Expression and function of junctional adhesion molecule-C in human and experimental arthritis

Male C57BL/6 mice were obtained from Janvier (Le Genest-St-Isle, France) and used between 9 and 11 weeks of age. KRN T-cell receptor transgenic mice, developed in the laboratory of Diane Mathis and Christophe Benoist, were kindly provided by the Institut de Génétique et de Biologie Moléculaire et Cellulaire (Strasbourg, France) [14] and were maintained on a C57BL/6 background (K/B). Progeny bearing the V^β6 transgenic T-cell receptor were identified by cytofluorometry of peripheral blood lymphocytes using antibodies labeled with anti-CD4 phycoerythrin (clone L3T4; BD Pharmingen, San Diego, CA, USA) and anti-V^β6 fluorescein isothiocyanate (clone RR4-7; BD Pharmingen). NOD/Lt mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). All mice were housed under conventional conditions, and water and standard laboratory chow were provided ad libitum. All animal experiments were approved by the Animal Ethics Committee of the Geneva University School of Medicine and by the Geneva Veterinarian Office.

Mice were injected intradermally at the base of the tail with 100 μg of methylated bovine serum albumin (mBSA) (Fluka, part of Sigma-Aldrich, St. Louis, MO, USA), emulsified in complete Freund's adjuvant (Difco Laboratories Inc., now part of Becton Dickinson and Company, Franklin Lakes, NJ, USA) containing 5 mg/ml Mycobacterium tuberculosis. Heat-killed Bordetella pertussis organisms (0.2 × 10⁹) (Berna, Bern, Switzerland) were injected intraperitoneally as an additional adjuvant. On day 7, a booster injection of 100 μg of mBSA in incomplete Freund's adjuvant (Becton Dickinson and Company) was given at the base of the tail. On day 21, arthritis was induced by intra-articular injection of 100 μg of mBSA in 10 μl of phosphate-buffered saline (PBS) into the left knee joint of mBSA-immunized mice, the right knee being injected with sterile PBS alone. The monoclonal anti-JAM-C antibody H36 [15] or an isotype-matched control antibody (9B5, rat IgG2a anti-human CD44) was injected (150 μg/mouse, intraperitoneal [i.p.]) 1 hour before intra-articular injection of mBSA into the left knee and of PBS into the right knee. Mice were sacrificed 4 or 8 days after induction of arthritis, the latter group receiving a second injection of antibodies on day 4. The development of arthritis was followed by measuring technetium-99m (Tc) uptake in the knees on days 1, 3, and 7 after intra-articular mBSA injection as previously described [16].

Arthritic K/BxN mice were obtained by crossing K/B mice with NOD/Lt (N) animals. Arthritic adult K/BxN mice were bled and the sera were pooled. Recipient C57BL/6 mice were injected with pooled serum (100 μl of serum i.p. on days 0 and 6). The monoclonal anti-JAM-C antibody H36 or an isotype-matched control antibody (9B5) was injected (150 μg/mouse, i.p.) 1 hour before the first injection of serum on day 0 and then again on days 4 and 8. Mice were sacrificed on day 13. The development of arthritis was assessed daily, and the severity of arthritis was scored in a blinded fashion for each paw on a 3-point scale, in which 0 = normal appearance, 1 = localized edema/erythema on one digit or over one surface of the paw, 2 = edema/erythema involving more than one surface of the paw, and 3 = marked edema/erythema involving the whole paw. The scores of all four paws were added for a composite score.

At sacrifice, the knees (AIA) or the paws (K/BxN serum transfer-induced arthritis) were dissected and fixed in 10% buffered formalin for 7 days. Fixed tissues were decalcified for 3 weeks in 15% EDTA (ethylenediaminetetraacetic acid), dehydrated, and embedded in paraffin. Sagittal sections (5 μm) of the whole joint were stained with safranin O and counterstained with fast green/iron hematoxilin. Histological sections were graded independently by two observers unaware of the treatment group by using the following parameters. For AIA, synovial membrane thickness, which reflects the degree of synovial inflammation and hyperplasia, was scored on a scale of 0 to 6 (0 = normal thickness to 6 = maximal thickness). Cartilage proteoglycan depletion, reflected by loss of safranin O staining intensity, was scored on a scale of 0 (fully stained cartilage) to 6 (totally unstained cartilage). For the K/BxN serum transfer-induced arthritis, synovial membrane thickness around the ankle, exudates in the ankle region, and ankle edema were scored on a scale of 0 to 3 (0 = normal to 3 = maximal).

Specimens of synovial tissue from osteoarthritis (OA) and RA patients undergoing joint surgery of the knee or the hip were obtained from the Department of Orthopedics of the Centre Hospitalier Universitaire Vaudois (Lausanne, Switzerland). Patients with RA fulfilled at least four of the seven American College of Rheumatology revised criteria for RA. All tissues were cut into small pieces and immediately frozen in precooled hexane and stored at -70°C until use. All subsequent analyses were performed on consecutive cryostat sections (one representative piece analyzed per patient). Samples were obtained after appropriate informed consent, and their use for research was approved by the Ethics Committee.

JAM-C expression was studied by immunohistochemistry (IHC) using rabbit polyclonal antibodies against human or murine JAM-C on cryostat sections of human or murine synovial tissues, respectively. Antibodies against mouse and human JAM-C have been previously described [17, 18]. Lymphocyte, macrophage, and neutrophil infiltrations into mouse synovium were detected by IHC using anti-CD3, anti-MAC-2, or anti-MPO antibodies, respectively, on paraffin-embedded sections. Briefly, frozen or deparaffinized and rehydrated sections were incubated for 30 minutes at room temperature with 5% BSA and 20% normal serum. Endogenous peroxidase activity was blocked with 3% H₂O₂ for 10 minutes. Slides were then overlaid with the primary antibody for 1 hour at room temperature. Bound antibody was visualized using the avidin-biotin-peroxidase complex (Vectastain Elite ABC kit; Vector Laboratories, Burlingame, CA, USA). The color was developed by 3,3'-diaminobenzidine (Sigma-Aldrich) containing 0.01% H₂O₂. After extensive washing in water, slides were counterstained with Papanicolaou (Merck AG, Dietikon, Switzerland) and mounted in Merckoglass (Merck AG, Dietikon, Switzerland). Staining specificity was confirmed using preimmune serum (for anti-JAM-C IHC), isotype-matched antibodies (for anti-CD3 and anti-MAC-2 IHC), or matched serum (for anti-MPO IHC) as primary antibodies. An incubation without the first antibody served as a negative control. Infiltration of lymphocytes, macrophages, and neutrophils was assessed by semi-quantitative visual scoring of CD3, MAC-2, and MPO immunostaining in the synovial membrane. For each marker, staining was graded independently by two observers (unaware of animal treatment) on a scale of 0 (no staining at all) to 3 (maximal staining). To estimate JAM-C levels in human synovial tissues, sections were magnified 400 times through a microscope (Olympus, Mont-sur-Lausanne, Switzerland), scanned using a JVC TK-C1381 color video camera (Olympus) and analyzed using Semper 6P image analysis software (Synoptics Ltd, Cambridge, UK). The results were expressed as the ratio between the number of pixels associated to immunoreactive regions and between the number of pixels of the total area examined.

Human synovial fibroblasts were isolated by collagenase digestion as reported previously [19]. Murine synovial fibroblasts were prepared from synovial tissue dissected from AIA knee joints. The tissue was finely minced and then digested in 0.1% collagenase (Gibco-BRL, now part of Invitrogen Corporation, Carlsbad, CA, USA) in Dulbecco's modified Eagle's medium (DMEM) for 2 hours at 37°C. Undigested material was removed, and cells were collected by centrifugation. Both human and murine synovial fibroblasts were kept in primary culture, at 37°C, in a humidified atmosphere containing 5% CO₂, in DMEM supplemented with 10% fetal calf serum and 100 U/ml penicillin and 100 μg/ml streptomycin. Non-adherent cells were removed by repeated washings during cell culture, and cells were used at the third passage.

Total RNA was isolated from human synovial tissue samples, knees of mice with AIA or collagen-induced arthritis [20], paws of mice with K/BxN serum transfer-induced arthritis, and human and mouse synovial fibroblasts using the TRIzol reagent (Invitrogen Corporation). Total RNA (1 to 3 μg) was digested with DNAse I (Promega AG, Wallisellen, Switzerland) and reverse-transcribed using avian myeloblastosis virus reverse transcriptase (RT) (Promega AG) and random hexamer primers. Polymerase chain reaction (PCR) amplification (40 cycles for JAM-C and 30 cycles for β-actin) was performed using Taq DNA polymerase (Qiagen AG, Hombrechtikon, Switzerland) and the following primers: murine Jam-C forward primer 5'-TGC TGC TGC TCT TCA GGG GC-3' and murine Jam-C reverse primer 5'-GAC AGG GGT CAC TGG CTT C-3' (GenBank accession number: NM023844), human JAM-C forward primer 5'-CTG GGG AAG ACA TCC CTG AAG-3' and human JAM-C reverse primer 5'-AGT GCG GAT GTA GTT AAC TCC-3' (GenBank accession number: NM032801), and β-actin forward primer 5'-CCAAGGCCAACCGCGAGAAGATGAC-3' and β-actin reverse primer 5'-AGGGTACATGGTGGTGCCGCCAGAC-3' (GenBank accession number: M10277). Annealing temperatures were 55°C for JAM-C and 60°C for β-actin. The absence of DNA contamination in RNA preparations was tested by including RNA samples, which had not been reverse-transcribed, and distilled water was used as a negative control for PCR amplification. The identity of the amplified products was confirmed by DNA sequencing.

Blood was taken at the end of experiment by cardiac puncture. Serum levels of serum amyloid A (SAA) were determined using a direct enzyme-linked immunosorbent assay (ELISA) as previously described [21]. The detection limit for this test is 13 μg/ml.

Spleen cells were harvested on day 4 after intra-articular mBSA injection and seeded at 4 × 10⁵ cells per well in 96-well plates in 200 μl of RPMI 1640 medium containing 100 IU/ml penicillin, 100 μg/ml streptomycin, 5 × 10^-5 M β-mercaptoethanol, and 1% mouse serum. Cells were incubated at 37°C with 5% CO₂ for 72 hours without or with 10 μg/ml mBSA or 5 μg/ml concanavalin A (ConA) (Amersham Pharmacia Biotech, now part of GE Healthcare, Little Chalfont, Buckinghamshire, UK). During the final 18 hours of incubation, ³H-thymidine was added at 1 μCi/well. Cells were harvested and radioactivity was counted to determine ³H-thymidine incorporation into DNA as a measure of cell proliferation.

Spleen cells were harvested on day 4 after intra-articular mBSA injection and cultured for 72 hours without or with 10 μg/ml mBSA or 5 μg/ml ConA. Culture supernatants were harvested, and interferon-gamma (IFN-γ) and interleukin-10 (IL-10) levels were quantified by ELISA by means of a commercial DuoSet ELISA Development System (R&D Systems, Abingdon, UK) for IFN-γ and a BD OptEIA Set (BD Biosciences, Heidelberg, Germany) for IL-10. The detection limits for both tests are 31 pg/ml. Serum levels of total anti-mBSA antibodies were measured as described previously [16].

We investigated JAM-C expression in human OA and RA synovial samples by IHC using polyclonal antibodies against human JAM-C. Interestingly, expression of JAM-C was mainly found in the lining layer and was associated with vessels in the sublining layer (Figure 1a,b) in OA and in RA synovial biopsies. JAM-C expression, as quantified by histomorphometry, was significantly higher in RA than in OA samples (Figure 1c). RT-PCR analysis showed JAM-C mRNA expression in OA and RA synovial biopsies as well as in purified human synovial fibroblasts (Figure 1d), suggesting that these cells account for at least part of the JAM-C expression observed in the lining layer by IHC.

To confirm our findings and extend our observations to another species, we then studied the pattern of JAM-C expression in two different mouse models of experimental arthritis: AIA and K/BxN serum transfer-induced arthritis (Figure 2a,b). In both cases, IHC using polyclonal antibodies against mouse JAM-C showed expression of JAM-C in a pattern similar to that found in human arthritic tissues. Expression of the protein was mainly found in the lining layer and was associated with vessels in the sublining layer. RT-PCR analysis showed JAM-C mRNA expression both in mouse synovial tissue samples and in cultured mouse synovial fibroblasts (Figure 2c).

To investigate a potential role for JAM-C in arthritis, we tested the effect of the monoclonal anti-JAM-C antibody H36 [17] on the course of AIA. Injection of an isotype-matched control antibody did not affect the development of AIA as compared to treatment with PBS (data not shown). In contrast, treatment of mice with the anti-JAM-C antibody significantly decreased the severity of arthritis, as quantified by Tc uptake on day 3 (Figure 3a). Histology showed lower inflammation in knees of mice treated with the anti-JAM-C antibody as compared to isotype-matched control antibody-treated mice (Figure 3b,c) on day 8, whereas cartilage damage was similar in the two groups. The lack of effect of JAM-C antibody on cartilage might be accounted for, in part, by the lack of sensitivity of safranin O staining, which cannot detect subtle changes in cartilage degradation. Serum SAA levels, which reflect the systemic inflammatory response and are maximal on day 4, were significantly decreased in anti-JAM-C antibody-treated mice (Figure 3d).

Proliferation of spleen cells isolated from mice treated with the isotype-matched control antibody was significantly increased upon restimulation with mBSA or ConA in vitro (Figure 4a). This proliferative response was significantly reduced in mice treated with the anti-JAM-C antibody (Figure 4a). Similarly, IFN-γ production was significantly increased by stimulation with mBSA or ConA in cells isolated from isotype-matched control antibody-treated mice, and this IFN-γ response was also markedly reduced in cells of anti-JAM-C antibody-treated mice (Figure 4b). IL-10 production was not detectable in spleen cell supernatants. A similar, though less marked, diminution of cell proliferation and IFN-γ production was also observed in draining lymph node cell cultures obtained from anti-JAM-C antibody-treated mice (data not shown). Lymph node cells obtained from isotype-matched control antibody-treated mice produced detectable levels of IL-10 upon stimulation with mBSA, and this IL-10 production was also reduced in cells isolated from anti-JAM-C antibody-treated mice. Total anti-BSA IgG and anti-BSA IgG2a levels were comparable in anti-JAM-C antibody-treated and isotype-matched control antibody-treated mice on days 4 and 8 (Figure 4c,d). Anti-BSA IgG1 levels were transiently elevated at day 4, but not at day 8, in the anti JAM-C antibody-treated mice.

Neutrophil infiltration in the inflamed synovium was selectively reduced in anti-JAM-C antibody-treated mice, whereas lymphocyte and macrophage infiltrations per field were similar in anti-JAM-C antibody and isotype-matched control antibody-treated mice (Figure 5).

To specifically investigate a potential role of JAM-C in the effector phase of arthritis, we tested the impact of anti-JAM-C antibody treatment in a model of K/BxN serum transfer-induced arthritis. Treatment with the anti-JAM-C antibody delayed the onset of K/BxN serum transfer-induced arthritis, as indicated by a significantly lower incidence of arthritis in anti-JAM-C antibody-treated mice as compared to isotype-matched control antibody-treated mice on day 3 (Figure 6a). In contrast, the severity of arthritis was not affected by the treatment in this model (Figure 6b). We also verified that injection of the isotype-matched control antibody 9B5 per se had no effect on the development of serum transfer-induced arthritis (data not shown).

Histological analysis of the paws showed no difference between anti-JAM-C antibody-treated and control mice at day 13 (Figure 6c), although we cannot exclude the possibility that histological differences, and in particular differences in neutrophil infiltration, could have been detectable at an earlier time point. The serum SAA levels, which are not significantly increased in this model, were not affected by treatment with the anti-JAM-C antibody. In contrast, ConA-induced spleen cell proliferation in vitro was decreased in anti-JAM-C antibody-treated arthritic mice as compared to isotype-matched antibody-treated controls (fold increase in proliferation over unstimulated control cultures: 11.3 ± 0.67 for anti-JAM-C antibody-treated mice, n = 10; 37.1 ± 4.49 for control mice, n = 10; p < 0.05).