• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Jun 11, 2002; 99(12): 8259–8264.
Published online May 28, 2002. doi:  10.1073/pnas.122033199
PMCID: PMC123055
Medical Sciences

Basic FGF mediates an immediate response of articular cartilage to mechanical injury


The extracellularly regulated kinase (ERK), one of the three types of mitogen-activated kinases, was rapidly activated after cutting porcine articular cartilage either when maintained as explants or in situ. Cutting released a soluble ERK-activating factor from the cartilage, which was purified and identified by MS as basic fibroblast growth factor (bFGF). Experiments with neutralizing Abs to bFGF and an FGFR1 tyrosine kinase inhibitor showed that this growth factor was the major ERK-activating factor released after injury. Treating cartilage with the heparin-degrading enzyme heparitinase also caused release of bFGF, suggesting the presence of an extracellular store that is sequestered in the matrix and released upon damage. Basic FGF induced the synthesis of a number of chondrocyte proteins including matrix metalloproteinases 1 and 3, tissue inhibitor of metalloproteinases-1, and glycoprotein 38, which were identified by MS. The strong induction of matrix metalloproteinases and tissue inhibitor of metalloproteinases-1 suggests that bFGF could have a role in remodeling damaged tissue.

Cartilage is an avascular tissue in which cells are isolated from one another by an extensive extracellular matrix of fibers (largely type II collagen), soluble polymers (principally proteoglycans), and water. Articular chondrocytes are long-lived and rarely divide after maturity. Their role is to maintain the matrix, but how this is achieved and its rate of turnover is not well understood.

The common primary disease of cartilage is osteoarthritis (OA). Its characteristic features are initial superficial fissuring, which extends into the depth of the tissue, clustering of chondrocytes in enlarged lacunae (perhaps as a result of proliferation), and loss of proteoglycan and collagen. Eventually the tissue degenerates, exposing the underlying bone (1).

Mechanical injury predisposes cartilage to OA, which occurs prematurely, for example, after sporting injuries. Cuts or defects made on the articular surface do not heal unless they extend to the subchondral bone, when repair takes place by formation of fibrous tissue from the underlying bone. Superficial cutting of the articular surface of rabbit knees leads, after 3–4 months, to degenerative changes similar to those observed in human OA (2). Larger defects in articular cartilage result initially in a zone of necrosis next to the cut edge. After a week or so, cells adjacent to this area proliferate and increased metabolic activity is detected (reviewed in ref. 3).

While investigating mitogen-activated protein kinase signaling pathways in cartilage, we serendipitously found that cutting the tissue rapidly activated the extracellularly regulated kinases (ERKs 1 and 2) in the chondrocytes. This pathway can be activated by many stimuli, including growth factors, cytokines, chemokines, and integrin engagement (46).

Little is known about how tissues respond to mechanical damage other than the well known phenomena associated with haemostasis and epithelial regeneration. This response, together with the fact that mechanical injury predisposes cartilage to OA, led us to investigate the basis of the ERK activation. We have found that ERK activation is caused by release of basic fibroblast growth factor (bFGF) from a sequestered extracellular pool in which it is possibly bound to heparan sulphate (HS) proteoglycan. The growth factor causes sustained ERK activation and has striking effects on chondrocyte gene expression. Our findings suggest that release of bFGF signals to the cells that damage has occurred, a mechanism that may be common to other connective tissues. Our findings also raise the possibility that bFGF plays a homeostatic role in articular cartilage. Exaggerated or inappropriate responses after mechanical injury might therefore contribute to the development of OA.

Materials and Methods

Materials and Tissue.

Porcine articular cartilage was from either the metacarpophalangeal or the knee joints of 3-month-old pigs obtained from a local abattoir 4–8 h after slaughter. Human articular was obtained from surgical specimens.

Abs to ERK (no. 9122) and phosphorylated ERK (pERK) (no. 9101S) were from New England Biolabs. Abs to bFGF, insulin-like growth factor, and platelet-derived growth factor (PDGF) were from Upstate Biotechnology U.K. Secondary Abs were from Dako. Recombinant bFGF (146-aa length) was from PeproTech (London), and heparitinase was from Calbiochem. FGFR1 inhibitor (SB402451) (7) was from S. Skaper (GlaxoSmithKline Pharmaceuticals). PDGF, enhanced chemiluminescence reagents, and radioisotopes ([35S]methionine/[35S]cysteine and 35SO4) were from Amersham Pharmacia Biotech. Resource S, heparin-agarose, MonoS, and Superdex 75 were from Amersham Pharmacia Biotech. Pronase E was from BDH. DMEM and FCS were from PAA Laboratories (Somerset, U.K.). All other reagents were best available grade from Sigma.

Tissue Culture.

Cartilage explants (0.125 g) were cultured in DMEM (500 μl) with 25 mmol of Hepes, penicillin/streptomycin, and amphotericin. Isolated chondrocytes were incubated in medium supplemented with 10% FCS. Freeze-thawed cartilage was frozen in liquid nitrogen for 2 min and thawed in a water bath at 37°C 3 times, after which the explants were washed thoroughly and further incubated as required.

Isolation of Chondrocytes.

Cartilage was incubated with pronase E (1 mg/ml/1 g of cartilage) for 30 min at 37°C, followed by collagenase (1 mg/ml/1 g of cartilage) for 5 h at 37°C. The digest was strained then centrifuged at 500 × g for 5 min. Pellets were washed then resuspended in DMEM containing 10% FCS. Cells were counted and plated on 24-well plates (1.5-cm diameter) at a density of 1 million cells per well (100% confluent).

Generation of Conditioned Medium (CM).

Cartilage was dissected and incubated for 48 h in serum-free medium or was freeze-thawed after dissection and washed thoroughly. Medium was conditioned by either uncut (rested) or cut cartilage (1 g/2 ml) for 30 min, after which time the medium was removed for immediate use or stored at −20°C. For larger volumes of cartilage medium was conditioned by fragmentation of cartilage in a polytron blender.

Western Blot Assay for pERK.

Five hundred microliters (explants) or 250 μl (cells) of ice-cold dissociative lysis was added to each sample. Lysates were removed after 45 min and clarified by centrifugation (10,000 × g). Samples were mixed with 4X sample buffer, boiled (5 min), then run on 12.5% SDS/PAGE and transferred to poly(vinylidene difluoride) membrane. Membranes were blocked (30 min) in 5% dried milk, then treated with primary Ab (1 h), washed 3 times (PBS/0.05% Tween 20), and incubated (1 h) with secondary horseradish peroxidase-conjugated Ab. Signal was developed with enhanced chemiluminescence and visualized by autoradiography.

Protein Kinase Assays.

Immunoprecipitation assays for ERK, p38, and c-jun N-terminal kinase activity were carried out as described (8).

Precipitation of Glycosaminoglycan-Containing Proteoglycans with Cetyl Pyridinium Chloride (CPC).

Three milligrams of 10% CPC was added per ml of CM, incubated (room temperature, 10 min) to form precipitates, put on ice (1 h), and spun at 4°C. Supernatants were removed and dialyzed.

Electrophoresis and Silver Staining.

Proteins were resolved by SDS/PAGE either in Laemmli buffer or in an ammediol buffer system. Gels were fixed and stained by a silver method compatible with mass spectrometric identification of proteins.

Protein Identification by MS.

In-gel digestion.

In-gel digestion of silver bands with trypsin was performed according to published methods (912). Matrix-assisted laser desorption ionization mass spectra were recorded with a TofSpec 2E spectrometer (Micromass, Manchester, U.K.). The resulting peptide mass fingerprints were searched against the National Center for Biotechnology Information nonredundant database (http://www3.ncbi.nlm.nih.gov) with the protein probe search engine (Micromass).

Electrospray MS.

Samples were desalted and dissolved in 1–2 μl of 50% methanol/0.1% aqueous formic acid. They were mounted in the source of a Q-Tof hybrid quadrupole/orthogonal acceleration time-of-flight spectrometer (Micromass). Amino acid sequences were determined from the daughter ion spectra semimanually with the assistance of the Micromass peptide-sequencing program, pepseq. Deduced sequence was searched against the NCBI nonredundant database with blast (13).

Metabolic Labeling.

Chondrocytes were washed and then stimulated overnight (without serum). The following day they were incubated with 15 μCi per well of [35S]methionine/[35S]cysteine in methionine/cysteine-free medium. After 6 h medium was removed, centrifuged (10,000 × g for 5 min), mixed with sample buffer, run on 12.5% SDS/PAGE, and silver-stained. Newly synthesized proteins were visualized by autoradiography.

Proteoglycan Synthesis.

This synthesis was carried out as described (14). Briefly, rested cartilage explants were weighed (≈0.125 g) then incubated for 48 h with serum-free medium, IL-1, bFGF, or IL-1 and bFGF (in triplicate). For the final 6 h they were pulsed with 25 μCi/ml of 35SO4. Explants and media were separately digested (2 h, 65°C) in 0.25 ml of 25 μg/ml papain. Glycosaminoglycans were precipitated with 10% CPC and washed with 3% CPC. The pellets were dissolved in formic acid, mixed with 3 ml of scintillant, and counted in a liquid scintillation counter.


Explanting Articular Cartilage Activates ERK.

Explanting cartilage from porcine metacarpophalangeal joints into serum-free medium caused strong activation of ERK as shown by an increase in the phosphorylated form of the enzyme (Fig. (Fig.11A). Little pERK was detected in the tissue put into lysis buffer directly after dissection (Fig. (Fig.11A, shown as 0 h). A strong signal was seen by 0.5 h, which persisted for several hours. The activity did not return to resting levels until after about 48 h. The medium was changed after overnight incubations (indicated by the arrows in Fig. Fig.11A) and this may have contributed to the return to baseline level by removing ERK-activating factors. The blot shown in Fig. Fig.11A was stripped and stained with an Ab to ERK protein (Fig. (Fig.1A1A Lower), which showed that the amount of ERK in the cells was similar at the times tested. In other experiments, ERK activation could be detected as early as 5 min (data not shown). To check that the ERK detected by the Ab to the phosphorylated form was active, the kinase was also immunoprecipitated from lysates of freshly dissected tissue, or tissue incubated for 1 h, and assayed for its ability to phosphorylate myelin basic protein. Strong activation of ERK was seen (Fig. (Fig.11B). Immunoprecipitation kinase assays were also carried out for p38 mitogen-activated protein kinase and c-jun N-terminal kinase (JNK). Explantation caused two-fold activation of p38 and four-fold transient activation of JNK, which returned to baseline within 1 h (data not shown).

Figure 1
Explantation of articular cartilage activates ERK. (A) Cartilage was dissected from metacarpophalangeal joints of porcine trotters either directly into lysis buffer (0 h) or into serum-free medium for the periods specified before lysis. Arrows indicate ...

Cutting Cartilage Activates ERK.

ERK activation could have been caused by exposure of the tissue to the culture medium (although it contained no serum) or to the physical damage of cutting. We investigated the effect of cutting cartilage in vitro. Cartilage was explanted and maintained in serum-free medium for 48 h to allow the ERK activity to return to a low level (as shown at 49 and 66 h in Fig. Fig.11A). The tissue was then treated with IL-1, a known activator of ERK, or was recut with a sharp scalpel. IL-1 and recutting both activated ERK within 10 min (Fig. (Fig.2A2A Upper). Levels of ERK protein did not change (Fig. (Fig.2A2A Lower). Similar experiments were performed on normal human articular cartilage. Recutting human explants also resulted in activation of ERK (Fig. (Fig.22B) but no activation of p38 or c-jun N-terminal kinase (data not shown). To create injury in situ we made closely spaced (1 mm), parallel superficial cuts with a scalpel on the articular surfaces of a metacarpophalangeal joint to see whether that would activate ERK. The porcine joint was opened and cartilage scored at time 0. Cartilage was dissected directly into cold lysis buffer at 5, 10, 15, and 20 min after scoring. Lysates were immunoblotted for pERK (Fig. (Fig.3).3). Activation was seen 15 min after scoring. Activation was not seen in cartilage of an opened joint whose surface was not scored (data not shown).

Figure 2
Recutting rested cartilage explants reactivates ERK. (A) Cartilage explants were dissected into serum-free medium and allowed to rest for 48 h (allowing the original ERK activation to reach undetectable levels). Explants were incubated with serum-free ...
Figure 3
Scoring the articular surfaces “in vivo ” activates ERK. At time 0 a trotter was opened and scored with a scalpel on all articular surfaces. After times 5, 10, 15, and 20 min, scored cartilage was dissected into lysis buffer. Cell lysates ...

Activation of ERK Is Caused by Release of a Soluble Factor from Recut Cartilage.

Either mechanical stress of the cells directly activated the kinase or a soluble factor was released upon physical injury. To investigate the second possibility, culture medium was conditioned either by rested cartilage explants for 1 h or by recut cartilage. Medium was removed 5 min after recutting, stored, and replaced with fresh serum-free medium. Medium was then conditioned for further periods over the next 48 h. The CM samples were put onto rested cartilage explants (30 min), which were assayed for activation of ERK (Fig. (Fig.4).4). An ERK-activating factor was released in the first 5 min after cutting and there was little further release after this. This result suggested that the ERK-activating factor (or factors) is present in the tissue and that de novo synthesis may not be required for its release. Recutting experiments were also performed with cartilage explants that had been frozen and thawed 3 times then washed in several changes of medium. Medium in which either this dead cartilage or live tissue was recut was equally active (data not shown). Taken together, these results suggested that the ERK-activating factor was present preformed in the explant, and live cells were not required for its release. In addition, these results suggested that the factor may be extracellular, because intracellular contents would largely be lost on freezing, thawing, and washing.

Figure 4
A factor is rapidly released from recut cartilage, which activates ERK. Medium was conditioned by rested cartilage for 1 h (C) or from recut cartilage over a 48-h period. Medium was removed at the end of each indicated period and replaced with fresh medium. ...

Purification and Identification of the ERK-Activating Factor.

Many growth factors, cytokines, hormones, and other agents activate ERK in cells. Preliminary experiments showed the factor was heat-labile and nondialyzable. To determine whether there were several factors or just a single entity, we decided to purify and identify the factor by MS. Porcine knee cartilage (100 g) was freeze-thawed 3 times, washed thoroughly, and rested for 3 days. Fresh serum-free medium (250 ml) was added, and the cartilage was fragmented in a polytron blender (rather than recut) because of the large amount. After extraction for 30 min the medium was centrifuged and the supernatant was removed. The CM was treated with CPC to remove glycosaminoglycans and was dialyzed overnight against S buffer for cation-exchange. This procedure was not associated with any loss of activity. The sample was chromatographed on a 6-ml Resource-S column that was eluted with a 0–1 M salt gradient. Active fractions eluted between 0.6 and 0.7 M NaCl. These were pooled, diluted 5 times into Tris chromatography buffer, and loaded on to a 1-ml heparin-agarose column. A 1.25 M NaCl elution step removed most of the protein. The active material was eluted by 2.5 M NaCl. It was diluted 10-fold into S chromatography buffer and chromatographed on a 0.1-ml MonoS column (Fig. (Fig.55A). The active fractions (11 and 12) were pooled and run on a gel filtration column (Fig. (Fig.55B). The active material eluted in fractions 7 and 8 (Fig. (Fig.5B5B Upper), and on SDS/PAGE these fractions contained a diffuse silver-stained band, migrating at 17 kDa (Fig. (Fig.5B5B Lower). The gel slice containing the band was treated with trypsin, and the peptides of the digest were analyzed by MS. Database searching identified the protein as bFGF (also called FGF2). Fifteen peptides had mass values corresponding to potential tryptic peptides in the sequence of bFGF (Fig. (Fig.55C). These gave 70% coverage of the amino acid sequence. The presence of a signal at m/z 2121.1, corresponding to the predicted N-terminal-containing peptide PALPEDGGSGAFPPGHFKDPK, indicated that bFGF was the fully processed form. The identification was confirmed by tandem MS sequencing of a doubly charged peptide ion at m/z 755.3 from which the sequence CVTDECFFFER was deduced (data not shown). A blast search of this peptide against the NCBI nonredundant database retrieved FGF2 as the best hit (expect value = 3 × 10−4). The prominent (15 kDa) band in fractions 9 and 10 (Fig. (Fig.5B5B Lower) was identified as ribonuclease 4. The presence of bFGF was confirmed by immunoblotting media conditioned by both recut porcine and recut human cartilage (data not shown).

Figure 5
Purification and identification of ERK-activating factor. Purification as described in Results. (A) MonoS fractions were assayed (1/50) on isolated chondrocytes. Active fractions 11 and 12 (100 μl each) were combined and run on a gel filtration ...

Neutralization of FGF Inhibited Activity in Recut CM.

One major chromatographable factor was identified by the purification procedure. To test whether bFGF accounted for all or part of the activity in the medium from the recut cartilage, the effect of a neutralizing Ab was tested. The neutralizing Ab inhibited the activity of the medium by 75% (Fig. (Fig.66A). The inhibition was greater when the medium had been treated with CPC. To exclude the involvement of other known connective tissue growth factors, Abs to platelet-derived growth factor and insulin-like growth factors I and II were also tested, but neither reduced the activity in the CM (data not shown).

Figure 6
Medium conditioned by recut cartilage can be neutralized with Abs to bFGF and inhibited by an FGFR inhibitor. (A) Twenty ng per ml of recombinant FGF (rFGF), CM by recut cartilage (CM), and CM after treatment with CPC (CPC CM) were preincubated either ...

FGF Receptor Blockade Inhibits ERK Activation After Recutting.

The experiments with the neutralizing Ab strongly suggested that bFGF was the major active factor in the medium conditioned by recutting cartilage. Abs penetrate the cartilage matrix poorly, therefore to inhibit the activation of ERK after cutting cartilage, an FGF receptor tyrosine kinase inhibitor (SB402451) was used. Rested cartilage was preincubated for 4 h with either vehicle or with 50 nM of SB402451. Cartilage explants were then treated in one of four different ways: (i) kept in serum-free medium, (ii) treated with IL-1 (an agonist that does not activate ERK through a receptor tyrosine kinase), (iii) treated with medium conditioned by re-cut cartilage, or (iv) recut. Cartilage that had been incubated with SB402451 responded to IL-1 but not to the CM or to recutting (Fig. (Fig.66B).

Heparitinase Treatment of Cartilage Explants Causes Release of Biologically Active bFGF.

Further experiments were made to establish the location of the bFGF in cartilage. The growth factor is present in the cytosol of many cell types, including chondrocytes in culture (data not shown). Release of bFGF after injury could therefore be from a cytosolic pool. However, bFGF has a high affinity for heparin and HS and is also thought to reside in the matrix bound to HS-containing proteoglycans. To investigate the possible existence of an extracellular pool of bFGF in cartilage, live explants were treated for 4 h with medium, with or without heparitinase. The medium samples were tested for their ability to activate ERK in isolated primary chondrocytes (Fig. (Fig.77A). Medium from the heparitinase-treated explants (HT CM) activated ERK, whereas medium from the control cultures lacking the enzyme (B CM) did not. Heparitinase itself did not activate ERK in the chondrocytes (Fig. (Fig.77A). Explants were also treated with chondroitinase ABC (an enzyme that degrades chondroitin sulphate), but this treatment did not release an ERK-activating factor. The ERK-activating factor released by heparitinase was inhibited by the Abs to bFGF (Fig. (Fig.77B). Its identity was also confirmed by Western blot (data not shown).

Figure 7
Heparitinase treatment liberates FGF. Cartilage explants were incubated with buffer alone, heparitinase (1 mμ/ml), or chondroitinase ABC (5 mμ/ml). After 4 h each CM was removed (B CM, HT CM, and Ch CM, respectively) and ...

bFGF Regulates Chondrocyte Gene Expression in Chondrocytes.

We examined the effect of bFGF on chondrocyte gene expression. Newly synthesized proteins were visualized by autoradiography (Fig. (Fig.8).8). bFGF strongly induced seven protein bands. Four were identified by MS (see Fig. Fig.8);8); they were tissue inhibitor of metalloproteinases 1 (TIMP-1), matrix metalloproteinases (MMPs) 1 and 3, and gp38. The latter is the porcine homologue of human gp39 (also known as YKL40). The growth factor had no effect on synthesis of type II collagen as judged by expression of the C-terminal propeptide (Fig. (Fig.8,8, band F). As expected, stimulation of chondrocytes with medium conditioned by recut cartilage (RC CM) also strongly induced TIMP-1, gp38, and MMP-1 (Fig. (Fig.8).8). However, the pattern of protein expression was not identical to that induced by bFGF, suggesting that other factors may be present in the CM.

Figure 8
FGF-regulated gene expression. Isolated primary porcine chondrocytes were incubated overnight with either serum-free medium (control) or 100 ng/ml of recombinant FGF. The following day cells were pulsed for 6 h with 15 μCi per well (150 ...

bFGF Has No Effect on Sulphated Proteoglycan Synthesis.

The effect of bFGF on sulphated proteoglycan synthesis was determined by measuring the incorporation of radio-sulphate into CPC-precipitable material. bFGF (200 ng/ml) had no significant effect on proteoglycan synthesis by chondrocytes in either explants or primary monolayer culture (data not shown). In addition, the inhibitory effect of IL-1 on proteoglycan synthesis was not counteracted by costimulation with bFGF.


Mechanical injury to cartilage causes a rapid cellular response because of release of a pool of bFGF. Our results are a demonstration of bFGF playing an immediate role in injury, and suggest a possible function for stored extracellular bFGF, namely to act as a signal of tissue injury.

bFGF is one of a family of nine (15) ubiquitous growth factors that have a broad range of activities (1618). Both acidic and bFGF have been implicated in the response of tissues to injury. In the skin, mRNA for bFGF was induced 1–5 days after incisional injury (19, 20), and levels of bFGF were increased in wound fluids (21). mRNA for bFGF was increased in postnatal rat retinas 1 day after needle incision (22), and immunostaining for bFGF was detected 1 week after focal mechanical lesions had been administered to rat brains (23). None of these studies demonstrated the immediate release of preexisting bFGF protein after injury.

The results of the chromatography of cartilage extract, together with the effects of the neutralizing Ab and the FGF receptor kinase inhibitor, were consistent with bFGF being the major ERK-activating factor released from the cartilage. The neutralizing Ab to bFGF inhibited by 75% the ERK activation caused by the medium conditioned by re-cut cartilage. The inhibition was greater when the CM had been treated with CPC to remove sulphated proteoglycans. This finding suggests that the bFGF in the crude CM was partly inaccessible to the Ab. Removal of glycosaminoglycans may have had the effect of freeing the bFGF, thus making it more accessible to neutralization by the Ab. This result would also explain the fact that the crude CM was more active after removal of proteoglycans.

Having established that bFGF was the major factor responsible for the ERK activation it was important to decide to what extent the growth factor was being released from inside cells, or from an extracellular reservoir. Five isoforms of bFGF are made from different initiation sites of translation (reviewed in ref. 24). The smallest form (18 kDa) is cytosolic. The mechanism of its release from cells is unknown. Extracellularly it is likely to be sequestered in the matrix bound to HS-containing proteoglycans such as perlecan, a component of basement membranes (25), or to cell surface proteoglycans such as syndecan (26). Despite the fact that chondrocytes lack a basement membrane, high levels of perlecan are found throughout the extracellular matrix of articular cartilage (27). Whether this colocalizes with bFGF remains to be determined. Binding of bFGF to HS is essential for full activation through the high affinity FGF receptor (28), and perlecan was identified as the proteoglycan responsible for this cofactor activity in vivo (29).

Extracellular bFGF has been extracted from basement membrane (30) and can be freed from the artificial matrices of isolated cells by treatment with heparin-degrading enzymes or by displacement with heparin (31). The concept that enzymatic activity can release bound growth factors from the matrix in vivo, is not new. In bone, osteoclasts activate latent matrix-bound transforming growth factor-β, a process that is blocked in the presence of a protease inhibitor (32). Our demonstration that heparitinase treatment mobilizes bFGF in articular cartilage is compatible with the release of an extracellular pool of bFGF after mechanical injury, by enzymatic cleavage.

Heparitinase is a commercially available, flavobacterium-derived enzyme. It cleaves HS at low-sulphated regions (33), allowing release of highly sulphated sequences upon which FGF is bound. In the past 10 years a family of mammalian heparin-degrading enzymes has been identified (reviewed in ref. 34). The most important of these include connective tissue activating peptide III and HpaI heparanase, a 50-kDa protein first purified from placenta and platelets. HpaI has been cloned and is highly expressed in a number of different tumors where it is thought to play an important role in the size and spread of the cancer by the release of growth and angiogenic factors from the matrix (35). Expression of these enzymes has not yet been investigated in articular cartilage.

FGFs have been extensively studied in the developing skeleton where they play a crucial role in the growth plate. Point mutations in FGF receptor 3, which result in ligand-independent activation of the receptor, cause a number of skeletal abnormalities including achondroplasia and the cranial dysostoses (reviewed in ref. 36). Interestingly, a similar skeletal phenotype is shared by the null perlecan mouse, which has short axial and long bones and disrupted growth plates (37).

Less data are available on the role of bFGF in mature articular cartilage. We have confirmed that isolated articular chondrocytes proliferate in response to bFGF (38), although whether this is the case for chondrocytes within cartilage tissue is unknown. If this is the case, bFGF may be involved in the clonal expansion of chondrocytes that is seen in OA cartilage. It has been reported that bFGF promotes synthesis of sulphated proteoglycans in isolated rabbit costal chondrocytes and it may be important in maintaining their phenotype (39). However, in our experiments bFGF had no effect on proteoglycan synthesis as judged by radiolabeled sulphate incorporation into proteoglycans either in cartilage explants or in isolated chondrocytes. bFGF had no significant effect on synthesis of type II collagen by chondrocytes but strongly induced a number of genes, some of which were identified. The strong induction of both MMPs and TIMP-1 suggests a role in tissue remodeling. Increased expression of both MMPs and TIMP-1 have been observed in human OA (40, 41).

The extracellular matrix of cartilage has always been considered a specialized tissue designed to withstand compressive forces and to absorb shock. Our results suggest that it may act as a sensor of mechanical injury by releasing sequestered bFGF. This response to injury may be important in a tissue such as cartilage where the cell density is low and paracrine signaling is impeded by the distance between cells. The possibility exists that bFGF also plays a role in normal tissue homeostasis, perhaps as a trophic factor or by maintaining matrix turnover. It remains to be answered whether a similar role for bFGF exists in other tissues and whether in cartilage it is involved in the development of OA.


This work was supported by grants from the Wellcome Trust, the Medical Research Council of the U.K., and the Arthritis Research Campaign.


basic fibroblast growth factor
heparan sulphate
extracellularly regulated kinase
cetyl pyridinium chloride
conditioned medium
matrix metalloproteinase
phosphorylated ERK
tissue inhibitor of metalloproteinases 1


This paper was submitted directly (Track II) to the PNAS office.


1. Collins D H. The Pathology of Articular and Spinal Diseases. Baltimore: Williams & Wilkins; 1949.
2. Meachim G. J Bone Jt Surg. 1963;45B:150–161.
3. Campbell C J. Clin Orthop Relat Res. 1969;64:45–63. [PubMed]
4. Seger R, Krebs E G. FASEB J. 1995;9:726–735. [PubMed]
5. Scherle P A, Pratta M A, Feeser W S, Tancula E J, Arner E C. Biochem Biophys Res Commun. 1997;230:573–577. [PubMed]
6. Geng Y, Valbracht J, Lotz M. J Clin Invest. 1996;98:2425–2430. [PMC free article] [PubMed]
7. Skaper S D, Kee W J, Facci L, Macdonald G, Doherty P, Walsh F S. J Neurochem. 2000;75:1520–1527. [PubMed]
8. Finch A, Davis W, Carter W G, Saklatvala J. Biochem J. 2001;353:275–281. [PMC free article] [PubMed]
9. Jeno P, Mini T, Moes S, Hintermann E, Horst M. Anal Biochem. 1995;224:75–82. [PubMed]
10. Wilm M, Shevchenko A, Houthaeve T, Breit S, Schweigerer L, Fotsis T, Mann M. Nature (London) 1996;379:466–469. [PubMed]
11. Shevchenko A, Wilm M, Vorm O, Mann M. Anal Chem. 1996;68:850–858. [PubMed]
12. Wait R, Gianazza E, Eberini I, Sironi L, Dunn M J, Gemeiner M, Miller I. Electrophoresis. 2001;22:3043–3052. [PubMed]
13. Altschul S F, Madden T L, Schaffer A A, Zhang J, Zhang Z, Miller W, Lipman D J. Nucleic Acids Res. 1997;25:3389–3402. [PMC free article] [PubMed]
14. Saklatvala J. Nature (London) 1986;322:547–549. [PubMed]
15. Fernig D G, Gallagher J T. Prog Growth Factor Res. 1994;5:353–377. [PubMed]
16. Mason I J. Cell. 1994;78:547–552. [PubMed]
17. Nugent M A, Iozzo R V. Int J Biochem Cell Biol. 2000;32:115–120. [PubMed]
18. Okada-Ban M, Thiery J P, Jouanneau J. Int J Biochem Cell Biol. 2000;32:263–267. [PubMed]
19. Werner S, Peters K G, Longaker M T, Fuller-Pace F, Banda M J, Williams L T. Proc Natl Acad Sci USA. 1992;89:6896–6900. [PMC free article] [PubMed]
20. Shukla A, Dubey M P, Srivastava R, Srivastava B S. Biochem Biophys Res Commun. 1998;244:434–439. [PubMed]
21. Penc S F, Pomahac B, Winkler T, Dorschner R A, Eriksson E, Herndon M, Gallo R L. J Biol Chem. 1998;273:28116–28121. [PubMed]
22. Cao W, Li F, Steinberg R H, Lavail M M. Exp Eye Res. 2001;72:591–604. [PubMed]
23. Finklestein S P, Apostolides P J, Caday C G, Prosser J, Philips M F, Klagsbrun M. Brain Res. 1988;460:253–259. [PubMed]
24. Delrieu I. FEBS Lett. 2000;468:6–10. [PubMed]
25. Noonan D M, Fulle A, Valente P, Cai S, Horigan E, Sasaki M, Yamada Y, Hassell J R. J Biol Chem. 1991;266:22939–22947. [PubMed]
26. Gallagher J T. Curr Opin Cell Biol. 1989;1:1201–1218. [PubMed]
27. SundarRaj N, Fite D, Ledbetter S, Chakravarti S, Hassell J R. J Cell Sci. 1995;108:2663–2672. [PubMed]
28. Yayon A, Klagsbrun M, Esko J D, Leder P, Ornitz D M. Cell. 1991;64:841–848. [PubMed]
29. Aviezer D, Hecht D, Safran M, Eisinger M, David G, Yayon A. Cell. 1994;79:1005–1013. [PubMed]
30. Folkman J, Klagsbrun M, Sasse J, Wadzinski M, Ingber D, Vlodavsky I. Am J Pathol. 1988;130:393–400. [PMC free article] [PubMed]
31. Bashkin P, Doctrow S, Klagsbrun M, Svahn C M, Folkman J, Vlodavsky I. Biochemistry. 1989;28:1737–1743. [PubMed]
32. Oursler M J. J Bone Miner Res. 1994;9:443–452. [PubMed]
33. Turnbull J E, Fernig D G, Ke Y, Wilkinson M C, Gallagher J T. J Biol Chem. 1992;267:10337–10341. [PubMed]
34. Bame K J. Glycobiology. 2001;11:91R–98R. [PubMed]
35. Vlodavsky I, Friedmann Y, Elkin M, Aingorn H, Atzmon R, Ishai-Michaeli R, Bitan M, Pappo O, Peretz T, Michal I, et al. Nat Med. 1999;5:793–802. [PubMed]
36. Burke D, Wilkes D, Blundell T L, Malcolm S. Trends Biochem Sci. 1998;23:59–62. [PubMed]
37. Costell M, Gustafsson E, Aszodi A, Morgelin M, Bloch W, Hunziker E, Addicks K, Timpl R, Fassler R. J Cell Biol. 1999;147:1109–1122. [PMC free article] [PubMed]
38. Guerne P A, Sublet A, Lotz M. J Cell Physiol. 1994;158:476–484. [PubMed]
39. Kato Y, Gospodarowicz D. J Cell Biol. 1985;100:477–485. [PMC free article] [PubMed]
40. Chevalier X, Conrozier T, Gehrmann M, Claudepierre P, Mathieu P, Unger S, Vignon E. Osteoarthritis Cartilage. 2001;9:300–307. [PubMed]
41. Tetlow L C, Adlam D J, Woolley D E. Arthritis Rheum. 2001;44:585–594. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem Substance links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...