- Research article
- Open Access
Electrophoretic characterization of species of fibronectin bearing sequences from the N-terminal heparin-binding domain in synovial fluid samples from patients with osteoarthritis and rheumatoid arthritis
Arthritis Res Thervolume 5, Article number: R329 (2003)
Fragments of fibronectin (FN) corresponding to the N-terminal heparin-binding domain have been observed to promote catabolic chondrocytic gene expression and chondrolysis. We therefore characterized FN species that include sequences from this domain in samples of arthritic synovial fluid using one-and two-dimensional (1D and 2D) Western blot analysis. We detected similar assortments of species, ranging from ~47 to greater than 200 kDa, in samples obtained from patients with osteoarthritis (n = 9) versus rheumatoid arthritis (n = 10). One of the predominant forms, with an apparent molecular weight of ~170 kDa, typically resolved in 2D electrophoresis into a cluster of subspecies. These exhibited reduced binding to gelatin in comparison with a more prevalent species of ~200+ kDa and were also recognized by a monoclonal antibody to the central cell-binding domain (CBD). When considered together with our previous analyses of synovial fluid FN species containing the alternatively spliced EIIIA segment, these observations indicate that the ~170-kDa species includes sequences from four FN domains that have previously, in isolation, been observed to promote catabolic responses by chondrocytes in vitro: the N-terminal heparin-binding domain, the gelatin-binding domain, the central CBD, and the EIIIA segment. The ~170-kDa N-terminal species of FN may therefore be both a participant in joint destructive processes and a biomarker with which to gauge activity of the arthritic process.
Fibronectins (FNs), a family of multifunctional adhesion proteins that differ from one another through alternative splicing of a pre-mRNA derived from a single gene, are found as soluble dimeric molecules in the blood and as insoluble multimers within the extracellular matrix of tissues, where they are concentrated in basement membranes and blood vessel walls [1–3]. They bind to cell-surface integrin receptors and participate in a variety of cellular processes, including adhesion, migration, transformation, and apoptosis, as well as wound healing, fibrosis, and hemostasis [1–5]. FN is deposited in cartilage from osteoarthritis (OA) [3, 6–9], and fragmented forms of FN have been detected in synovial fluid (SF) and articular cartilage from patients with OA and patients with rheumatoid arthritis (RA) [10–17]. On the basis of such findings, plasma-derived FN (pFN) and specific purified pFN fragments have been tested for their capacity to regulate the function of chondrocytes in vitro. Whereas intact, soluble pFN has been observed to exert little or no effect, several purified, proteolytically derived pFN fragments have proved to be active [18–26]. Additionally, mixtures of fragments derived from OA cartilage have been observed to promote chondrolysis in vitro .
Although fragments corresponding to the 29-kDa (also referred to as 30-kDa) amino-terminal (N-terminal) heparin-binding domain (HBD) have been studied most extensively, species derived from sites spanning most of the FN molecule have been observed to trigger catabolic gene expression in chondrocytes [18–26]. For example, purified fragments of pFN corresponding to the 120- to 140-kDa central cell-binding domain (CBD), the 50-kDa gelatin-binding domain (GBD), and the 40-kDa C-terminal HBD have each been observed to trigger release of proteoglycans from cartilage slices in vitro, as has a recombinant version of the alternatively spliced EIIIA segment (Fig. 1) [18, 22, 25–27]. In addition, the 29-kDa N-terminal HBD has been observed to trigger gene expression for stromelysin, inducible nitric oxide synthetase, hyaluronan receptor proteins, and other biologically active molecules in cultured chondrocytes [20, 21, 23–26]. Chondrolysis triggered by FN fragments occurs in association with local release of catabolic cytokines, including tumor necrosis factor α, interleukin-1β, and interleukin-1α . Furthermore, intra-articular injection of either N-terminal or central CBD fragments into rabbit joints triggers loss of cartilage proteoglycan, whereas injection of intact, dimeric pFN does not [28, 29].
Our goal in this study was to characterize and compare the assortments of N-terminal SF FN species in samples from OA versus RA patients with respect to their domain structures and ligand-binding properties. We have found that, among the two predominant species of SF FN that bear sequences from the N-terminal HBD in patients with OA or RA, the smaller, ~170-kDa species binds less readily to gelatin and to a monoclonal antibody (mAb) specific for the GBD than does the larger, ~200+-kDa species. Furthermore, 2D electrophoretic analysis reveals the ~170-kDa species to be comprised of distinct subspecies, most of which extend sufficiently toward the carboxy terminus (C terminus) to include the 10th type III repeat within the central CBD. In addition to prominent ~200+- and ~170-kDa species, several additional forms of FN that bear sequences from the N-terminal HBD were detected in OA and RA samples. Each of the soluble species identified in this study, in addition to its possible roles in the promotion of arthritic joint injury, is a candidate as a biomarker for the arthritic disease process.
Materials and methods
Synovial fluid samples
This research was conducted according to the principles of the Declaration of Helsinki and was approved by committees overseeing human experimentation at the relevant institutions. After informed consent had been obtained, SF was taken from patients with OA or active RA who were undergoing diagnostic and/or therapeutic arthrocentesis at Long Island Jewish Medical Center, New Hyde Park, NY, or at Winthrop-University Hospital, Minneola, NY, USA. Fluid was drawn into plastic syringes and placed directly into tubes containing EDTA, phenylmethylsulfonyl fluoride (PMSF), and aprotinin at final concentrations 5.7 mM, 1 mM, and 500 U/ml, respectively (Sigma Chemical Co, St Louis, MO, USA). The fluids were centrifuged at 500 g, and supernatants were frozen at -80°C with the exception of short periods at -20°C. The 9 OA samples (numbered 1–9) and 10 RA samples (numbered 10–19) that were analyzed in this study were previously examined for their content of species of FN bearing the alternatively spliced EIIIA segment .
Antibodies and purified FN fragments
Purified anti-FN mAbs specific for the N-terminal HBD (mAb 1936) (hereafter referred to as anti-N-terminal mAb) [1–3, 30] and the GBD (mAb 1892) were from Chemicon (Temecula, CA, USA) (Fig. 1). mAb A2C2, which recognizes the 10th type III repeat of FN , was a gift (as ascites) from Dr Richard Hynes, Massachusetts Institute of Technology, Cambridge, MA, USA. Purified proteolytic fragments from human pFN corresponding to the 30-kDa (equivalent to 29-kDa) N-terminal HBD, and a 45-kDa stretch from the GBD (Fig. 1), were from Sigma.
Affinity isolation of synovial fluid FNs using immobilized gelatin
To block nonspecific binding sites, 25 μl of gelatin–Sepharose (Amersham Pharmacia, Piskataway, NJ, USA) was rocked with 400 μl 1% bovine serum albumin (RIA grade, Sigma) in phosphate-buffered saline (PBS) for 30 min at room temperature. SF (50 μl) plus 1% BSA/PBS (225 μl) were then added to individual bead pellets, followed by PMSF, aprotinin, leupeptin, and EDTA, to give final concentrations of 2 mM, 9.9 U/ml, 13.3 μg/ml, and 4 mM, respectively. After rocking for 2 h, supernatant ('flow-through' [FT]) fractions were collected and bead pellets were washed four times with PBS containing 2 mM EDTA. Gelatin beads were boiled in 40 μl reduced sample buffer (40 mM Tris, pH 6.8, containing 4.3% SDS, 21.5% glycerol, 1 mM EDTA, and 0.2 M dithiothreitol) for 5 min prior to SDS–PAGE .
Preparation from OA synovial fluid of a fraction enriched in the ~170-kDa species
This fraction was prepared from OA SF sample 6 as described, using sequential gelatin and heparin affinity chromatography, step-gradient NaCl elution of ~170-kDa N-terminal FN fragments from the heparin column, and Centriprep (Amicon, Beverly, MA, USA) concentration of the 250 mM NaCl fraction .
One-dimensional (1D) and two-dimensional (2D) electrophoresis was performed as described elsewhere . Six volumes of FT, diluted six-fold during affinity isolation, were submitted to 1D SDS–PAGE alongside one volume of the corresponding SF. The weights of molecular standards (Gibco BRL, Rockville, MD, USA) are those reported by the manufacturer for prestained proteins. For 2D analysis, 5 μl of SF or FT obtained after affinity isolation from SF was added to 100 μl rehydration solution consisting of 2% immobilized pH gradient buffer, 8 M urea, and 2% CHAPS. Dithiothreitol (18.2 mM), PMSF (2 mM), and aprotinin (0.1 U/ml) were added to the rehydration solution just before the sample, and the mixture was centrifuged at 14,000 gfor 15 min and applied via sample cup to a 7-cm isoelectric focusing strip (pre-equilibrated overnight in rehydration solution) for focusing at 20°C in a Multiphor II apparatus (Amersham Pharmacia) at 200 V for 1 min followed by 3500 V for 170 min. After storage at -75°C, strips were incubated for 15 min in 50 mM Tris, pH 8.8, plus 6 M urea, 30% glycerol, 2% SDS, trimmed to exclude ~7 mm from the anodic end, and submitted to 5% SDS–PAGE with an overlay of 0.5% agarose in 25 mM Tris, 192 mM glycine, and 0.1% SDS.
Western blot analysis
Proteins in 1D and 2D gels were electrophoretically transferred and stained as described elsewhere . Nitrocellulose membranes that had been blocked, stained with primary antibodies, and washed with triethanolamine-buffered saline plus 0.05% Tween 20 (TBST) were incubated for 2 hours in TBST containing 125I-labeled donkey Fab' fragments specific for rabbit IgG, or whole rabbit IgG specific for mouse IgG (Amersham Pharmacia) at 0.15 to 0.5 μCi/ml; or horseradish peroxidase (HRP)-conjugated affinity-purified goat anti-mouse IgG (Jackson ImmunoResearch, West Grove, PA, USA). Membranes that had been incubated with iodinated antibodies were washed, dried, and exposed to XAR film (Kodak, Rochester, NY, USA) with an intensifying screen before development. Membranes that had been exposed to HRP conjugates were washed and overlaid with enhanced chemiluminescence (ECL) reagent and then exposed to Hyperfilm ECL (Amersham Pharmacia) for periods of 10 s to 10 min before development. Control membranes were stained with secondary antibodies only.
Quantitation, data presentation, and statistical analysis
Band densities were measured using a Phosphorimager (Molecular Dynamics, Sunnyvale, CA, USA). Quantitative data for OA versus RA samples is expressed as the average ± SD for each group. Statistical comparisons between groups were made with Student's t-test using Sigma Stat Version 2.0 statistical software. P values less than 0.05 were considered significant.
The ~170-kDa N-terminal species of synovial fluid FN typically exhibits reduced affinity for gelatin in comparison with the N-terminal ~200+-kDa species
Since the potential for a particular species of FN to regulate chondrocyte function may be related both to its capacity to be recognized by cell-surface receptors and to its ability to bind to other components of the extracellular matrix, we wished to compare the capacities of the various N-terminal species of SF FN to bind to gelatin (denatured collagen). This was assessed by comparison of the content of N-terminal species of FN in SF samples before and after exposure to immobilized gelatin. As we reported previously, 1D Western blot analysis of unprocessed SF reveals two predominant species in most OA and RA samples, possessing apparent molecular weights of ~200+ and ~170 kDa . Given the proximity of the GBD to the N terminus (Fig. 1) [1–3], both of these large N-terminal species would be expected to include gelatin-binding sequences. However, when samples of SF were subjected to affinity isolation on gelatin beads, the ~200+-kDa species was routinely observed to bind more readily than the ~170-kDa N-terminal species (Fig. 2). For example, in all nine OA samples, the ratio of staining intensities for ~200+- to ~170-kDa bands decreased in the gelatin-bead FT fraction in comparison with the starting material (Fig. 2, top panel). Furthermore, the average intensity of the ~170-kDa band in the FT was 61.5 ± 44.2% of the corresponding value in the starting material for the seven OA samples shown in Fig. 2, whereas FT fractions lacked visible staining for the ~200+-kDa species.
Similarly, although RA samples 16 and 19 lacked sufficient staining of ~170-kDa forms to permit assessment, the ratio of staining intensities for ~200+- to ~170-kDa bands decreased in the FT fractions in comparison with the starting material in the remaining eight RA samples, with little or no staining for ~200+-kDa species in the FT fractions (Fig. 2, bottom). The average intensity of staining of the ~170-kDa band in the FT fractions averaged 81.5 ± 49.6% of the corresponding value in the starting material for the eight RA samples in which two major bands were detected by anti-N-terminal-HBD mAb (excluding samples 16 and 19).
Reflecting the preferential gelatin-binding capacity of the ~200+-kDa species as opposed to that of the ~170-kDa species of SF FN, gelatin isolates from both OA and RA samples were uniformly enriched in the former as compared with the latter species. In addition to ~200+-kDa forms, fragments smaller than ~170 kDa were detected in gelatin isolates derived from 6 of the 8 OA samples and 9 of the 10 RA samples shown in Fig. 2. Specifically, gelatin-binding N-terminal fragments with apparent molecular weights of ~100, ~60, ~50, and ~47 kDa were detected in both OA and RA samples (Fig. 2).
Similar to the staining pattern previously obtained on these same samples with anti-total-FN antibody , anti-N-terminal-HBD mAb was observed to produce preferential staining of the ~200+- as compared with the ~170-kDa species in both OA and RA SF samples. The ratio of staining intensities for the ~200+-kDa bands as compared with the ~170-kDa bands was significantly greater in the 10 RA than in the 9 OA samples (22.6 ± 35.0 and 3.8 ± 6.9, respectively; P < 0.05). Although the magnitude of this difference could largely be attributed to RA samples 16 and 19, which exhibited negligible staining for the ~170-kDa species, the average ratio for the remaining eight RA samples (6.2 ± 3.4) was also significantly greater than for the OA group (P < 0.05). Despite the use of gradient gels with the capacity to resolve species as small as ~15 kDa, little or no staining of forms of FN smaller than ~170 kDa was detected in unconcentrated SF samples by anti-N-terminal-HBD mAb (or anti-total-FN polyclonal antibody; not shown) after autoradiogram exposure times of 5 days (Fig. 2).
Analysis of species of synovial fluid FN bearing sequences from the N-terminal HBD under nonreducing conditions
Since FN exists in nature as dimers that are disulfide-bonded near their C termini (Fig. 1), information regarding the state of such bonds is not forthcoming in reduced electrophoretic analysis. When OA SF sample 6 was subjected to nonreduced SDS–PAGE, species bearing an N-terminal HBD sequence with migration expected of FN dimers and monomers predominated, in addition to a major species that migrated at a position expected for a reduced, ~140-kDa protein (Fig. 3a). The latter species appeared to equate with the ~170-kDa species seen in reduced electrophoresis, since an ~140-kDa band also predominated in the fraction enriched in the ~170-kDa species derived from the same sample , whether staining was achieved with anti-N-terminal-HBD or anti-CBD mAbs (not shown).
SF was also analyzed under nonreducing conditions using an anti-GBD mAb (mAb 1892), which does not recognize FN under reducing conditions (manufacturer's information and unpublished observations, J Peters). In contrast to the anti-N-terminal-HBD mAb, which stained a major fragment species in addition to dimers and monomers, mAb 1892 produced staining of dimeric and monomeric species but did not recognize a faster-migrating fragment species, either in unprocessed SF or in gelatin FT (Fig. 3a). The failure of mAb 1892 to stain species of SF FN smaller than monomers did not stem from an inability to recognize the GBD in FN fragments, since this antibody retained the capacity to produce specific staining of a 45-kDa GBD-containing fragment (Fig. 3b).
Analysis of species of synovial fluid FN bearing sequences from the N-terminal HBD using two-dimensional Western blot analysis
To provide greater electrophoretic resolution of N-terminal species of SF FN, each SF sample was submitted to 2D Western blot analysis using a pH 4–7 isoelectric focusing gradient in the first dimension, followed by reduced 5% SDS–PAGE in the second. Three major species of SF FN were typically detected in samples from both types of patient: a ~200+-kDa cluster of staining, corresponding to the ~200+-kDa species in 1D electrophoresis; a series of ~170-kDa spots corresponding to the ~170-kDa band; and a more faintly stained series of ~185-kDa spots (Figs 4 and 5; Table 1).
In SF samples from both patient groups, the ~200+-kDa N-terminal species was typically detected as a cluster that spanned a broad pI range (~4.9 to ~5.9) (Figs 4 and 5; Table 1). In 8 of 10 RA and 4 of 9 OA samples, a separate cluster of ~200+-kDa staining could also be detected migrating closer to the anode (pI ~4.0 to ~4.4) than the major cluster (Figs 4 and 5; Table 1). This 'extra' material was streaked vertically upward in the second dimension in four of eight RA (Fig. 5) and two of four OA (Fig. 4) samples.
In most samples, the ~170-kDa species resolved into a cluster of one to six spots arrayed nearly horizontally in the second dimension, with pIs ranging from ~6.0 to ~7.0 (Figs 4 and 5). In the two SF samples for which gelatin FT fractions were submitted to 2D analysis (OA samples 1 and 3), these subspecies persisted in the absence of ~200+-kDa species (not shown). An additional ~170-kDa spot that migrated farther toward the anode (pI ~4.3) was detected by anti-N-terminal mAb in two OA samples (including sample 4 in Fig. 4; Table 1) and one RA (sample 17 in Fig. 5; Table 1) sample. Additionally, a spot that migrated slightly more slowly in the second dimension (~180 kDa, pI ~5.5–5.8) was detected in OA samples 1, 3, 4, 7, and 8, as well as RA samples 11–15 (denoted by diagonal arrows pointing upward and to the left in Figs 4 and 5; Table 1).
In 4 of the 9 OA samples (1 and 3 in Fig. 4, also samples 5 and 9 in blots not shown; Table 1) and 6 of the 10 RA samples (10, 11, 13–15, and 18 in Fig. 5; Table 1), mAb to the N-terminal HBD produced staining of an ~185-kDa species that migrated slightly farther toward the anode (pIs ranging from ~5.8 to ~6.4) than the ~170-kDa cluster (Figs 4 and 5). A faint band corresponding to this species could also be detected in 1D Western blots subjected to long autoradiographic exposures (not shown) . Similar to the ~170-kDa cluster, the ~185-kDa subspecies persisted in the gelatin FT fraction from OA sample 1 (no staining of an ~185-kDa species was evident in the aliquot of OA sample 3 that was submitted to gelatin isolation), despite the absence of ~200+-kDa species from this fraction (not shown).
In addition to pH 4–7 gradients, OA samples 3, 5, 8, and 9 were analyzed using pH 3–10 first-dimension gradients. In each case, a cluster of staining that migrated at ~130-kDa in the second dimension, and too close to the anode (pI ~9.1) to be evident in pH 4–7 gradients, was detected (sample number 9, Fig. 4i). Faintly stained ~130-kDa species were also detected in long exposures of 1D blots from these four samples (Fig. 1) .
Anti-N-terminal-HBD and anti-CBD antibodies recognize the same ~170-kDa FN subspecies in 2D Western blot analysis of OA and RA synovial fluid samples
Anti-CBD and anti-N-terminal-HBD mAbs were observed to stain the same ~170-kDa spots in 2D analysis of OA sample number 6 (Fig. 6). Additionally, each of the two mAbs exhibited corecognition of ~170- and ~185-kDa species in RA sample 18 (not shown).
Despite dramatic clinical and pathologic differences between the OA and RA, we have detected qualitatively similar arrays of N-terminal species of FN in SF samples from patients with the two disorders. Specifically, although there was a greater preponderance of ~200+-kDa as compared with ~170-kDa forms in RA versus OA samples, generally similar assortments of such species, ranging from ~47 to ~200+ kDa, were detected in the two types of sample both by 1D and 2D Western blot analysis. Therefore, the similarity in 2D electrophoretic resolution patterns that we previously reported for one patient with OA and another with RA  appears to be more generally applicable.
2D Western blot analysis revealed that samples of SF from OA and RA joints share in common at least six N-terminal species of FN (Table 1). One of the most prevalent, with a molecular weight of ~170 kDa, was found to usually be comprised of subspecies which, by antibody mapping in this and a previous study , appear to include sequences from four domains that have previously, in the context of small purified fragments or partial recombinant FNs, proved to be potent in the regulation of chondrocyte function, namely, the N-terminal HBD [18–20, 23, 24], the GBD , the central CBD [18–20], and the alternatively spliced EIIIA segment . The ~170-kDa species therefore appears structurally similar to a placenta-derived FN fragment that was previously observed to trigger expression of matrix metalloproteinase by synovial cells .
Although the mechanisms by which FN fragments regulate chondrocyte function remain uncertain [25, 26], a close physical interaction has been detected between central cell-binding FN fragments and the α5 integrin subunit on chondrocytes in vitro, suggesting that surface-expressed integrins could constitute intermediaries in the transduction of catabolic signals from such fragments to chondrocytes . Such signal transmission could also emanate from sequences near the N terminus of FN, based upon the observation that N-terminal fragments lacking central CBD sequences are recognized by α5β1 integrins on fibroblasts . Although similar observations have not yet been reported for chondrocytes, α5β1 integrins are prevalent on the surfaces of chondrocytes in vivo and in vitro [35, 36]. Therefore, the ~170-kDa forms of FN described in this study could potentially be recognized by chondrocyte α5β1 integrins via sequences situated in both the central CBD and the N-terminal HBD. Elucidation of the primary sequence of all of the FN species detected in this study will provide more clues to their functions.
Qualitatively similar assortments of FN species bearing sequences from the N-terminal HBD are present in SF samples from patients with OA and RA. One of the predominant species, possessing a molecular weight of ~170-kDa, is composed of distinct subspecies that have lesser capacities for gelatin binding than larger N-terminal species of SF FN. Since the ~170-kDa species has central CBD sequences, yet exhibits reduced binding to denatured collagen, it could potentially represent a soluble agent with the capacity to disrupt FN-mediated interactions between chondrocytes and their insoluble extracellular matrix. Based upon their potential roles in the pathogenesis of arthritis, the species described in this study also constitute candidate soluble biomarkers for the joint-destructive process in OA and RA.
bovine serum albumin
triethanolamine-buffered saline plus 0.05% Tween 20
Hynes RO: Fibronectins. New York, NY: Springer-Verlag Inc. 1990
Mosher DF: Assembly of fibronectin into extracellular matrix. Curr Opin Struct Biol. 1993, 3: 214-222. 10.1016/S0959-440X(05)80155-1.
Burton-Wurster N, Lust G, Macleod JN: Cartilage fibronectin isoforms: in search of functions for a special population of matrix glycoproteins. Matrix Biol. 1997, 15: 441-454. 10.1016/S0945-053X(97)90018-4.
Zhang Z, Vuori K, Reed JC, Ruoslahti E: The α5β1 integrin supports survival of cells on fibronectin and up-regulates Bcl-2 expression. Proc Natl Acad Sci USA. 1995, 92: 6161-6165.
Sakai T, Johnson KJ, Murozono M, Sakai K, Magnuson MA, Wieloch T, Cronberg T, Isshiki A, Erickson HP, Fassler R: Plasma fibronectin supports neuronal survival and reduces brain injury following transient focal cerebral ischemia but is not essential for skin-wound healing and hemostasis. Nature Med. 2001, 7: 324-330. 10.1038/85471.
Wurster NB, Lust G: Fibronectin in osteoarthritic canine articular cartilage. Biochem Biophys Res Comm. 1982, 109: 1094-1101.
Burton-Wurster N, Butler M, Harter S, Colombo C, Quintavalla J, Swartzendurber D, Arsenis C, Lust G: Presence of fibronectin in articular cartilage in two animal models of osteoarthritis. J Rheumatol. 1986, 13: 175-182.
Rees JA, Ali SY, Brown RA: Ultrastructural localization of fibronectin in human osteoarthritic cartilage. Ann Rheum Dis. 1987, 46: 816-822.
Jones KL, Brown M, Ali SY, Brown RA: Immunohistochemical study of fibronectin in human osteoarthritic and disease free articular cartilage. Ann Rheum Dis. 1987, 46: 810-815.
Clemmensen I, Andersen RB: Different molecular forms of fibronectin in rheumatoid synovial fluid. Arthritis Rheum. 1982, 25: 25-31.
Carsons S, Lavietes BB, Diamond HS, Kinney SG: The immunoreactivity, ligand, and cell binding characteristics of rheumatoid synovial fluid fibronectin. Arthritis Rheum. 1985, 28: 601-612.
Griffiths AM, Herber KE, Perrett D, Scott DL: Fragmented fibronectin and other synovial fluid proteins in chronic arthritis: their relation to immune complexes. Clin Chim Acta. 1989, 184: 133-146. 10.1016/0009-8981(89)90283-0.
Xie D-L, Meyers R, Homandberg GA: Fibronectin fragments in osteoarthritic synovial fluid. J Rheumatol. 1992, 19: 1448-1452.
Chevalier X, Claudepierre P, Groult N, Zardi L, Hornebeck W: Presence of ED-A containing fibronectin in human articular cartilage from patients with osteoarthritis and rheumatoid arthritis. J Rheumatol. 1996, 23: 1022-1030.
Peters JH, Carsons S, Kalunian K, McDougall S, Yoshida M, Ko F, van der Vliet-Hristova M, Hahn TJ: Preferential recognition of a fragment species of osteoarthritic synovial fluid fibronectin by antibodies to the alternatively spliced EIIIA segment. Arthritis Rheum. 2001, 44: 2572-2585. 10.1002/1529-0131(200111)44:11<2572::AID-ART438>3.0.CO;2-Y.
Chevalier X, Groult N, Hornebeck W: Increased expression of the Ed-B-containing fibronectin (an embryonic isoform of fibronectin) in human osteoarthritic cartilage. Br J Rheumatol. 1996, 35: 407-415.
Homandberg G, Wen C, Hui F: Cartilage damaging activities of fibronectin fragments derived from cartilage and synovial fluid. Osteoarthritis Cartilage. 1998, 6: 231-244. 10.1053/joca.1998.0116.
Homandberg GA, Meyers R, Xie D-L: Fibronectin fragments cause chondrolysis of bovine articular cartilage slices in culture. J Biol Chem. 1992, 267: 3597-3604.
Arner EC, Tortorella MD: Signal transduction through chondrocyte integrin receptors induces matrix metalloproteinase synthesis and synergizes with interleukin-1. Arthritis Rheum. 1995, 38: 1304-1314.
Bewsey KE, Wen C, Purple C, Homandberg G: Fibronectin fragments induce the expression of stromelysin-1-mRNA and protein in bovine chondrocytes in monolayer culture. Biochim Biophys Acta. 1996, 1317: 55-64. 10.1016/0925-4439(96)00037-3.
Homandberg GA, Hui F, Wen C, Purple C, Bewsey K, Koepp H, Huch K, Harris A: Fibronectin-fragment-induced cartilage chondrolysis is associated with release of catabolic cytokines. Biochem J. 1997, 321: 751-757.
Yasuda T, Poole A: A fibronectin fragment induces type II collagen degradation by collagenase through an interleukin-1-mediated pathway. Arthritis Rheum. 2002, 46: 138-148. 10.1002/1529-0131(200201)46:1<138::AID-ART10051>3.0.CO;2-K.
Gemba T, Valbracht J, Alsalameh S, Lotz M: Focal adhesion kinase and mitogen-activated protein kinases are involved in chondrocyte activation by the 29-kDa amino-terminal fibronectin fragment. J Biol Chem. 2001, 277: 907-911. 10.1074/jbc.M109690200.
Chow G, Knudson CB, Homandberg G, Knudson W: Increased expression of CD44 in bovine articular chondrocytes by catabolic cellular mediators. J Biol Chem. 1995, 270: 27734-27741. 10.1074/jbc.270.46.27734.
Barilla M-L, Carsons SE: Fibronectin fragments and their role in inflammatory arthritis. Semin Arthritis Rheum. 2000, 29: 252-265. 10.1016/S0049-0172(00)80012-8.
Peters JH, Loredo GA, Benton HP: Is osteoarthritis a "fibronectin-integrin imbalance disorder"?. Osteoarthritis Cartilage. 2002, 10: 831-835. 10.1053/joca.2002.0845.
Saito S, Yamaji N, Yasunaga K, Saito T, Matsumoto S-I, Katoh M, Kobayashi S, Masuho Y: The fibronectin extra domain A activates matrix metalloproteinase gene expression by an interleukin-1-dependent mechanism. J Biol Chem. 1999, 274: 30756-30763. 10.1074/jbc.274.43.30756.
Homandberg GA, Meyers R, Williams JM: Intraarticular injection of fibronectin fragments causes severe depletion of cartilage proteoglycanse in vivo. J Rheumatol. 1993, 20: 1378-1382.
Williams JM, Zhang J, Kang H, Ummadi V, Homandberg GA: The effects of hyaluronic acid on fibronectin fragment mediated cartilage chondrolysis in skeletally mature rabbits. Osteoarthritis Cartilage. 2003, 11: 44-49. 10.1053/joca.2002.0864.
Grant MB, Cabellero S, Tarnuzzer RW, Bass KE, Ljubimov AV, Spoerri PE, Galardy RE: Matrix metalloproteinase expression in human retinal microvascular cells. Diabetes. 1998, 47: 1311-1317.
Gardner JM, Hynes RO: Interaction of fibronectin with its receptor on platelets. Cell. 1985, 42: 439-448. 10.1016/0092-8674(85)90101-1.
Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature. 1970, 227: 680-685.
Homandberg GA, Costa V, Wen C: Fibronectin fragments active in chondrocytic chondrolysis can be chemically cross-linked to the alpha5 integrin receptor subunit. Osteoarthritis Cartilage. 2002, 10: 938-949. 10.1053/joca.2002.0854.
Hocking DC, Sottile J, McKeown-Longo PJ: Activation of distinct α5β1-mediated signaling pathways by fibronectin's cell adhesion and matrix assembly domains. J Cell Biol. 1998, 141: 241-253. 10.1083/jcb.141.1.241.
Salter DM, Hughes DE, Simpson R, Gardner DL: Integrin expression by human articular chondrocytes. Brit J Rheum. 1992, 31: 231-234.
Woods VL, Schreck PJ, Gesink DS, Pacheco HO, Amiel D, Akeson WH, Lotz M: Integrin expression by human articular chondrocytes. Arthritis Rheum. 1994, 37: 537-544.
JHP was supported by a UCLA Claude Pepper Older Americans Independence Center, NIA #P60 AG10415, a gift from the Charles B See Foundation, and a Career Development and a Merit Review Award, both from the Department of Veterans Affairs; TJH was supported by a VA Merit Review Grant; and SC was supported in part by the Arthritis Foundation, Long Island Chapter. We wish to thank Dr Richard Hynes for his generous gifts of antibodies, Dr Livingston Van De Water for his critical reading of the manuscript, and Jerry Sproul of the West LA VAMC Geriatric Research, Education and Clinical Center for his fine assistance with computer graphics.