Growth of and metalloproteinase production by fibroblast-like synoviocytes (FLSs) in patients with rheumatoid arthritis (RA) contribute to cartilage and bone destruction associated with development of the expanding inflammatory tissue referred to as pannus. Increased levels of extracellular matrix (ECM) proteins in the pannus suggest that intracellular signals generated through integrin receptors might control these processes. We developed a cell culture system permitting accurate assessment of the effect of cell adhesion to various ECM proteins on FLS phenotype. We show that FLS proliferation to platelet-derived growth factor requires a second signal provided by adhesion to an ECM protein. Fibronectin, vitronectin, collagen, or laminin could provide the second signal and was similarly required for the proliferation of FLSs from RA or osteoarthritis patients. Adhesion to fibronectin, collagen, or Arg-Gly-Asp peptide down-regulated collagenase expression. Primarily αv integrin receptors mediated this down-regulation upon adhesion to fibronectin. Loss of cell adhesion and TNF-α stimulation synergistically increased collagenase expression. Increased collagenase expression upon nonadherence was mimicked by treatment with cytochalasin B, suggesting that the loss of cytoskeletal structure associated with a change in cell shape mediates increased collagenase in nonadherent cells. Thus, although increased fibronectin in the lining layer in RA might be expected to inhibit collagenase expression, the change in cell shape associated with this multilayer structure might actually lead to increased collagenase expression.

In rheumatoid synovial tissues, fibroblast-like synoviocytes (FLS)3 probably mediate joint destruction as final effector cells in a cytokine/cell cascade (1, 2). Cells in the hyperplastic synovial lining layer, fibroblast-like and macrophage-like synoviocytes, express high levels of metalloproteinases, including collagenase (metalloproteinase-1) and stromelysin (metalloproteinase-3) (3, 4, 5). FLSs at the cartilage-pannus also express elevated collagenase levels (6). These and other metalloproteinases are probably primary mediators of bone and cartilage degradation in the rheumatoid joint.

Synovial tissue expansion creating the pannus in rheumatoid tissue results in part from the proliferation of FLSs (7). Transformed-like cellular growth has been proposed as a model for erosion of bone by the expanding pannus (2, 8, 9). In vitro studies of synovial fibroblast transformation have relied on assay of anchorage-independent cell growth. These studies showed that synovial fibroblasts require serum factors, notably platelet-derived growth factor (PDGF), to grow without anchorage. PDGF similarly stimulated cells derived from normal or arthritic joints to grow under anchorage-independent conditions (8). On the basis of these studies a concept of cytokine-dependent cellular transformation was proposed (8).

Cellular transformation as assayed by anchorage-independent cell growth measures cell proliferation not requiring adhesion to plastic. Cell adhesion to plastic relies on the binding of the serum proteins fibronectin and vitronectin to plastic, followed by the binding of cells to these proteins through surface integrin receptors (10). Signaling through integrin receptors is thus an important stimulus of cell proliferation in normal cells that is bypassed or constitutively activated in neoplastic cells (11, 12). Why FLSs can grow under anchorage-independent conditions in response to PDGF is not known, but might reflect local fibronectin matrix formation within the agarose gel (13, 14).

Integrin receptors also regulate metalloproteinase expression. In rabbit synovial fibroblasts engagement of the α5β1 integrin increases and that of the α4β1 integrin inhibits collagenase (15), and in dermal fibroblasts engagement of the α2β1 receptor increases collagenase (16). Thus, integrin interactions with ECM might be expected to regulate metalloproteinase of FLSs in RA synovial tissues.

Several histological studies have examined ECM and integrin expression in rheumatoid synovium. Several groups have observed increased levels of fibronectin in rheumatoid synovial tissues (17, 18, 19, 20, 21). This is particularly pronounced in the synovial lining layer and the luminal surface of endothelial cells, but also occurs in the sublining layer. Vitronectin is increased in the synovial lining layer in both OA and RA tissues compared with that in normal tissues (22). Expression of several integrin receptor subunits is also increased in situ; rheumatoid synovial tissues express high levels of the α3, α4, α5, α6, and β1 integrin subunits, but low levels of the αv integrin subunit (23, 24, 25).

We hypothesized that cell interactions with ECM control FLS phenotype in RA and developed a system to address the impact of cell adhesion on FLS growth and metalloproteinase expression. We found that FLS cell adhesion to any of several ECM components is essential for cell division. Fibronectin, vitronectin, collagen, and laminin were each capable of providing a signal complementary to a signal provided by PDGF to induce FLS proliferation. Adhesion to fibronectin, collagen, or an Arg-Gly-Asp (RGD) peptide down-regulated collagenase expression. Cell adhesion also affected the sensitivity of synovial fibroblasts to TNF-α and TGF-β. TNF-α and nonadherence or cytoskeletal disruption synergistically increased collagenase expression. Our data indicate that FLS/ECM interactions likely regulate FLS collagenase production and proliferation in RA synovial tissues.

Synovial fibroblasts were obtained from patients with rheumatoid arthritis (RA) at the time of elective surgery for knee replacement in accordance with an approved institutional review board protocol. Tissue (0.2–0.5 g) was minced finely, digested with collagenase (4 mg/ml; Worthington Biochemical, Freehold, NJ) for 4–6 h, centrifuged, and resuspended in DMEM supplemented with 10% FBS and penicillin (100 U/ml)/streptomycin (100 μg/ml). Cells were passaged upon confluence and were used between passages 4–8. Porcine TGF-β1 was provided by Dr. Michael Sporn, and human recombinant PDGF-BB was obtained from Life Technologies (Gaithersburg, MD). mAb to human integrin αv (Chemicon, Temecola, CA; MAB1980), human integrin α5 (Becton Dickinson, Mountain View, CA; mAb16), human integrin β1 (Coulter, Hialeah, FL; 4B4), and control Ab (Organon Teknika, Durham, NC; MOPC 141) were used at a concentration of 10 μg/ml for blocking experiments.

To eliminate attachment of cells to culture dishes, wells were coated with poly(2-hydroxyethyl methacrylate) (HEMA). One-half milliliter of a solution of 0.12 g/ml HEMA in 95% ethanol was added to each well of a six-well cluster dish, or 0.05 ml was added to each well of a 96-well dish, and the ethanol was allowed to dry over several days. ECM ligands and peptides were covalently linked to activated agarose beads (AminoLink Coupling Gel, Pierce, Rockland, IL) according to the manufacturer’s instructions. Briefly, an equal volume of washed beads was added to 1.0 mg/ml of ECM protein or peptide in 0.1 M NaPO4 (pH 7.0). A 0.02 vol of 1.0 M NaCNBH3 was added, and the suspension was incubated at room temperature on a rotator for 2 h. The resulting gel was washed, incubated with 1.0 M Tris, pH 7.4, to block unreacted sites, and washed again in complete medium three times before use. ECM proteins used as ligands were type I collagen (from calf skin, Sigma, St. Louis, MO), fibronectin (from human plasma, Life Technologies), or vitronectin (from human plasma, provided by C. Peterson). Peptide containing the RGD tripeptide (Gly-Arg-Gly-Asp-Ser-Pro-Lys) and control peptide (Gly-Arg-Ala-Asp-Ser-Pro-Lys) were purchased from Sigma.

Cells (∼2 × 104) were added to 96-well plates coated with HEMA or left uncoated in DMEM supplemented with 10% FBS. In other experiments about 106 cells were added to HEMA-coated wells of six-well plates, and ligand-coated beads (20 μl packed volume) were added to each well. The cells were allowed to adhere overnight to the ligand-coated bead or plastic surfaces. The next day the cells were washed three times with serum-free DMEM. For the cells adherent to plastic, medium was removed and replaced in the well (three times). For cells in HEMA-coated wells, the complex of beads and cells were transferred to a 15-ml conical tube and centrifuged at 1000 rpm for 5 min, the supernatant was aspirated, and serum-free DMEM was added (wash repeated three times). Cells were then left in serum-free medium for 3 days before adding PDGF or TGF-β, and [3H]thymidine. After the addition of cytokines, the cells were left for another 3 days, and then [3H]thymidine incorporation was assayed by collection and lysis of cells on a cell harvester, and counting on a scintillation counter.

To assay nonadherent cell viability, FLSs cultured on HEMA-coated plates were treated with trypsin to separate the cells, the trypsin was neutralized by addition of medium supplemented with serum, trypan blue was added to 0.02%, and 100 cells were counted for viability by dye exclusion.

Total RNAs were prepared from cells by direct lysis in 400 μl of RT-lysis buffer and purification according to the procedure described for isolation of total RNA (RNeasy Total RNA kit, Qiagen, Chatsworth, CA). RNA was eluted from the columns in a volume of 40 μl and was stored at −70°C. RNAs were analyzed on agarose/formaldehyde gels and blotted to nitrocellulose as previously described (26). Blotted RNAs were serially hybridized to cDNA probes of collagenase (metalloproteinase-1) and collagen and then to a probe specific for 18S ribosomal RNA (rRNA) using the method of Church et al. (27). An 18S human rRNA probe was generated by RT followed by PCR using the 5′ primer (5′-ACGTCTGCCCTATCAACTTTCGA-3′; bp 450–472) and the 3′ primer (5′-CCTCACTAAACCATCCAATCGG-3′; bp 1817–1838), yielding an amplified product of 1388 bp. For each hybridization signals were quantified by phosphorimaging and normalized to rRNA expression.

Data from proliferation studies were analyzed by one- or two-way analysis of variance and Tukey’s method for multiple comparison post-tests (using Prism software, GraphPad, San Diego, CA).

To understand more completely the potential contribution of cell adhesion to phenotypic abnormalities expressed by RA FLSs, proliferation and collagenase expression of FLSs were assayed under adherent and nonadherent conditions. Primary cell cultures were passaged to plastic wells or wells coated with HEMA, a polymer that inhibits cell attachment to plastic (28). Nonadherent cells formed cell clumps (Fig. 1, C and D). In the absence of adhesion, cell proliferation declined and became unresponsive to PDGF, a potent mitogen for FLSs (Fig. 1,A). As previously described, PDGF enhanced FLS proliferation of cells adherent to plastic (29). In contrast, FLSs cultured under nonadherent conditions highly increased collagenase mRNA expression (Fig. 1 B). We have seen similar increases in collagenase in FLSs derived from three other patients with RA and from five patients with OA and in dermal fibroblast cell lines. Although loss of adhesion induces cell death or ankoisis in some cell types, synovial fibroblasts remained fully viable during these experiments by trypan blue exclusion. This is consistent with the observation that nontransformed fibroblasts are resistant to apoptosis induced by nonadherence (30).

FIGURE 1.

Nonadherence inhibits proliferation and stimulates collagenase expression of RA FLSs. Primary synovial fibroblasts from a patient with RA were passaged to tissue culture plastic (PLASTIC) or HEMA-coated wells (HEMA) in DMEM supplemented with 10% FBS. A, About 2 × 104 cells were passaged to wells of a 96-well plate. After overnight incubation the cells were washed in serum-free DMEM, and culture was continued in serum-free DMEM supplemented with BSA (0.1 mg/ml). After an additional 24 h 1 μCi [3H]thymidine was added to each well, and the wells were either left untreated (control) or were treated with PDGF (20 ng/ml). Cells were harvested after 2 days, and [3H]thymidine incorporation was measured on a scintillation counter. Each bar represents the mean of quadruplicate wells. Error bars represent the SDs. B, About 2 × 106 cells were passaged to wells of a six-well plate. After overnight incubation the cells were washed in serum-free DMEM, and culture was continued in serum-free DMEM. Cells were harvested after 2 days and were analyzed for the expression of collagenase mRNA by Northern blot. Blotted mRNAs were hybridized sequentially to collagen, collagenase, and β-actin probes. C and D, Cell morphology of synovial fibroblasts grown on plastic (C) or HEMA-coated (D) plates (magnification, ×100).

FIGURE 1.

Nonadherence inhibits proliferation and stimulates collagenase expression of RA FLSs. Primary synovial fibroblasts from a patient with RA were passaged to tissue culture plastic (PLASTIC) or HEMA-coated wells (HEMA) in DMEM supplemented with 10% FBS. A, About 2 × 104 cells were passaged to wells of a 96-well plate. After overnight incubation the cells were washed in serum-free DMEM, and culture was continued in serum-free DMEM supplemented with BSA (0.1 mg/ml). After an additional 24 h 1 μCi [3H]thymidine was added to each well, and the wells were either left untreated (control) or were treated with PDGF (20 ng/ml). Cells were harvested after 2 days, and [3H]thymidine incorporation was measured on a scintillation counter. Each bar represents the mean of quadruplicate wells. Error bars represent the SDs. B, About 2 × 106 cells were passaged to wells of a six-well plate. After overnight incubation the cells were washed in serum-free DMEM, and culture was continued in serum-free DMEM. Cells were harvested after 2 days and were analyzed for the expression of collagenase mRNA by Northern blot. Blotted mRNAs were hybridized sequentially to collagen, collagenase, and β-actin probes. C and D, Cell morphology of synovial fibroblasts grown on plastic (C) or HEMA-coated (D) plates (magnification, ×100).

Close modal

We next investigated whether the ECM proteins shown to be responsible for cell binding to plastic, fibronectin and vitronectin, could reproduce proliferation dependent on adhesion to plastic. The addition of these proteins in a soluble form did not induce proliferation when added alone or together with PDGF (data not shown). Since proliferation might require cytoskeletal formation dependent on cell spreading, FLSs were then cultured in HEMA-coated wells with fibronectin- or vitronectin-coated beads. Cells attached to beads coated with either fibronectin (Fig. 2,B) or vitronectin (not shown), but not to control beads (Fig. 2,A). Cell proliferation was measured after the addition of PDGF, TGF-β, or both growth factors. Experiments were conducted using concentrations of PDGF and TGF-β previously found to maximally affect FLS growth and collagenase expression (8, 31). In early experiments the addition of ligand-coated beads without any exogenous growth factor(s) increased proliferation (for example, see Fig. 3,B). However, upon extended serum starvation the addition of either ligand-coated beads or PDGF alone did not stimulate cell proliferation (Fig. 3,A; p > 0.05 for the difference between untreated (without cytokine) cells with control beads vs untreated cells on each of the ECM-coated beads; also p > 0.05 for the difference between untreated vs PDGF-treated cells with control beads). Cell proliferation increased after addition of both ligand-coated beads and PDGF (Fig. 3,A; p < 0.05 for the difference between untreated vs PDGF-treated cells on fibronectin- and vitronectin-coated beads). TGF-β synergistically stimulated the proliferation of PDGF-stimulated cells (Fig. 3,A; p < 0.001 for the difference between PDGF vs PDGF + TGF-β for fibronectin- and vitronectin-coated beads; p > 0.05 for control beads). TGF-β had little effect alone (Fig. 3 A; p > 0.05 for the difference between TGF-β vs untreated cells on all ECM-coated and control beads).

FIGURE 2.

RA FLSs adhere to ECM proteins covalently bound to agarose beads. Fibronectin-coated (B), collagen-coated (C), or control (A) beads were added to early passage FLSs in HEMA-coated wells in DMEM supplemented with 10% FBS. The cells were allowed to adhere overnight (magnification, ×100).

FIGURE 2.

RA FLSs adhere to ECM proteins covalently bound to agarose beads. Fibronectin-coated (B), collagen-coated (C), or control (A) beads were added to early passage FLSs in HEMA-coated wells in DMEM supplemented with 10% FBS. The cells were allowed to adhere overnight (magnification, ×100).

Close modal
FIGURE 3.

Adhesion of RA FLSs to fibronectin, vitronectin, collagen, or laminin synergistically with PDGF activates synovial fibroblast proliferation. A, Early passage FLSs were passaged to HEMA-coated wells with fibronectin-, vitronectin-, collagen-, or laminin-coated or control beads in DMEM supplemented with 10% FBS. After overnight incubation the cells were washed in serum-free DMEM, and culture was continued in DMEM supplemented with BSA (0.1 mg/ml). After 3 more days 1 μCi of [3H]thymidine was added to each well, and the wells were either left untreated (control) or were treated with TGF-β (2.0 ng/ml), PDGF (20 ng/ml), or TGF-β (2.0 ng/ml) and PDGF (20 ng/ml). Cells were harvested after 3 days, and [3H]thymidine incorporation was measured on a scintillation counter. Each bar represents the mean of quadruplicate wells. The error bars represent the SDs. Statistically significant differences between cytokine and untreated wells are indicated (∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001). B, Early passage FLSs were cultured as described for A, except that the cells indicated (with serum) were not washed and cultured in serum-free medium. As well as addition of TGF-β and PDGF as described in A, the indicated cells received treatment with TNF-α (10 ng/ml) at the same time as the growth factor additions. C, Early passage FLSs were analyzed as described in A, except that cells were passed to fibronectin-coated wells.

FIGURE 3.

Adhesion of RA FLSs to fibronectin, vitronectin, collagen, or laminin synergistically with PDGF activates synovial fibroblast proliferation. A, Early passage FLSs were passaged to HEMA-coated wells with fibronectin-, vitronectin-, collagen-, or laminin-coated or control beads in DMEM supplemented with 10% FBS. After overnight incubation the cells were washed in serum-free DMEM, and culture was continued in DMEM supplemented with BSA (0.1 mg/ml). After 3 more days 1 μCi of [3H]thymidine was added to each well, and the wells were either left untreated (control) or were treated with TGF-β (2.0 ng/ml), PDGF (20 ng/ml), or TGF-β (2.0 ng/ml) and PDGF (20 ng/ml). Cells were harvested after 3 days, and [3H]thymidine incorporation was measured on a scintillation counter. Each bar represents the mean of quadruplicate wells. The error bars represent the SDs. Statistically significant differences between cytokine and untreated wells are indicated (∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001). B, Early passage FLSs were cultured as described for A, except that the cells indicated (with serum) were not washed and cultured in serum-free medium. As well as addition of TGF-β and PDGF as described in A, the indicated cells received treatment with TNF-α (10 ng/ml) at the same time as the growth factor additions. C, Early passage FLSs were analyzed as described in A, except that cells were passed to fibronectin-coated wells.

Close modal

Studies of integrin signaling have suggested that some, but not all, integrin receptors can stimulate cell proliferation. In these studies HUVECs up-regulated integrin association of Shc upon activation of α1, α5, and αvβ3 integrins, but not α2, α3, or α6 receptors (32). Related to this integrin-specific activation, proliferation increased upon cell binding to fibronectin or vitronectin, but not laminin. Therefore, to define the variety of ECM proteins able to synergistically activate FLS proliferation with PDGF, beads coated with collagen or laminin were also tested. Cells bound readily to beads coated with collagen (Fig. 2,C) or laminin (not shown) and responded to PDGF and TGF-β similarly to cells attached to fibronectin or vitronectin (Fig. 3 A; p < 0.01 and p < 0.05 for the difference between control vs PDGF-treated cells on, respectively, collagen- and laminin-coated beads; p < 0.001 for the difference between control vs PDGF- and TGF-β-treated cells on collagen- or laminin-coated beads).

To confirm that FLSs maintained viability under these conditions of prolonged serum starvation, untreated or TGF-β- or PDGF-treated FLSs cultured under nonadherent, serum-free conditions were analyzed after 1, 2, 3, 4, 5, 7, and 10 days for cell death by trypsin separation of the cell aggregates and trypan blue exclusion. No cell death was detected at any time point (<1%). After 10 days under these conditions the cells were placed in medium supplemented with serum on plastic tissue culture dishes. The cells attached to the surface overnight (at ∼10% confluence) and grew to confluence.

Results for cells growing on fibronectin-coated beads were similar to those for cells grown on fibronectin-coated wells (Fig. 3 C). However, the ECM ligand(s) providing the signal to cells grown on fibronectin-coated wells is not as well defined, since cells can bind to tissue culture plastic in the absence of the fibronectin coating.

Studies of FLSs from two other patients with RA and two patients with OA showed similar matrix and cytokine requirements. Thus, OA FLSs also required both adhesion to an ECM protein and PDGF to stimulate proliferation, indicating that OA and RA FLSs have the same requirements for proliferation (see Table I).

Table I.

Summary comparison of RA versus OA cell lines proliferation and collagenase expression in response to adhesion

Clinical DiagnosisFold Increase in [3H]Thymidine Incorporation Compared to Untreated Nonadherent CellsaFold Change in Collagenase Expression Compared to Nonadherent Cellsb
ControlTGF-βPDGFTGF-β + PDGF
Fibronectin beads      
RA 1.10 ± 0.05 1.58 ± 0.11 3.97 ± 1.74 11.2 ± 7.2 0.189 ± 0.129* 
OA 1.07 ± 0.12 1.03 ± 0.59 2.83 ± 0.018 4.94 ± 0.92** 0.203 ± 0.120* 
      
Collagen beads      
RA 1.31 ± 0.13 1.33 ± 0.04 4.6 ± 1.8 12.05 ± 0.78** 0.196 ± 0.076** 
OA 0.92 ± 0.027 1.37 ± 0.54 4.05 ± 0.80 8.13 ± 2.79* 0.180 ± 0.068** 
Clinical DiagnosisFold Increase in [3H]Thymidine Incorporation Compared to Untreated Nonadherent CellsaFold Change in Collagenase Expression Compared to Nonadherent Cellsb
ControlTGF-βPDGFTGF-β + PDGF
Fibronectin beads      
RA 1.10 ± 0.05 1.58 ± 0.11 3.97 ± 1.74 11.2 ± 7.2 0.189 ± 0.129* 
OA 1.07 ± 0.12 1.03 ± 0.59 2.83 ± 0.018 4.94 ± 0.92** 0.203 ± 0.120* 
      
Collagen beads      
RA 1.31 ± 0.13 1.33 ± 0.04 4.6 ± 1.8 12.05 ± 0.78** 0.196 ± 0.076** 
OA 0.92 ± 0.027 1.37 ± 0.54 4.05 ± 0.80 8.13 ± 2.79* 0.180 ± 0.068** 
a

Values are the mean ± SD of the fold increase in [3H]thymidine incorporation of two different cell lines for each diagnosis. Fold increases for each cell line were determined by averaging the CPM of triplicate wells treated as indicated (as shown in Fig. 3) and dividing by the average of untreated nonadherent cells from the same patient. Proliferation data were analyzed by two-way ANOVA. Values for RA vs OA are not statistically different (p > 0.05 for fibronectin and collagen beads), but values between cytokine treatments are statistically different (p = 0.0089 for fibronectin and p < 0.00001 for collagen beads). RA and OA values were further analyzed by one-way ANOVA and Tukey’s post test with significance indicated (∗, p < 0.5; ∗∗, p < 0.01).

b

Values are the mean ± SD of results from three different cell lines from each diagnosis. Each experimental value was determined by dividing the value obtained by phosphorimaging of the blot after hybridization to collagenase mRNA by the value obtained by phosphorimaging of the blot after hybridization to 18S rRNA. Each value was then normalized by dividing by the value of nonadherent cells. Collagenase expression was analyzed by Student’s t-test. No significant differences were found between RA and OA cell lines, but adherent cells were statistically different from nonadherent cells as indicated (∗, p < 0.5; ∗∗, p < 0.01).

Combinations of ligand-coated beads were also tested to determine whether stimulation of multiple integrin receptors could replace the requirement for PDGF and/or TGF-β. For these experiments, fibronectin-, vitronectin-, collagen-, and laminin-coated beads were added in pairwise combinations, or all four ECM-coated beads were added to cells. Cultures containing ECM-coated beads of more than one type did not proliferate more than control cultures containing ECM-coated beads of only one type (data not shown). Further, cultures containing two or more ECM-coated beads types proliferated in response to TGF-β and PDGF similarly to cultures containing ECM-coated beads of only one type.

To better understand the role of serum in this system, cell proliferation was studied under serum-free or serum containing conditions. In control nonadherent cells serum had little effect. The presence of serum resulted in increased proliferation upon the addition of collagen beads, but blunted the effect of TGF-β and PDGF (Fig. 3 B). This is probably due to the presence of TGF-β and PDGF in serum.

Although shown to markedly stimulate collagenase expression (see below), TNF-α had no significant effect on proliferation of FLS bound to ECM-coated beads under serum-free conditions (Fig. 3,B and data not shown). TNF-α slightly inhibited the proliferation of FLSs bound to ECM beads in the presence of serum; however, this did not reach statistical significance (Fig. 3 B and data not shown). This is similar to the effect of IL-1 in the presence of serum. In the case of IL-1, inhibition of FLS proliferation is mediated by PGE2, which is induced by both IL-1 and TNF-α and inhibits PDGF-stimulated proliferation of FLSs (33).

To further define the importance of integrin engagement, collagenase expression by FLSs was measured under nonadherent conditions or after adhesion to ligand-coated beads. Adhesion to plastic, fibronectin, or collagen decreased collagenase expression compared with that of nonadherent cells (Fig. 4). TGF-β inhibits basal and stimulated FLS collagenase expression (31). The addition of TGF-β decreased collagenase expression by both adherent and nonadherent cells, but did not completely inhibit the high levels of collagenase induced in nonadherent cells (Fig. 4). Cell integrin receptors bind to several different peptide motifs on ECM proteins, including the arginine-glycine-aspartatic acid (RGD) tripeptide found in fibronectin and vitronectin (34). Collagen molecules, although not containing this tripeptide, bind through a conformational analogue to this motif (35, 36). An RGD-containing peptide known to bind highly to fibronectin and vitronectin receptors was covalently bound to beads, and similar to the full-length proteins stimulated cell binding (not shown) and inhibited collagenase mRNA expression (Fig. 4). Control peptide beads slightly inhibited collagenase expression compared with control beads (Fig. 4). This correlated with slight binding of the cells to these beads compared with the complete absence of binding to control beads. Apparently integrin receptors are able to bind weakly to the mutated peptide. We have seen similar inhibition of collagenase expression by all cell lines tested from patients with RA and OA (Table I). Although there is some variability, collagen beads and fibronectin beads generally inhibited collagenase expression to a similar degree (Table I).

FIGURE 4.

Adhesion to collagen, fibronectin, or and RGD-containing peptide inhibits collagenase expression by RA FLSs. Early passage FLSs from a patient with RA were passaged to tissue culture plastic (PLASTIC) or HEMA-coated wells (HEMA, COL-BEADS, FN-BEADS, CON PEP, RGD PEP). Agarose beads with covalently linked collagen (COL-BEADS), fibronectin (FN-BEADS), control peptide (CON PEPT), or RGD-containing peptide (RGD PEPT) were added to wells, and the cells were allowed to adhere overnight in DMEM supplemented with 10% FBS. The following day the cells (and beads) were washed three times in DMEM, placed in serum-free DMEM supplemented with 0.1 mg/ml BSA, and left untreated (−) or were treated with TGF-β1 (+; 2.0 ng/ml). Cells were harvested after 2 additional days and were analyzed for expression of collagenase mRNA by Northern blot. Blotted mRNAs were hybridized to the collagenase probe and then to a probe specific for 18S ribosomal rRNA. Signals were quantified by phosphorimaging and were normalized to rRNA expression.

FIGURE 4.

Adhesion to collagen, fibronectin, or and RGD-containing peptide inhibits collagenase expression by RA FLSs. Early passage FLSs from a patient with RA were passaged to tissue culture plastic (PLASTIC) or HEMA-coated wells (HEMA, COL-BEADS, FN-BEADS, CON PEP, RGD PEP). Agarose beads with covalently linked collagen (COL-BEADS), fibronectin (FN-BEADS), control peptide (CON PEPT), or RGD-containing peptide (RGD PEPT) were added to wells, and the cells were allowed to adhere overnight in DMEM supplemented with 10% FBS. The following day the cells (and beads) were washed three times in DMEM, placed in serum-free DMEM supplemented with 0.1 mg/ml BSA, and left untreated (−) or were treated with TGF-β1 (+; 2.0 ng/ml). Cells were harvested after 2 additional days and were analyzed for expression of collagenase mRNA by Northern blot. Blotted mRNAs were hybridized to the collagenase probe and then to a probe specific for 18S ribosomal rRNA. Signals were quantified by phosphorimaging and were normalized to rRNA expression.

Close modal

The preincubation of Abs against various integrin receptors permitted assignment of the integrins responsible for the observed effects. FLSs preincubated with blocking Abs to the αv, and α5 or β1 integrin subunits no longer attached to fibronectin-coated beads, but preincubation with αv, α5, or β1 Abs alone had no observable effect on cell binding to the beads (Figs. 5 and 6A). These results indicate that FLSs bind to fibronectin by α5β1 and αvβ1, αvβ3, and/or αvβ5 integrin receptors. Other experiments showed that binding of FLSs to vitronectin was blocked by Ab to αv (but did not require Ab to the α5 subunit; data not shown), indicating that FLSs bind to vitronectin by αvβ1, αvβ3, and/or αvβ5 integrin receptors (Fig. 6,A). The receptors used by FLSs are thus similar to those reported for adhesion of dermal fibroblasts to fibronectin and vitronectin (37). These results were consistent in all OA (five of five) and most RA (four of five) FLS lines tested. Ab to only the αv integrin subunit was sufficient to block binding to fibronectin (and vitronectin) by one RA cell line, highlighting the importance of αv receptors in binding to both fibronectin and vitronectin. Antibody to the β1 subunit (but not by any combination of the α receptor blocking Abs) blocked binding to collagen (Fig. 6 A), indicating that FLSs bind to collagen through undefined β1 receptors (probably α1β1 and/or α2β1) (38).

FIGURE 5.

RA FLSs adhere to fibronectin by αv and α5 integrin receptors. Early passage FLSs from a patient with RA were passaged to HEMA-coated wells in DMEM supplemented with 10% FBS, and blocking Abs to human integrin α5 (B), human integrin αv (C), or human integrin αv and α5 (D) or no Ab (A) were added. After a 30-min incubation, agarose beads with covalently linked fibronectin were added to wells. The cells were cultured overnight in DMEM supplemented with 10% FBS (magnification, ×100).

FIGURE 5.

RA FLSs adhere to fibronectin by αv and α5 integrin receptors. Early passage FLSs from a patient with RA were passaged to HEMA-coated wells in DMEM supplemented with 10% FBS, and blocking Abs to human integrin α5 (B), human integrin αv (C), or human integrin αv and α5 (D) or no Ab (A) were added. After a 30-min incubation, agarose beads with covalently linked fibronectin were added to wells. The cells were cultured overnight in DMEM supplemented with 10% FBS (magnification, ×100).

Close modal
FIGURE 6.

RA collagenase regulation in FLSs by fibronectin depends on the αv integrin. Early passage FLSs from a patient with RA were passaged to HEMA-coated wells in DMEM supplemented with 10% FBS, and blocking Abs to human integrin α5 (anti-α5), human integrin αv (anti-αv), human integrin αv and α5 (anti-α5 + αv), or no Ab (CONTROL) were added. After a 30-min incubation, agarose beads with covalently linked fibronectin (B), fibronectin, vitronectin, or collagen (A) were added to wells as indicated. The cells were cultured overnight (A) or for 2 days (B) and then analyzed for cell binding (A) or expression of collagenase mRNA and 18S rRNA by Northern blot (B). Collagenase expression was quantified by phosphorimaging, normalized to rRNA expression, and then normalized to control expression. Results in B show the mean and SD of values obtained in two different experiments.

FIGURE 6.

RA collagenase regulation in FLSs by fibronectin depends on the αv integrin. Early passage FLSs from a patient with RA were passaged to HEMA-coated wells in DMEM supplemented with 10% FBS, and blocking Abs to human integrin α5 (anti-α5), human integrin αv (anti-αv), human integrin αv and α5 (anti-α5 + αv), or no Ab (CONTROL) were added. After a 30-min incubation, agarose beads with covalently linked fibronectin (B), fibronectin, vitronectin, or collagen (A) were added to wells as indicated. The cells were cultured overnight (A) or for 2 days (B) and then analyzed for cell binding (A) or expression of collagenase mRNA and 18S rRNA by Northern blot (B). Collagenase expression was quantified by phosphorimaging, normalized to rRNA expression, and then normalized to control expression. Results in B show the mean and SD of values obtained in two different experiments.

Close modal

Blocking Abs to αv or β1 integrin subunits increased collagenase expression by FLSs bound to fibronectin, but blocking Ab to α5 alone had no effect (Fig. 6,B). The combination of blocking Abs to both αv and α5 increased collagenase expression to a level slightly higher than that seen in fibronectin-adherent FLSs treated with Ab to αv alone (Fig. 6 B). Control mAb MOPC 141 had no effect on collagenase expression (data not shown).

Blocking of TNF-α leads to dramatic clinical improvement of patients with RA, implicating a central role for this cytokine in RA pathogenesis (39, 40). It has been long recognized that TNF-α potently induces collagenase expression by FLSs (41). To determine the potential contribution of TNF-α and ECM adhesion to collagenase expression in vivo, nonadherent FLSs or FLSs adherent to fibronectin or collagen beads were treated with TNF-α or TGF-β, and collagenase expression was analyzed. As described above, adhesion to collagen or fibronectin beads inhibited collagenase expression relative to nonadherent cells (Figs. 4 and 7). TGF-β further inhibited collagenase levels in both adherent and nonadherent cells. TNF-α stimulated increased collagenase expression by both adherent and nonadherent FLSs (Figs. 4 and 7). Similar effects were seen in two other cell lines. In other experiments FLSs were treated with PDGF or TGF-β plus PDGF under adherent and nonadherent conditions. PDGF had no significant effect on collagenase expression under these conditions (data not shown).

FIGURE 7.

Nonadherence and TNF-α synergistically stimulate collagenase expression by RA FLSs. Early passage FLSs from a patient with RA were passaged to HEMA-coated wells, and agarose beads with covalently linked collagen (collagen beads), fibronectin (fibronectin beads), or no protein (control beads) were added to wells (A and B) or cells were passaged to plastic (B). The cells were allowed to adhere overnight in DMEM supplemented with 10% FBS, and the following day the cells (and beads) were washed three times in DMEM, placed in serum-free DMEM supplemented with 0.1 mg/ml BSA, and left untreated or treated with TGF-β1 or TNF-α as indicated. Cells were harvested after 2 additional days and analyzed for expression of collagenase mRNA by Northern blot. Blotted mRNAs were hybridized to the collagenase probe and then to a probe specific for 18S ribosomal rRNA. Signals were quantified by phosphorimaging, normalized to rRNA expression, and then normalized to control expression.

FIGURE 7.

Nonadherence and TNF-α synergistically stimulate collagenase expression by RA FLSs. Early passage FLSs from a patient with RA were passaged to HEMA-coated wells, and agarose beads with covalently linked collagen (collagen beads), fibronectin (fibronectin beads), or no protein (control beads) were added to wells (A and B) or cells were passaged to plastic (B). The cells were allowed to adhere overnight in DMEM supplemented with 10% FBS, and the following day the cells (and beads) were washed three times in DMEM, placed in serum-free DMEM supplemented with 0.1 mg/ml BSA, and left untreated or treated with TGF-β1 or TNF-α as indicated. Cells were harvested after 2 additional days and analyzed for expression of collagenase mRNA by Northern blot. Blotted mRNAs were hybridized to the collagenase probe and then to a probe specific for 18S ribosomal rRNA. Signals were quantified by phosphorimaging, normalized to rRNA expression, and then normalized to control expression.

Close modal

Integrin engagement stimulates the formation of the actin-based cytoskeleton that terminates in focal adhesions composed of several intracellular cytoskeletal proteins, including α-actinin, talin, vinculin, paxillin, and tensin. These proteins bind to the β1 integrin tail at the interior of the cell membrane. The observation that adhesion to any of several different integrins leads to decreased collagenase expression suggested that formation of the cytoskeleton might mediate this signal. The addition of cytochalasin B, an inhibitor of cytoskeletal formation, up-regulated collagenase expression by FLSs. This up-regulation was similar in magnitude to that seen upon culture of the cells under nonadherent conditions (Fig. 8). The addition of TNF-α to cytochalasin B-treated FLSs resulted in a synergistic increase in collagenase expression similar to that seen upon treatment of nonadherent cells with TNF-α (Fig. 8). Similar effects were seen on two other cell lines.

FIGURE 8.

Disruption of the cytoskeleton and TNF-α synergistically stimulate collagenase expression by RA FLSs. Early passage FLSs from a patient with RA were passaged to HEMA-coated wells or tissue culture plastic wells, then treated as indicated with TNF-α (10 ng/ml) and/or cytochalasin B (4.0 μg/ml). Cells were harvested after 3 additional days and were analyzed for expression of collagenase mRNA by Northern blot. Blotted mRNAs were hybridized to the collagenase probe and then to a probe specific for 18S ribosomal rRNA. Signals were quantified by phosphorimaging normalized to rRNA expression and then normalized to control (HEMA) expression.

FIGURE 8.

Disruption of the cytoskeleton and TNF-α synergistically stimulate collagenase expression by RA FLSs. Early passage FLSs from a patient with RA were passaged to HEMA-coated wells or tissue culture plastic wells, then treated as indicated with TNF-α (10 ng/ml) and/or cytochalasin B (4.0 μg/ml). Cells were harvested after 3 additional days and were analyzed for expression of collagenase mRNA by Northern blot. Blotted mRNAs were hybridized to the collagenase probe and then to a probe specific for 18S ribosomal rRNA. Signals were quantified by phosphorimaging normalized to rRNA expression and then normalized to control (HEMA) expression.

Close modal

We show that adhesion to ECM regulates both FLS proliferation and metalloproteinase expression. Changes in FLS cell adhesion and shape probably contribute to the formation of the destructive pannus through the generation of intracellular signals mediated by integrin engagement (or disengagement). Further, cytokine signals generated by infiltrating inflammatory cells are dependent on the state of cell adhesion. ECM and cytokine signals thus coordinately regulate FLS phenotype and destructive potential.

We show that FLS proliferation requires a signal provided by ECM in addition to a signal provided by a mitogenic cytokine, the most potent being PDGF. These data reinforce earlier observations that FLSs from RA and OA patients have the same requirements for anchorage-dependent and anchorage-independent growth (8, 29). Mutations in genes regulating cellular growth and transformation should change the dependence of FLS proliferation on adhesion and/or cytokines. By these measures RA FLSs need the same stimuli as OA FLSs for proliferation. Our data do not support the idea of a mutational event or a transforming virus contributing to the pathogenesis of synovial hyperplasia in RA (2).

Any of several different ECM proteins, including fibronectin, collagen, vitronectin, or laminin can provide the ECM signal required for FLS proliferation. The integrin receptors used by different ECM proteins are different, but overlapping, and depend on the cell type analyzed (38). Through blocking experiments we found that synovial fibroblasts bind to fibronectin through α5β1 and αv integrins, to vitronectin through αv integrins, and to collagen though β1 integrins. Our results do not support an integrin receptor specificity of a signal for FLS cell cycle entry as has been described for HUVEC (32). These cells do not proliferate or activate a Fos promoter-luciferase construct upon cultivation on laminin 1, in contrast to cultivation on fibronectin or vitronectin. The difference between our results and these findings was not due to differences in the type or the preparation of laminin because we used the same source of laminin 1 as these investigators (32). Possibly FLSs bind to laminin through receptors different from those found on HUVEC. Alternatively, a broader array of integrin receptors on FLSs are able to activate the required costimulus for proliferation.

Previous analyses have shown that fibroblast adhesion permits progression through the G1/S phase of the cell cycle (11). In this regard FLS cell growth is similar to NIH-3T3 or BALB/c 3T3 fibroblasts, which require an ECM stimulus for proliferation. In contrast, treatment with TGF-β can bypass this need for an ECM stimulus in NRK cells, permitting the cells to grow without anchorage (42). TGF-β stimulation of anchorage-independent growth by NRK cells has been attributed to its stimulation of ECM formation (13). In contrast, TGF-β inhibits anchorage-independent growth of FLSs (8), even though it stimulates FLS collagen formation and fibrillar fibronectin formation (14). PDGF can also promote fibrillar fibronectin matrix formation by FLSs, and thus local formation of ECM may contribute to PDGF-stimulated anchorage-independent growth of FLSs (14). Apparently other intracellular signals elicited by TGF-β selectively inhibit FLS growth under anchorage-independent conditions.

The ECM highly regulates collagenase expression. In rheumatoid synovial tissues collagenase expression is particularly pronounced in the synovial lining layer. In RA, the cells in this region form a multilayer structure of macrophages and FLSs with little intervening extracellular space. Despite this structure, this region stains highly with Abs to fibronectin. We suggest that the high levels of collagenase in this region may be related to this structure. We show that ECM engagement without attendant cytoskeletal formation leads to high level collagenase expression associated with a change in cell shape. Probably cytokines, including TNF-α secreted by macrophages in this region, may act synergistically with cytoskeletal changes to increase FLS metalloproteinase expression (43, 44).

Werb et al. have shown that collagenase expression by rabbit synovial fibroblasts is cooperatively regulated by binding of the α4 and α5 integrins to fibronectin (15). Our results show human RA FLSs do not adhere to fibronectin by α4β1, but use αv and α5β1 integrin receptors. In contrast to rabbit synovial fibroblasts in which α5 engagement increases collagenase expression, α5 engagement modestly inhibits RA FLS collagenase expression. Engagement of αv integrin receptors is more important in the down-regulation of collagenase expression upon FLS binding to fibronectin. This result is consistent with similar collagenase down-regulation upon FLS binding to vitronectin; binding to vitronectin only required αv receptors.

Recently, Wang et al. have reported data suggesting that αv, α5, and α4 receptors are important in IL-1β-stimulated invasion of cartilage (45). However, α4 blocking Abs had no effect on unstimulated FLSs. Possibly, IL-1 stimulated FLS to express increased α4 integrin receptors, and thus this receptor may contribute to regulation of proliferation and/or collagenase expression in IL-1-stimulated FLSs. Further studies will be needed to clarify this possibility.

Integrins use at least two intracellular signaling pathways and a plethora of signaling molecules to deliver signals intracellularly. Early studies identified FAK as a potential mediator of integrin signaling. Inhibition of FAK is associated with disruption or reorganization of the actin cytoskeleton (46, 47), and a constitutively activated form of FAK confers resistance to anoikis on epithelial cell lines (48). Intracellular pathways not involving FAK can be activated upon integrin stimulation. Integrin engagement activates mitogen-activated protein (MAP) kinases, Erk1 and Erk2, through a Ras-dependent pathway (49). Although previously thought to be downstream from FAK activation, Ras-dependent integrin signals represent a separate signaling pathway (32, 49). The adaptor protein, Shc, may mediate this pathway by recruiting Grb2 to β1 integrins (32). Surprisingly, triggering this pathway does not require the β1 integrin cytoplasmic tail that binds to cytoskeletal proteins to form the focal adhesion complex. However, formation of the cytoskeleton plays a key role in signal transduction, since phosphorylation of both MAP kinases (50) and FAK (51) upon integrin engagement depends on cytoskeletal formation. Clustering of Src and MAP kinases and Ras upon integrin aggregation also depends on the formation of a cytoskeleton (52, 53). Apparently cytoskeletal formation is requisite for propagation of Ras-dependent signals, although it is not required for early aspects of signal transmission. Thus, our data do not distinguish between Ras- and FAK-dependent pathways in mediating integrin signaling intracellularly on FLS collagenase expression. Our data do suggest that one of these intracellular signaling pathways interacts with TNF-α signals to synergistically regulate collagenase gene expression.

We thank Drs. Richard Scott and Richard Paul for providing synovial tissue samples.

1

This work was supported by grants from the Arthritis Foundation and the National Institutes of Health (Shannon Award; to R.L.).

3

Abbreviations used in this paper: FLS, fibroblast-like synoviocytes; PDGF, platelet-derived growth factor; ECM, extracellular matrix; RA, rheumatoid arthritis; OA, osteoarthritis; RGD, Arg-Gly-Asp; HEMA, poly(2-hydroxyethyl methacrylate); rRNA, ribosomal ribonucleic acid; MAP, mitogen-activated protein.

1
Burmester, G. R., B. Stuhlmuller, G. Keyszer, R. W. Kinne.
1997
. Mononuclear phagocytes and rheumatoid synovitis: mastermind or workhorse in arthritis?.
Arthritis Rheum.
40
:
5
2
Firestein, G. S..
1996
. Invasive fibroblast-like synoviocytes in rheumatoid arthritis. Passive responders or transformed aggressors?.
Arthritis Rheum.
39
:
1781
3
Gravallese, E. M., J. M. Darling, A. L. Ladd, J. N. Katz, L. H. Glimcher.
1991
. In situ hybridization studies of stromelysin and collagenase messenger RNA expression in rheumatoid synovium.
Arthritis Rheum.
34
:
1076
4
Firestein, G. S., M. M. Paine, B. H. Littman.
1991
. Gene expression (collagenase, tissue inhibitor of metalloproteinases, complement, and HLA-DR) in rheumatoid arthritis and osteoarthritis synovium.
Arthritis Rheum.
34
:
1094
5
McCachren, S. S..
1991
. Expression of metalloproteinases and metalloproteinase inhibitor in human arthritic synovium.
Arthritis Rheum.
34
:
1085
6
Wooley, D. E., M. J. Crossley, J. M. Evanson.
1977
. Collagenase at sites of cartilage erosion in the rheumatoid joint.
Arthritis Rheum.
20
:
1231
7
Qu, Z., C. H. Garcia, L. M. O’Rourke, S. R. Planck, M. Kohli, J. T. Rosenbaum.
1994
. Local proliferation of fibroblast-like synoviocytes contributes to synovial hyperplasia.
Arthritis Rheum.
37
:
212
8
Lafyatis, R., E. F. Remmers, A. B. Roberts, D. E. Yocum, M. B. Sporn, R. L. Wilder.
1989
. Anchorage-independent growth of synoviocytes from arthritic and normal joints: stimulation by exogenous platelet-derived growth factor and inhibition by transforming growth factor-β and retinoids.
J. Clin. Invest.
83
:
1267
9
Fassbender, H. G., M. Simmling-Annefeld, E. Stofft.
1980
. Transformation der Synovialzellen bei rheumatoider Arthritis.
Verh. Dtsch. Ges. Pathol.
64
:
193
10
Steele, J. G., B. A. Dalton, G. Johnson, P. A. Underwood.
1993
. Polysterene chemistry affects vitronectin activity: an explanation for cell attachment to tissue culture polystyrene but not to unmodified polystyrene.
J. Biomed. Mater. Res.
27
:
927
11
Guadagno, T. M., R. K. Assoian.
1991
. G1/S control of anchorage-independent growth in the fibroblast cell cycle.
J. Cell Biol.
115
:
1419
12
Benecke, B.-J., A. Ben-Ze’ev, S. Penman.
1978
. The control of mRNA production, translation and turnover in suspended and reattached anchorage-dependent fibroblasts.
Cell
14
:
931
13
Ignotz, R. A., J. Massague.
1986
. Transforming growth factor-β stimulates expression of fibronectin and collagen and their incorporation into extracellular matrix.
J. Biol. Chem.
261
:
4337
14
Sarkissian, M., R. Lafyatis.
1998
. Transforming growth factor-β and platelet-derived growth factor regulation of fibrillar fibronectin matrix formation by synovial fibroblasts.
J. Rheumatol.
25
:
613
15
Huhtala, P., M. J. Humphries, J. B. McCarthy, P. M. Tremble, Z. Werb, C. H. Damsky.
1995
. Cooperative signaling by α5β1 and α4β1 integrins regulates metalloproteinase gene expression in fibroblasts adhering to fibronectin.
J. Cell Biol.
129
:
867
16
Riikonen, T., J. Westermarck, L. Koivisto, A. Broberg, V.-M. Kahari, J. Heino.
1995
. Integrin α2β1 is a positive regulator of collagenase (MMP-1) and collagen α1(I) gene expression.
J. Biol. Chem.
22
:
13548
17
Clemmensen, I., B. Holund, R. B. Andersen.
1983
. Fibrin and fibronectin in rheumatoid synovial membrane and rheumatoid synovial fluid.
Arthritis Rheum.
26
:
479
18
Haynes, B. F., L. P. Hale, K. L. Patton, M. E. Martin, R. M. McCallum.
1991
. Measurement of an adhesion molecule as an indicator of inflammatory disease activity.
Arthritis Rheum.
34
:
1434
19
Scott, D. L., A. C. Wainwright, K. W. Walton, N. Williamson.
1981
. Significance of fibronectin in rheumatoid arthritis and osteoarthrosis.
Ann. Rheum. Dis.
40
:
142
20
Vartio, T., A. Vaheri, R. V. Essen, H. Isomaki, S. Stenman.
1981
. Fibronectin in synovial fluid and tissue in rheumatoid arthritis.
Eur. J. Clin. Invest.
11
:
207
21
Elices, M. J., V. Tsai, D. Strahl, A. S. Goel, V. Tollefson, T. Arrhenius, E. A. Wayner, F. C. A. Gaeta, J. D. Fikes, G. S. Firestein.
1994
. Expression and functional significance of alternatively spliced CS1 fibronectin in rheumatoid arthritis microvasculature.
J. Clin. Invest.
93
:
405
22
Nikkari, L., K. Haapasalmi, H. Aho, A. Torvinen, D. Sheppard, H. Larjava, J. Heino.
1995
. Localization of the αv subfamily of integrins and their putative ligands in the synovial cell layer.
J. Rheum.
22
:
16
23
Nikkari, L., H. Aho, T. Yli-Jama, H. Larjava, M. Jalkanen, J. Heino.
1993
. Expression of integrin family of cell adhesion receptors in rheumatoid synovium.
Am. J. Pathol.
142
:
1019
24
Ishikawa, H., S. Hirata, Y. Nishibayashi, S. Imura, H. Kubo, O. Ohno.
1994
. The role of adhesion molecules in synovial pannus formation in rheumatoid arthritis.
Clin. Orthop.
300
:
297
25
Rinaldi, N., T. Barth, C. Henne, G. Mechterscheimer, P. Moller.
1994
. Synoviocytes in chronic synovitis in situ and cytokine stimulated synovial cells in vitro neo-express α1, α3 and α5 chains of β1 integrins.
Virchows Arch.
425
:
171
26
Davis, L. G., M. D. Dibner, J. E. Battery.
1986
.
Basic Methods in Molecular Biology
130
Elsevier, New York.
27
Church, G. M., W. Gilbert.
1984
. Genomic sequencing.
Proc. Natl. Acad. Sci. USA
81
:
1991
28
Folkman, J., A. Moscona.
1978
. Role of cell shape in growth control.
Nature
273
:
345
29
Remmers, E. F., R. Lafyatis, J. P. Case, A. B. Roberts, M. B. Sporn, R. L. Wilder.
1990
. Cytokines and growth regulation of synoviocytes from patients with rheumatoid arthritis and rats with streptococcal cell wall arthritis.
Growth Factors
2
:
179
30
McGill, G., A. Shimamura, R. C. Bates, R. E. Savage, D. E. Fisher.
1997
. Loss of matrix adhesion triggers rapid transformation-selective apoptosis in fibroblasts.
J. Cell Biol.
138
:
901
31
Lafyatis, R., N. L. Thompson, E. F. Remmers, K. C. Flanders, N. S. Roche, S. J. Kim, J. P. Case, M. B. Sporn, A. B. Roberts, R. L. Wilder.
1989
. Transforming growth factor-β production by synovial tissues from rheumatoid patients and streptococcal cell wall arthritic rats: studies on secretion by synovial fibroblast-like cells and immunohistologic localization.
J. Immunol.
143
:
1142
32
Wary, K. K., F. Mainiero, S. J. Isakoff, E. E. Marcantonio, F. G. Giancotti.
1996
. The adaptor protein Shc couples a class of integrins to the control of cell cycle progression.
Cell
87
:
733
33
Kumkumian, G. K., R. Lafyatis, E. F. Remmers, J. P. Case, S. J. Kim, R. L. Wilder.
1989
. Platelet-derived growth factor and IL-1 interactions in rheumatoid arthritis: regulation of synoviocyte proliferation, prostaglandin production, and collagenase transcription.
J. Immunol.
143
:
833
34
Yamada, K. M..
1991
. Adhesive recognition sequences.
J. Biol. Chem.
266
:
12809
35
Pignatelli, M., W. F. Bodmer.
1988
. Genetics and biochemistry of collagen binding-triggered glandular differentiation in a human carcinoma cell line.
Proc. Natl. Acad. Sci. USA
85
:
5561
36
Staatz, W. D., K. F. Fok, M. M. Zutter, S. P. Adams, B. A. Rodriguez, S. A. Santoro.
1991
. Identification of a tetrapeptide recognition sequence for the α2β1 integrin in collagen.
J. Biol. Chem.
266
:
7363
37
Gailit, J., R. A. F. Clark.
1996
. Studies in vitro on the role of αv and β1 integrins in the adhesion of human dermal fibroblasts to provisional matrix proteins fibronectin, vitronectin, and fibrinogen.
J. Invest. Dermatol.
106
:
102
38
Hynes, R. O..
1992
. Integrins: versatility, modulation, and signaling in cell adhesion.
Cell
69
:
11
39
Elliott, M. J., R. N. Maini, M. Feldman, A. Long-Fox, P. Charles, P. Katsikis, F. M. Brennan, J. Walker, H. Bijl, J. Grayheb, et al
1993
. Treatment of rheumatoid arthritis with chimeric antibodies to tumor necrosis factor α.
Arthritis Rheum.
36
:
1681
40
Elliott, M. J., R. N. Maini, M. Feldman, J. R. Kalden, C. Antoni, J. S. Smolen, B. Leeb, F. C. Breedveld, J. D. Macfarlane, H. Bijl, et al
1994
. Randomized double-blind comparison of chimeric monoclonal antibody to tumor necrosis factor α (cA2) versus placebo in rheumatoid arthritis.
Lancet
344
:
1105
41
Dayer, J. M., B. Beutler, A. Cerami.
1985
. Cachectin/tumor necrosis factor stimulates collagenase and prostaglandin E2 production by human synovial cells and dermal fibroblasts.
J. Exp. Med.
162
:
2163
42
Han, E. K.-H., T. M. Guadagno, S. L. Dalton, R. K. Assoin.
1993
. A cell cycle and mutational analysis of anchorage-independent growth: cell adhesion and TGF-β1 control G1/S transit specifically.
J. Cell Biol.
122
:
461
43
MacNaul, K. L., N. I. Hutchinson, J. N. Parsons, E. K. Bayne, M. J. Tocci.
1990
. Analysis of IL-1 and TNF-α gene expression in human rheumatoid synoviocytes and normal monocytes by in situ hybridization.
J. Immunol.
1454
:
154
44
Chu, C. Q., M. Field, M. Feldmann.
1991
. Localization of tumor necrosis factor-α in synovial tissues and at the cartilage-pannus junction in patients with rheumatoid arthritis.
Arthritis Rheum.
34
:
1125
45
Wang, A. Z., J. C. Wang, G. W. Fisher, H. S. Diamond.
1997
. Interleukin-1β-stimulated invasion of articular cartilage by rheumatoid synovial fibroblasts is inhibited by antibodies to specific integrin receptors and by collagenase inhibitors.
Arthritis Rheum.
40
:
1298
46
Ilic, D., Y. Furuta, S. Kanazawa, N. Takeda, K. Sobue, N. Nakatsuji, S. Nomura, J. Fujimoto, M. Okado, T. Yamamoto, S. Alzawa.
1995
. Reduced cell motility and enhanced focal adhesion formation in cells from FAK-deficient mice.
Nature
377
:
539
47
Richardson, A., J. T. Parsons.
1996
. A mechanism for the regulation of the adhesion-associated protein kinase pp125FAK.
Nature
380
:
538
48
Frisch, S. M., K. Vuori, E. Ruoslahti, P.-Y. Chan-Hui.
1996
. Control of adhesion-dependent cell survival by focal adhesion kinase.
J. Cell Biol.
134
:
793
49
Clark, E. A., R. O. Hynes.
1996
. Ras activation is necessary for integrin-mediated activation of extracellular signal-regulated kinase 2 and cytosolic phospholipase A2 but not cytoskeletal organization.
J. Biol. Chem.
271
:
14814
50
Chen, Q., M. S. Kinch, T. H. Lin, K. Burridge, R. L. Juliano.
1994
. Integrin-mediated cell adhesion activates mitogen-activated protein kinases.
J. Biol. Chem.
269
:
26602
51
Lipfert, L., B. Haimovich, M. D. Schaller, B. S. Cobb, J. T. Parsons, J. S. Brugge.
1992
. Integrin-dependent phosphorylation and activation of the protein tyrosine kinase pp125FAK in platelets.
J. Cell Biol.
122
:
473
52
Miyamoto, S., H. Teramoto, O. A. Coso, J. S. Gutkind, P. D. Burbelo, S. K. Akiyama, K. M. Yamada.
1995
. Integin function: molecular hierarchies of cytoskeletal and signaling molecules.
J. Cell Biol.
131
:
791
53
Miyamoto, S., S. K. Akiyama, K. M. Yamada.
1995
. Synergistic roles for receptor occupancy and aggregation in integrin transmembrane function.
Science
267
:
883