ARTHRITIS & RHEUMATISM Vol. 62, No. 12, December 2010, pp 3595–3606 DOI 10.1002/art.27701 © 2010, American College of Rheumatology
CD248 and Its Cytoplasmic Domain A Therapeutic Target for Arthritis Margarida Maia,1 Astrid de Vriese,2 Tom Janssens,2 Michaël Moons,2 Kristel Van Landuyt,3 Jan Tavernier,4 Rik J. Lories,3 and Edward M. Conway5 Objective. CD248 is a transmembrane glycoprotein expressed on the surface of activated perivascular and fibroblast-like cells. This study was undertaken to explore the function of CD248 and its cytoplasmic domain in arthritis. Methods. Synovial tissue biopsy samples from healthy controls, from patients with psoriatic arthritis (PsA), and from patients with rheumatoid arthritis (RA) were stained for CD248. Transgenic mice that were CD248-deficient (CD248-knockout [CD248KO/KO]) or mice with CD248 lacking the cytoplasmic domain (CD248CyD/CyD) were generated. Collagen antibody– induced arthritis (CAIA) was induced in these mice and in corresponding wild-type (WT) mice as controls. Clinical signs and histologic features of arthritis were evaluated. Cytokine levels were determined by enzymelinked immunosorbent assay, and the number of infiltrating inflammatory cells was quantified by immuno-
histochemistry. In vitro studies were performed with fibroblasts from CD248-transgenic mouse embryos to explain the observed effects on inflammation. Results. Immunostaining of synovium from patients with PsA and patients with RA and that from mice after the induction of CAIA revealed strong CD248 expression in perivascular and fibroblast-like stromal cells. CD248KO/KO and CD248CyD/CyD mice had less severe arthritis, with lower plasma levels of proinflammatory cytokines, as compared with WT controls. Moreover, the joints of these mice had less synovial hyperplasia, reduced accumulation of inflammatory cells, and less articular cartilage and bone damage. Tumor necrosis factor ␣ –induced monocyte adhesion to CD248CyD/CyD fibroblasts was impaired. CD248CyD/CyD fibroblasts exhibited reduced expression of hypoxia-inducible factor 1␣, placental growth factor, vascular endothelial growth factor, and matrix metalloproteinase 9 activity in response to transforming growth factor . Conclusion. CD248 contributes to synovial hyperplasia and leukocyte accumulation in inflammatory arthritis, the effects of which are mediated partly via its cytoplasmic domain. CD248 is therefore a potential new target in the treatment of arthritis.
Ms Maia’s work was supported by the VIB International PhD Program, Belgium. Dr. Van Landuyt is an Aspirant Fellow of the Flanders Research Foundation (FWO Vlaanderen). Dr. Lories is the recipient of a FWO Vlaanderen postdoctoral fellowship. Dr. Conway’s work was supported by a FWO Vlaanderen grant. Dr. Conway also is an adjunct scientist with the Canadian Blood Services and holds a CSL-Behring Research Chair and a Canada Research Chair in Endothelial Cell Biology at the University of British Columbia. 1 Margarida Maia, BSc: Katholieke Universiteit–Leuven, Flanders Interuniversity Institute for Biotechnology (VIB)–Leuven, VIB–Ghent, and Ghent University, Ghent, Belgium; 2Astrid de Vriese, BSc, Tom Janssens, BSc, Michae¨l Moons, BSc: Katholieke Universiteit–Leuven, and VIB–Leuven, Leuven, Belgium; 3Kristel Van Landuyt, MD, Rik J. Lories, MD, PhD: Katholieke Universiteit– Leuven, Leuven, Belgium; 4Jan Tavernier, PhD: Ghent University, and VIB–Ghent, Ghent, Belgium; 5Edward M. Conway, MD, PhD: Katholieke Universiteit–Leuven and VIB–Leuven, Leuven, Belgium, and University of British Columbia, Vancouver, British Columbia, Canada. Address correspondence and reprint requests to Edward M. Conway, MD, PhD, Centre for Blood Research, University of British Columbia, 2350 Health Sciences Mall, LSC 4306, Vancouver BC V6T 1Z3, Canada. E-mail:
[email protected]. Submitted for publication April 5, 2010; accepted in revised form August 5, 2010.
In diseases such as rheumatoid arthritis (RA) and psoriatic arthritis (PsA), the synovium transforms from a thin paucicellular tissue into an invasive and jointdestructive tissue that is characterized by hyperplasia, angiogenesis, immune and mesenchymal cell infiltration, and development of secondary lymphoid structures. Increasing evidence suggests that resident and infiltrating mesenchymal cell populations contribute to the influx of inflammatory cells, their retention in the synovium, the angiogenic response, and, ultimately, the chronicity of the disease (1–3). The resultant changes in 3595
3596
the synovial stromal code are also hypothesized to account for recurrence of disease after interruption of successful therapies (4). The most prominent stromal player in the progression of joint inflammation and destruction is the fibroblast-like synoviocyte. These cells synthesize matrix components (fibronectin, collagens, tenascin, proteoglycans, and laminin) as well as cytokines and proteinases that are necessary for matrix remodeling and cellular movement. Fibroblast-like synoviocytes can secrete serine proteases, metalloproteinases, and cathepsins and release hyaluronic acid and the synovial fluid joint lubricant lubricin (5). Activated fibroblast-like synoviocytes also secrete vascular endothelial growth factor (VEGF) and placental growth factor (PlGF), thereby enhancing the pathologic processes of angiogenesis, vascular permeability, activated leukocyte recruitment, synovial hyperplasia, and inflammation (6,7). Unchecked proliferation and activation of these cells will promote further joint inflammation and damage. In arthritis, this occurs in parallel with the activation and alteration of the phenotype of other cells of mesenchymal origin, including synovial perivascular cells, which are in close apposition to synovial fibroblasts in the joint space. Perivascular cells interact with the subintimal stroma of the synovium and regulate the release of numerous factors implicated in modulating joint damage, including, for example, platelet-derived growth factor type B (8), transforming growth factor  (TGF) (9), and VEGF (10). Only a few disease-specific molecules that are expressed by perivascular and/or fibroblast-like synovial cells and that are amenable to therapeutic targeting have been identified. CD248 is one such molecule that might play a role in inflammatory arthritis and, thus, could be a drug target. CD248 is a highly sialylated, cell surface, type I transmembrane glycoprotein that was originally detected with an antibody against fetal fibroblasts (11). It has since been identified in rheumatoid fibroblast-like synovial cells (12) and in a subset of fibroblasts that participate in the remodeling of lymphoid tissue after infection (13). Expression of CD248 is, in fact, restricted to proliferating tissues, i.e., during fetal development, and postnatally in tumors and in inflammatory lesions, where it is found on the surface of perivascular cells and stromal fibroblast-like cells (12,14–16). Since expression of CD248 has been observed in almost all carcinomas in humans (17–19), much attention has been directed to the therapeutic targeting of CD248 for the treatment of cancer. A role for CD248 in tumorigenesis has been confirmed by generating CD248-deficient (CD248-knockout
MAIA ET AL
[CD248KO/KO]) mice. These mice develop normally but exhibit reduced tumor growth, invasiveness, and metastatic dissemination (20). In spite of the expression pattern of CD248 in activated synovial fibroblasts, little notice has been paid to its potentially important role in inflammatory disorders such as arthritis. In this study, we demonstrate that CD248 is prominently expressed by perivascular and fibroblastlike synovial cells in the synovium from patients with PsA and patients with RA, but not in synovium from healthy individuals. We also show that mice deficient in CD248 (CD248KO/KO) or lacking its cytoplasmic domain (CD248CyD/CyD)—a structure that may mediate intracellular signaling—develop less severe arthritis, with less synovial hyperplasia, reduced infiltration with leukocytes, and lower levels of cytokines in the plasma. Moreover, cultured fibroblasts from CD248CyD/CyD mice release lower amounts of active matrix metalloproteinase 9 (MMP-9) and exhibit a more quiescent phenotype in terms of reduced leukocyte adhesion. These results indicate that synovial cell CD248, via its cytoplasmic domain, functions as a proinflammatory mediator in arthritis. The findings highlight the potential role of CD248 as a novel target in the suppression of inflammation. PATIENTS AND METHODS Patients and controls. Synovial biopsy samples were obtained from patients with PsA and patients with RA at the University Hospitals of Leuven in Belgium. Written informed consent was obtained from all patients, and procedures were approved by the Ethics Committee of Katholieke Universiteit– Leuven. Control synovial tissue samples were obtained from patients with noninflammatory knee problems who were undergoing diagnostic arthroscopy in the Division of Orthopedic Surgery at the University Hospitals. All male and female control patients (ages 48–61 years) showed meniscal and cartilage lesions during this procedure. Synovial biopsy samples from patients with chronic arthritis were obtained by needle arthroscopy in the Division of Rheumatology at the University Hospitals. All patients were male. Patients with RA (ages 40–51 years) had a disease duration of ⬍5 months, with the exception of 1 patient (age 17 years) who had a disease duration of 72 months. The 2 patients with PsA (ages 42 years and 59 years) had a disease duration of 2 months and 9 months, respectively. Patients with RA fulfilled the American College of Rheumatology (formerly, the American Rheumatism Association) revised criteria for RA (21). Patients with PsA fulfilled the CASPAR Study Group criteria (22). One of the patients with RA was undergoing therapy with sulfasalazine. None of the other patients were being treated with immunosuppressive drugs at the time of arthroscopy. Generation of CD248CyD/CyD and CD248lox/lox mice. A 13-kb Xba1 fragment of the murine CD248 gene containing the
CD248 AND ITS CYTOPLASMIC DOMAIN IN ARTHRITIS
coding region was subcloned into a pBS vector (Invitrogen). With the use of polymerase chain reaction (PCR)–based mutagenesis, the wild-type (WT) coding region of CD248 was replaced with one that encodes for a CD248 variant lacking the cytoplasmic domain. Two PCRs were performed, using the forward/reverse primer pairs F1/R2 (5⬘-CAGCTGTAGCCCTGCAGGAGCCATGGGT-3⬘ and 5⬘-GGCATCTGCACCCCATCACACAATGCCCAGGGC-3⬘, respectively) and F2/R1 (5⬘-GCCCTGGGCATTGTGTGATGGGGTGCAGATGCC-3⬘ and 5⬘-CTACAACCCTAGGGTGGCTGCATCCTT-3⬘, respectively). The resultant amplicons were used as the target for recombinant PCR with the primers F3 (5⬘-CTTCCGGCCAGCTGAGGAGGATGATCCACAC-3⬘) and R3 (5⬘-GAACCGGCGTCTCCCTCTTCCAGGTCCCGT-3⬘). The recombined PCR product was used to replace the WT coding region of CD248, using the restriction sites Sse 8387I and Avr II. The truncated protein (CD248CyD) retained 2 intracellular juxtamembranous amino acids and lacked amino acid residues 717–764. The knock-in vector retained the in-frame stop codon. A targeting vector containing a floxed cassette encompassing the coding region of CD248CyD (loxP-FRT-PGKNeo-FRT-loxP) and a neomycin resistance gene flanked by Flp recombinase target sites was constructed. A 3⬘ diphtheria toxin A gene was used for negative selection. To target the CD248 gene locus, 5⬘ and 3⬘ homology arms of 3.25 kb and 1.98 kb, respectively, were used. The linearized targeting vector was electroporated into embryonic stem (ES) cells from 129Sv: BL/6 mice, and homologous recombination was confirmed by Southern blotting and PCR using the primers F4 (5⬘CCCAAGGGAAGGAGTTCCCAGTCC-3⬘) and R4 (5⬘CAGATGCCTGAGCGGTAGATGGGGGACA-3⬘). The neomycin cassette was excised in vitro by electroporation of a complementary DNA (cDNA) expression vector encoding Flp recombinase. Correctly targeted ES cell clones were aggregated and introduced into pseudopregnant C57BL/6 mice. ES cell clone 36 contained a targeted allele with the deleted cytoplasmic domain, and this ES cell clone was used for the generation of CD248CyD/CyD mice. ES cell clone 37 contained a targeted allele in which the coding region remained intact, and this ES cell clone was used for the generation of CD248lox/lox mice. Chimeric mice were mated to select for germline transmission and to generate heterozygous offspring. These offspring were intercrossed and maintained on a C57BL/6:129Sv (75:25) background. Genotyping of CD248CyD/CyD mice was performed by PCR analyses of tail genomic DNA using the primers F4 and R5 (5⬘-CCCAAGGGAAGGAGTTCCCAGTCC-3⬘ and 5⬘CTCTGTCAGCTGGGCAGCCCCCATAA-3⬘, respectively), which span the cytoplasmic domain and yield fragments of 442 bp and 297 bp in WT and cytoplasmic tail–deleted alleles, respectively. Genotyping of CD248lox/lox mice was performed using the primers F5 and R6 (5⬘-GTTCTCAAGGAGGTTATCAAGTTGATTC-3⬘ and 5⬘-TCTTCCAGAGAGACTCCCGAAAGC-3⬘, respectively). The amplicon spans the first loxP site and generates fragments of 331 bp and 380 bp in WT CD248 and floxed CD248, respectively. The integrity of the entire coding region of WT and mutant CD248 was further confirmed by DNA sequencing of PCR-generated amplicons
3597
(full details of these analyses are available from the corresponding author upon request). Generation of CD248KO/KO mice. In vivo excision of the CD248 gene in mice was achieved by cross-breeding the CD248lox/lox mice with mice homozygous for ubiquitous expression of Cre recombinase under the control of the phosphoglucokinase promoter (23). Confirmation of excision was accomplished by genotyping tail genomic DNA of CD248KO/KO mice by PCR analysis using the primer pair F5/R6 to amplify a 331-bp band in WT CD248 mice and primer pair F5/R3 to amplify a 580-bp band in CD248KO/KO mice (details available from the corresponding author upon request). CD248KO/KO and CD248CyD/CyD mice were bred in parallel with their corresponding WT counterparts using a breeding scheme devoid of sibling matings. All experimental animal procedures were approved by the Institutional Animal Care and Research Advisory Committee of Katholieke Universiteit–Leuven. Generation of collagen antibody–induced arthritis (CAIA) in mice. For induction of CAIA, the ArthroMAB kit (Millipore) was used, in accordance with the manufacturer’s instructions. Briefly, 6–8-week-old male mice were injected intraperitoneally (IP) with 8 mg of an arthritogenic cocktail of 4 monoclonal antibodies that recognize type II collagen, followed 3 days later by IP injection of 50 g of lipopolysaccharide from Escherichia coli O111:B4. Clinical evidence of arthritis was evaluated daily by 2 blinded investigators (de Vriese and Moons), as previously reported (24). Each of the paws was scored according to the extent of redness and swelling, with a score range of 0–4, resulting in a maximum arthritis score of 16 per mouse. Mice that did not develop any arthritis were excluded from the study. Thirteen days after the injection of the anticollagen antibodies, the knee joints were dissected free of skin and muscles and fixed overnight in 2% paraformaldehyde in phosphate buffered saline (PBS). After a 30-minute wash in PBS, the bones were decalcified by incubation in 0.5M EDTA for 10 days, with daily renewal of the buffer. Paraffin-embedded 7-m sections were cut, deparaffinized, and stained with hematoxylin and eosin (H&E) or Safranin O, or were used for immunohistochemistry (as described below). H&E-stained sections were analyzed for characteristic changes associated with arthritis and scored using a previously described scoring system for histologic changes (24), as follows: synovial membrane hyperplasia (score 0–4), infiltration of synovium with inflammatory cells (score 0–4), and pannus formation (score 0–4). Normal joints were scored 0. Cartilage erosion was analyzed on Safranin O–stained sections and graded on a scale of 0–3, as follows: smooth cartilage surface (score 0), localized cartilage erosions (score 1), cartilage erosion and proteoglycan depletion (score 2), and severe cartilage destruction (score 3). In addition, detection of bone erosions on Safranin O–stained sections was scored 0–4. Analyses were performed in a blinded manner by 2 independent observers. Scoring by each investigator was similar, and thus the average of the scores of the 2 investigators was used as the final score. Immunohistochemistry. Knee joint sections were deparaffinized and incubated for 20 minutes at 80°C in antigen retrieval solution (Dako). Sections were quenched of endogenous peroxidase activity with 0.3% H2O2 in methanol for 20
3598
MAIA ET AL
Figure 1. CD248 expression in normal and arthritic human knee joints. Representative synovial biopsy tissue samples from subjects without arthritis (A and D), patients with psoriatic arthritis (PsA) (B and E), and patients with rheumatoid arthritis (RA) (C and F) were stained with specific anti-CD248 antibodies (A–C) or with nonspecific isotype control immunoglobulins (NSIg) (D–F). CD248 is absent or weakly expressed in normal synovial tissue (A) but is strongly expressed in the perivascular region (open arrowheads) and sublining zone (solid arrowheads) in the synovium from patients with PsA (B) and patients with RA (C). The absence of signal with NSIg (D–F) confirms the specificity of the anti-CD248 antibodies. Bars ⫽ 100 m.
minutes, blocked with 5% preimmune serum in 0.1M Tris HCl, pH 7.5, 0.15M NaCl with 0.5% blocking reagent, and incubated overnight with specific primary antibodies as follows: rabbit anti-CD248 (1 g/ml; a gift from Philippe Gasque, Cardiff University, Cardiff, UK) (20), rat anti-mouse CD45 (2 g/ml; Becton Dickinson), or nonspecific immunoglobulin as an isotype control. The highly specific, affinity-purified anti-CD248 antibodies, which were raised in rabbits against recombinant C-type lectin-like domain, cross-react with murine CD248 (20). After washing with 0.1M Tris HCl, pH 7.5, 0.15M NaCl, 0.05% Tween 20, the corresponding biotinylated secondary antibody was added for 1 hour and the signal was amplified with the Vectastain Elite ABC kit (Vector Laboratories). Signal detection was achieved with 3,3⬘-diaminobenzidine (Sigma-Aldrich). Sections were then dehydrated and mounted with DPX. Morphometric analyses were performed using a Zeiss Imager Z1 or AxioPlan 2 microscope with KS300 image analysis software. Enzyme-linked immunosorbent assay (ELISA). The concentration of interleukin-1 (IL-1) and VEGF in plasma or cell lysates was quantified using Quantikine ELISA kits for mouse IL-1/IL-1F2 or murine VEGF (R&D Systems). Isolation of mouse embryonic fibroblasts (MEFs). MEFs were isolated from E13.5 embryos. Embryos were washed in PBS, the head and internal organs were removed, and the carcass was minced in 0.25% trypsin containing 40 l/ml DNase I (Roche) and incubated for 10 minutes at 37°C. MEFs were cultured in Dulbecco’s modified Eagle’s medium with 10% fetal calf serum and used at passages 2–5. To assess the response to exogenous growth factors, 1.5 ⫻ 105 MEFs were seeded in 6-well plates. After 18 hours of serum starvation, cells were stimulated with 0.1 g/ml recombinant murine tumor necrosis factor ␣ (TNF␣; BD PharMingen) for 3 hours
or 3 ng/ml recombinant human TGF1 (R&D Systems) for 72 hours. Cells were processed for protein or RNA extraction, as required. U937 cell adhesion. MEFs (5 ⫻ 104) were seeded in 24-well plates. After 24 hours, cells were left unstimulated or stimulated for 3 hours with 0.1 g/ml recombinant mouse TNF␣. Cells were washed twice with a buffer containing 1% bovine serum albumin, 1.2 mM CaCl2, and 1.2 mM MgCl2, and 7.5 ⫻ 104 U937 monocytes labeled with BCECF-AM (Invitrogen) were added to each well. After 1 hour, nonadherent U937 monocytes were removed by washing. The number of adherent cells was quantified by counting 8 optical fields at 10⫻ magnification using a Zeiss AxioVert 100M inverted microscope. For each condition, 3–4 wells were analyzed. Quantitative reverse transcription–PCR (RT-PCR). RNA was extracted from cells using the Qiagen RNeasy Mini kit and from synovial tissue using the Qiagen RNeasy Micro kit. Total RNA (0.5–1 g) was analyzed by RT with the QuantiTect Reverse Transcription kit (Qiagen). Quantitative RT-PCR was performed using TaqMan Fast Universal PCR Master Mix (Applied Biosystems) and commercially available or homemade primers and probes for the genes of interest (details available from the corresponding author upon request). Analyses were performed using the Applied Biosystems 7500 Fast Real-Time PCR System. Gene transcription was calculated as the number of messenger RNA (mRNA) copies relative to the number of mRNA copies of the housekeeping genes -actin or hypoxanthine guanine phosphoribosyltransferase (HPRT). Gelatin zymography. Equal numbers of MEFs were cultured to confluence and placed in serum-free medium with or without TGF. After 72 hours, equal volumes of serum-free
CD248 AND ITS CYTOPLASMIC DOMAIN IN ARTHRITIS
conditioned media were separated by sodium dodecyl sulfate (SDS)–polyacrylamide gel electrophoresis under nonreducing conditions in gels containing 0.1% porcine skin gelatin. Gels were renatured by exchanging SDS with a solution of 2.5% Triton X-100, and then incubated for 18 hours at 37°C in 50 mM Tris HCl, pH 7.5, 7 mM CaCl2, 200 mM NaCl, 0.02% Brij-35, followed by staining with 0.5% Coomassie R250 in a solution containing 45% methanol and 10% acetic acid for 3 hours and, finally, destaining in 45% ethanol/10% acetic acid. Statistical analysis. Results are expressed as the mean ⫾ SEM values from experiments performed at least in duplicate. Statistical significance was calculated by t-test or two-way analysis of variance with repeated measurements (Prism version 5.0; GraphPad Software). P values less than 0.05 were considered statistically significant.
RESULTS CD248 expression in human arthritic joints. We first tested whether the expression pattern of CD248 in human synovium could provide the rationale to study functional aspects of the molecule in transgenic mice. An exploratory analysis of synovial tissue from control individuals without arthritis and from patients with RA or PsA was therefore performed (Figures 1A–F). Sections of synovial tissue samples from control patients without arthritis revealed a normal synovial architecture with a thin lining layer and a paucicellular sublining zone. In contrast, RA and PsA samples exhibited a thickened lining layer, diffuse inflammatory cell infiltration into the sublining zone, and hypervascularity. In cultures with specific anti-CD248 antibodies (20), only very weak CD248 staining was detected in the synovial tissue of healthy joints (Figure 1A). In contrast, CD248 was readily detected in the synovium from patients with PsA and patients with RA (Figures 1B and C) (further details available from the corresponding author upon request). Both in PsA and in RA synovial tissue, staining was prominent in the stromal infiltrate, with an intense signal also observed in the perivascular cells and stromal fibroblasts surrounding the synovial vasculature. Staining of the sections with an isotype-matched nonspecific immunoglobulin did not reveal any signal for detection of CD248 (Figures 1D–F). The findings are consistent with the idea that the mesenchymal cell population in inflamed joints becomes CD248 positive. Reduced clinical signs of arthritis in CD248KO/KO and CD248CyD/CyD mice. We reasoned that CD248 plays an important physiologic role in arthritis and predicted that the cytoplasmic domain of CD248, which has structural motifs that could participate in cell
3599
signaling, might mediate its function (25,26). To test this, we generated, by homologous recombination in ES cells, mice that entirely lack CD248 (CD248KO/KO) and mice lacking the cytoplasmic domain of CD248 (CD248CyD/CyD). Deletion of the cytoplasmic domain was confirmed by direct sequencing of genomic DNA derived from targeted ES cells and by RT-PCR with cDNA derived from RNA extracted from MEFs. Consistent with the characteristics of another strain of CD248KO/KO mice (20), our CD248KO/KO and CD248CyD/CyD mice were also viable and fertile, with normal growth and development. In primary MEFs from CD248CyD/CyD embryos, transcript levels of CD248 lacking the cytoplasmic domain were not significantly different from WT MEFs (results not shown). Moreover, the cell surface expression pattern was not altered in CD248CyD/CyD MEFs, as shown by immunostaining (results not shown). MEFs from CD248KO/KO embryos did not express detectable levels of CD248. (Further details on the characteristics of these generated mice are available from the corresponding author upon request.) To examine the response of the CD248transgenic mice to induction of arthritis, we used a well-characterized murine model of rapid-onset arthritis, the CAIA model, which was initiated by IP injection of mouse paws with a mixture of 4 anti–type II collagen monoclonal antibodies. The clinical severity of arthritis in CD248KO/KO mice was compared with that in their WT counterparts (CD248WT/WT), and in separate experiments, arthritis in CD248CyD/CyD mice was compared with that in CD248WT/WT mice. Clinical signs of arthritis were monitored every day from the onset of disease up to day 13, which showed that 87% of CD248WT/WT mice, 75% of CD248CyD/CyD mice, and 74% of CD248KO/KO mice developed arthritis. CD248KO/KO mice (n ⫽ 12) exhibited significantly milder clinical signs of arthritis as compared with CD248WT/WT mice (n ⫽ 10) (P ⬍ 0.001) (Figure 2A). CD248CyD/CyD mice (n ⫽ 12) also developed less severe arthritis than that in their WT counterparts (n ⫽ 9) (P ⬍ 0.001) (Figure 2B), supporting the notion that the cytoplasmic domain of CD248 is critical for normal function of the molecule. Consistent with these findings, plasma IL-1 levels at day 7 were lower in CD248KO/KO and CD248CyD/CyD mice compared with WT mice, with the difference approaching statistical significance (mean ⫾ SEM 71 ⫾ 15 pg per ml plasma in CD248WT/WT mice versus 39 ⫾ 9 pg per ml plasma in CD248KO/KO mice [n ⫽ 10 each], P ⫽ 0.0761; mean ⫾ SEM 64 ⫾ 10 pg per ml plasma in CD248WT/WT mice versus 38 ⫾ 8 pg per ml plasma in CD248CyD/CyD mice [n ⫽ 8 each], P ⫽ 0.0577).
3600
MAIA ET AL
perivascular cells of the inflamed knee joints of CD248WT/WT mice. Less inflammation was observed in the joints of CD248CyD/CyD mice after the induction of arthritis, and CD248CyD/CyD joints also expressed lower levels of CD248. As expected, no CD248 was detectable in the joints of CD248KO/KO mice (results not shown). Since experiments with CD248 CyD/CyD and CD248KO/KO mice were performed at different times, direct comparisons between them could not be made. When independently compared with WT mice, H&E analyses of the joints revealed 2- to 3-fold less synovial
Figure 2. Reduced clinical signs of collagen antibody–induced arthritis (CAIA) in CD248-knockout mice (CD248KO/KO) and mice lacking the cytoplasmic domain of CD248 (CD248CyD/CyD) as compared with wild-type mice (CD248WT/WT). CAIA was initiated by administration of anti–type II collagen antibodies on day 0, followed by 50 g lipopolysaccharide on day 3. Clinical signs of arthritis were scored daily as described in Patients and Methods. The clinical severity of arthritis was significantly lower in A, CD248KO/KO mice (n ⫽ 12) versus CD248WT/WT mice (n ⫽ 10) and in B, CD248CyD/CyD mice (n ⫽ 12) versus CD248WT/WT mice (n ⫽ 9). Bars show the mean ⫾ SEM. ⴱⴱⴱ ⫽ P ⬍ 0.001, by two-way analysis of variance with repeated measurements.
Reduced synovial inflammation in CD248KO/KO and CD248CyD/CyD mice. We histologically evaluated the knee joints of the mice at day 13 after injection of the arthritogenic antibodies. Joint sections were immunostained with specific anti-CD248 antibodies. As compared with normal nonarthritic joints in which minimal expression of CD248 was evident in some synovial fibroblasts, CAIA induction resulted in marked upregulation of CD248 in the stromal fibroblasts and
Figure 3. Reduced synovial hyperplasia and inflammation in CD248knockout mice (CD248KO/KO) (A) and in mice lacking the cytoplasmic domain of CD248 (CD248CyD/CyD) (B) as compared with wild-type mice (CD248WT/WT). In A and B, knee sections were stained with hematoxylin and eosin (H&E) (panels a and b) or were stained for CD45 to assess leukocyte infiltration (arrows) (panels c and d). Bars in panels a and b ⫽ 100 m; bars in panels c and d ⫽ 50 m. Line segments in panels a and b indicate the thickness of the synovial lining.
CD248 AND ITS CYTOPLASMIC DOMAIN IN ARTHRITIS
3601
Table 1. Scoring of synovial hyperplasia, pannus formation, cartilage degradation, and bone damage in CD248-transgenic mice after the induction of collagen antibody–induced arthritis*
WT/WT
Hyperplasia
Pannus
Cartilage
Bone
1.8 ⫾ 0.3 0.8 ⫾ 0.1 0.0082
1.0 ⫾ 0.2 0.2 ⫾ 0.1 0.0019
2.5 ⫾ 0.2 1.5 ⫾ 0.2 0.002
1.5 ⫾ 0.3 0.6 ⫾ 0.2 0.017
2.2 ⫾ 0.3 0.6 ⫾ 0.2 0.0003
1.2 ⫾ 0.2 0.4 ⫾ 0.2 0.0022
1.9 ⫾ 0.3 0.9 ⫾ 0.2 0.024
1.3 ⫾ 0.3 0.6 ⫾ 0.2 0.059
KO/KO
CD248 WT/WTvs. CD248 CD248 CD248KO/KO P CD248WT/WT vs. CD248CyD/CyD CD248WT/WT CD248CyD/CyD P
* Values are the mean ⫾ SEM score of the severity of arthritis in mouse knee joints as quantified from histologic sections stained with hematoxylin and eosin (synovial hyperplasia and pannus formation; n ⫽ 10 knees from 5 mice per group) and Safranin O (cartilage degradation and bone damage; n ⫽ 8 knees from 4 mice per group). Higher scores represent worse disease. P values were determined by t-test. WT ⫽ wild-type; KO ⫽ knockout; CyD ⫽ cytoplasmic domain.
hyperplasia and 3- to 5-fold less pannus formation in the joints of the CD248KO/KO and CD248CyD/CyD mice (Table 1 and Figures 3A and B, panels a and b). Moreover, infiltration of CD45-positive leukocytes into the synovium was significantly dampened in CD248KO/KO and CD248CyD/CyD mice (mean ⫾ SEM CD45-positive cells/ mm2, 630 ⫾ 40 in CD248WT/WT mice versus 420 ⫾ 40 in CD248KO/KO mice [n ⫽ 8 each], P ⫽ 0.0024; 650 ⫾ 70 in CD248WT/WT mice versus 430 ⫾ 50 in CD248CyD/CyD mice [n ⫽ 10 each], P ⫽ 0.014) (Figures 3A and B, panels c and d). We also determined, from H&E and Safranin O staining of tissue sections, that there was reduced cartilage destruction and bone damage in the joints of CD248KO/KO and CD248CyD/CyD mice compared with WT mice (Table 1 and details available from the corresponding author upon request). Consistent with our observation that there was less severe arthritis in the CD248CyD/CyD mice, transcript levels of the leukocyte- and fibroblast-derived proinflammatory cytokines IL-1 and TNF␣ were significantly reduced in synovial tissue from inflamed CD248CyD/CyD joints (mean ⫾ SEM copies of IL-1 per 1,000 copies of HPRT, 57,600 ⫾ 5,780 in CD248WT/WT mice versus 27,200 ⫾ 8,850 in CD248CyD/CyD mice [n ⫽ 12 each], P ⫽ 0.0089; mean ⫾ SEM copies of TNF␣ per 1,000 copies of HPRT, 470 ⫾ 26 in CD248WT/WT mice versus 269 ⫾ 61 in CD248CyD/CyD mice [n ⫽ 12 each], P ⫽ 0.0045). Similarly, transcript levels of intercellular adhesion molecule 1 (ICAM-1) were significantly reduced in the CD248CyD/CyD synovium (mean ⫾ SEM copies of ICAM-1 per 1,000 copies of HPRT, 920 ⫾ 47 in CD248WT/WT mice versus 630 ⫾ 80 in CD248CyD/CyD mice [n ⫽ 12 each], P ⫽ 0.005). Overall, the findings indicate that CD248 facilitates the development of inflammatory arthritis in vivo and that the proinflamma-
tory properties of CD248 are mediated, at least in part, via its cytoplasmic domain. Impaired leukocyte adhesion to CD248CyD/CyD fibroblasts. We examined potential mechanisms by which the cytoplasmic domain of CD248 regulates leukocyte infiltration and retention in the synovium. Since fibroblast-like synoviocytes display and secrete an array of adhesion molecules and cytokines that stimulate leukocyte migration and adhesion (27), we used MEFs from the CD248-transgenic mice as a fibroblast model to
Figure 4. Decreased adhesion of monocytes to mouse embryonic fibroblasts (MEFs) from mice lacking the cytoplasmic domain of CD248 (CD248CyD/CyD). MEFs were grown to confluence and left unstimulated or stimulated with mouse recombinant tumor necrosis factor ␣ (TNF␣) for 3 hours. Fluorescently labeled U937 monocytes were added for 1 hour. Nonadherent cells were removed by washing and the number of adherent cells was quantified. The number of monocytes adhering to wild-type (CD248 WT/WT ) MEFs or CD248CyD/CyD MEFs significantly increased after stimulation with TNF␣. However, TNF␣-induced adhesion was significantly dampened in CD248CyD/CyD MEFs compared with CD248WT/WT MEFs (n ⫽ 8 each). Bars show the mean and SEM. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01, by t-test.
3602
MAIA ET AL
Figure 5. Decreased expression of growth factors and hypoxia-inducible factors (HIFs) in mouse embryonic fibroblasts (MEFs) from mice lacking the cytoplasmic domain of CD248 (CD248CyD/CyD). MEFs were grown to confluence and left unstimulated or stimulated with human transforming growth factor  (TGF) for 72 hours or mouse recombinant tumor necrosis factor ␣ (TNF␣) for 3 hours. A, Transcript levels of placental growth factor (PlGF) and vascular endothelial growth factor receptor 1 (VEGFR-1) were significantly reduced at baseline in CD248CyD/CyD MEFs as compared with wild-type (CD248WT/WT) MEFs (n ⫽ 3–4 each; P ⫽ 0.001). B, CD248CyD/CyD MEFs, compared with CD248WT/WT MEFs, had significantly reduced levels of cell-associated VEGF after stimulation with TGF, as measured by enzyme-linked immunosorbent assay (n ⫽ 9 each). Each symbol represents an individual animal; bars show the mean ⫾ SEM. C, Similarly, VEGF transcription was significantly reduced in CD248CyD/CyD MEFs (n ⫽ 9 each; P ⫽ 0.007). D, Transcript levels of HIF-1␣ were significantly reduced in CD248CyD/CyD MEFs, as compared with CD248WT/WT MEFs, at baseline (n ⫽ 6 each; P ⫽ 0.036) and after TNF␣ stimulation (n ⫽ 6 each; P ⫽ 0.017). Values in A, C, and D are the mean and SEM. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01; ⴱⴱⴱ ⫽ P ⬍ 0.001, by t-test. HPRT ⫽ hypoxanthine guanine phosphoribosyltransferase.
assess CD248-dependent changes. For all studies, at least 3 independent clones from mice of each genotype were tested, and results were comparable. All MEFs had a similar appearance under direct microscopic visualization. We investigated the role of the cytoplasmic domain of CD248 in promoting the adhesion of monocytic U937 cells to MEFs, in experiments comparing MEFs from CD248WT/WT and CD248CyD/CyD embryos. Under baseline conditions, the number of U937 cells adhering to CD248WT/WT MEFs did not differ from that adhering to CD248CyD/CyD MEFs (mean ⫾ SEM number of U937 cells per field, 50 ⫾ 10 in CD248WT/WT mice versus 53 ⫾ 4 in CD248CyD/CyD mice [n ⫽ 8 each]; P not significant)
(Figure 4). As expected, stimulation of MEFs with the proinflammatory cytokine TNF␣ resulted in increased adhesion of the U937 cells to CD248WT/WT MEFs. However, the response was significantly dampened when CD248CyD/CyD MEFs were stimulated with TNF␣ (mean ⫾ SEM number of U937 cells per field, 103 ⫾ 11 in CD248WT/WT mice versus 74 ⫾ 7 in CD248CyD/CyD mice [n ⫽ 8 each]; P ⫽ 0.043) (Figure 4). To determine whether other signaling pathways that are modulated by the cytoplasmic domain of CD248 are involved in promoting and perpetuating synovial inflammation, we performed additional in vitro experiments. PlGF is highly expressed by synovial fibroblasts in RA patients, in whom it exerts angiogenic and proin-
CD248 AND ITS CYTOPLASMIC DOMAIN IN ARTHRITIS
flammatory effects (6). PlGF promotes inflammation via several mechanisms, such as inducing cytokine release and chemotaxis in bone marrow–derived monocytes (7,28). We measured the transcript levels of PlGF and its receptor, VEGF receptor 1 (VEGFR-1), in the MEFs and determined that their levels were significantly reduced in CD248CyD/CyD MEFs (Figure 5A). These findings are consistent with the observation of reduced recruitment of inflammatory cells into the joints of CD248CyD/CyD mice after the onset of CAIA. Synovial fibroblasts also synthesize VEGF, which has numerous angiogenic and proinflammatory properties (6). Similar to the findings with regard to PlGF, VEGF is up-regulated by hypoxia and several soluble factors, including, for example, TGF. Cell-associated VEGF levels at baseline were not significantly different between CD248WT/WT and CD248CyD/CyD MEFs (Figures 5B and C). The addition of TGF caused VEGF levels to increase in cell lysates of CD248WT/WT MEFs, as measured by ELISA and quantitative RT-PCR. In contrast, the response of CD248CyD/CyD MEFs was notably dampened, since VEGF levels remained unchanged from baseline (Figures 5B and C). Since hypoxia-inducible factors (HIFs) mediate the hypoxia-induced transcriptional up-regulation of VEGF, PlGF, and VEGFR-1 (29), we also measured the transcript levels of HIF-1␣ in the MEFs. As predicted, CD248CyD/CyD MEFs had significantly lower HIF-1␣ transcript levels, both under baseline conditions and after stimulation with TNF␣ (Figure 5D).
Figure 6. CD248-dependent fibroblast release of matrix metalloproteinases (MMPs). Conditioned medium of fibroblasts collected from wild-type mice (CD248WT/WT) or mice lacking the cytoplasmic domain of CD248 (CD248CyD/CyD) was left unstimulated or stimulated with transforming growth factor  (TGF) at 3 ng/ml for 72 hours. Equal volumes from 3 independent fibroblast clones for each genotype were analyzed by gelatin zymography. Representative results are shown, with the bands indicating that there was a significant reduction in the amount of active MMP-9 released by TGF-treated CD248CyD/CyD fibroblasts.
3603
Impaired release of MMP-9 by CD248CyD/CyD fibroblasts. Activated fibroblasts release extracellular matrix–degrading proteinases, such as MMPs, that contribute to destruction of articular cartilage and underlying subchondral bone (30). Overexpression of CD248 in CHO cells has been shown to enhance MMP-9 activity (31). To determine whether the cytoplasmic domain of CD248 regulates MMP-9 release and activation, we collected conditioned medium from equal numbers of cultured CD248WT/WT and CD248CyD/CyD MEFs. The rate of growth of the cells was not different between the 2 strains of mice (results not shown). When the cells were quiescent, no discernible differences in MMP activity in the conditioned medium were observed by gelatin zymography (Figure 6). Since MMP release and activation may be induced by various cytokines (32), we also collected conditioned medium from TGF  stimulated MEFs. Under these conditions, release of active MMP-9 (but not MMP-2) from CD248CyD/CyD fibroblasts was significantly dampened (Figure 6). These findings are consistent with the less severe cartilage and bone damage exhibited in the arthritic joints of the CD248 CyD/CyD mice as compared with that in CD248WT/WT mice. DISCUSSION The central role of resident mesenchymal cell populations, particularly fibroblast-like synoviocytes and perivascular cells, in promoting joint inflammation is well established in RA and PsA (33–35) and has justified efforts to discover disease-specific molecules expressed by these cells. Unfortunately, very few molecules have been identified, hindering the development of targeted therapies. Based primarily on reports of its high cell surface expression by fibroblasts from RA synovium (12), we hypothesized that CD248 might be such a molecule. We therefore first confirmed that CD248 is expressed by human hyperplastic stromal fibroblasts in vivo, as shown by immunodetection in synovial biopsy samples from patients with RA and patients with PsA, but not in synovium from individuals without arthritis. These studies also showed that CD248 synthesis is up-regulated by synovial vessel perivascular cells during arthritis, findings that are in line with the notion that activation of perivascular cells, such as occurs in pathologic angiogenesis, is associated with enhanced expression of CD248 (12,15). This inflammation-specific mesenchymal cell pattern of CD248 expression prompted us to next assess the role of CD248 in arthritis. Using genetic approaches, we
3604
definitively showed, in an experimental arthritis model, that mice lacking CD248 or its cytoplasmic domain develop less severe arthritis, and that the intracellular domain of CD248 participates in signaling pathways that promote the inflammatory properties of fibroblast-like synoviocytes and synovial vessel perivascular cells. The findings are entirely novel and support the design of CD248-targeting strategies to reduce arthritis. The human CD248 gene is intronless and encodes a 95-kd multidomain protein of 757 amino acids. The mature protein comprises an N-terminal C-type lectin-like domain, a complement control protein domain, 3 epidermal growth factor (EGF)–like repeats, a mucin-like region, a single transmembrane segment, and a short cytoplasmic domain (25,26). Based on its C-type lectin, CD248 is a member of a family of cell surface glycoproteins that includes thrombomodulin (CD141) and the complement receptor C1qR (CD93) (for review, see ref. 36). All are involved in regulating innate immune responses, but only thrombomodulin and CD248 are known to be expressed by fibroblast-like synoviocytes (12,37,38). The organization of the ectodomains of these molecules is similar in that they have a structurally related N-terminal C-type lectin, a series of 3–6 EGFlike repeats, and a juxtamembranous mucin-like domain, and, at least in thrombomodulin, some of these structures have defined and distinctly different functions. We did not examine the role of the ectodomain of CD248, and our studies do not exclude the possibility that it participates in regulating inflammation. Indeed, other investigators have reported that the ectodomain of CD248 interacts with extracellular matrix proteins, which may promote activation of MMP-9 (31). Thus, it is possible that the critical role of the cytoplasmic domain of CD248 in promoting an inflammatory response, as we have shown herein, requires binding of one or more ligands to its ectodomain. Studies to address this clinically important question have been initiated. A shared feature of the intracellular domains of CD248 and CD93 is a PDZ binding motif (36,39). In CD93, this domain interacts with a G␣ adaptor protein (40), which is believed to modulate the function of CD93 in phagocytosis and adhesion. No such interactions have yet been reported for the cytoplasmic domain of CD248, which, in addition to the PDZ motif, contains 3 highly conserved potential sites for phosphorylation (cAMPand cGMP-dependent protein kinase, casein kinase II, and protein kinase C [41]). These could be important in mediating functionally important intracellular signals (26), and future mutational analyses will uncover the relevant pathways.
MAIA ET AL
In spite of a gap in our understanding of the intracellular signals that are regulated by CD248, our studies show that mice lacking the cytoplasmic domain have less severe disease than their WT counterparts. Primary fibroblasts derived from the CD248CyD/CyD mice exhibit a range of defects that are consistent with the observed diminished inflammatory and invasive response. Although many pathways are implicated in activating fibroblast-like synoviocytes and perivascular stromal fibroblasts in arthritis (for review, see ref. 42), we focused on a few that illustrate the important role of the cytoplasmic domain of CD248 in promoting synovial hyperplasia, leukocyte influx and retention in the inflamed synovium, and fibroblast-like synovial cell invasion. Expression levels of PlGF, VEGF, and VEGFR-1, well-documented for their roles in angiogenesis and in inflammation, were either diminished at baseline in CD248CyD/CyD MEFs or were resistant to up-regulation by TGF. PlGF and VEGF in the RA joint have well-recognized and distinct proinflammatory properties. PlGF recruits activated leukocytes from the bone marrow to sites of inflammation, stimulating chemotaxis in monocytes and increasing production of TNF␣, IL-1, IL-8, and monocyte chemotactic protein 1 (MCP-1) (43,44). Both PlGF and its receptor, VEGFR-1, are highly expressed in fibroblast-like synoviocytes in RA synovium (7). In preclinical models, blocking VEGFR-1 dampens the progression of arthritis in mice (7,28). VEGF is also expressed by synovial fibroblasts in patients with RA, in whom it recruits monocytes by inducing the release of TNF␣, IL-6, IL-8, and MCP-1 by resident synovial cells and prevents synovial cell apoptosis (45). The latter effect is probably mediated by VEGF binding to its receptor neuropilin 1 (NP-1) (46). This is supported by studies showing that NP-1 blockade reduces VEGF-mediated synoviocyte survival and arthritis in mouse models (46). Overall, PlGF, VEGF, and VEGFR-1 promote synovial hyperplasia and leukocyte activation and recruitment, and thereby induce a self-perpetuating cycle of inflammation (for review, see refs. 6 and 7). Hypoxia plays a pivotal role in RA joints by inducing transcriptional up-regulation of many genes involved in inflammation, angiogenesis, and joint destruction, including those encoding VEGF, PlGF and their receptors, and MMPs (47). In hypoxic conditions associated with RA, the key hypoxia-inducible transcription factor HIF-1␣ is up-regulated and stabilized primarily in the synovium. HIF-1␣ expression is required for joint inflammation in experimental models and is recog-
CD248 AND ITS CYTOPLASMIC DOMAIN IN ARTHRITIS
nized as a potential therapeutic target for arthritis (47). CD248CyD/CyD MEFs have lower transcript levels of HIF-1␣, which likely explains, in part, the reduced levels of VEGF, PlGF, VEGFR-1, and MMP-9 and, consequently, the less severe arthritis in CD248CyD/CyD mice. We speculate that intracellular signals transmitted via the cytoplasmic domain of CD248 may participate in up-regulating HIF-1␣, again highlighting the potential for CD248 as a novel therapeutic target. In summary, the objective of our study was to examine the role of CD248 in arthritis. We have shown that in human RA and PsA, this multidomain glycoprotein is specifically expressed in the inflamed synovium by fibroblast-like synoviocytes and activated perivascular cells. Importantly, our studies in transgenic mice demonstrate that inflammation may be significantly reduced by interfering with intracellular signaling pathways mediated by the cytoplasmic domain of CD248. It remains to be seen whether this can be accomplished by targeting the ectodomain of the molecule, or whether other pathways must be targeted.
3605
5.
6.
7.
8.
9.
10.
11.
ACKNOWLEDGMENTS We gratefully acknowledge Prof. Bellemans, Dr. Van Lauwe, Dr. Van den Neucker, Prof. Verschueren, and Dr. Van Landuyt for providing patient biopsy specimens. We thank all of the staff in the ES cell laboratories and the animal and histology facilities at our institutions for their invaluable support.
12.
13.
14.
AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Conway had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. Maia, Tavernier, Lories, Conway. Acquisition of data. Maia, de Vriese, Janssens, Moons, Van Landuyt, Lories, Conway. Analysis and interpretation of data. Maia, Lories, Conway.
REFERENCES 1. Flavell SJ, Hou TZ, Lax S, Filer AD, Salmon M, Buckley CD. Fibroblasts as novel therapeutic targets in chronic inflammation. Br J Pharmacol 2008;153 Suppl 1:S241–6. 2. 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. 3. Parsonage G, Filer AD, Haworth O, Nash GB, Rainger GE, Salmon M, et al. A stromal address code defined by fibroblasts. Trends Immunol 2005;26:150–6. 4. Luyten FP, Lories RJ, Verschueren P, de Vlam K, Westhovens R.
15.
16. 17.
18. 19.
20.
21.
Contemporary concepts of inflammation, damage and repair in rheumatic diseases. Best Pract Res Clin Rheumatol 2006;20: 829–48. Kiener HP, Watts GF, Cui Y, Wright J, Thornhill TS, Skold M, et al. Synovial fibroblasts self-direct multicellular lining architecture and synthetic function in three-dimensional organ culture. Arthritis Rheum 2010;62:742–52. Yoo SA, Kwok SK, Kim WU. Proinflammatory role of vascular endothelial growth factor in the pathogenesis of rheumatoid arthritis: prospects for therapeutic intervention. Mediators Inflamm 2008;2008:129873. Yoo SA, Yoon HJ, Kim HS, Chae CB, De Falco S, Cho CS, et al. Role of placenta growth factor and its receptor flt-1 in rheumatoid inflammation: a link between angiogenesis and inflammation. Arthritis Rheum 2009;60:345–54. Sandler C, Joutsiniemi S, Lindstedt KA, Juutilainen T, Kovanen PT, Eklund KK. Imatinib mesylate inhibits platelet derived growth factor stimulated proliferation of rheumatoid synovial fibroblasts. Biochem Biophys Res Commun 2006;347:31–5. Tsai SH, Sheu MT, Liang YC, Cheng HT, Fang SS, Chen CH. TGF- inhibits IL-1-activated PAR-2 expression through multiple pathways in human primary synovial cells. J Biomed Sci 2009;16:97. Paleolog EM, Young S, Stark AC, McCloskey RV, Feldmann M, Maini RN. Modulation of angiogenic vascular endothelial growth factor by tumor necrosis factor ␣ and interleukin-1 in rheumatoid arthritis. Arthritis Rheum 1998;41:1258–65. Rettig WJ, Garin-Chesa P, Healey JH, Su SL, Jaffe EA, Old LJ. Identification of endosialin, a cell surface glycoprotein of vascular endothelial cells in human cancer. Proc Natl Acad Sci U S A 1992;89:10832–6. MacFadyen JR, Haworth O, Roberston D, Hardie D, Webster MT, Morris HR, et al. Endosialin (TEM1, CD248) is a marker of stromal fibroblasts and is not selectively expressed on tumour endothelium. FEBS Lett 2005;579:2569–75. Lax S, Hou TZ, Jenkinson E, Salmon M, MacFadyen JR, Isacke CM, et al. CD248/endosialin is dynamically expressed on a subset of stromal cells during lymphoid tissue development, splenic remodeling and repair. FEBS Lett 2007;581:3550–6. Brady J, Neal J, Sadakar N, Gasque P. Human endosialin (tumor endothelial marker 1) is abundantly expressed in highly malignant and invasive brain tumors. J Neuropathol Exp Neurol 2004;63: 1274–83. Christian S, Winkler R, Helfrich I, Boos AM, Besemfelder E, Schadendorf D, et al. Endosialin (Tem1) is a marker of tumorassociated myofibroblasts and tumor vessel-associated mural cells. Am J Pathol 2008;172:486–94. Simonavicius N, Robertson D, Bax DA, Jones C, Huijbers IJ, Isacke CM. Endosialin (CD248) is a marker of tumor-associated pericytes in high-grade glioma. Mod Pathol 2008;21:308–15. Rouleau C, Curiel M, Weber W, Smale R, Kurtzberg L, Mascarello J, et al. Endosialin protein expression and therapeutic target potential in human solid tumors: sarcoma versus carcinoma. Clin Cancer Res 2008;14:7223–36. Carson-Walter EB, Winans BN, Whiteman MC, Liu Y, Jarvela S, Haapasalo H, et al. Characterization of TEM1/endosialin in human and murine brain tumors. BMC Cancer 2009;9:417. Davies G, Cunnick GH, Mansel RE, Mason MD, Jiang WG. Levels of expression of endothelial markers specific to tumourassociated endothelial cells and their correlation with prognosis in patients with breast cancer. Clin Exp Metastasis 2004;21:31–7. Nanda A, Karim B, Peng Z, Liu G, Qiu W, Gan C, et al. Tumor endothelial marker 1 (Tem1) functions in the growth and progression of abdominal tumors. Proc Natl Acad Sci U S A 2006;103: 3351–6. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987
3606
22.
23.
24.
25.
26.
27.
28.
29.
30. 31.
32.
33.
revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31:315–24. Taylor W, Gladman D, Helliwell P, Marchesoni A, Mease P, Mielants H, and the CASPAR Study Group. Classification criteria for psoriatic arthritis: development of new criteria from a large international study. Arthritis Rheum 2006;54:2665–73. Lallemand Y, Luria B, Haffner-Krausz R, Lonai P. Maternally expressed PGK-Cre transgene as a tool for early and uniform activation of the Cre site-specific recombinase. Transgenic Res 1998;7:105–12. Van de Wouwer M, Plaisance S, De Vriese A, Waelkens E, Collen D, Persson J, et al. The lectin-like domain of thrombomodulin interferes with complement activation and protects against arthritis. J Thromb Haemost 2006;4:1813–24. Christian S, Ahorn H, Koehler A, Eisenhaber F, Rodi HP, Garin-Chesa P, et al. Molecular cloning and characterization of endosialin, a C-type lectin-like cell surface receptor of tumor endothelium. J Biol Chem 2001;276:7408–14. Opavsky R, Haviernik P, Jurkovicova D, Garin MT, Copeland NG, Gilbert DJ, et al. Molecular characterization of the mouse Tem1/ endosialin gene regulated by cell density in vitro and expressed in normal tissues in vivo. J Biol Chem 2001;276:38795–807. Rabquer BJ, Pakozdi A, Michel JE, Gujar BS, Haines GK III, Imhof BA, et al. Junctional adhesion molecule C mediates leukocyte adhesion to rheumatoid arthritis synovium. Arthritis Rheum 2008;58:3020–9. Luttun A, Tjwa M, Moons L, Wu Y, Angelillo-Scherrer A, Liao F, et al. Revascularization of ischemic tissues by PlGF treatment, and inhibition of tumor angiogenesis, arthritis and atherosclerosis by anti-Flt1. Nat Med 2002;8:831–40. Gobble RM, Groesch KA, Chang M, Torry RJ, Torry DS. Differential regulation of human PlGF gene expression in trophoblast and nontrophoblast cells by oxygen tension. Placenta 2009; 30:869–75. Akhavani MA, Madden L, Buysschaert I, Sivakumar B, Kang N, Paleolog EM. Hypoxia upregulates angiogenesis and synovial cell migration in rheumatoid arthritis. Arthritis Res Ther 2009;11:R64. Tomkowicz B, Rybinski K, Foley B, Ebel W, Kline B, Routhier E, et al. Interaction of endosialin/TEM1 with extracellular matrix proteins mediates cell adhesion and migration. Proc Natl Acad Sci U S A 2007;104:17965–70. Stuelten CH, DaCosta Byfield S, Arany PR, Karpova TS, StetlerStevenson WG, Roberts AB. Breast cancer cells induce stromal fibroblasts to express MMP-9 via secretion of TNF-␣ and TGF-. J Cell Sci 2005;118:2143–53. Noss EH, Brenner MB. The role and therapeutic implications of fibroblast-like synoviocytes in inflammation and cartilage erosion in rheumatoid arthritis. Immunol Rev 2008;223:252–70.
MAIA ET AL
34. Ritchlin CT. Pathogenesis of psoriatic arthritis. Curr Opin Rheumatol 2005;17:406–12. 35. Robinson H, Kelly S, Pitzalis C. Basic synovial biology and immunopathology in psoriatic arthritis. J Rheumatol Suppl 2009; 83:14–6. 36. Greenlee MC, Sullivan SA, Bohlson SS. CD93 and related family members: their role in innate immunity. Curr Drug Targets 2008;9:130–8. 37. Crockard AD, Thompson JM, Malhotra R, McNeill TA. Increased expression of C1q receptors on neutrophils from inflammatory joint fluids. Immunol Lett 1993;36:195–201. 38. Conway EM, Nowakowski B. Biologically active thrombomodulin is synthesized by adherent synovial fluid cells and is elevated in synovial fluid of patients with rheumatoid arthritis. Blood 1993;81: 726–33. 39. Nourry C, Grant SG, Borg JP. PDZ domain proteins: plug and play! Sci STKE 2003;2003:RE7. 40. Bohlson SS, Zhang M, Ortiz CE, Tenner AJ. CD93 interacts with the PDZ domain-containing adaptor protein GIPC: implications in the modulation of phagocytosis. J Leukoc Biol 2005;77:80–9. 41. De Castro E, Sigrist CJ, Gattiker A, Bulliard V, LangendijkGenevaux PS, Gasteiger E, et al. ScanProsite: detection of PROSITE signature matches and ProRule-associated functional and structural residues in proteins. Nucleic Acids Res 2006;34: W362–5. 42. Maciejewska Rodrigues H, Jungel A, Gay RE, Gay S. Innate immunity, epigenetics and autoimmunity in rheumatoid arthritis. Mol Immunol 2009;47:12–8. 43. Clauss M, Weich H, Breier G, Knies U, Rockl W, Waltenberger J, et al. The vascular endothelial growth factor receptor Flt-1 mediates biological activities: implications for a functional role of placenta growth factor in monocyte activation and chemotaxis. J Biol Chem 1996;271:17629–34. 44. Selvaraj SK, Giri RK, Perelman N, Johnson C, Malik P, Kalra VK. Mechanism of monocyte activation and expression of proinflammatory cytochemokines by placenta growth factor. Blood 2003; 102:1515–24. 45. Kim WU, Kang SS, Yoo SA, Hong KH, Bae DG, Lee MS, et al. Interaction of vascular endothelial growth factor 165 with neuropilin-1 protects rheumatoid synoviocytes from apoptotic death by regulating Bcl-2 expression and Bax translocation. J Immunol 2006;177:5727–35. 46. Kong JS, Yoo SA, Kim JW, Yang SP, Chae CB, Tarallo V, et al. Anti–neuropilin-1 peptide inhibition of synoviocyte survival, angiogenesis, and experimental arthritis. Arthritis Rheum 2010;62: 179–90. 47. Westra J, Molema G, Kallenberg CG. Hypoxia-inducible factor-1 as regulator of angiogenesis in rheumatoid arthritis—therapeutic implications. Curr Med Chem 2010;17:254–63.
