Synoviocyte-Mediated Expansion of Inflammatory T Cells in Rheumatoid Synovitis Is Dependent on CD47-Thrombospondin 1 Interaction

Abbe N. Vallejo, Hongyu Yang, Piotr A. Klimiuk, Cornelia M. Weyand and Jörg J. Goronzy
J Immunol August 15, 2003, 171 (4) 1732-1740; DOI: https://doi.org/10.4049/jimmunol.171.4.1732

Abstract

Fibroblast-like synoviocytes (FLS) from patients with rheumatoid arthritis elicit spontaneous proliferation of autologous T cells in an HLA-DR and CD47 costimulation-dependent manner. T cell costimulation through CD47 is attributed to specific interaction with thrombospondin-1 (TSP1), a CD47 ligand displayed on FLS. CD47 binding by FLS has broad biological impact that includes adhesion and the triggering of specific costimulatory signals. TSP1+ FLS are highly adhesive to T cells and support their aggregation and growth in situ. Long-term cultures of T cells and FLS form heterotypic foci that are amenable to propagation without exogenous growth factors. T cell adhesion and aggregate formation on TSP1+ FLS substrates are inhibited by CD47-binding peptides. In contrast, FLS from arthroscopy controls lack adhesive or T cell growth-promoting activities. CD47 stimulation transduces a costimulatory signal different from that of CD28, producing a gene expression profile that included induction of ferritin L chain, a component of the inflammatory response. Ferritin L chain augments CD3-induced proliferation of T cells. Collectively, these results demonstrate the active role of FLS in the recruitment, activation, and expansion of T cells in a CD47-dependent manner. Because TSP1 is abundantly expressed in the rheumatoid synovium, CD47-TSP1 interaction is proposed to be a key component of an FLS/T cell regulatory circuit that perpetuates the inflammatory process in the rheumatoid joint.

Proliferation of fibroblast-like synoviocytes (FLS)5 and the infiltration and in situ expansion of lymphocytes characterize the inflamed joints of patients with rheumatoid arthritis (RA). Although FLS phenotypes are poorly defined, heterogeneity of FLS function has been reported (1, 2, 3). Among their properties are the abilities to induce autoproliferation and to prevent apoptosis of tissue-infiltrating T cells (4, 5), suggesting that FLS are active participants of immune activation in inflammatory synovitis.

Although the antigenic basis of FLS-induced activation of autologous T cells is unknown, we have reported that this phenomenon is HLA-DR restricted and costimulation dependent (6). Costimulation is brought about by the interaction of CD47 on responder T cells and thrombospondin-1 (TSP1) displayed on stimulatory FLS. TSP1 is likely anchored on CD36, a known TSP1R (7) that is expressed on FLS. Indeed, our studies have shown that T cell costimulation can be mediated through a trimolecular complex of CD47-TSP1-CD36 (6). This novel costimulatory complex elicits vigorous proliferative responses of autologous T cells to TSP1+ FLS and prevents their apoptosis. These findings corroborate and provide biological relevance to the observation that cross-linking of CD47 by specific Ab augments TCR/CD3-induced IL-2 production by both murine and human T cells (8, 9). Collectively, these observations authenticate CD47 as a costimulatory molecule. By augmenting and sustaining T cell proliferation, increasing IL-2 production, and preventing apoptosis, CD47 fulfills the three biological criteria of costimulation as exemplified by CD28, the prototypic costimulatory molecule (10). These findings also support the notion that the paradigm of costimulation (11) is applicable to both autoreactive and Ag-specific T cells (12, 13).

CD47 and CD28 are constitutively expressed on a vast majority of CD4+ T cells (Refs. 6, 8 , and 10 ; our unpublished observations), unlike other known costimulatory molecules that are expressed subsequent to activation (14, 15). Because both molecules are potent costimulators of T cell growth (6, 8, 9), coexpression of CD47 and CD28 raises two issues. The first is the physiological context in which either CD47 or CD28 (or both) may be the relevant costimulator. The second is whether or not CD47- and CD28-mediated costimulation elicit different biological outcomes. Because the CD47 ligand, TSP1, is expressed primarily on injured and/or repairing tissue (16), we postulated that CD47-TSP1 interaction is a dominant costimulatory interaction that regulates T cell recruitment and tissue infiltration. This hypothesis predicts that up-regulation of TSP1 in chronically inflamed tissue, such as the rheumatoid synovium, is a critical factor in the inflammatory process. Hence, the pattern(s) of TSP1 expression in the inflamed joint was examined. We assessed the T cell growth-promoting activities of TSP1-expressing cells and examined whether the specific ligation of CD47 leads to distinctive growth-enhancing signals.

Materials and Methods

Tissue collection

Blood and synovial tissue were obtained from patients with RA who met the 1987 American College of Rheumatology criteria for the diagnosis of RA (17). Synovial tissue samples were collected from patients with RA who were undergoing joint replacement surgery, and control synovial tissue samples were obtained from arthroscopic biopsies of trauma patients. All donors provided informed consent, and biological specimens were prepared/stored according to Mayo Clinic Institutional Review Board-approved protocols.

Immunohistochemical studies

Synovial tissues were embedded in OCT compound (Sakura Finetek, Torrance, CA) and processed for immunohistochemical staining as described previously (18). Briefly, 0.5-μm cryostat sections were labeled with mouse mAbs specific for human TSP1 (P12), CD36 (FA6-152), or CD68 (all Beckman Coulter, Miami, FL), washed, treated with biotinylated anti-mouse Ig (DAKO, Carpinteria, CA), and developed using streptavidin-peroxidase (1:250) and diaminobenzidine (Sigma-Aldrich, St. Louis, MO). Negative controls using an IgG isotype were processed in parallel. Sections were counterstained with hematoxylin and mounted in Cytoseal-280 (Stephens Scientific, Riverdale, NJ).

For two-color immunohistochemistry, tissue staining with the second mAb was detected with VectaStain ABC (avidin/biotin complex)-alkaline phosphatase and Vector Red (Vector Laboratories, Burlingame, CA). To verify the specificity of staining for each of the Ags in the two-color slides, adjacent serial sections were separately stained with the individual mAbs and developed reciprocally with peroxidase/diaminobenzidine and Vector Red.

Densities of the cellular infiltrate and the topographical organization of mononuclear cells were classified as described previously (18).

Cell culture

CD4+ T cell lines and clones were established as described previously (19, 20). Cells were propagated in RPMI 1640 medium supplemented with 10% FCS (both BioWhittaker, Walkersville, MD), 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin sulfate (Invitrogen, Carlsbad, CA), and 20 U/ml rIL-2 (Proleukin; Chiron, Emeryville, CA). T cell clones and lines used in these studies included both CD4+CD28+ and CD4+CD28null cells.

FLS were isolated from synovial tissue samples by plastic adherence of collagenase-treated tissue and maintained as described previously (6). To propagate FLS, monolayers were exposed to 0.25% trypsin-EDTA solution (Sigma-Aldrich) for 2–5 min at 37°C. Detached cells were washed twice in culture medium, seeded to new T75 culture flasks at 1000 cells/ml, and grown to confluence. Neither the viability nor the phenotype of FLS was affected by trypsinization as determined by trypan blue staining and by flow cytometry (see below), respectively. For these studies, FLS were used between the third and eighth passages.

The human foreskin fibroblast line Hs69 was obtained from American Type Culture Collection (Manassas, VA). Cells were maintained in the RPMI 1640 culture medium used for T cell culture without IL-2 supplementation.

Phenotyping of cells

T cells were routinely analyzed by flow cytometry for the expression of CD3, CD4, CD28, and CD47 using fluorochrome-conjugated specific mAbs (BD Biosciences, San Diego, CA). Flow cytometric data were acquired by a FACSCalibur cytometer (BD Biosciences) and were analyzed offline using the WinMDI program (J. Trotter, Scripps Research Institute, La Jolla, CA). CD3, CD4, CD28, and CD47 expression on all of the T cell clones and lines used in these studies were stable (data not shown).

FLS and dermal fibroblast lines were similarly routinely analyzed for the expression of HLA-DR, CD36, and TSP1. HLA-DR and CD36 expression were detected by direct immunostaining with fluorochrome-conjugated mAbs (BD Biosciences). TSP1 expression was detected by indirect immunostaining (P12 mAb; Beckman Coulter) followed by a fluorochrome-conjugated goat anti-mouse IgG (BD Biosciences). TSP1 expression was also verified by immunocytochemical staining and by RT-PCR assays as described previously (6). Based on these analyses, all fibroblast lines used in the present studies expressed HLA-DR. All of the FLS lines isolated from patients with RA were ≥90% TSP1+ and CD36+, whereas the control FLS were minimally positive (≤3%) for either Ag. TSP1 and CD36 expression were stable and were unaffected by trypsinization and by treatment with TNF-α at the indicated concentrations (data not shown).

Adhesion assays

FLS were seeded at 1 × 104 cells/well in 24-well tissue culture plates and incubated overnight in the presence or absence of 1–10 ng/ml recombinant human TNF-α (R&D Systems, Minneapolis, MN). These concentrations of TNF-α were empirically determined to be nontoxic either to FLS or to T cells (21). The wells were washed, and CD4+CD28+ or CD4+CD28null T cells were added at a 5:1 T cell:FLS ratio. In the appropriate wells, the synthetic peptide 4N1K, which corresponds to the CD47-binding domain of TSP1, or its mutated variant, 4NGG (6, 22), were added at the indicated concentrations (both peptides were synthesized at the Mayo Protein Core Facility). The plates were incubated at 37°C on a rocking platform for 8–16 h. Culture supernatants were carefully aspirated, the wells were gently washed, and T cell adhesion on the FLS substrates was examined by phase-contrast microscopy. Adhesion assays were conducted in triplicate cultures in at least two independent experiments for each FLS line examined.

To facilitate visualization and counting of the adherent T cells, the T cells were labeled with CFSE (Molecular Probes, Eugene, OR) (23) before the assay. CFSE-labeled cells were visualized by UV fluorescence microscopy.

Differential display (DD)

To examine whether costimulation through CD47 is distinct from that through CD28, DD experiments were performed. Approximately 2 × 107 CD4+CD28+ T cells were incubated with a suboptimal concentration (50 ng/ml) of anti-CD3 mAb (Orthoclone; Ortho Diagnostics, Raritan, NJ) and 500 ng/ml anti-CD47 (clone 2E11; provided by Dr. F. Lindberg, Washington University in St. Louis, St. Louis, MO) (8, 24) or anti-CD28 (clone 28-2; BD PharMingen, San Diego, CA) (25) mAb immobilized on rabbit anti-mouse Ig-coated plates (ICN Pharmaceuticals, Costa Mesa, CA) (6). Controls included cells incubated with anti-CD47 or anti-CD28 mAb alone, with a suboptimal dose of anti-CD3 mAb alone, or with rabbit anti-mouse Ig alone. After overnight incubation, the cells were harvested and mRNA was extracted and subjected to restriction fragment DD-PCR assays using the Display-PROFILE kit and 64 sets of amplification primers (Display Systems Biotech, Vista, CA). DD-PCR products were labeled with [α-32P]dCTP (Amersham Pharmacia Biotech, Piscataway, NJ), and the cDNA fragments were fractionated by polyacrylamide gel electrophoresis and visualized by autoradiography.

Where differentially expressed cDNA bands were indicated in the autoradiograms, such bands were excised from the gel. The cDNA was extracted, amplified by PCR using the appropriate primer sets used in the DD-PCR, and the products were cloned and transformed into bacterial hosts using the TA cloning system (Invitrogen). Recombinant plasmids were prepared, the cloned cDNA fragment was sequenced, and the encoding gene was identified by sequence homology database searches using GCG software (Accelrys, Madison, WI).

T cell/FLS cocultures

Long-term cultures of FLS and autologous CD4+28+ or CD4+CD28null T cell lines and/or clones were conducted as indicated. Briefly, FLS were seeded in six-well tissue culture plates and grown to confluence. T cells were added at 1 × 106 cells/well with or without 20 U/ml rIL-2 (Proleukin). The cultures were monitored daily for T cell proliferation, and the cell densities were estimated using a hemocytometer. When T cell numbers increased to ≥5 × 106/ml, the culture supernatants, which contained the nonadherent T cells, were carefully aspirated, leaving the FLS in the wells. The nonadherent T cells were resuspended in fresh medium and added back to the same culture plates at 1 × 106 T cells/well. FLS/T cell cocultures were maintained in the presence of 100 μM peptide 4N1K or its mutated variant, 4NGG (6, 22). Cell morphology of the cultures was examined periodically with a phase-contrast microscope. Long-term autologous FLS/T cell cocultures were maintained for 12 wk. Similar cultures with FLS and heterologous T cells did not progress beyond 10 days of culture, because the T cells rapidly died as determined by trypan blue staining (data not shown).

T cell activation assay and Western blotting

A suboptimal concentration of anti-CD3 mAb (50 ng/ml) and 500 ng/ml anti-CD28 mAb, anti-CD47 mAb, or mouse IgG (R&D Systems) were immobilized on rabbit anti-mouse Ig-coated plates (6). T cells were added at 1 × 106/well and incubated overnight at 37°C. Cells from triplicate cultures were harvested and lysed in a hypotonic buffer. Cell lysates were prepared and subjected to Western blotting by standard procedures (20, 26). The presence of ferritin L chain (L-ferritin) was detected with the mAb LF03 (provided by Drs. P. Arosio and P. Santambrogio, University of Brescia, Brescia, Italy) (27, 28).

In activation assays, triplicate cultures of 2 × 105 T cells were incubated with a suboptimal concentration of anti-CD3 mAb (50 ng/ml) immobilized on rabbit anti-mouse Ig-coated plates (6) in the presence or absence of human recombinant L-ferritin (Calbiochem, San Diego, CA). Empirical studies showed that concentrations of ≤50 ng/ml L-ferritin were nontoxic. CD25 expression was determined by flow cytometry after 24 h.

T cells used in the activation studies were newly established short-term cell lines from peripheral blood or were previously established cell lines/clones (6, 19, 20, 26). Fresh T cells were isolated from sheep RBC-rosetted PBMCs. T cell lines were established from these rosetted cells by standard procedure as described previously (6, 20, 26). The phenotypes of the short-term T cell lines were determined by flow cytometry. Assays with the clones/lines were conducted using cells 10 days after the last stimulation.

T cell proliferation assays have been described previously (6, 20). Ten days after the last stimulation, triplicate cultures of T cells were incubated with optimal (50 ng/ml) or suboptimal (5 μg/ml) concentrations of anti-CD3 mAb immobilized on rabbit anti-mouse Ig-coated plates (6) in the presence or absence of 10 ng/ml human recombinant L-ferritin (Calbiochem). As controls, parallel cultures were incubated with immobilized mouse IgG (R&D Systems) or L-ferritin alone. After 2 days of culture, 1 μCi of [3H]thymidine (NEN, Boston, MA) was added, and the cultures were incubated for another 12–16 h. Thymidine incorporation was measured by liquid scintillation spectrometry.

Results

TSP1 is expressed in the rheumatoid synovium

Unlike the nearly acellular and avascular normal joint, the rheumatoid synovium is a fairly vascularized tissue composed of a conglomerate of infiltrating cells, such as T cells, B cells, and macrophages, and of resident FLS of mesenchymal origin (29). Although TSP1 is a transient component of the extracellular matrix in injured or repairing tissues (16), immunohistochemical studies showed that it was abundantly expressed in the rheumatoid synovium. TSP1 distribution was characteristically focal rather than of a diffuse or generalized distribution. TSP1 was found predominantly on the endothelial lining of synovial capillaries. Moreover, it colocalized with CD36, a known TSP1R (7). Colocalization of this receptor-ligand pair on endothelial cells was manifested by the high intensity staining of dense capillary beds (Fig. 1⇓, A and B). In some tissues, the synovial capillaries appeared to be organized as vascular islands that also showed high staining intensity for TSP1 and CD36 (Fig. 1⇓C).

FIGURE 1.
FIGURE 1.

TSP1, a CD47 ligand, is predominantly expressed on endothelial cells and FLS in the rheumatoid synovium. Synovial tissues from patients with RA were processed for immunohistochemical staining for TSP1 and its receptor CD36, and the macrophage marker CD68. Photomicrographs shown are representative of synovial tissues examined from five patients with RA. Colocalization of TSP1 (red) and CD36 (brown) was found predominantly on the endothelial linings of capillaries (ca) (A and B) and in vascular islands (vi) (C), as well as some spindle-shaped fibroblast-like cells (>) in the tissue parenchyma (A, C, and D). In contrast, cells that doubly stained with TSP1 (brown; ∧) and the macrophage marker CD68 (red, arrows) were infrequent (E).

TSP1 and CD36 were also coexpressed in a subpopulation of FLS that were scattered throughout the tissue parenchyma (Fig. 1⇑, A, C, and D). TSP1+CD36+ FLS were conspicuously seen as elongated- to stellate-shaped cells situated outside vascular islands. The colocalization of TSP1 and CD36 among FLS and on endothelial linings was confirmed by two-color reciprocal immunodetection to visualize both Ags. This colocalization was observed in all of the synovial tissues samples examined.

In contrast, two-color immunohistochemical staining for TSP1 (or CD36) and the macrophage marker CD68 (30) did not show significant colocalization of Ags. As shown in Fig. 1⇑E, cells that expressed CD68 were readily identified, but they rarely showed appreciable staining for either CD36 or TSP1.

FLS lines were established from synovial tissue biopsies. FLS were plastic-adherent spindle- or stellate-shaped cells that generally grew as confluent monolayers (Fig. 2⇓A), characteristics typical of fibroblasts (1). They uniformly lacked CD68 as examined by immunocytochemical staining (data not shown). However, they expressed TSP1 and CD36 (Fig. 2⇓B). Two-color flow cytometric analysis showed that FLS coexpressed equivalent levels of TSP1 and CD36. They also expressed high levels of HLA-DR. TSP1, CD36, and HLA-DR expression were stable in all FLS lines examined, derived from five patients with RA. Consistent with previous studies (6), relatively high levels of TSP1 and CD36 expression were maintained even at the eighth or tenth passage, as measured by flow cytometry and by immunocytochemical staining for the protein, and by RT-PCR for specific transcripts.

FIGURE 2.
FIGURE 2.

Rheumatoid FLS lines express TSP1 and CD36. A, FLS lines were established from synovial tissue biopsies of patients with RA. At least two separate lines were isolated from each of five patients. Photomicrograph shown was a typical monolayer of spindle- to stellate-shaped CD68-negative FLS at the second passage. B, Phenotypes of FLS lines were examined periodically by immunofluorescence staining and flow cytometry. Histograms shown were representative of FLS subjected to two-color staining for HLA-DR, CD36, and TSP1. The level of specific expression for each Ag is marked by R1.

CD47-dependent adhesion and expansion of T cells on TSP1+ FLS substrates

Because TSP1 is a matricellular protein that facilitates cell adhesion and motility (16), we examined whether TSP1+ FLS are effective adhesion substrates for T cells. Results of adhesion assays showed that unstimulated FLS derived from patients with RA were naturally adhesive for T cells (Fig. 3⇓A), and these adhesive properties increased significantly when the FLS were preincubated with TNF-α in a dose-dependent fashion. In contrast, FLS derived from arthroscopy controls, which largely lack TSP1 (data not shown), were minimally adhesive. However, the adhesive property of control FLS also increased when preincubated with TNF-α, but the adhesiveness only attained the levels observed with unstimulated RA-derived FLS. Such low adhesive properties of control FLS were recapitulated in experiments using the dermal fibroblast line Hs69 (data not shown). It is important to note that the concentrations of TNF-α (1 and 10 ng/ml) used in these studies were empirically determined to be nontoxic (21). Fibroblast viability remained at ≥97% as determined by trypan blue exclusion and by annexin V staining.

FIGURE 3.
FIGURE 3.

T cells adhere to rheumatoid FLS in a CD47-dependent manner. A, Rheumatoid FLS (FLS#1, FLS#2) and control FLS (NF) were seeded in 24-well tissue culture plates at 1 × 104 cells/well and incubated overnight with nontoxic concentrations of 1 and 10 ng/ml TNF-α. The plates were washed, and CFSE-labeled T cells were added to each well at a 5:1 T cell: FLS ratio. Cultures were incubated at 37°C on a rocking platform for 8 h, the supernatants were aspirated, and the wells were washed and examined by UV fluorescence microscopy. Data shown were the number of FLS-adherent T cells in 12 randomly selected low-power fields from three T cell/FLS cultures. B, Adhesion assays (as in A) were conducted with unstimulated rheumatoid FLS (FLS) or control FLS (NF) and CSFE-labeled T cells in the presence of 100 μM synthetic peptides corresponding to the CD47-binding domain of TSP1 (4N1K) or its mutated variant, 4NGG (22 ). The number of FLS-adherent T cells was determined from digitized images captured by combined UV fluorescence/phase-contrast microscopy.

Because the C-terminal end of TSP1 is a binding domain of CD47 (22), one of the TSP1Rs that is constitutively expressed on T cells (6, 8), experiments were conducted to determine the basis for T cell adhesion on RA-derived FLS substrates. Results showed that the synthetic peptide 4N1K, corresponding to the CD47-binding domain of TSP1 (22), significantly inhibited T cell adhesion on FLS (Fig. 3⇑B). In contrast, the mutated peptide variant, 4NGG, did not affect T cell adhesion. Similarly, the low but detectable levels of T cell adhesion on control FLS or dermal Hs69 fibroblasts were also specifically inhibited by the 41NK, but not the 4NGG, peptide.

The morphology of T cell adhesion on FLS was characteristically focal. As examined by phase-contrast and dark-field microscopy (Fig. 4⇓A), T cells that did not adhere to FLS generally appeared as spherical structures with an opaque border under transmitted light. Such light-opaque borders were interrupted in the T cells that were in physical contact with the irregular- to stellate-shaped FLS. These T cell/FLS contact points appeared as a coalescence of the intercellular junction. Such contact points were not observed in T cell/FLS cultures in the presence of the 4N1K peptide. T cell/FLS contact was not altered in the presence of the mutated 4NGG peptide.

FIGURE 4.
FIGURE 4.

T cell/FLS adhesion involves focal contact that results in the formation of heterotypic foci and the long-term growth of T cells. A, Parallel mixed-cell cultures of T cells and rheumatoid FLS (F) were set up using unlabeled cells. After 8 h of incubation with gentle agitation, the nonadherent T cells were carefully aspirated, and the morphology of T cell/FLS adhesion was examined by phase-contrast microscopy. Photomicrographs are representative of six T cell/FLS cocultures examined. Arrows in the insets point to T cells that formed focal contacts on the surfaces of spindle- or stellate-shaped FLS. B, Unstimulated RA-derived FLS (F) were seeded in six-well plates and grown to confluence. Autologous T cell clones (↑) were subsequently added at a 2:1 T cell:FLS ratio. Cultures were maintained for 8 wk without additional T cell stimulation or exogenous growth factors except for 100 μM CD47-binding peptide, 4N1K, or its mutated variant, 4NGG. T cell/FLS foci () were commonly seen in cultures with medium alone or with the 4NGG peptide. The phase-contrast photomicrograph was taken after 15 days of culture and was representative of six cultures examined. C, Individual T cell/FLS foci in B were isolated, deaggregated by trypsinization, and cultured in fresh medium. The photomicrograph shown depicts two daughter foci formed after the third passage of one of the original heterotypic foci. Data shown are representative of five T cell/FLS foci-propagation cultures examined.

Because TSP+ FLS derived from patients with RA are stimulatory for autologous T cells in standard proliferation assays (6), the ability of these synoviocytes to sustain long-term T cell growth was evaluated. Results showed that FLS elicited T cell expansion and survival for a period of at least 8 wk even in the absence of exogenous IL-2 or any other form of stimulation. In prolonged autologous T cell-FLS cocultures (>15 days), there was a characteristic aggregation of T cells anchored on a few FLS that formed foci of growth (Fig. 4⇑B). Isolation of these T cell/FLS aggregates and subsequent culture of deaggregated cells resulted in the formation of new heterotypic foci (Fig. 4⇑C).

Formation of these aggregates was observed in autologous, but not heterologous, FLS/T cell cocultures. Formation of heterotypic foci and the long-term growth of T cells were observed in FLS/T cell cocultures in medium alone or in the presence of the 4NGG peptide (Fig. 4⇑B). The addition of the CD47-binding peptide, 4N1K, to the cultures prevented foci formation and limited the expansion of T cells to no more than 10 days of culture (data not shown).

The intrinsic and TNF-α-enhanced adhesive (Fig. 3⇑) and T cell growth-promoting properties (Fig. 4⇑) of FLS were seen in all FLS lines from at least five patients with RA who were examined. FLS consistently supported the growth of autologous CD28+ and CD28null T cells in situ without additional exogenous growth factors.

Differential gene expression through CD47-mediated costimulation

To examine whether costimulation of T cells through CD47 is distinct from that through CD28, gene-profiling studies were conducted. T cells were stimulated with a suboptimal concentration of anti-CD3 mAb and optimal concentrations of anti-CD28 (25) or anti-CD47 mAbs (6) and subjected to DD-PCR assays. Results showed that the profile of genes induced through CD47 differed significantly from that seen with CD28-mediated costimulation. As shown in Fig. 4⇑, some genes were clearly more inducible by CD47 costimulation, but not by CD28 costimulation, and vice versa. As expected, there were also numerous genes that were equivalently induced by the ligation of either receptor. In three DD-PCR assays, there was a prominent band of cDNA induced by CD47 costimulation (Fig. 5⇓, arrow) that hybridized with 12 of 64 random display probes. All 12 cDNA bands were cloned and sequenced. Results showed that they had identical DNA sequences that corresponded to L-ferritin. This molecule is a noncatalytic protein that normally complexes with the catalytic ferritin H chain (H-ferritin) forming the iron-binding ferritin heteropolymer (31). However, as a homopolymer, L-ferritin is a component of the acute-phase response (31, 32, 33).

FIGURE 5.
FIGURE 5.

Costimulation of T cells through CD47 or CD28 induces a distinct gene profile. T cells were stimulated with a suboptimal concentration (50 ng/ml) of anti-CD3 mAb and with 200 ng/ml anti-CD47 or anti-CD28 mAb for 24 h (6 ,25 ). Total RNA was extracted and subjected to restriction fragment DD-PCR experiments. The autoradiograph is the gene induction profile of cells costimulated through CD47 or CD28 as detected by 7 of 64 random primers (P2 to P8) used. The arrow indicates the most commonly expressed cDNA fragment from CD47-stimulated cells. DNA cloning and sequencing revealed that these were identical cDNAs, which encoded L-ferritin.

CD47-mediated production of L-ferritin and the augmentation of T cell proliferation by L-ferritin

Because of the prominence of L-ferritin in the gene-profiling studies, we examined whether L-ferritin production by T cells was indeed costimulation dependent. In these studies, CD4+CD28+ T cell clones or lines were stimulated with a suboptimal concentration of anti-CD3 mAb in the presence of an optimal concentration of anti-CD47 mAb. As shown in Fig. 6⇓A, coligation of CD3 and CD47 resulted in the significant induction of L-ferritin as detected in Western blots. The CD47-induced L-ferritin showed a slightly higher molecular size than the recombinant controls, which may likely be due to posttranslational modification as has been reported in a variety of cell types (27, 28, 34). In contrast, T cells that were costimulated with anti-CD28 mAb did not show any detectable L-ferritin. Stimulation of T cells with anti-CD3 alone also failed to induce L-ferritin production (data not shown). It should be noted that the isotypes and concentrations of anti-CD3 and anti-CD47 (or anti-CD28) mAb used in these experiments were those that elicited high levels of T cell proliferation in our previous studies (6).

FIGURE 6.
FIGURE 6.

Costimulation of T cells through CD47, but not CD28, induces L-ferritin, which activates T cells and augments T cell growth. A, CD4+CD28+ T cells were stimulated with a suboptimal concentration of anti-CD3 mAb and with 500 ng/ml anti-CD47 or anti-CD28 mAb for 24 h. Total cell lysates were prepared and subjected to Western blotting with anti-human L-ferritin Ab. rhLF, recombinant L-ferritin. B, T cell lines were stimulated with a suboptimal concentration (50 ng/ml) of plate-immobilized anti-CD3 mAb (OKT3) in the presence or absence of human recombinant L-ferritin (10 ng/ml). After 24 h of incubation, CD25 expression was examined by flow cytometry. C, Previously activated T cells were incubated in suboptimal (sOKT3) or optimal (hOKT3) concentrations of anti-CD3 mAb in the presence or absence of human recombinant L-ferritin. [3H]Thymidine incorporation was measured after 3 days of culture. Data shown are the means ± SD cpm of triplicate cultures of a T cell clone (□) and a short-term T cell line (•), which were representative of five clones and two lines examined.

T cells express receptors for L-ferritin (34, 35). Therefore, T cell activation assays were conducted to examine the biological significance of the induction of L-ferritin. As depicted in Fig. 6⇑B, incubation of previously activated T cells with a suboptimal concentration of anti-CD3 mAb in the presence of human recombinant L-ferritin induced high levels of CD25 expression. In contrast, neither recombinant L-ferritin nor suboptimal anti-CD3 mAb alone promoted significant expression of CD25.

The stimulatory activity of L-ferritin was further evaluated in T cell proliferation assays. As shown in Fig. 6⇑C, recombinant L-ferritin was found to augment [3H]thymidine incorporation in response to a suboptimal concentration of anti-CD3 mAb. Neither L-ferritin nor a suboptimal concentration of anti-CD3 mAb alone was stimulatory. As expected, high/optimal levels of anti-CD3 mAb induced vigorous proliferation.

Discussion

Data presented here support the idea that FLS actively participate in immune activation in inflammatory synovitis (4, 5, 6). They are highly efficient adhesion substrates (Fig. 3⇑) that actively support T cell growth (Fig. 4⇑). Whether or not the ability of FLS to support T cell expansion in situ is Ag driven is unclear. Nevertheless, our data demonstrate the central role of the CD47 costimulatory pathway in FLS-dependent autoproliferation of inflammatory T cells. Costimulation of T cells through CD47 is triggered through its interaction with TSP1, which is displayed on FLS. Consistent with our previous findings (6), the present data clearly show that both T cell adhesion on FLS (Fig. 3⇑) and long-term FLS-mediated T cell growth (Fig. 4⇑, B and C) are inhibited by the 4N1K peptide, which corresponds to the CD47-binding domain of TSP1 (22). Because CD47 is constitutively expressed on T cells (6), these studies suggest that CD47-TSP1 interaction can substitute and/or complement CD28-B7 interaction in the activation of naive and memory T cells (6, 8, 9). Because RA is associated with high frequencies of highly oligoclonal CD28null T cells (19), CD47-TSP1 interaction could also provide a basis for the infiltration of this unusual lymphocyte subset into the synovium (36). Parenthetically, CD28null T cells have also been found in other inflammatory conditions, such as Wegener’s granulomatosis, ankylosing spondylitis, multiple sclerosis, and atherosclerotic coronary artery disease (37, 38, 39, 40), suggesting that CD47-TSP1 interaction could be a common pathway for T cell infiltration and expansion in chronic inflammatory lesions.

TSP1 is a transient component of the extracellular matrix (16). Thus, CD47-TSP1 interactions are expected to be relevant in biological situations in which there is sustained TSP1 expression. Because TSP1 has been recognized as one of the unusual matrix proteins in inflammatory synovitis (41), we examined the pattern of TSP1 expression in rheumatoid synovial tissues. Unlike the rather diffuse deposition of TSP1 in the extracellular matrix of repairing tissue (16), we found consistent focal expression of TSP1 in the inflamed synovium, primarily on endothelial cells (Fig. 1⇑) and on a subpopulation of FLS (Fig. 2⇑). Such an expression pattern suggests a model for the role of TSP1-CD47 interaction in the maintenance of synovial inflammation. As depicted in Fig. 7⇓, expression of TSP1 on endothelial cells can facilitate synovial tissue infiltration by CD47+ T cells. Indeed, studies have shown that T cell arrest on vascular endothelium and their subsequent diapedesis can be regulated by CD47. Deficiency in CD47 expression (or mutations in the transmembrane region of CD47) severely impairs endothelial adhesion of T cells under flow (42). In the tissue parenchyma, interaction of TSP1+ FLS and CD47+ T cells can lead to the expansion of infiltrating T cells. This is supported by studies showing heterotypic foci formation and long-term growth of T cells cultured with autologous FLS in the absence of exogenous growth factors (Fig. 4⇑). Inasmuch as FLS-induced T cell proliferation is also HLA-DR dependent (6), the resulting expanded T cells in situ are likely to be autoreactive clonotypes. TSP+ FLS/CD47+ T cell interaction may also influence the maintenance or growth of FLS (1, 4, 43).

FIGURE 7.
FIGURE 7.

CD47-TSP1 interaction is critical to T cell infiltration and expansion in the rheumatoid synovium. A model is proposed in which CD47/TSP1-mediated processes play a key role in the inflammatory process of RA. The key features of the model are as follows. First, T cell infiltration of the synovium is facilitated by the up-regulated expression of TSP1 on endothelial cells (EC). T cell adhesion on the endothelium is brought about by the interaction of CD47 on the T cell surface with TSP1, which occurs in an integrin-independent manner. Second, within the synovial tissue parenchyma, the FLS is a potent nonclassical APC that displays the immunogen bound to HLA class II molecules. FLS also express TSP1, which facilitates focal adhesion of T cells and triggers costimulatory signals through its interaction with CD47. Third, CD47-mediated costimulation of T cells results in their proliferation and aggregation in situ. This is facilitated by the CD47-dependent production of L-ferritin, which elicits autocrine and paracrine stimulation of T cells. Fourth, T cell aggregation is also facilitated by FLS-derived chemokines such as RANTES. Finally, FLS/T cell interaction also induces the differentiation of FLS into a phenotype that promotes/sustains T cell expansion/aggregation.

Costimulation of T cells through CD47 is distinct from that through CD28, the prototypic costimulatory molecule. Previous studies have shown that CD47-mediated costimulation is driven by adhesion through CD47 itself (8, 9). Indeed, the present data show the marked inhibition of T cell adhesion by the CD47-binding peptide 4N1K, and consequently, in situ expansion and aggregation on TSP+ FLS substrates (Figs. 3⇑ and 4⇑). Ligation of CD47, but not CD28, has been demonstrated to elicit T cell spreading (44), which is thought to facilitate and strengthen T cell/APC contact. Although productive T cell activation, proliferation, and IL-2 production require coengagement of CD47 and the TCR, CD47-mediated spreading occurs independent of TCR ligation (44, 45). CD47 ligation itself induces actin polymerization and protein kinase Cθ translocation, which morphologically manifest as polarized amoeboid-like cytoplasmic structures. However, in conjunction with TCR ligation, CD47 induces raft formation that leads to synergy of TCR- and CD47-mediated protein kinase Cθ/cytoskeletal association. Collectively, these data indicate that CD47 costimulation is triggered, at least in part, by a unique second signal that is not evident with CD28 costimulation (46, 47). Therefore, it is of interest to examine the biochemical nature of the CD47 cosignal (45, 48), which could provide further insight into CD47/TSP1-regulated expansion of autoreactive/inflammatory T cells in the rheumatoid joint.

The present data show a unique biological outcome of CD47-mediated costimulation. Indeed, the array of genes induced by CD47 can be distinguished from those induced by CD28 (Fig. 5⇑). CD47 costimulation of T cells invariably results in the induction of L-ferritin (Fig. 6⇑A), a molecule involved in the storage and metabolism of iron. L-ferritin itself does not have ferridoxidase activity, but it complexes with H-ferritin to form the high-affinity catalytic iron-binding molecule ferritin (31). Interestingly, L-ferritin has been implicated in inflammatory responses as a component of the acute-phase response (33) and as one of the up-regulated serum proteins in certain inflammatory conditions such as adult-onset Still’s disease (49) and hyperferritinemia cataract syndrome (50).

The role of ferritins in the regulation of immune responses has been suggested with the identification of receptors for both l- and H-ferritins on T cells (35). H-ferritin, but not L-ferritin, has immunosuppressive effects by inhibiting T cell proliferation and by the induction of IL-10 production (51, 52). In contrast, the present work shows that L-ferritin is in fact stimulatory to T cells and is capable of augmenting proliferation (Fig. 6⇑, B and C), consistent with its suggested proinflammatory activity (32, 33). Thus, accumulation of L-ferritin in sites of inflammation could contribute to tissue injury through its paracrine growth effects. Along these lines, the finding that L-ferritin is overexpressed in the rheumatoid joint (53) suggests its role in the perpetuation of inflammation. Through the costimulatory interaction of CD47 with TSP1+ FLS (Figs. 4⇑ and 5⇑A), CD47 induces T cells to produce L-ferritin, thereby causing tissue-destructive T cells to expand and to be maintained (Figs. 2⇑ and 3⇑) (6).

In summary, the present work supports the idea that TSP1/CD47 interaction contributes to the perpetuation of inflammation in rheumatoid synovitis (Fig. 7⇑). Up-regulated expression of TSP1 on the vascular endothelium in the inflamed joint facilitates the infiltration of CD47+ T cells. In the synovial tissue, T cell expansion is promoted by FLS/T cell interaction through the TSP1/CD47 costimulatory pathway (6). This leads to the preferential expression of CD47-inducible proinflammatory molecules such as L-ferritin, which has paracrine and autocrine growth effects to promote T cell expansion in situ. FLS/T cell interaction also activates FLS (2, 4), which produce humoral factors, including chemokines like RANTES/CC chemokine ligand 5 (Ref. 54 ; our unpublished data) that facilitate further T cell infiltration, and various proteinases (3) that perpetuate tissue injury. Therefore, the inhibition of TSP1-CD47 interaction and/or the down-regulation of TSP1 expression could be potential mechanisms to dampen the inflammatory cascade in the rheumatoid joint.

Acknowledgments

We thank Dr. Paolo Arosio (Faculty of Medicine, University of Brescia) and Dr. Paolo Santambrogio (Instituti di Ricovero e Cura a Carattere Scientifico San Raffaele, Milan, Italy) for providing the anti-L-ferritin Ab LF03, and Dr. Frederik Lindberg (Washington University in St. Louis) for the 2E11 mAb. We thank James W. Fulbright for assistance in preparing the manuscript and Linda H. Arneson for secretarial support.

Footnotes

  • ↵1 This work was supported by National Institutes of Health Grants R03-AR45830, R01-AG15043, and R01-AR41974, and by the Mayo Foundation.

  • ↵2 Address correspondence and reprint requests to Dr. Abbe N. Vallejo, Mayo Clinic, Guggenheim 401, 200 First Street Southwest, Rochester, MN 55905. E-mail address: vallejo.abbe{at}mayo.edu

  • ↵3 Current address: Department of Pathology, University of Missouri Medical School, Columbia, MO 65212.

  • ↵4 Current address: Department of Rheumatology, Klinika Reumatologii-SPSK, Sklodowskiej 24a, 15-276 Bialystok, Poland.

  • ↵5 Abbreviations used in this paper: FLS, fibroblast-like synoviocyte; RA, rheumatoid arthritis; TSP1, thrombospondin-1; DD, differential display; L-ferritin, ferritin L chain; H-ferritin, ferritin H chain.

  • Received March 24, 2003.
  • Accepted June 18, 2003.

References

  1. ↵
    Ritchlin, C.. 2000. Fibroblast biology: effector signals released by the synovial fibroblast in arthritis. Arthritis Res. 2:356.
    OpenUrlCrossRefPubMed
  2. ↵
    Cho, C. S., M. L. Cho, S. Y. Min, W. U. Kim, D. J. Min, S. S. Lee, S. H. Park, J. Choe, H. Y. Kim. 2000. CD40 engagement on synovial fibroblast up-regulates production of vascular endothelial growth factor. J. Immunol. 164:5055.
    OpenUrlAbstract/FREE Full Text
  3. ↵
    DiBattista, J. A., J. P. Pelletier, M. Zafarullah, N. Fujimoto, K. Obata, J. Martel-Pelletier. 1995. Coordinate regulation of matrix metalloproteases and tissue inhibitor of metalloproteinase expression in human synovial fibroblasts. J. Rheumatol. 43:(Suppl.):123.
    OpenUrl
  4. ↵
    McInnes, I. B., B. P. Leung, F. Y. Liew. 2000. Cell-cell interactions in synovitis: interactions between T lymphocytes and synovial cells. Arthritis Res. 2:374.
    OpenUrlCrossRefPubMed
  5. ↵
    Salmon, M., D. Scheel-Toellner, A. P. Huissoon, D. Pilling, N. Shamsadeen, H. Hyde, A. D. D’Angeac, P. A. Bacon, P. Emery, A. N. Akbar. 1997. Inhibition of T cell apoptosis in the rheumatoid synovium. J. Clin. Invest. 99:439.
    OpenUrlCrossRefPubMed
  6. ↵
    Vallejo, A. N., L. O. Mügge, P. A. Klimiuk, C. M. Weyand, J. J. Goronzy. 2000. Central role of thrombospondin-1 in the activation and expansion of inflammatory T cells. J. Immunol. 164:2947.
    OpenUrlAbstract/FREE Full Text
  7. ↵
    Asch, A. S., S. Silbiger, E. Heimer, R. L. Nachman. 1992. Thrombospondin sequence motif (CSVTCG) is responsible for CD36 binding. Biochem. Biophys. Res. Commun. 182:1208.
    OpenUrlCrossRefPubMed
  8. ↵
    Reinhold, M. I., F. P. Lindberg, G. J. Kersch, P. M. Allen, E. J. Brown. 1997. Costimulation of T cell activation by integrin-associated protein (CD47) is an adhesion-dependent, CD28-independent signaling pathway. J. Exp. Med. 185:1.
    OpenUrlAbstract/FREE Full Text
  9. ↵
    Ticchioni, M., M. Deckert, F. Mary, G. Bernard, E. J. Brown, A. Bernard. 1997. Integrin-associated protein (CD47) is a comitogenic molecule on CD3-activated human T cells. J. Immunol. 158:677.
    OpenUrlAbstract
  10. ↵
    June, C. H., J. A. Ledbetter, P. S. Linsley, C. B. Thompson. 1990. Role of the CD28 receptor in T-cell activation. Immunol. Today 11:211.
    OpenUrlCrossRefPubMed
  11. ↵
    Chambers, C. A.. 2001. The expanding world of costimulation: the two signal-model revisited. Trends Immunol. 22:217.
    OpenUrlCrossRefPubMed
  12. ↵
    Whitmire, J. K., R. Ahmed. 2000. Costimulation in antiviral immunity: differential requirements for CD4+ and CD8+ T cell responses. Curr. Opin. Immunol. 12:448.
    OpenUrlCrossRefPubMed
  13. ↵
    Salomon, B., J. A. Bluestone. 2001. Complexities of CD28/B7:CTLA-4 costimulatory pathways in autoimmunity and transplantation. Annu. Rev. Immunol. 19:225.
    OpenUrlCrossRefPubMed
  14. ↵
    Watts, T. H., M. A. DeBenedette. 1999. T cell costimulatory molecules other than CD28. Curr. Opin. Immunol. 11:286.
    OpenUrlCrossRefPubMed
  15. ↵
    Mueller, D. L.. 2000. T cells: a proliferation of costimulatory molecules. Curr. Biol. 10:R227.
    OpenUrlCrossRefPubMed
  16. ↵
    Adams, J. C.. 2001. Thrombospondins: multifunctional regulators of cell interactions. Annu. Rev. Cell. Dev. Biol. 17:25.
    OpenUrlCrossRefPubMed
  17. ↵
    Arnett, F. C., S. M. Edworthy, D. A. Bloch, D. J. McShane, J. F. Fries, N. S. Cooper, L. A. Healey, S. R. Kaplan, M. H. Liang, H. S. Luthra. 1988. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum. 31:315.
    OpenUrlCrossRefPubMed
  18. ↵
    Klimiuk, P. A., J. J. Goronzy, J. Bjornsson, R. D. Beckenbaugh, C. M. Weyand. 1997. Tissue cytokine patterns distinguish variants of rheumatoid synovitis. Am. J. Pathol. 151:1311.
    OpenUrlPubMed
  19. ↵
    Schmidt, D., J. J. Goronzy, C. M. Weyand. 1996. CD4+CD7−CD28− T cells are expanded in rheumatoid arthritis and are characterized by autoreactivity. J. Clin. Invest. 97:2027.
    OpenUrlCrossRefPubMed
  20. ↵
    Schirmer, M., A. N. Vallejo, C. M. Weyand, J. J. Goronzy. 1998. Resistance to apoptosis and elevated expression of Bcl-2 in clonally expanded CD4+CD28− T cells from rheumatoid arthritis patients. J. Immunol. 161:1018.
    OpenUrlAbstract/FREE Full Text
  21. ↵
    Bryl, E., A. N. Vallejo, C. M. Weyand, J. J. Goronzy. 2001. Down-regulation of CD28 expression by TNF-α. J. Immunol. 167:3231.
    OpenUrlAbstract/FREE Full Text
  22. ↵
    Gao, A. G., F. P. Lindberg, M. B. Finn, S. D. Blystone, E. J. Brown, W. A. Frazier. 1996. Integrin-associated protein is a receptor for the COOH-terminal domain of thrombospondin. J. Biol. Chem. 271:21.
    OpenUrlAbstract/FREE Full Text
  23. ↵
    Weston, S. A., C. R. Parish. 1990. New fluorescent dyes for lymphocyte migration studies: analysis by flow cytometry and fluorescence microscopy. J. Immunol. Methods 133:87.
    OpenUrlCrossRefPubMed
  24. ↵
    Brown, E. J., L. Hooper, T. Ho, H. D. Gresham. 1990. Integrin-associated protein: a 50kD plasma membrane antigen physically and functionally associated with integrins. J. Cell Biol. 111:2785.
    OpenUrlAbstract/FREE Full Text
  25. ↵
    Vallejo, A. N., J. C. Brandes, C. M. Weyand, J. J. Goronzy. 1999. Modulation of CD28 expression: distinct regulatory pathways during activation and replicative senescence. J. Immunol. 162:6572.
    OpenUrlAbstract/FREE Full Text
  26. ↵
    Vallejo, A. N., M. Schirmer, C. M. Weyand, J. J. Goronzy. 2000. Clonality and longevity of CD4+CD28null T cells are associated with defects in apoptotic pathways. J. Immunol. 165:6301.
    OpenUrlAbstract/FREE Full Text
  27. ↵
    Santambrogio, P., S. Levi, A. Cozzi, E. Rovida, A. Albertini, P. Arosio. 1993. Production and characterization of recombinant heteropolymers of human ferritin H and L chains. J. Biol. Chem. 268:12744.
    OpenUrlAbstract/FREE Full Text
  28. ↵
    Corsi, B., F. Perrone, M. Bourgeois, C. Beaumont, M. C. Panzeri, A. Cozzi, R. Sangregorio, P. Santambrogio, A. Albertinini, P. Arosio, S. Levi. 1998. Transient over-expression of human H- and L-ferritin chains in COS cells. Biochem. J. 330:315.
    OpenUrlAbstract/FREE Full Text
  29. ↵
    Tak, P. P., B. Bresnihan. 2000. The pathogenesis and prevention of joint damage in rheumatoid arthritis: advances from synovial biopsy and tissue analysis. Arthritis Rheum. 43:2619.
    OpenUrlCrossRefPubMed
  30. ↵
    Martinez-Pomares, L., N. Platt, A. J. McKnight, R. P. da Silva, S. Gordon. 1996. Macrophage membrane molecules: markers of tissue differentiation and heterogeneity. Immunobiology 195:407.
    OpenUrlCrossRefPubMed
  31. ↵
    Harrison, P. M., P. Arosio. 1996. The ferritins: molecular properties, iron storage function and cellular regulation. Biochim. Biophys. Acta 1275:161.
    OpenUrlCrossRefPubMed
  32. ↵
    Torti, S. V., F. M. Torti. 1994. Iron and ferritin in inflammation and cancer. Adv. Inorg. Biochem. 10:119.
    OpenUrlPubMed
  33. ↵
    Leggett, B. A., L. M. Fletcher, G. A. Ramm, L. W. Powell, J. W. Halliday. 1993. Differential regulation of ferritin H and L subunit mRNA during inflammation and long-term iron overload. J. Gastroentrol. Hepatol. 8:21.
    OpenUrlCrossRefPubMed
  34. ↵
    Cairo, G., P. Vezzoni, L. Bardella, L. Schiaffonati, E. Rappociolo, S. Levi, P. Arosio, A. Bernelli-Zazzera. 1986. Regulation of ferritin synthesis in malignant and non-malignant lymphoid cells. Biochem. Biophys. Res. Commun. 139:652.
    OpenUrlCrossRefPubMed
  35. ↵
    Anderson, G. J., W. P. Faulk, P. Arosio, D. Moss, L. W. Powell, J. W. Halliday. 1989. Identification of H- and L-ferritin subunit binding sites on human T and B lymphoid cells. Br. J. Haematol. 73:260.
    OpenUrlCrossRefPubMed
  36. ↵
    Warrington, K. J., S. Takemura, J. J. Goronzy, C. M. Weyand. 2001. CD4+, CD28− T cells in rheumatoid arthritis patients combine features of the innate and adaptive immune systems. Arthritis Rheum. 44:13.
    OpenUrlCrossRefPubMed
  37. ↵
    Moosig, E., E. Csernok, G. Wang, W. L. Gross. 1998. Costimulatory molecules in Wegener’s granulomatosis (WG): lack of expression of CD28 and preferential upregulation of its ligands B7-1 (CD80) and B7-2 (CD86) on T cells. Clin. Exp. Immunol. 114:113.
    OpenUrlCrossRefPubMed
  38. ↵
    Schirmer, M., C. Goldberger, R. Wurzner, C. Duftner, K. P. Pfeiffer, J. Clausen, G. Neumayr, A. Falkenbach. 2002. Circulating cytotoxic CD8+CD28− T cells in ankylosing spondylitis. Arthritis Res. 4:71.
    OpenUrlPubMed
  39. ↵
    Markovic-Plese, S., I. Cortese, K. P. Wandinger, H. F. McFarland, R. Martin. 2001. CD4+CD28− costimulation-independent T cells in multiple sclerosis. J. Clin. Invest. 108:1185.
    OpenUrlCrossRefPubMed
  40. ↵
    Liuzzo, G., J. J. Goronzy, H. Yang, S. L. Kopecky, D. R. Holmes, R. L. Frye, C. M. Weyand. 2000. Monoclonal T-cell proliferation and plaque instability in acute coronary syndromes. Circulation 101:2883.
    OpenUrlAbstract/FREE Full Text
  41. ↵
    Koch, A. E., J. Friedman, J. C. Burrows, G. K. Haines, N. P. Bouck. 1993. Localization of the angiogenesis inhibitor thrombospondin in human synovial tissues. Pathobiology 61:1.
    OpenUrlCrossRefPubMed
  42. ↵
    Ticchioni, M., V. Raimondi, L. Lamy, J. Wijdenes, F. P. Lindberg, E. J. Brown, A. Bernard. 2001. Integrin-associated protein (CD47/IAP) contributes to T cell arrest on inflammatory vascular endothelium under flow. FASEB J. 15:341.
    OpenUrlAbstract/FREE Full Text
  43. ↵
    Burger, D., R. Rezzonico, J. M. Li, C. Modoux, R. A. Pierce, H. G. Welgus, J. M. Dayer. 1998. Imbalance between interstitial collagenase and tissue inhibitor of metalloproteinase 1 in synoviocytes and fibroblasts upon direct contact with stimulated T lymphocytes: involvement of membrane-associated cytokines. Arthritis Rheum. 41:1748.
    OpenUrlCrossRefPubMed
  44. ↵
    Reinhold, M. I., J. M. Greenberg, F. P. Lindberg, M. Ticchioni, E. J. Brown. 1999. Cell spreading distinguishes the mechanism of augmentation of T cell activation by integrin-associated protein/CD47 and CD28. Int. Immunol. 11:707.
    OpenUrlAbstract/FREE Full Text
  45. ↵
    Rebres, R. A., J. M. Green, M. I. Reinhold, M. Ticchioni, E. J. Brown. 2001. Membrane raft association of CD47 is necessary for actin polymerization andprotein kinase C θ translocation in its synergistic activation of T cells. J. Biol. Chem. 276:7672.
    OpenUrlAbstract/FREE Full Text
  46. ↵
    Kane, L. P., J. Lin, A. Weiss. 2002. It’s all Rel-ative: NFκB and CD28 costimulation of T cell activation. Trends Immunol. 23:413.
    OpenUrlCrossRefPubMed
  47. ↵
    Viola, A., S. Schroeder, Y. Sakakibara, A. Lanzavecchia. 1999. T lymphocyte costimulation mediated by reorganization of membrane microdomains. Science 283:680.
    OpenUrlAbstract/FREE Full Text
  48. ↵
    Brown, E. J., W. A. Frazier. 2001. Integrin-associated protein (CD47) and its ligands. Trends Cell Biol. 11:130.
    OpenUrlCrossRefPubMed
  49. ↵
    Fautrel, B., G. Le Moel, B. Saint-Marcoux, P. Taupin, S. Vignes, S. Rozenberg, A. C. Koeger, O. Meyer, L. Guillevin, J. C. Piette, P. Bourgeois. 2001. Diagnostic value of ferritin and glycosylated ferritin in adult onset Still’s disease. J. Rheumatol. 28:322.
    OpenUrlAbstract/FREE Full Text
  50. ↵
    Cazzola, M., G. Bergamaschi, L. Tonon, E. Arbustini, M. Grasso, E. Vercesi, G. Barosi, P. E. Bianchi, G. Cairo, P. Arosio. 1997. Hereditary hyperferritinemia-cataract syndrome: relationship between phenotypes and specific mutations in the iron-responsive element of ferritin light-chain mRNA. Blood 90:814.
    OpenUrlAbstract/FREE Full Text
  51. ↵
    Gray, C. P., P. Arosio, P. Hersey. 2002. Heavy chain ferritin activates regulatory T cells by induction of changes in dendritic cells. Blood 99:3326.
    OpenUrlAbstract/FREE Full Text
  52. ↵
    Gray, C. P., A. V. Franco, P. Arosio, P. Hersey. 2001. Immunosuppressive effects of melanoma-derived heavy-chain ferritin are dependent on stimulation of IL-10 production. Int. J. Cancer 92:843.
    OpenUrlCrossRefPubMed
  53. ↵
    Heller, R. A., M. Schena, A. Chai, D. Shalon, T. Bedilion, J. Gilmore, D. E. Woolley, R. W. Davis. 1997. Discovery and analysis of inflammatory disease-related genes using cDNA microarrays. Proc. Natl. Acad. Sci. USA 94:2150.
    OpenUrlAbstract/FREE Full Text
  54. ↵
    Rathanaswami, P., M. Hachicha, M. Sadick, T. J. Schall, S. R. McColl. 1993. Expression of the cytokine RANTES in human rheumatoid synovial fibroblasts: differential regulation of RANTES and interleukin-8 genes by inflammatory cytokines. J. Biol. Chem. 268:5834.
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools

Jump to section