ARTHRITIS & RHEUMATISM Vol. 52, No. 10, October 2005, pp 3004–3014 DOI 10.1002/art.21301 © 2005, American College of Rheumatology Pathogenic Role of the CXCL16–CXCR6 Pathway in Rheumatoid Arthritis Toshihiro Nanki,1 Takeshi Shimaoka,2 Kenji Hayashida,3 Ken Taniguchi,4 Shin Yonehara,2 and Nobuyuki Miyasaka1 Objective. Rheumatoid arthritis (RA) is a chronic inflammatory disease associated with massive T cell infiltration into the synovium. The accumulated T cells express type 1 cytokines, such as interferon-␥ (IFN␥) and tumor necrosis factor ␣, and activated markers of inflammation, such as CD154 and inducible costimulator (ICOS). It is thought that chemokines contribute to T cell accumulation in the synovium. In this study, we examined the role of CXCL16 and CXCR6 in T cell migration and stimulation in RA synovium. Methods. Expression of CXCL16 and CXCR6 was analyzed by immunohistochemistry, reverse transcription–polymerase chain reaction, Western blotting, and/or flow cytometry. Migration activity was assessed using a chemotaxis chamber. IFN␥ production was analyzed by enzyme-linked immunosorbent assay. The effect of anti-CXCL16 monoclonal antibody on murine collagen-induced arthritis (CIA) was evaluated. Results. CXCL16 was expressed in RA synovium. CXCR6 was expressed more frequently on synovial T cells than in peripheral blood. Moreover, CXCR6positive synovial T cells more frequently expressed CD154 and ICOS than did CXCR6-negative T cells. Stimulation with interleukin-15 (IL-15) up-regulated the expression of CXCR6 on peripheral blood T cells, and then stimulation with CXCL16 induced migration of IL-15–stimulated T cells and enhanced IFN␥ production. Furthermore, anti-CXCL16 monoclonal antibody significantly reduced the clinical arthritis score and reduced infiltration of inflammatory cells and bone destruction in the synovium of mice with CIA. Conclusion. Our results indicate that CXCL16 plays an important role in T cell accumulation and stimulation in RA synovium and suggest that CXCL16 could be a target molecule in new therapies for RA. Chemokines induce leukocyte migration and are generally classified as either C, CC, CXC, or CX3C chemokines, based on the conserved cysteine motifs (1). The majority of chemokines are secreted as small, soluble molecules. Two exceptional chemokines, fractalkine/CX3CL1 and CXCL16, are expressed on the cell surface as membrane-bound molecules (2–5). Membrane-bound CXCL16 induces firm adhesion of cells that express CXCR6 (6), a unique receptor for CXCL16. The membrane-bound form of CXCL16 is cleaved by ADAM-10 (7,8), and the cleaved soluble form of CXCL16 induces migration of activated T cells (4). It has been reported that CXCL16 is expressed on dendritic cells, macrophages, B cells, T cells, smooth muscle cells, and umbilical endothelial cells (4,5,9–11). CXCL16 was also independently identified as a scavenger receptor for phosphatidylserine and oxidized lipoprotein (SR-PSOX) (9). In contrast, CXCR6 is expressed on a subset of type 1 polarized peripheral blood T cells, natural killer (NK) T cells, NK cells, and B cells (12–15). CXCL16/SR-PSOX is considered to play a pathogenic role in atherogenesis (16,17), facilitation of the uptake of various pathogens (18), and development of inflammatory cardiac valvular disease (19) and experimental autoimmune encephalomyelitis (20). Rheumatoid arthritis (RA) is characterized by Supported in part by grants-in-aid from the Ministry of Health, Labor, and Welfare and the Ministry of Education, Science, Sports, and Culture, Japan. 1 Toshihiro Nanki, MD, Nobuyuki Miyasaka, MD: Tokyo Medical and Dental University, Tokyo, Japan; 2Takeshi Shimaoka, PhD, Shin Yonehara, PhD: Kyoto University, Kyoto, Japan; 3Kenji Hayashida, MD: Hoshigaoka Koseinenkin Hospital, Osaka, Japan; 4 Ken Taniguchi, MD: Tokyo Metropolitan Bokutoh Hospital, Tokyo, Japan. Address correspondence and reprint requests to Toshihiro Nanki, MD, Department of Medicine and Rheumatology, Graduate School, Tokyo Medical and Dental University, 1-5-45, Yushima, Bunkyo-ku, Tokyo 113-8519, Japan. E-mail: firstname.lastname@example.org. Submitted for publication January 4, 2005; accepted in revised form June 23, 2005. 3004 PATHOGENIC ROLE OF CXCL16 IN RA chronic inflammation of multiple joints. The affected synovium shows accumulation of T cells, B cells, macrophages, and dendritic cells (21,22). It was previously reported that synovial T cells are oligoclonally expanded (23), suggesting that T cells are antigenically stimulated in the synovium. Synovial T cells express mainly type 1 cytokines, such as interferon-␥ (IFN␥) and tumor necrosis factor ␣ (TNF␣) (24), and activated surface markers such as CD154 and inducible costimulator (ICOS) (25,26). Therefore, the accumulated synovial T cells might play a role in the pathogenesis of RA. T cell migration into the synovium is thought to involve the interaction of chemokines and chemokine receptors, and the roles of chemokines in RA have been analyzed extensively (27). The present study was designed to determine the role of CXCL16 and CXCR6 in the pathogenesis of RA. We observed abundant expression of CXCL16 and accumulation of CXCR6-positive CD4 and CD8 T cells in RA synovium compared with that in peripheral blood. Furthermore, CXCL16 induced T cell migration and enhanced IFN␥ production by IL-15–stimulated T cells. Moreover, blockade of CXCL16 reduced the severity of collagen-induced arthritis (CIA) in mice. These results indicate that the CXCL16–CXCR6 interaction may play an important role in T cell accumulation and stimulation in RA synovium, suggesting that CXCL16 could be a potentially useful therapeutic target in RA. MATERIALS AND METHODS Samples. All subjects provided informed consent. Peripheral blood samples were obtained from 6 female patients with RA, which was diagnosed according to the 1987 criteria of the American College of Rheumatology (formerly, the American Rheumatism Association) (28). Among the patients with RA, the mean ⫾ SEM age was 63.2 ⫾ 4.1 years, and the mean ⫾ SEM disease duration was 12.5 ⫾ 2.0 years. All 6 patients were seropositive for rheumatoid factor and had a mean ⫾ SEM C-reactive protein concentration of 3.8 ⫾ 0.7 mg/dl. Four of the patients with RA were receiving prednisolone, 5 were receiving methotrexate, 2 were being treated with sulfasalazine, and 1 was receiving D-penicillamine. Synovial tissue samples from patients with RA and patients with osteoarthritis (OA) were obtained at the time of total knee joint replacement surgery (signed consent forms were obtained before the operations). The experimental protocol was approved in advance by the ethics committee of the Tokyo Medical and Dental University. Immunohistochemical analysis. Immunohistochemical analysis was conducted on OCT compound–embedded sections of frozen synovial tissue. Briefly, 8-m–thick cryostat sections were fixed in 4% paraformaldehyde for 10 minutes, and then the samples were rehydrated in phosphate buffered saline (PBS) for 5 minutes, 3 times. Nonspecific binding was 3005 blocked with 1.5% H2O2 in methanol for 15 minutes, and then with 5% normal goat serum in PBS for 30 minutes. Serial sections were then incubated overnight at 4°C with rabbit polyclonal antiserum against human CXCL16 (9) or normal rabbit serum as a control. The incubation included 5% normal goat serum in PBS. The samples were then washed twice for 5 minutes in PBS, and expression was detected using the Envision⫹ kit (DakoCytomation, Carpinteria, CA). Sections were counterstained with hematoxylin for 5 seconds and washed in tap water for 10 minutes. Immunocytochemical analysis. Synovial tissue was minced and then incubated with 0.3 mg/ml of collagenase (Sigma, St. Louis, MO) for 1 hour at 37°C in Dulbecco’s modified Eagle’s medium (DMEM; Sigma). Partially digested pieces of tissue were pressed through a metal screen to obtain single-cell suspensions. In the next step, 1 ⫻ 105 cells were cultured overnight on 4-well chamber slides (Lab-Tek II Chamber Slide System; Nalge Nunc International, Naperville, IL). The chamber slides were fixed in 4% paraformaldehyde for 20 minutes, and then the samples were rehydrated in PBS for 5 minutes, 3 times. Nonspecific binding was blocked with nonfat milk for 60 minutes. The slides were incubated for 2 hours at room temperature with chicken anti-human CXCL16 IgY (GenWay Biotech, San Diego, CA) or normal chicken IgY as a control antibody at 61.5 g/ml in PBS–0.2% nonfat milk. The samples were then washed twice for 5 minutes in PBS– 0.2% nonfat milk and incubated with Alexa Fluor 488–conjugated goat anti-chicken IgY (Molecular Probes, Eugene, OR) at 20 g/ml for 1 hour at room temperature. Next, the slides were incubated with 10% normal goat serum in PBS for 60 minutes. For staining with CD68 or prolyl 4-hydroxylase, after washing twice in PBS for 5 minutes the slides were incubated with mouse anti-human CD68 monoclonal antibody (mAb) (KP1; DakoCytomation) at 15.6 g/ml, mouse anti-human prolyl 4-hydroxylase mAb (5B5; DakoCytomation) at 6.2 g/ ml, or mouse IgG1 for 2 hours. Subsequently, the samples were washed twice for 5 minutes in PBS and incubated with Alexa Fluor 647–conjugated goat anti-mouse IgG1 (Molecular Probes) at 20 g/ml for 1 hour at room temperature. The slides were examined using fluorescence microscopy (FluoView 300; Olympus, Tokyo, Japan). The number of Alexa Fluor 488– labeled CXCL16-positive cells among Alexa Fluor 647–labeled CD68-positive or prolyl 4-hydroxylase–positive cells was counted. CXCL16 expression on fibroblast-like synoviocytes (FLS). FLS were obtained from RA synovial tissue, as described previously (29). FLS were maintained in high-glucose DMEM with 10% fetal calf serum (FCS; Sigma) and used for experiments after 5 passages. RA FLS were cultured separately in 60-mm culture dishes at a density of 8 ⫻ 105 cells/well in DMEM with 10% FCS for 24 hours. The cells were then incubated in a medium supplemented, as indicated, with TNF␣ (5 ng/ml), interleukin-1␤ (IL-1␤; 5 ng/ml), IFN␥ (5,000 units/ ml), IL-18 (5 ng/ml), transforming growth factor ␤1 (TGF␤1; 1 ng/ml), or IL-10 (5 ng/ml) (PeproTech, Rocky Hill, NJ). After incubation at 37°C for 24 hours, total RNA was prepared from RA FLS using Isogen (Qiagen, Valencia, CA) and treated with DNase I (Invitrogen, Carlsbad, CA). Firststrand complementary DNA (cDNA) was synthesized using oligo(dT)12-18 primers (Amersham Pharmacia Biotech, Piscat- 3006 away, NJ) and SuperScript II reverse transcriptase (Invitrogen). The amount of cDNA for amplification was adjusted to the amount of RNA measured by optical density meter as well as ␤-actin polymerase chain reaction (PCR) products, using serially diluted cDNA. The cDNA was amplified with primers for CXCL16 (5⬘-CTGACTCAGCCAGGCAATGG and 3⬘TGAGTGGACTGCAAGGTGGA) or ␤-actin (5⬘-GTCCTCTCCCAAGTCCACACA and 3⬘-CTGGTCTCAAGTCAGTGTACAGGTAA). The PCR conditions have been described previously (30). The PCR products were then separated by electrophoresis through 1.5% agarose. After incubation at 37°C for 3 days, FLS were collected and lysed with extraction buffer (20 mM HEPES, 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 10 g/ml aprotinin). Synovial tissue obtained from patients with RA and patients with OA were also lysed with the extraction buffer. After 30 minutes at 4°C, debris was eliminated by centrifugation at 14,000 rpm for 30 minutes, and the supernatant was collected. After measuring the protein concentration with a protein assay kit (Bio-Rad, Richmond, CA), cell lysates were mixed with 6⫻ sample loading buffer containing 6% 2-methoxyestradiol and 10% sodium dodecyl sulfate (SDS), and stored at 4°C until analyzed. Twenty micrograms of protein was separated by 10% SDS–polyacrylamide gel electrophoresis, transferred to polyvinylidene difluoride membranes, and blocked overnight with 0.05% Tween 20 and 5% bovine serum albumin (BSA). The immunoblots were incubated with rabbit polyclonal antiserum against human CXCL16 (9) in PBS with 1% BSA for 1 hour. Subsequently, the immunoblots were incubated with horseradish peroxidase (HRP)–conjugated goat anti-rabbit IgG antibody (DakoCytomation) for 1 hour. The ␤-actin protein was detected by anti–␤-actin mAb (AC-15; Sigma) as first antibody and HRP-conjugated rabbit anti-mouse IgG antibody (DakoCytomation) as second antibody. All immunoblots were detected by enhanced chemiluminescence (Amersham Pharmacia Biotech). Signal intensities were quantified with NIH Image 1.63 (National Institutes of Health, Bethesda, MD), and the relative intensity of CXCL16 (the signal intensity of CXCL16 protein divided by the signal intensity of ␤-actin protein) was calculated. The fold increase in the relative intensity of CXCL16 protein (the relative intensity of CXCL16 with cytokine stimulation divided by the relative intensity of CXCL16 without stimulation) was calculated. Fluorescence-activated cell sorting (FACS) analysis. The following antibodies were used for FACS analysis: phycoerythrin (PE) Cy5–conjugated anti-CD4 mAb (13B8.2; Beckman Coulter, San Jose, CA), PE Cy5–conjugated anti-CD8 mAb (B9.11; Beckman Coulter), PE-conjugated anti-CXCR6 mAb (56811; R&D Systems, Minneapolis, MN), fluorescein isothiocyanate (FITC)–conjugated anti-CD154 mAb (24-31; eBioscience, San Diego, CA), FITC-conjugated anti-ICOS mAb (ISA-3; eBioscience), and isotype-matched control mAb. Peripheral blood mononuclear cells or synovial tissue cells were incubated with mAb for 20 minutes and rinsed with PBS–3% FCS. The stained cells were analyzed with a FACSCalibur (Becton Dickinson, San Jose, CA). Migration and stimulation by CXCL16. Peripheral blood mononuclear cells were isolated by Ficoll-Hypaque (Immuno-Biological Laboratories, Gunma, Japan) gradient NANKI ET AL centrifugation. Peripheral blood CD4 and CD8 T cells were purified by magnetic-activated cell sorting microbead-coupled mAb and magnetic cell separation columns (Miltenyi Biotec, Auburn, CA). The purified CD4 and CD8 T cells were incubated with 30 ng/ml of IL-15 (PeproTech) for 8 days in RPMI 1640 (Sigma) containing 10% FCS. Migration of the T cells through ECV304 cells was conducted as described previously (31). Briefly, 2 ⫻ 105 ECV304 cells were cultured in 24-well chemotaxis chambers (6.5-mm diameter, 5-m pore polycarbonate transwell culture insert; Costar, Cambridge, MA) for 48–72 hours in Medium 199 (Sigma) with 10% FCS. The migration medium (RPMI 1640:Medium 199 ⫽ 1:1; 0.5% BSA), supplemented where indicated with various concentrations of CXCL16 (PeproTech), was added to the lower wells. ECV304 cell–coated chemotaxis chambers were placed in each well, and 5 ⫻ 105 IL-15–stimulated CD4 or CD8 T cells suspended in migration medium were added to the upper wells. After a 2-hour incubation, the membrane was removed, the number of migrated cells was counted using FACSCalibur, and the migration index (the number of migrated cells divided by the number of cells that migrated without CXCL16) was calculated. IL-15–stimulated CD4 and CD8 T cells (2 ⫻ 105) were transferred to 96-well microtiter plates, which were precoated with anti-CD3 mAb (500 ng OKT3; Janssen Pharmaceutical, Tokyo, Japan) and CXCL16 (0, 2.5, or 25 pg per microwell), and supplemented where indicated with soluble CXCL16 (150 ng/ml). After incubation at 37°C for 24 hours, the concentration of IFN␥ in the culture supernatant was measured using an enzyme-linked immunosorbent assay (ELISA) kit (BioSource International, Camarillo, CA), and the fold increase in the IFN␥ concentration (the concentration of IFN␥ in the culture supernatant divided by the concentration of IFN␥ without CXCL16) was calculated. Induction of CIA. Eight-week-old male DBA1/J mice were purchased from Oriental Yeast (Tokyo, Japan). Bovine type II collagen (Collagen Research Center, Tokyo, Japan) was dissolved in 0.05M acetic acid at 4 mg/ml and emulsified in an equal volume of Freund’s complete adjuvant (Difco, Detroit, MI). Mice were immunized intracutaneously with 100 l of the emulsion (day 1). After 21 days (day 22), the same amount of bovine type II collagen emulsified in Freund’s complete adjuvant was injected as a booster immunization. The experimental protocol was approved by the Institutional Animal Care and Use Committee of Tokyo Medical and Dental University. The expression of CXCL16 and CXCR6 in tissue specimens obtained from around the ankle joints of mice with CIA was analyzed using reverse transcription (RT)–PCR, as described above. The primers used were as follows: for CXCL16, 5⬘-GGGAAGAGTTTTCACCACCA and 3⬘-GGTTGGGTGTGCTCTTTGTT; for CXCR6, 5⬘-GCTTGCTCATTTGGGTGGTC and 3⬘-TCTGGGTCAGCAGGAACACA; and for GAPDH, 5⬘-ACCCAGAAGACTGTGGATGG and 3⬘-GTCATCATCCTTGGCAGGTT. CXCL16 expression was also analyzed by Western blotting, as described above. Rat anti-mouse CXCL16 mAb (142417; R&D Systems) was used as first antibody, and anti-rat IgG–HRP (DakoCytomation) was used as second antibody. Expression of ␤-actin as a loading control was also analyzed, as described above. PATHOGENIC ROLE OF CXCL16 IN RA 3007 Figure 1. Expression of CXCL16 in rheumatoid arthritis (RA) synovium. A–C, Synovial tissue samples obtained from 6 patients with RA were stained with polyclonal antiserum against human CXCL16 (A and C) or with normal rabbit serum as a control (B). All sections were counterstained with hematoxylin. Photomicrographs representative of those with similar staining patterns are shown (original magnification ⫻ 100 in A and B; ⫻ 200 in C). D and E, Single-cell suspensions from RA synovial tissue were cultured overnight on 4-well chamber slides. The slides were double stained for CD68, a macrophage marker (D), or prolyl 4-hydroxylase, a fibroblast marker (E), and CXCL16, and analyzed with fluorescence microscopy. Representative staining results from 4 patients with RA (original magnification ⫻ 200) are shown. F, The number of CXCL16-positive cells among CD68-positive or prolyl 4-hydroxylase–positive cells in samples from 4 patients with RA was counted. The numbers of CXCL16-positive cells/numbers of counted CD68-positive or prolyl 4-hydroxylase–positive cells are shown. G, The expression of CXCL16 in synovial tissue obtained from 6 patients with RA and 3 patients with osteoarthritis (OA) was analyzed by Western blotting. Treatment with anti-mouse CXCL16 mAb. Five hundred micrograms of rat anti-mouse CXCL16 mAb (20) or control antibody (rat IgG; ICN Pharmaceuticals, Aurora, OH) was injected into the peritoneal cavity of mice 3 times weekly, from day 15 to day 39. Anti-mouse CXCL16 mAb inhibited migration of murine CXCR6-expressed L1.2 cells induced by murine CXCL16 (20). Mice were observed for signs of arthritis. Disease severity for each limb was recorded as follows: 0 ⫽ normal, 1 ⫽ erythema and mild swelling confined to the midfoot (tarsal joints) or ankle joint, 2 ⫽ erythema and mild swelling extending from the ankle to the midfoot, 3 ⫽ erythema and moderate swelling extending from the ankle to the metatarsal joint, and 4 ⫽ erythema and severe swelling of the ankle, foot, and digits. The clinical arthritis score was defined as the sum of the scores for all 4 paws of each mouse. On day 41, the thickness of the hind paw was measured. On days 18, 27, and 41, the ankle joints were harvested and examined histologically, after hematoxylin and eosin staining. The severity of arthritis was evaluated using the following 2 parameters: synovial inflammation (0 ⫽ normal, 1 ⫽ focal inflammatory infiltrates, and 2 ⫽ inflammatory infiltrate that dominates the cellular histology) and bone destruction (0 ⫽ normal, 1 ⫽ small areas of bone destruction, 2 ⫽ widespread bone destruction), as described previously (32,33). 3008 NANKI ET AL Figure 2. Expression of CXCL16 on rheumatoid arthritis (RA) fibroblast-like synoviocytes (FLS). RA FLS were incubated in 60-mm dishes with tumor necrosis factor ␣ (TNF␣; 5 ng/ml), interleukin-1␤ (IL-1␤; 5 ng/ml), interferon-␥ (IFN␥; 5,000 units/ml), IL-18 (5 ng/ml), transforming growth factor ␤1 (TGF␤1; 1 ng/ml), or IL-10 (5 ng/ml). After a 24-hour stimulation, the expression of CXCL16 mRNA was analyzed using reverse transcription–polymerase chain reaction (PCR). A, PCR products were separated by electrophoresis through 1.5% agarose. B, After a 3-day stimulation, cell lysates were analyzed by Western blotting with antiserum against CXCL16. Representative results from 2–3 independent experiments are shown. C, Signal intensities of CXCL16 protein were normalized to those of ␤-actin protein, and the fold increase in relative CXCL16 intensities was calculated. Top, Relative CXCL16 intensity with cytokine stimulation divided by the relative CXCL16 intensity without stimulation. Bottom, Relative CXCL16 intensity with TGF␤1 ⫹ TNF␣ or IL-1␤ stimulation divided by the relative CXCL16 intensity with TNF␣ or IL-1␤ stimulation alone. Values are the mean ⫾ SEM. ⴱ ⫽ P ⬍ 0.05. Statistical analysis. Wilcoxon’s signed rank test was used to compare paired samples of peripheral blood and synovium from the same subjects for the expression of CXCR6, CD154, and ICOS on CD4 and CD8 T cells; the expression of CD154 and ICOS on CXCR6-negative versus CXCR6-positive synovial CD4 and CD8 T cells; and the expression of CXCR6 with versus without IL-15 stimulation. Differences in the migration index and IFN␥ production induced by CXCL16 stimulation and in the arthritis score and histologic score between control antibody– and anti-mouse CXCL16 mAb–treated mice were examined for statistical significance using the Mann-Whitney U test. All data are expressed as the mean ⫾ SEM. P values less than 0.05 were considered significant. RESULTS Expression of CXCL16 in RA synovium. The expression of CXCL16 in synovial tissue from patients with RA was assessed by immunohistochemical analysis. CXCL16 expression was widespread in all of the RA synovium samples analyzed, with dense staining of the synovial lining cells and moderate staining of the sublining cells (Figures 1A and C). To identify CXCL16expressing synoviocytes, a single-cell suspension of RA synoviocytes was cultured overnight on chamber slides. The cells were then double stained for CXCL16 and CD68, a macrophage marker, or prolyl 4-hydroxylase, a fibroblast marker. Both CD68-positive macrophage-like synoviocytes and prolyl 4-hydroxylase–positive FLS expressed CXCL16 (Figures 1D and E). The frequency of CXCL16-positive cells in macrophage-like synoviocytes or FLS from 4 patients with RA was analyzed. On average, 43% of macrophage-like synoviocytes and 49% of FLS expressed CXCL16 (Figure 1F). We next compared the expression level of CXCL16 in synovial tissue from patients with RA and patients with OA. CXCL16 was highly expressed in RA synovium compared with OA synovium, in which it was weakly expressed (Figure 1G). In order to examine the effect of cytokine stimulation on the expression of CXCL16, the expression of CXCL16 messenger RNA (mRNA) and protein on RA FLS was analyzed using RT-PCR and Western blotting. FLS were stimulated with TNF␣, IL-1␤, IFN␥, IL-18, TGF␤1, or IL-10, all of which are expressed in RA synovium. CXCL16 was expressed on unstimulated FLS in vitro. Stimulation with TNF␣ and IL-1␤ significantly up-regulated the expression of CXCL16 mRNA (Figure 2A) and protein (Figures 2B and C). In contrast, stimulation with TGF␤1 down-regulated the expression of CXCL16. Moreover, PATHOGENIC ROLE OF CXCL16 IN RA Figure 3. Expression of CXCR6 on CD4 and CD8 T cells. A, The proportion of CXCR6-expressing CD4 and CD8 T cells in peripheral blood (PB) and synovium (SY) was assessed by fluorescence-activated cell sorting (FACS) (n ⫽ 6 samples). B, The proportion of CD154expressing and inducible costimulator (ICOS)–expressing CXCR6negative and CSCR6-positive synovial CD4 and CD8 T cells was analyzed by FACS (n ⫽ 6). Each symbol represents an individual subject. TGF␤1 decreased the expression of CXCL16 on TNF␣and IL-1␤–stimulated FLS. Stimulation with IFN␥, IL18, and IL-10 did not change the expression of CXCL16. These results indicate that macrophage-like synoviocytes and FLS in RA synovium express CXCL16, and that the production of CXCL16 by FLS is regulated by TNF␣, IL-1␤, and TGF␤1. 3009 Expression of CXCR6 by CD4 and CD8 T cells. The expression of CXCR6 on peripheral blood and synovial CD4 and CD8 T cells from patients with RA was analyzed by flow cytometry. A few peripheral blood CD4 and CD8 T cells expressed CXCR6. The proportion of CXCR6-positive CD4 and CD8 T cells in synovium was significantly higher than the proportion of these cells in peripheral blood (Figure 3A). It was previously reported that activated T cells express CD154 and ICOS (34,35). We used flow cytometry to analyze CD154 and ICOS expression on peripheral blood and synovial CD4 and CD8 T cells from patients with RA. Few peripheral blood CD4 and CD8 T cells expressed CD154 (mean ⫾ SEM 0.4 ⫾ 0.2% and 1.5 ⫾ 0.6%, respectively; n ⫽ 6 samples) and ICOS (0.8 ⫾ 0.2% and 0.9 ⫾ 0.3%, respectively; n ⫽ 6). In contrast, synovial CD4 and CD8 T cells more frequently expressed CD154 (mean ⫾ SEM 14.4 ⫾ 4.6% and 14.5 ⫾ 5.2%, respectively [P ⬍ 0.05]; n ⫽ 6) and ICOS (15.3 ⫾ 2.4% and 10.8 ⫾ 2.2%, respectively [P ⬍ 0.05]; n ⫽ 6) compared with peripheral blood CD4 and CD8 T cells. To analyze the association between CXCR6 and the expression of CD154 or ICOS, we compared the proportions of CD154-expressing and ICOS-expressing CXCR6-negative and CXCR6-positive synovial CD4 and CD8 T cells. The proportions of CD154-expressing and ICOS-expressing CXCR6-positive CD4 and CD8 T cells were significantly higher than the proportions of CXCR6-negative CD4 and CD8 T cells, respectively (Figure 3B). These results showed that CXCR6 was more frequently expressed on synovial T cells, and that in RA synovium, CXCR6-positive T cells might be more activated than CXCR6-negative T cells. T cell stimulation by CXCL16. Previous studies demonstrated the expression of IL-15 on synoviocytes and synovial endothelial cells in RA (36,37). Although a few peripheral blood T cells expressed CXCR6 (Figure 3A), stimulation with 30 ng/ml of IL-15 for 8 days significantly increased CXCR6 expression in vitro (for CD4, mean ⫾ SEM 30.2 ⫾ 6.5%; for CD8, 16.9 ⫾ 1.4% [P ⬍ 0.05]; n ⫽ 5) compared with expression without IL-15 stimulation (for CD4, 4.0 ⫾ 1.0%; for CD8, 7.1 ⫾ 2.0%; n ⫽ 5); these findings are similar to those previously reported (38). Although IL-15 stimulation upregulated CXCR6 expression by both CD154-negative and CD154-positive CD4 and CD8 T cells (for CD4, mean ⫾ SEM 18.1 ⫾ 5.7% and 32.8 ⫾ 10.5%, respectively; for CD8, 14.9 ⫾ 0.9% and 42.1 ⫾ 7.8%, respectively [P ⬍ 0.05]; n ⫽ 3) compared with unstimulated T cells (for CD4, 0.5 ⫾ 0.1% and 2.0 ⫾ 2.0%, respectively; 3010 Figure 4. CXCL16-induced T cell migration and stimulation. A, Peripheral blood CD4 and CD8 T cells were stimulated with interleukin-15 (IL-15) for 8 days. Next, the T cells were cultured with various concentrations of CXCL16 for 2 hours. The number of cells migrating through ECV304-coated transwells was assessed by fluorescence-activated cell sorting. The migration index (the number of cells that migrated divided by the number of cells without CXCL16 that migrated) for 4 donors was calculated. B, IL-15–stimulated CD4 and CD8 T cells were transferred to microtiter plates coated with 500 g of anti-CD3 monoclonal antibody and various amounts of CXCL16, and were cultured with or without soluble CXCL16 for 24 hours. The interferon-␥ (IFN␥) concentration in the culture supernatant was analyzed by enzyme-linked immunosorbent assay. The fold increase in the IFN␥ concentration (IFN␥ concentration divided by the concentration of IFN␥ without CXCL16) for 3–5 donors was calculated. Values are the mean ⫾ SEM. ⴱ ⫽ P ⬍ 0.05. for CD8, 1.7 ⫾ 0.3% and 5.4 ⫾ 2.5%, respectively; n ⫽ 3), IL-15-stimulated CXCR6-positive T cells expressed CD154 more frequently than IL-15-stimulated CXCR6negative T cells (for CD4, mean ⫾ SEM 3.0 ⫾ 1.2% and 17.4 ⫾ 6.1%, respectively; for CD8, 0.4 ⫾ 0.3% and 3.2 ⫾ 0.5%, respectively [P ⬍ 0.05]; n ⫽ 3). Similarly, CXCR6 expression was also upregulated by both ICOS-negative and ICOS-positive CD4 and CD8 T cells (for CD4, mean ⫾ SEM 20.3 ⫾ 7.6% and 59.4 ⫾ 8.3%, respectively; for CD8, 15.7 ⫾ 1.0% and 56.1 ⫾ 6.7%, respectively [P ⬍ 0.05]; n ⫽ 3) compared with unstimulated T cells (for CD4, 0.4 ⫾ 0.2% and 1.6 ⫾ 1.6%, respectively; for CD8, 1.7 ⫾ 0.5% and 0.0 ⫾ 0.0%, respectively; n ⫽ 3). Stimulated CXCR6positive T cells expressed ICOS more frequently than did CXCR6-negative T cells (for CD4, 16.2 ⫾ 1.5% and 30.4 ⫾ 4.1%, respectively; for CD8, 0.8 ⫾ 0.6% and 3.4 ⫾ 0.7%, respectively [P ⬍ 0.05]; n ⫽ 3). NANKI ET AL In order to analyze the functional effects of CXCL16 on IL-15–stimulated CXCR6-positive T cells, we analyzed the effects of CXCL16 on migration activity and IFN␥ production. Soluble CXCL16 induced the migration of both peripheral blood CD4 T cells and peripheral blood CD8 T cells, which were cultured with IL-15 for 8 days (Figure 4A). Without stimulation of precoated anti-CD3 mAb during ELISA, IL-15– stimulated CD4 or CD8 T cells did not produce IFN␥ when incubated with soluble CXCL16 or precoated CXCL16. Although stimulation with anti-CD3 mAb induced IFN␥ production by IL-15–stimulated CD4 and CD8 T cells, culture with soluble CXCL16 did not change the level of IFN␥ production by the cells. However, culture in wells precoated with CXCL16 enhanced IFN␥ production by anti-CD3 mAb-stimulated peripheral blood CD4 and CD8 T cells stimulated with IL-15 (Figure 4B). Soluble CXCL16 blocked the enhancement of IFN␥ production by precoated CXCL16. These results suggest that peripheral blood T cells could respond to stimulation by IL-15 in the synovium. Stimulation with IL-15 up-regulates CXCR6 expression, and CXCL16 induces migration and enhances IFN␥ production by CXCR6-positive T cells. Effect of anti-mouse CXCL16 mAb on mice with CIA. We first analyzed the expression of CXCL16 and CXCR6 in the synovium of mice with CIA. The expression of CXCL16 was higher in synovium from mice with CIA compared with that in synovium from normal mice (Figures 5A and B). Furthermore, the expression of CXCR6 mRNA was also up-regulated in synovium from mice with CIA (Figure 5A). In order to analyze the effect of inhibition of the CXCL16–CXCR6 interaction on mice with CIA, 500 g of anti-mouse CXCL16 mAb or control antibody was injected into the peritoneal cavity 3 times per week, from day 15 to day 39. Clinical scores for arthritis were recorded. Treatment with anti-CXCL16 mAb significantly reduced the clinical arthritis score compared with treatment with control antibody (Figure 5C). On day 41, the thickness of the hind paw was measured. The paw thickness was also significantly suppressed by treatment with antiCXCL16 mAb compared with control antibody treatment (Figure 5D). Finally, on day 18, day 27, and day 41 the ankle joints were harvested, stained with eosin and hematoxylin, and examined microscopically. On day 18, mild focal inflammatory infiltrates were observed in joints from a few mice, without a significant difference between the mice treated with anti-CXCL16 mAb and those that PATHOGENIC ROLE OF CXCL16 IN RA 3011 Figure 5. Expression of CXCL16 and CXCR6 in mice with collagen-induced arthritis (CIA), and inhibition of clinical arthritis in mice with CIA by treatment with anti-CXCL16 monoclonal antibody (mAb). A and B, The expression of CXCL16 and CXCR6 mRNA in synovium from 3 normal mice and 3 mice with CIA was analyzed by reverse transcription–polymerase chain reaction (PCR). PCR products were separated by electrophoresis through 1.5% agarose (A), and the expression of CXCL16 in the synovium was analyzed by Western blotting with anti-CXCL16 antibody (Ab) (B). C and D, Five hundred micrograms of rat anti-mouse CXCL16 mAb or control antibody was injected into the peritoneal cavity 3 times per week from day 15 to day 39. Disease severity was recorded as an arthritis score until day 41 (C), and hind paw thickness was measured on day 41 (D). Values are the mean ⫾ SEM of 20 animals in 2 separate experiments. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01. received control antibody treatment (Figures 6A and B). On day 27, mild-to-moderate mononuclear cell infiltration was observed in the joints from some mice (Figures 6C and D). Treatment with anti-CXCL16 mAb tended to reduce the histology score compared with control antibody treatment, although the difference was not statistically significant (Figure 6G). On day 41, the joints from control antibody–treated mice showed massive mononuclear cell infiltration and bone destruction (Figure 6E). In comparison, the joints from mice treated with anti-CXCL16 mAb showed milder histologic changes (Figure 6F). Analysis of the histology scores indicated that anti-CXCL16 mAb treatment significantly reduced both synovial inflammation and bone destruction in the ankle joint (Figures 6G and H). These results indicate that treatment with anti-CXCL16 mAb reduces the clinical arthritis score and histologic changes in mice with CIA. DISCUSSION The major findings of the present study were as follows: 1) expression of CXCL16 was abundant in synovium from patients with RA, 2) the proportion of CXCR6-positive CD4 and CD8 T cells in the synovium was higher than the proportion in peripheral blood, 3) the expression of CXCR6 on peripheral blood T cells was up-regulated by stimulation with IL-15, 4) CXCL16 induced migration of T cells and enhanced IFN␥ production by IL-15–stimulated T cells, and 5) blockade of 3012 Figure 6. Histologic changes induced by treatment with anti-CXCL16 monoclonal antibody (mAb). A–F, Ankle joints harvested on day 18 (A and B), day 27 (C and D), and day 41 (E and F) were stained with hematoxylin and eosin. Representative photomicrographs from mice treated with control antibody (A, C, and E) and mice treated with anti-CXCL16 mAb (B, D, and F) are shown. Mononuclear cell (MC) infiltration and bone erosion (BE) are indicated. (Original magnification ⫻ 40.) G, Scores for synovial inflammation. H, Scores for bone destruction. Values are the mean ⫾ SEM scores in 7–20 samples. ⴱ ⫽ P ⬍ 0.05. CXCL16–CXCR6 interaction ameliorated murine CIA. These results indicate that CXCL16 may play an important role in T cell migration and stimulation in the synovium of patients with RA. The proportion of CXCR6-positive CD4 and CD8 T cells in RA synovium was higher than that in peripheral blood. Synovial CXCR6-positive CD4 and NANKI ET AL CD8 T cells, respectively, expressed CD154 and ICOS, which are markers of T cell activation, more frequently than did CXCR6-negative CD4 and CD8 T cells. Moreover, previous studies showed that CXCR6 was expressed by peripheral blood memory T cells, which could express type 1 cytokines (12). Taken together, these results indicate that CXCL16 could induce migration of CXCR6-positive T cells (which are activated, memory, and type 1 cytokine–producing T cells) from peripheral blood into synovial tissue. It is also possible that migrated CD4 and CD8 T cells are stimulated with IL-15, which is expressed in the RA synovium (36,37), and then express CXCR6. Consequently, CXCL16 could retain the CXCR6-positive CD4 and CD8 T cells in the synovial tissue. CXCL16 could also enhance IFN␥ production by IL-15–stimulated CD4 and CD8 T cells. We could not measure the CXCL16 concentration in RA synovial tissue. However, chemokines bind surface proteoglycans (39), and they could therefore be sequestered and presented to target cells at high concentrations within the local microenvironment. Consequently, we suggest that the CXCL16–CXCR6 interaction might contribute to CD4 and CD8 T cell accumulation and stimulation in the RA synovium. A high proportion of CCR5-positive T cells in RA synovium has also been reported (30,40–43). Furthermore, Unutmaz et al (38) showed that IL-15 stimulation increased the number of CXCR6 and CCR5 double-positive CD4 and CD8 T cells in vitro. These results suggest that CCR5 could also have an important role in T cell migration into the synovium. However, the severity of murine CIA was not altered in genetically engineered CCR5-deficient mice (44). Moreover, some patients with RA do not have functional CCR5 because of a homozygous ⌬32 deletion (45,46). Thus, blockade of CCR5 only may not be sufficient in T cell target therapy for RA. Based on results of the present study, we believe that CXCR6 is potentially a better target molecule for RA treatment aimed at T cell inhibition, because inhibition of the CXCL16–CXCR6 interaction was effective in CIA. Our results showed that CXCL16 is expressed on synovial macrophage-like synoviocytes and FLS in patients with RA and that its production by FLS is regulated by TNF␣, IL-1␤, and TGF␤1. Previous studies demonstrated the expression of CXCL16 on dendritic cells, macrophages, T cells, B cells, smooth muscle cells, and umbilical endothelial cells (4,5,9–11). FLS from patients with RA are additional and new members of the group of cells producing CXCL16. Because membranous CXCL16 is expressed on the cell surface (4,5,9), PATHOGENIC ROLE OF CXCL16 IN RA macrophage-like synoviocytes and FLS could activate T cells by cell–cell contact via the CXCL16–CXCR6 interaction in the synovial tissue. Furthermore, membranebound CXCL16 can be cleaved by ADAM-10 (7,8), and thus, soluble CXCL16 subsequently induces the migration and stimulation of CXCR6-positive T cells in the RA synovium. Treatment with anti-CXCL16 mAb reduced the clinical arthritis score and the histopathology score in mice with CIA. Blockade of the CXCL16–CXCR6 interaction might inhibit T cell migration into the synovium and inhibit T cell stimulation by CXCL16, suggesting that CXCL16 inhibition could decrease the number of activated T cells in the synovium. It is interesting that anti-CXCL16 mAb reduced not only cellular infiltration, but also bone destruction in mice with CIA. Kotake et al (47) reported that RANKL is expressed on activated T cells in the RA synovium, and that such expression induces differentiation and activation of osteoclasts. It is possible that blockade of the CXCL16–CXCR6 interaction can reduce RANKL expression on synovial T cells by inhibiting T cell stimulation or reducing the number of activated T cells in the RA synovium. Therefore, treatment with CXCL16 inhibitors could also prevent bone destruction, but further study of this is necessary. Our group previously reported that treatment with anti-fractalkine (CX3CL1) mAb ameliorated CIA and reduced macrophage migration into the synovium (48). Matthys et al (49) showed that CXCR4 antagonist inhibited CIA in IFN␥ receptor–deficient mice. Inhibition of fractalkine and CX3CR1, as well as the stromal cell–derived factor 1–CXCR4 interaction, might prevent the migration of not only T cells, but also macrophages into the synovium. In contrast, CXCL16 inhibition may affect only T cells and not macrophages. In fact, antiCXCL16 mAb reduced the arthritis score up to 25%, although inhibition of the fractalkine–CX3CR1 interaction reduced the arthritis score up to 50%. Collectively, these results suggest that the combination of macrophage inhibition and blockade of the CXCL16–CXCR6 interaction may be a better treatment strategy for RA compared with either of these approaches individually. In conclusion, in the present study, we observed expression of CXCL16 in RA synovium and demonstrated that such expression might induce the migration and stimulation of synovial T cells. Moreover, we showed that inhibition of CXCL16 significantly improved the clinical arthritis score and the histopathology score in mice with CIA. Thus, blockade of the CXCL16– CXCR6 interaction could be a new, potentially useful therapy for RA. 3013 ACKNOWLEDGMENT We thank Fumiko Inoue for excellent technical support. REFERENCES 1. Yoshie O, Imai T, Nomiyama H. Chemokines in immunity. Adv Immunol 2001;78:57–110. 2. Bazan JF, Bacon KB, Hardiman G, Wang W, Soo K, Rossi D, et al. A new class of membrane-bound chemokine with a CX3C motif. Nature 1997;385:640–4. 3. Pan Y, Lloyd C, Zhou H, Dolich S, Deeds J, Gonzalo JA, et al. Neurotactin, a membrane-anchored chemokine upregulated in brain inflammation. Nature 1997;387:611–7. 4. Matloubian M, David A, Engel S, Ryan JE, Cyster JG. A transmembrane CXC chemokine is a ligand for HIV-coreceptor Bonzo. Nat Immunol 2000;1:298–304. 5. Wilbanks A, Zondlo SC, Murphy K, Mak S, Soler D, Langdon P, et al. Expression cloning of the STRL33/BONZO/TYMSTRligand reveals elements of CC, CXC, and CX3C chemokines. J Immunol 2001;166:5145–54. 6. Shimaoka T, Nakayama T, Fukumoto N, Kume N, Takahashi S, Yamaguchi J, et al. Cell surface-anchored SR-PSOX/CXC chemokine ligand 16 mediates firm adhesion of CXC chemokine receptor 6-expressing cells. J Leukoc Biol 2004;75:267–74. 7. Gough PJ, Garton KJ, Wille PT, Rychlewski M, Dempsey PJ, Raines EW. A disintegrin and metalloproteinase 10-mediated cleavage and shedding regulates the cell surface expression of CXC chemokine ligand 16. J Immunol 2004;172:3678–85. 8. Abel S, Hundhausen C, Mentlein R, Schulte A, Berkhout TA, Broadway N, et al. The transmembrane CXC-chemokine ligand 16 is induced by IFN-␥ and TNF-␣ and shed by the activity of the disintegrin-like metalloproteinase ADAM10. J Immunol 2004;172: 6362–72. 9. Shimaoka T, Kume N, Minami M, Hayashida K, Kataoka H, Kita T, et al. Molecular cloning of a novel scavenger receptor for oxidized low density lipoprotein, SR-PSOX, on macrophages. J Biol Chem 2000;275:40663–6. 10. Shashkin P, Simpson D, Mishin V, Chesnutt B, Ley K. Expression of CXCL16 in human T cells. Arterioscler Thromb Vasc Biol 2003;23:148–9. 11. Hofnagel O, Luechtenborg B, Plenz G, Robenek H. Expression of the novel scavenger receptor SR-PSOX in cultured aortic smooth muscle cells and umbilical endothelial cells. Arterioscler Thromb Vasc Biol 2002;22:710–1. 12. Kim CH, Kunkel EJ, Boisvert J, Johnston B, Campbell JJ, Genovese MC, et al. Bonzo/CXCR6 expression defines type 1-polarized T-cell subsets with extralymphoid tissue homing potential. J Clin Invest 2001;107:595–601. 13. Motsinger A, Haas DW, Stanic AK, van Kaer L, Joyce S, Unutmaz D. CD1d-restricted human natural killer T cells are highly susceptible to human immunodeficiency virus 1 infection. J Exp Med 2002;195:869–79. 14. Thomas SY, Hou R, Boyson JE, Means TK, Hess C, Olson DP, et al. CD1d-restricted NKT cells express a chemokine receptor profile indicative of Th1-type inflammatory homing cells. J Immunol 2003;171:2571–80. 15. Sharron M, Pohlmann S, Price K, Lolis E, Tsang M, Kirchhoff F, et al. Expression and coreceptor activity of STRL33/Bonzo on primary peripheral blood lymphocytes. Blood 2000;96:41–9. 16. Minami M, Kume N, Shimaoka T, Kataoka H, Hayashida K, Akiyama Y, et al. Expression of SR-PSOX, a novel cell-surface scavenger receptor for phosphatidylserine and oxidized LDL in human atherosclerotic lesions. Arterioscler Thromb Vasc Biol 2001;21:1796–800. 3014 17. Wuttge DM, Zhou X, Sheikine Y, Wagsater D, Stemme V, Hedin U, et al. CXCL16/SR-PSOX is an interferon-␥-regulated chemokine and scavenger receptor expressed in atherosclerotic lesions. Arterioscler Thromb Vasc Biol 2004;24:750–5. 18. Shimaoka T, Nakayama T, Kume N, Takahashi S, Yamaguchi J, Minami M, et al. Cutting edge: SR-PSOX/CXC chemokine ligand 16 mediates bacterial phagocytosis by APCs through its chemokine domain. J Immunol 2003;171:1647–51. 19. Yamauchi R, Tanaka M, Kume N, Minami M, Kawamoto T, Togi K, et al. Upregulation of SR-PSOX/CXCL16 and recruitment of CD8⫹ T cells in cardiac valves during inflammatory valvular heart disease. Arterioscler Thromb Vasc Biol 2004;24:282–7. 20. Fukumoto N, Shimaoka T, Fujimura H, Sakoda S, Tanaka M, Kita T, et al. Critical roles of CXC chemokine ligand 16/scavenger receptor that binds phosphatidylserine and oxidized lipoprotein in the pathogenesis of both acute and adoptive transfer experimental autoimmune encephalomyelitis. J Immunol 2004;173:1620–7. 21. Morimoto C, Romain PL, Fox DA, Anderson P, DiMaggio M, Levine H, et al. Abnormalities in CD4⫹ T-lymphocyte subsets in inflammatory rheumatic diseases. Am J Med 1988;84:817–25. 22. Kinne RW, Brauer R, Stuhlmuller B, Palombo-Kinne E, Burmester GR. Review: macrophages in rheumatoid arthritis. Arthritis Res 2000;2:189–202. 23. Yamamoto K, Sakoda H, Nakajima T, Kato T, Okubo M, Dohi M, et al. Accumulation of multiple T cell clonotypes in the synovial lesions of patients with rheumatoid arthritis revealed by a novel clonality analysis. Int Immunol 1992;4:1219–23. 24. Nanki T, Lipsky PE. Cytokine, activation marker, and chemokine receptor expression by individual CD4⫹ memory T cells in rheumatoid arthritis synovium. Arthritis Res 2000;2:415–23. 25. MacDonald KP, Nishioka Y, Lipsky PE, Thomas R. Functional CD40 ligand is expressed by T cells in rheumatoid arthritis. J Clin Invest 1997;100:2404–14. 26. Okamoto T, Saito S, Yamanaka H, Tomatsu T, Kamatani N, Ogiuchi H, et al. Expression and function of the co-stimulator H4/ICOS on activated T cells of patients with rheumatoid arthritis. J Rheumatol 2003;30:1157–63. 27. Szekanecz Z, Kim J, Koch AE. Chemokines and chemokine receptors in rheumatoid arthritis. Semin Immunol 2003;15:15–21. 28. 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. 29. Nanki T, Nagasaka K, Hayashida K, Saita Y, Miyasaka N. Chemokines regulate IL-6 and IL-8 production by fibroblast-like synoviocytes from patients with rheumatoid arthritis. J Immunol 2001;167:5381–5. 30. Nanki T, Hayashida K, El-Gabalawy HS, Suson S, Shi K, Girschick HJ, et al. Stromal cell-derived factor-1-CXC chemokine receptor 4 interactions play a central role in CD4 T cell accumulation in rheumatoid arthritis synovium. J Immunol 2000;165:6590–8. 31. Imai T, Hieshima K, Haskell C, Baba M, Nagira M, Nishimura M, et al. Identification and molecular characterization of fractalkine receptor CX3CR1, which mediates both leukocyte migration and adhesion. Cell 1997;91:521–30. 32. Gong JH, Ratkay LG, Waterfield JD, Clark-Lewis I. An antagonist of monocyte chemoattractant protein 1 (MCP-1) inhibits arthritis in the MRL-lpr mouse model. J Exp Med 1997;186:131–7. 33. Bruhl H, Cihak J, Schneider MA, Plachy J, Rupp T, Wenzel I, et al. Dual role of CCR2 during initiation and progression of collagen-induced arthritis: evidence for regulatory activity of CCR2⫹ T cells. J Immunol 2004;172:890–8. NANKI ET AL 34. Nanki T, Lipsky PE. Cutting edge: stromal cell-derived factor-1 is a costimulator for CD4⫹ T cell activation. J Immunol 2000;164: 5010–4. 35. Hutloff A, Dittrich AM, Beier KC, Eljaschewitsch B, Kraft R, Anagnostopoulos I, et al. ICOS is an inducible T-cell co-stimulator structurally and functionally related to CD28. Nature 1999;397: 263–6. 36. McInnes IB, Al-Mughales J, Field M, Leung BP, Huang FP, Dixon R, et al. The role of interleukin-15 in T-cell migration and activation in rheumatoid arthritis. Nat Med 1996;2:175–82. 37. Oppenheimer-Marks N, Brezinschek RI, Mohamadzadeh M, Vita R, Lipsky PE. Interleukin 15 is produced by endothelial cells and increases the transendothelial migration of T cells in vitro and in the SCID mouse-human rheumatoid arthritis model in vivo. J Clin Invest 1998;101:1261–72. 38. Unutmaz D, Xiang W, Sunshine MJ, Campbell J, Butcher E, Littman DR. The primate lentiviral receptor Bonzo/STRL33 is coordinately regulated with CCR5 and its expression pattern is conserved between human and mouse. J Immunol 2000;165: 3284–92. 39. Tanaka Y, Adams DH, Hubscher S, Hirano H, Siebenlist U, Shaw S. T-cell adhesion induced by proteoglycan-immobilized cytokine MIP-1␤. Nature 1993;361:79–82. 40. Qin S, Rottman JB, Myers P, Kassam N, Weinblatt M, Loetscher M, et al. The chemokine receptors CXCR3 and CCR5 mark subsets of T cells associated with certain inflammatory reactions. J Clin Invest 1998;101:746–54. 41. Suzuki N, Nakajima A, Yoshino S, Matsushima K, Yagita H, Okumura K. Selective accumulation of CCR5⫹ T lymphocytes into inflamed joints of rheumatoid arthritis. Int Immunol 1999;11: 553–9. 42. Mack M, Bruhl H, Gruber R, Jaeger C, Cihak J, Eiter V, et al. Predominance of mononuclear cells expressing the chemokine receptor CCR5 in synovial effusions of patients with different forms of arthritis. Arthritis Rheum 1999;42:981–8. 43. Loetscher P, Uguccioni M, Bordoli L, Baggiolini M, Moser B, Chizzolini C, et al. CCR5 is characteristic of Th1 lymphocytes. Nature 1998;391:344–5. 44. Quinones MP, Ahuja SK, Jimenez F, Schaefer J, Garavito E, Rao A, et al. Experimental arthritis in CC chemokine receptor 2-null mice closely mimics severe human rheumatoid arthritis. J Clin Invest 2004;113:856–66. 45. Cooke SP, Forrest G, Venables PJ, Hajeer A. The D32 deletion of CCR5 receptor in rheumatoid arthritis [concise communication]. Arthritis Rheum 1998;41:1135–6. 46. Garred P, Madsen HO, Petersen J, Marquart H, Hansen TM, Freiesleben Sorensen S, et al. CC chemokine receptor 5 polymorphism in rheumatoid arthritis. J Rheumatol 1998;25:1462–5. 47. Kotake S, Udagawa N, Hakoda M, Mogi M, Yano K, Tsuda E, et al. Activated human T cells directly induce osteoclastogenesis from human monocytes: possible role of T cells in bone destruction in rheumatoid arthritis patients. Arthritis Rheum 2001;44: 1003–12. 48. Nanki T, Urasaki Y, Imai T, Nishimura M, Muramoto K, Kubota T, et al. Inhibition of fractalkine ameliorates murine collageninduced arthritis. J Immunol 2004;173:7010–6. 49. Matthys P, Hatse S, Vermeire K, Wuyts A, Bridger G, Henson GW, et al. AMD3100, a potent and specific antagonist of the stromal cell-derived factor-1 chemokine receptor CXCR4, inhibits autoimmune joint inflammation in IFN-␥ receptor-deficient mice. J Immunol 2001;167:4686–92.