0%). Cocultures of NK cells and FLS separated by a 0.4-m Transwell did not result in a significant increase in NK cell secretion of IFN␥ (2.3 ⫾ 0.8% in SpA, 2.0 ⫾ 0.3% in RA, and 1.8 ⫾ 0.8% in OA), showing that NK cell/FLS contact is important in this interaction. Figure 3B shows that a proportion of CFSE-DA– labeled NK cells from SpA and RA patients divide upon culture with FLS. Over 20% of SpA and RA NK cells cultured with FLS underwent cell division, as compared with 5% of OA NK cells (Figure 3B). There was also a greater than 10-fold rise in the percentage of SpA and NK cells undergoing cell division when cultured with FLS as compared with media alone. The number of SpA and RA NK cells undergoing proliferation was reduced, but not abolished, when contact with FLS was lost through the presence of a Transwell. Promotion of NK cell survival by SpA and RA FLS and dependence upon IL-15. We next sought to establish whether the survival of NK cells is promoted by FLS. NK cell survival was promoted at a higher level in SpA and RA patients as compared with OA patients. A representative experiment from 6 SpA, 4 RA, and 8 OA samples is shown in Figure 4A. In all 3 forms of arthritis, NK cells cultured in media alone underwent high levels of apoptosis (mean ⫾ SD 33.7 ⫾ 3.6% [n ⫽ 6] in SpA, 31.5 ⫾ 4.4% [n ⫽ 4] in RA, and 36.7 ⫾ 3.4% [n ⫽ 8] in OA), and this was further increased in the presence of blocking anti–IL-15 mAb (68.7 ⫾ 2.7% in SpA, 68.2 ⫾ 1.5% in RA, and 71.1 ⫾ 3.6% in OA). In OA patients, 19.1 ⫾ 2.7% of NK cells cultured with FLS were undergoing apoptosis (annexin V⫹, 7-AAD⫹), as compared with 5.9 ⫾ 1.3% in SpA patients and 6.5 ⫾ 1.3% in RA patients (P ⬍ 0.01 for both comparisons). We wished to determine if IL-15 was implicated in the promotion of NK cell survival by FLS. We found that survival of NK cells was significantly inhibited by a combination of blocking anti–IL-15 mAb and a polyclonal goat IgG anti–IL-15R␣ (22.5 ⫾ 3.7% in SpA, 22.3 ⫾ 5.8% in RA, and 28.5 ⫾ 5.5% in OA; P not significant for SpA or RA versus OA), but not by an irrelevant isotype–matched mAb (6.2 ⫾ 1.0% in SpA, 7.0 ⫾ 2.4% in RA, and 15.3 ⫾ 2.2% in OA; P ⬍ 0.01 for SpA or RA versus OA). Up-regulation of cell surface IL-15 expression upon contact of FLS with NK cells. The expression of IL-15 on the surface of FLS was determined by FACS CHAN ET AL analysis of nonpermeabilized cells (Figure 4B). At rest, the surface expression of IL-15 on FLS was higher in SpA patients (mean ⫾ SD 6.8 ⫾ 1.8% [n ⫽ 6]) and RA patients (8.9 ⫾ 1.5% [n ⫽ 4]) as compared with that in OA patients (2.3 ⫾ 0.9% [n ⫽ 8]) (P ⬍ 0.005 for both comparisons). Surface IL-15 expression increased with time upon coculture with NK cells. Thus, on day 7, surface expression of IL-15 by FLS was 16.9 ⫾ 2.2% in SpA patients, 20.9 ⫾ 3.3% in RA patients, and 4.1 ⫾ 1.4% in OA patients (P ⬍ 0.001 for SpA or RA versus OA). Production of a proinflammatory cytokine milieu following interactions between RA and SpA FLS with NK cells. We next determined the cytokine profile resulting from interactions between NK cells and FLS. A total of 18 cytokines and 9 stromal factors were measured in culture supernatants using the multiplex assay. Supernatants obtained on day 7 from cultures of NK cells or FLS alone or from cocultures of NK cells and FLS were analyzed (Table 2). NK cells from all 3 arthritis groups cultured alone produced low but detectable amounts of vascular endothelial growth factor (VEGF), RANTES, and MMP-9. Additionally, SpA NK cells produced IL-15, and both SpA and RA NK cells produced low levels of IL-1␤, IFN␥, and MIP-1␤. FLS cultured alone produced significant quantities of IL-6 (greatest in SpA patients), GM-CSF, and MCP-1, as well as low levels of IL-1␤ and IL-8. NK cell/FLS coculture resulted in the production of high levels of IL-6, IL-8, GM-CSF, and MCP-1. In addition, small but significant levels of TNF␣, IL-5, IL-15, granulocyte colony-stimulating factor (G-CSF), and IFN␥-inducible 10-kd protein were now produced. For example, TNF␣ levels were increased in all NK cell/FLS coculture supernatants. Thus, the mean ⫾ SD TNF␣ levels in supernatants from cocultures of NK cells and FLS were 7.6 ⫾ 0.3 pg/ml in SpA and 8.8 ⫾ 1.2 pg/ml in RA. In OA, the TNF␣ levels in NK cell/FLS supernatants were significantly lower at 4.1 ⫾ 1.0 pg/ml (P ⬍ 0.001 versus SpA and versus RA). Up-regulation of MMP-3 in cocultures of NK cells and FLS from SpA patients. In the stromal factor panel, the level of MMP-3 was very substantially increased in cocultures of NK cells and FLS from patients with SpA (Table 2). Levels were higher than those in the RA and OA NK cell/FLS cocultures, and were ⬃1,000 times higher than the level produced by cultures of FLS alone. In addition, coculture produced increased levels of MMP-9 and detectable levels of MMP-13. MUTUAL STIMULATION OF FLS AND NK CELLS IN INFLAMMATORY ARTHRITIS DISCUSSION Fibroblast-like synoviocytes have previously been shown to play an important role in RA through T cell activation as well as the promotion of both T lymphocyte and neutrophil survival (26). This study is the first to show that FLS from patients with arthritis interact with syngeneic NK cells resulting in mutual stimulation and the release of proinflammatory cytokines. NK cells physically interact with FLS, undergoing pseudoemperipolesis, activation, IFN␥ production, proliferation, and increased survival, in addition to maturation, as described recently (27). Our data do not prove that NK cell–FLS contact is essential for these effects. However, in addition to observing their direct interaction, we found that their separation in Transwell cultures prevented NK cell IFN␥ production and proliferation, showing that at least close proximity is required. We also demonstrated increased surface expression of IL-15 by FLS upon coculture with NK cells, suggesting one possible mechanism, although further detailed studies of the mechanics will be required. All of these features were significantly enhanced in both RA and SpA cells as compared with OA cells, implying a preexisting or induced functional enhancement of either FLS or peripheral blood NK cells (or both) in these diseases. We demonstrated not only that basal levels of surface expression of IL-15 are higher in SpA and RA FLS than in OA FLS, but that the interaction with NK cells results in further increases in surface IL-15 expression on FLS. We further provide evidence that IL-15 promotes increased NK cell survival, since the presence of IL-15–blocking antibody resulted in increased apoptosis. However, the survival of NK cells promoted by FLS is not completely IL-15–dependent, since IL-15 blockade in NK cell/FLS cocultures resulted in a lower level of NK cell apoptosis as compared with IL-15 blockade in NK cells cultured alone. Previous studies have shown that IL-15 is crucial in NK cell survival (32,33). This increased expression of IL-15 on FLS may also act in an autocrine manner to promote FLS activation, proliferation, and resistance to apoptosis (34). Thus, our data support an important role of IL-15 in NK cell–FLS interactions in inflammatory arthritis but not in OA. Our data also suggest that strategies aimed at blocking this interaction (e.g., with anti–IL-15 mAb) are likely to be effective in reducing both inflammation and joint damage. The finding of increased NK survival induced by FLS further extends the role of fibroblasts in supporting the survival of leukocytes, which has already been shown 715 for T lymphocytes (9), neutrophils, dendritic cells, B lymphocytes, and plasma cells (26). Maintaining the inappropriate survival of leukocyte subpopulations is likely to be an important contribution by FLS to the persistence of chronic inflammatory diseases. The interaction of NK cells and FLS also results in NK cell activation, proliferation, and production of a proinflammatory milieu consisting of high levels of IL-6, IL-8, GM-CSF, MCP-1, and MMP-3. Production was 5–10 times higher in RA and SpA patient-derived samples than in OA patient-derived samples, except for IL-6. Increased production of IL-15, IFN␥, and VEGF was also seen in cells from SpA and RA patients, albeit at slightly lower levels. IL-6 was abundantly produced by FLS and is likely to be important in NK stimulation (35–37). However, we were unable to demonstrate blockade of NK cell activation with anti–IL-6 mAb, perhaps because the action of IL-6 on NK cell activation is mediated by the production of IL-2 by T cells (36). The high levels of IL-8 production observed in NK cell/FLS cocultures were most marked in SpA and RA and could play a significant role in attracting neutrophils into the joints in all 3 diseases. Interestingly, we also showed the ability of NK cell/FLS cocultures to produce large amounts of MCP-1. This was consistently greater in cocultures of NK cells and FLS from RA and SpA patients as compared with cells from OA patients. MCP-1, GM-CSF, and TNF␣ are involved in monocyte/dendritic cell activation and maturation (38,39). Therefore, our data suggest that NK cells and FLS could be involved in a cumulative and/or synergistic effect with monocytes in the promotion of inflammation. Of note, the production of type I IFNs (IFN␣ and IFN␤) by FLS was not determined in our study, but has previously been shown to be important in the survival of T cells in RA patients (10). Although our data were obtained in vitro and we used peripheral blood NK cells, our findings are consistent with the previously reported finding of activated NK cells within the joints of patients with inflammatory arthritis (38). In the future, our data should be confirmed using synovial fluid NK cells. Further study of NK receptor–ligand interactions, such as natural killer cell receptor group 2D/major histocompatibility complex class I chain–related molecule (40), KIR/HLA (19), and lymphocyte function⫺associated antigen 1/intercellular adhesion molecule 1, between NK cells and FLS, respectively, also further elucidate the mechanisms underlying the proinflammatory response seen in our study. Overall, our data suggest that similar modes and levels of immune activation occur following NK interac- 716 CHAN ET AL tions with FLS in RA and SpA patients. Although IL-13, IFN␥, G-CSF, RANTES, and MMP-3 production were all increased most markedly in SpA and more so in RA than in OA, it is uncertain, given the small sample size, whether these differences could be of clinical significance. It is more likely that the NK cell–FLS interaction amplifies and/or prevents resolution of the inflammatory processes in both SpA and RA. Our results also show differential production of MMP-3 by activated FLS, correlating with the elevated serum and synovial fluid levels of MMP-3 in patients with ankylosing spondylitis (41,42). In summary, our experiments show that in both RA and SpA patients, the interaction of FLS with NK cells results in enhanced NK cell survival and in the elaboration of multiple proinflammatory and chemotactic products, which may contribute to the persistence of inflammation. 9. 10. 11. 12. 13. 14. 15. AUTHOR CONTRIBUTIONS Drs. Chan and Bowness had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study design. Chan, Kollnberger, Buckley, Bowness. Acquisition of data. Chan, Filer, Parsonage, Gundle. Analysis and interpretation of data. Chan, Filer, Kollnberger, Buckley, Bowness. Manuscript preparation. Chan, Buckley, Bowness. Statistical analysis. Chan. REFERENCES 1. Kontoyiannis D, Kollias G. Fibroblast biology. Synovial fibroblasts in rheumatoid arthritis: leading role or chorus line? Arthritis Res 2000;2:342–3. 2. Mulherin D, FitzGerald O, Bresnihan B. Synovial tissue macrophage populations and articular damage in rheumatoid arthritis. Arthritis Rheum 1996;39:115–24. 3. Pap T, Muller-Ladner U, Gay RE, Gay S. Fibroblast biology: role of synovial fibroblasts in the pathogenesis of rheumatoid arthritis. Arthritis Res 2000;2:361–7. 4. Buckley CD, Pilling D, Lord JM, Akbar AN, Scheel-Toellner D, Salmon M. Fibroblasts regulate the switch from acute resolving to chronic persistent inflammation. Trends Immunol 2001;22: 199–204. 5. Buckley CD, Filer A, Haworth O, Parsonage G, Salmon M. Defining a role for fibroblasts in the persistence of chronic inflammatory joint disease. Ann Rheum Dis 2004;63 Suppl 2:ii92–5. 6. Sciaky D, Brazer W, Center DM, Cruikshank WW, Smith TJ. Cultured human fibroblasts express constitutive IL-16 mRNA: cytokine induction of active IL-16 protein synthesis through a caspase-3-dependent mechanism. J Immunol 2000;164:3806–14. 7. Pierer M, Rethage J, Seibl R, Lauener R, Brentano F, Wagner U, et al. Chemokine secretion of rheumatoid arthritis synovial fibroblasts stimulated by Toll-like receptor 2 ligands. J Immunol 2004;172:1256–65. 8. Buckley CD, Amft N, Bradfield PF, Pilling D, Ross E, ArenzanaSeisdedos F, et al. Persistent induction of the chemokine receptor 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. CXCR4 by TGF-␤1 on synovial T cells contributes to their accumulation within the rheumatoid synovium. J Immunol 2000; 165:3423–9. Bradfield PF, Amft N, Vernon-Wilson E, Exley AE, Parsonage G, Rainger GE, et al. Rheumatoid fibroblast-like synoviocytes overexpress the chemokine stromal cell–derived factor 1 (CXCL12), which supports distinct patterns and rates of CD4⫹ and CD8⫹ T cell migration within synovial tissue. Arthritis Rheum 2003;48: 2472–82. Pilling D, Akbar AN, Girdlestone J, Orteu CH, Borthwick NJ, Amft N, et al. Interferon-␤ mediates stromal cell rescue of T cells from apoptosis. Eur J Immunol 1999;29:1041–50. Jones SA. Directing transition from innate to acquired immunity: defining a role for IL-6. J Immunol 2005;175:3463–8. McLoughlin RM, Jenkins BJ, Grail D, Williams AS, Fielding CA, Parker CR, et al. IL-6 trans-signaling via STAT3 directs T cell infiltration in acute inflammation. Proc Natl Acad Sci U S A 2005;102:9589–94. Seaman WE. Natural killer cells and natural killer T cells [review]. Arthritis Rheum 2000;43:1204–17. Robertson MJ, Caligiuri MA, Manley TJ, Levine H, Ritz J. Human natural killer cell adhesion molecules: differential expression after activation and participation in cytolysis. J Immunol 1990;145:3194–201. Farag SS, Caligiuri MA. Human natural killer cell development and biology. Blood Rev 2006;20:123–37. Loza MJ, Zamai L, Azzoni L, Rosati E, Perussia B. Expression of type 1 (interferon ␥) and type 2 (interleukin-13, interleukin-5) cytokines at distinct stages of natural killer cell differentiation from progenitor cells. Blood 2002;99:1273–81. Loza MJ, Perussia B. Differential regulation of NK cell proliferation by type I and type II IFN. Int Immunol 2004;16:23–32. Robertson MJ. Role of chemokines in the biology of natural killer cells. J Leukoc Biol 2002;71:173–83. Chan AT, Kollnberger SD, Wedderburn LR, Bowness P. Expansion and enhanced survival of natural killer cells expressing the killer immunoglobulin-like receptor KIR3DL2 in spondylarthritis. Arthritis Rheum 2005;52:3586–95. Dougados M, van der Linden S, Juhlin R, Huitfeldt B, Amor B, Calin A, et al, and the European Spondylarthropathy Study Group. The European Spondylarthropathy Study Group preliminary criteria for the classification of spondylarthropathy. Arthritis Rheum 1991;34:1218–27. Van der Linden S, Valkenburg HA, Cats A. Evaluation of diagnostic criteria for ankylosing spondylitis: a proposal for modification of the New York criteria. Arthritis Rheum 1984;27:361–8. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31:315–24. Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K, et al. Development of criteria for the classification and reporting of osteoarthritis: classification of osteoarthritis of the knee. Arthritis Rheum 1986;29:1039–49. Scaife S, Brown R, Kellie S, Filer A, Martin S, Thomas AM, et al. Detection of differentially expressed genes in synovial fibroblasts by restriction fragment differential display. Rheumatology (Oxford) 2004;43:1346–52. Parsonage G, Falciani F, Burman A, Filer A, Ross E, Bofill M, et al. Global gene expression profiles in fibroblasts from synovial, skin and lymphoid tissue reveals distinct cytokine and chemokine expression patterns. Thromb Haemost 2003;90:688–97. Filer A, Parsonage G, Smith E, Osborne C, Thomas AM, Curnow SJ, et al. Differential survival of leukocyte subsets mediated by synovial, bone marrow, and skin fibroblasts: site-specific versus activation-dependent survival of T cells and neutrophils. Arthritis Rheum 2006;54:2096–108. MUTUAL STIMULATION OF FLS AND NK CELLS IN INFLAMMATORY ARTHRITIS 27. Chan A, Hong DL, Atzberger A, Kollnberger S, Filer AD, Buckley CD, et al. CD56bright human NK cells differentiate into CD56dim cells: role of contact with peripheral fibroblasts. J Immunol 2007;179:89–94. 28. Loza MJ, Perussia B. Final steps of natural killer cell maturation: a model for type 1-type 2 differentiation? Nat Immunol 2001;2: 917–24. 29. Warren HS. Using carboxyfluorescein diacetate succinimidyl ester to monitor human NK cell division: analysis of the effect of activating and inhibitory class I MHC receptors. Immunol Cell Biol 1999;77:544–51. 30. De Jager W, Hoppenreijs EP, Wulffraat NM, Wedderburn LR, Kuis W, Prakken BJ. Blood and synovial fluid cytokine signatures in patients with juvenile idiopathic arthritis: a cross-sectional study. Ann Rheum Dis 2006. E-pub ahead of print. 31. Fogli M, Costa P, Murdaca G, Setti M, Mingari MC, Moretta L, et al. Significant NK cell activation associated with decreased cytolytic function in peripheral blood of HIV-1-infected patients. Eur J Immunol 2004;34:2313–21. 32. Nguyen KB, Salazar-Mather TP, Dalod MY, Van Deusen JB, Wei XQ, Liew FY, et al. Coordinated and distinct roles for IFN-␣␤, IL-12, and IL-15 regulation of NK cell responses to viral infection. J Immunol 2002;169:4279–87. 33. Cooper MA, Bush JE, Fehniger TA, Van Deusen JB, Waite RE, Liu Y, et al. In vivo evidence for a dependence on interleukin 15 for survival of natural killer cells. Blood 2002;100:3633–8. 34. Kurowska M, Rudnicka W, Kontny E, Janicka I, Chorazy M, Kowalczewski J, et al. Fibroblast-like synoviocytes from rheumatoid arthritis patients express functional IL-15 receptor complex: endogenous IL-15 in autocrine fashion enhances cell proliferation and expression of Bcl-xL and Bcl-2. J Immunol 2002;169:1760–7. 35. Mihara M, Moriya Y, Kishimoto T, Ohsugi Y. Interleukin-6 (IL-6) 36. 37. 38. 39. 40. 41. 42. 717 induces the proliferation of synovial fibroblastic cells in the presence of soluble IL-6 receptor. Br J Rheumatol 1995;34:321–5. Luger TA, Krutmann J, Kirnbauer R, Urbanski A, Schwarz T, Klappacher G, et al. IFN-␤2/IL-6 augments the activity of human natural killer cells. J Immunol 1989;143:1206–9. Luger TA, Schwarz T, Krutmann J, Kirnbauer R, Neuner P, Kock A, et al. Interleukin-6 is produced by epidermal cells and plays an important role in the activation of human T-lymphocytes and natural killer cells. Ann N Y Acad Sci 1989;557:405–14. Dalbeth N, Gundle R, Davies RJ, Lee YC, McMichael AJ, Callan MF. CD56bright NK cells are enriched at inflammatory sites and can engage with monocytes in a reciprocal program of activation. J Immunol 2004;173:6418–26. Vitale M, Della Chiesa M, Carlomagno S, Pende D, Arico M, Moretta L, et al. NK-dependent DC maturation is mediated by TNF␣ and IFN␥ released upon engagement of the NKp30 triggering receptor. Blood 2005;106:566–71. Groh V, Bruhl A, El-Gabalawy H, Nelson JL, Spies T. Stimulation of T cell autoreactivity by anomalous expression of NKG2D and its MIC ligands in rheumatoid arthritis. Proc Natl Acad Sci U S A 2003;100:9452–7. Chen CH, Lin KC, Yu DT, Yang C, Huang F, Chen HA, et al. Serum matrix metalloproteinases and tissue inhibitors of metalloproteinases in ankylosing spondylitis: MMP-3 is a reproducibly sensitive and specific biomarker of disease activity. Rheumatology (Oxford) 2006;45:414–20. Vandooren B, Kruithof E, Yu DT, Rihl M, Gu J, De Rycke L, et al. Involvement of matrix metalloproteinases and their inhibitors in peripheral synovitis and down-regulation by tumor necrosis factor ␣ blockade in spondylarthropathy. Arthritis Rheum 2004; 50:2942–53.