Activation of the ERK1/2 and p38 Mitogen-activated Protein Kinase Pathways Mediates Fibroblast Growth Factor-induced Growth Arrest of Chondrocytes*

Fibroblast growth factors (FGFs) regulate long bone development by affecting the proliferation and differentiation of chondrocytes. FGF treatment inhibits the proliferation of chondrocytes both in vitro and in vivo, but the signaling pathways involved have not been clearly identified. In this report we show that both the MEK-ERK1/2 and p38 MAPK pathways, but not phospholipase Cγ or phosphatidylinositol 3-kinase, play a role in FGF-mediated growth arrest of chondrocytes. Chemical inhibitors of the MEK1/2 or the p38 MAPK pathways applied to rat chondrosarcoma (RCS) chondrocytes significantly prevented FGF-induced growth arrest. The retinoblastoma family members p107 and p130 were previously shown to be essential effectors of FGF-induced growth arrest in chondrocytes. The dephosphorylation of p107, one of the earliest events in RCS growth arrest, was significantly blocked by MEK1/2 inhibitors but not by the p38 MAPK inhibitors, whereas that of p130, which occurs later, was partially prevented both by the MEK and p38 inhibitors. Furthermore, by expressing the nerve growth factor (NGF) receptor, TrkA, and the epidermal growth factor (EGF) receptor, ErbB1, in RCS cells we show that NGF treatment of the transfected cells caused growth inhibition, whereas EGF did not. FGF- and NGF-induced growth inhibition is accompanied by a strong and sustained activation of ERK1/2 and p38 MAPK and a decrease of AKT phosphorylation, whereas EGF induces a much more transient activation of p38 and ERK1/2 and increases AKT phosphorylation. These results indicate that inhibition of chondrocyte proliferation by FGF requires both ERK1/2 and p38 MAPK signaling and also suggest that sustained activation of these pathways is required to achieve growth inhibition.

Fibroblast growth factors (FGFs) regulate long bone development by affecting the proliferation and differentiation of chondrocytes. FGF treatment inhibits the proliferation of chondrocytes both in vitro and in vivo, but the signaling pathways involved have not been clearly identified. In this report we show that both the MEK-ERK1/2 and p38 MAPK pathways, but not phospholipase C␥ or phosphatidylinositol 3-kinase, play a role in FGFmediated growth arrest of chondrocytes. Chemical inhibitors of the MEK1/2 or the p38 MAPK pathways applied to rat chondrosarcoma (RCS) chondrocytes significantly prevented FGF-induced growth arrest. The retinoblastoma family members p107 and p130 were previously shown to be essential effectors of FGF-induced growth arrest in chondrocytes. The dephosphorylation of p107, one of the earliest events in RCS growth arrest, was significantly blocked by MEK1/2 inhibitors but not by the p38 MAPK inhibitors, whereas that of p130, which occurs later, was partially prevented both by the MEK and p38 inhibitors. Furthermore, by expressing the nerve growth factor (NGF) receptor, TrkA, and the epidermal growth factor (EGF) receptor, ErbB1, in RCS cells we show that NGF treatment of the transfected cells caused growth inhibition, whereas EGF did not. FGF-and NGF-induced growth inhibition is accompanied by a strong and sustained activation of ERK1/2 and p38 MAPK and a decrease of AKT phosphorylation, whereas EGF induces a much more transient activation of p38 and ERK1/2 and increases AKT phosphorylation. These results indicate that inhibition of chondrocyte proliferation by FGF requires both ERK1/2 and p38 MAPK signaling and also suggest that sustained activation of these pathways is required to achieve growth inhibition.
Fibroblast growth factors (FGFs) 1 are potent regulators of cell proliferation, migration, and differentiation, and it has become evident that the most important role of FGF signaling is during development, including skeletal formation (1)(2)(3)(4).
Development of most skeletal elements occurs through the multistep process of endochondral ossification, in which a cartilage template is converted into bone. Chondrocytes transit through maturational stages of proliferation, prehypertrophy, and hypertrophy within the epiphyseal growth plate and eventually undergo apoptosis. FGF signaling inhibits endochondral bone growth by regulating the proliferation and differentiation of chondrocytes (4). Gain of function mutations of FGF receptor 3 (FGFR3), which is expressed in growth plate chondrocytes, cause several forms of human dwarfism (4,5), and conversely, mice lacking FGFR3 showed overgrowth of long bones (6,7). In line with this genetic evidence, we showed previously (8,9) that FGF inhibits proliferation of chondrocytes both in vivo and in vitro and that this phenomenon requires STAT1 function. Furthermore, we established that two members of the retinoblastoma (Rb) protein family, p107 and p130, but not pRb, are essential downstream mediators of FGF-induced growth arrest in chondrocytes (10). The function Rb proteins is regulated by phosphorylation (11), and a very early event in the growth arrest of chondrocytes following FGF treatment is the rapid dephosphorylation of p107, which took place even in the absence of new RNA or protein synthesis (12).
FGFs activate a complex network of signaling and transcriptional events in chondrocytes leading to growth arrest (8 -10, 12, 13). Although the studies summarized above identified STAT1 and Rb proteins as essential mediators of FGF-induced chondrocyte growth arrest, the signal transduction pathways that utilize STAT1 or lead to dephosphorylation of Rb proteins have not been clearly identified.
FGF binding to its cognate receptors induces receptor tyrosine kinase activation and transphosphorylation of the receptor dimers. The phosphorylated tyrosines function as binding sites for Src homology domain 2 or phosphotyrosine binding domains of a variety of downstream signaling enzymes and adaptor proteins (14,15). A key component of FGF signaling is the docking protein FRS2, which binds to the FGFRs and recruits several signal transducing molecules leading to the activation of the mitogen-activated protein kinase (MAPK) cascade and the PI3K-AKT antiapoptotic pathway (16 -18).
The MAPKs are focal points for diverse extracellular stimuli and regulate the activities of kinases or transcription factors downstream, thereby influencing gene expression and cellular responses (19). The MAPK family includes the following: 1) extracellular signal-regulated kinases (ERK1 and ERK2); 2) c-Jun N-terminal or stress-activated protein kinases; and 3) the p38 kinases ␣, ␤, ␥, and ␦ (20). In mammalian cells the Ras-Raf-MEK1/2-ERK1/2 cascade is activated by growth factors and is implicated in cell proliferation, differentiation, and survival (21)(22)(23)(24). Unlike the ERK signaling pathway, p38 and JNK are preferentially activated by inflammatory cytokines and cellular stresses, but recent studies have demonstrated that p38 is also activated by several growth factors including FGFs (25)(26)(27). Although many JNK/p38-activating stimuli are proapoptotic, the biological outcome of JNK/p38 activation is highly divergent and appears to be largely dependent on the cell type (28).
In this report we have investigated the role that FGF activation of MAPK signal transduction pathways plays in chondrocyte growth arrest. We show that ERK1/2 and p38 MAPK play a role in the FGF-induced growth arrest of RCS chondrocytes. By introducing exogenous receptors for NGF (TrkA) and EGF (ErbB1) into RCS cells, we could also show that NGF treatment causes growth arrest, whereas EGF does not. Both the FGF-and NGF-induced RCS growth arrest are accompanied by a strong and prolonged activation of ERK1/2 and p38 MAPK, whereas treatment with EGF activates ERKs and p38 only transiently. These findings indicate that both p38 MAPK and ERK signalings are required for the inhibition of the proliferation of chondrocytes induced by FGFs and suggest that the duration and the strength of ERK1/2 and p38 activation play a role in this process.
Cloning and Retroviral Infections-TrkA cDNA was subcloned into pLXSN retroviral expression vector that contains a neomycin resistance gene. The cDNA was cut from pCMV/TrkA expression vector (gift from Dr. M. Chao, New York University), by restriction digestion with EcoRI and cloned into pLXSN generating pLXSN/TrkA. The pLXSN/ErbB1 retroviral expression vector was a gift from Dr. J. Schlessinger (Yale University).
293 gag-pol-expressing cells were used for production of retroviruses. Cells were plated at 60% confluency in 100-mm plates and cotransfected with 5 g of pLXSN/TrkA or pLXSN/ErbB1 plasmids and 5 g of pCMV-VSV-G plasmid by calcium phosphate precipitation (30). The following day the medium was changed, and 48 h later the viral supernatant was collected. About 2 ϫ 10 5 RCS cells were infected with 1 ml of viral supernatant in the presence of 8 g/ml Polybrene for 5 h. The cells were selected with geneticin G418 (600 g/ml) (Sigma) for 2 weeks. Protein expression of TrkA or ErbB1 was further analyzed. Pools of selected cultures were used in the studies.
Proliferation Assay-2 ϫ 10 4 cells were seeded onto glass coverslips in 24-well plates (Costar) for 48 h in DMEM containing 10% FCS. Recombinant human FGF1 and heparin (Sigma, 10 g/ml), NGF, or EGF were added for 20 h at the indicated concentrations, and the cells were labeled with 1 g/ml bromodeoxyuridine (BrdUrd; Sigma) for 4 h at 37°C. The cells were fixed in 3.7% paraformaldehyde for 10 min at room temperature, permeabilized for 10 min with 0.5% Triton in phosphate-buffered saline, and treated for 15 min with 1.5 N HCl. After several washes, analysis of BrdUrd incorporation was performed using an anti-BrdUrd monoclonal antibody (Amersham Biosciences) followed by anti-mouse secondary antibody conjugated with Cy3 (1:400). Cell nuclei were stained with a solution of 0.5 g/ml of Hoechst (Sigma) in phosphate-buffered saline for 20 min prior to mounting on slides. The fluorescence was visualized using an Axioplan 2 Zeiss microscope equipped with a digital camera. The frequency of S phase cells was calculated as a ratio of BrdUrd-positive nuclei to the total Hoechststained nuclei.
Immunoprecipitation and Western Blot Analysis-500 g of protein extract was immunoprecipitated with specific antibodies and 40 l of Protein A Plus-agarose (Oncogene Science) over night at 4°C. The immunocomplexes were washed 3 times in lysis buffer and resuspended in Laemmli sample buffer. Immunoprecipitates or total protein extracts (20 -30 g) were boiled separately by SDS-PAGE and electrotransferred on nitrocellulose membranes (Protran®). The membranes were blocked in TBST (10 mM Tris, pH 7.4, 0.5 mM NaCl, 0.1% Tween 20) containing 3% bovine serum albumin (Sigma) or in TBST containing 5% dried milk. The blots were first probed with the indicated primary antibodies in 3% bovine serum albumin/TBST and then with horseradish peroxidaseconjugated secondary antibodies in TBST. Proteins were visualized by ECL detection system (Amersham Biosciences).
For proliferation assay and biochemical assay, the cells were pretreated with MEK1/2 inhibitors, PD98059 and U0126, and p38 MAPK inhibitors, SB202190 and SB203580, and the PLC␥ inhibitor, U73122, for 1 h before addition of growth factors. All the compounds were dissolved in dimethyl sulfoxide (Me 2 SO).
p38 MAPK Assay-The p38 kinase assay was performed according to the manufacturer's instructions (Cell Signaling). Treated cells were lysed (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerol phosphate, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, 1 mM sodium orthovanadate), and 20 l of immobilized agarose phospho-p38 MAPK antibody was incubated with 400 g of total protein lysate overnight at 4°C. The immunocomplexes were washed twice with the lysis buffer and once with the kinase buffer (25 mM Tris, pH 7.5, 5 mM ␤-glycerol phosphate, 2 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate). Immunoprecipitates were suspended in 50 l of kinase buffer supplemented with 200 M ATP and 2 g of ATF-2 fusion protein and incubated for 30 min at 30°C. The reaction was stopped with Laemmli sample buffer, and the samples were boiled for 5 min, separated by SDS-PAGE, transferred to a nitrocellulose membrane, and immunoblotted with an antibody anti-phospho-ATF2 (Thr-71).

RESULTS
For our experiments we used the RCS cell line that displays numerous chondrocytic characteristics, including expression of type II, type IX, and type XI collagen and Alcian blue-stainable cartilage-specific proteoglycans (29). We treated RCS cells with FGF1, a high affinity ligand for all known FGFR isoforms. In preliminary studies, we confirmed that RCS cells expressed FGFR3 and FGFR2, which are phosphorylated at tyrosine residues upon treatment with FGF1 (not shown). As reported previously (8,10,13), FGF1 treatment of RCS resulted in a drastic and dose-dependent growth inhibition, reflected in the reduction of the proportion of cells incorporating BrdUrd. Inhibition of DNA synthesis was rapid and reached a maximum by 16 -20 h after treatment.
FGF Stimulation of RCS Cells Leads to the Activation of ERK1/2, p38 MAPK, and PLC␥ but Not JNK and AKT-To determine which of the signal transduction pathways activated by FGFR in RCS cells may contribute to the growth-inhibitory response, we studied the activation of known FGF signaling pathways, including those of the MAPK family, PI3K and PLC␥, in FGF-treated RCS cells.
FGF1 (100 ng/ml) was added to RCS cells for different times, and cellular lysates were analyzed. Fig. 1A shows that FGF1 stimulation led to a strong phosphorylation of ERK1 and ERK2 that was detectable at 5 min and persisted up to 1 h. p38 MAPK was also phosphorylated upon FGF1 stimulation, whereas no activation of JNK1 (p46) and JNK2 (p54) was detected. To verify that the phosphorylation of p38 MAPK led to its activation, an in vitro kinase assay was performed. The transcription factor ATF-2 is a direct substrate for p38 MAPK in vitro and in vivo; therefore, its phosphorylation can serve as a measure of p38 activity (31). Fig. 1C shows that FGF1 is able to induce p38-mediated phosphorylation of ATF-2, which peaked at 15 min and decreased to basal levels in 3 h.
The activation of PI3K induced by FGF1 in RCS cells was determined by analyzing the phosphorylation of AKT, a downstream effector of PI3K signaling (32). Immunoblotting using an antibody against the phosphorylation site (Ser-473) in AKT shows that FGF1 did not activate AKT (Fig. 1A). Untreated serum-starved cells exhibit a basal level of AKT phosphorylation that was slightly reduced after 15 min of FGF1 treatment. This result is in contrast with the finding that FGF strongly induces phosphorylation of AKT in fibroblasts (33).
PLC␥ is phosphorylated on tyrosine residues following binding to activated FGFRs (34). PLC␥ was immunoprecipitated from cellular lysates using an anti-PLC␥ antibody followed by a Western blot using an anti-phosphotyrosine antibody. As shown in Fig. 1B, FGF1 strongly activated PLC␥ after 5 min of treatment. Thus, these results show that FGF1 stimulation of RCS leads to the activation of ERK1/2, p38 MAPK, and PLC␥ but not JNK1/2 and to a decrease of the basal levels of phosphorylated AKT.
MEK1/2 and p38 MAPK Inhibitors Block FGF-mediated Growth Arrest of RCS Cells-We used specific chemical inhibitors for ERK1/2, p38 MAPK, and PLC␥ to determine their effect on FGF-induced growth arrest. The compounds used were the MEK1/2 inhibitors, U0126, PD98059, SB202190, and SB203580, which inhibit p38␣ and p38␤ MAPK, and the PLC␥ inhibitor U73122. The MEK1/2 inhibitors selectively prevent the activation of MEK1/2 by Raf with a subsequent blocking of the ERK1/2 cascade. U0126 was shown to be able to inhibit both MEK1 and MEK2, whereas PD98095 inhibits MEK1 more potently than MEK2. SB202190 and SB203580 are able to block the kinase activity of p38␣ and p38␤ MAP kinases, and U73122 has been shown to inhibit PLC␥-associated phosphatidylinositol 4,5-bisphosphate hydrolysis (35)(36)(37).
The effect of MEK1/2 inhibitors on ERK1/2 activation and on RCS proliferation was tested. Treatment of RCS cells with PD98059 and U0126 produced a significant reduction in FGF1induced ERK1/2 activation. A complete block of ERK1/2 phosphorylation was seen using 100 M PD98059 and 20 M U0126 ( Fig. 3A and data not shown). These concentrations of the inhibitors were able to partially counteract the FGF-mediated growth arrest of RCS cells. BrdU incorporation in the presence of the inhibitors shows that the effect was detectable at concentrations of FGF1 ranging from 0.5 to 10 ng/ml ( Fig. 2A).
In order to determine whether p38␣ and p38␤ MAPKs also played a role in the response to FGF, we used two different concentrations of SB202190 and SB203580 (10 and 30 M), which inhibited p38 by the effect on its target MAPKAP kinase-2 activity (data not shown) (38,39). Both inhibitors were very effective at blocking the FGF1-mediated growth arrest with a maximal effect occurring at 30 M (Fig. 2B). On the other hand, inhibition of PLC␥ activity in the FGF-treated RCS cells using U73122, at concentrations of 1 and 5 M, did not reverse the FGF-induced growth arrest (Fig. 2C). Inhibition of PI3K activity by wortmannin also had no effect on FGF-mediated growth arrest (data not shown). Thus, these results indicate that p38 MAPK and ERK1/2 signaling play a role in FGFmediated growth arrest of RCS cells.
The MEK1/2 Inhibitors Affect p38 MAPK Activity-Recent studies (40) have shown that cross-talk between signaling cascades can modulate the response to growth factors. To determine whether inhibition of MEK1/2 could affect the activity of p38 and vice versa, we studied the phosphorylation pattern of ERK1/2, and the enzymatic activity of p38 MAPK in the absence and presence of a MEK1/2 inhibitor (U0126) and a p38 MAPK inhibitor (SB202190). As expected, FGF-induced ERK1/2 phosphorylation was blocked by 20 M U0126 but not by 20 M SB202190, indicating that p38 MAPK is not required for ERK1/2 activation (Fig. 3A). We determined the effect of U0126 on p38 activity by an in vitro kinase assay using ATF-2 as substrate. Fig. 3B shows that the p38 activity induced by FIG. 1. FGF signal transduction in RCS cells. RCS cells were serum-starved in DMEM overnight and treated with 100 ng/ml FGF1 for the indicated times before cellular extracts were prepared. A, 30 g of total lysates was analyzed by SDS-PAGE and immunoblotting. Specific antibodies against phospho-ERK1/2, phospho-p38 MAPK, phospho-JNK1/2, and phospho-AKT were used. To ensure equal loading of total cellular extracts, the membranes were reprobed with antibodies detecting total ERK2, p38 MAPK, JNK1, and AKT as indicated. B, 700 g of cellular extract was subjected to immunoprecipitation (IP) with anti-PLC␥ antibody and blotted with anti-phosphotyrosine (4G10) and/or PLC␥ antibodies. C, in vitro p38 MAPK assay. Phospho-p38 MAPK was immunoprecipitated from the cellular extracts, and the kinase activity was assayed using ATF-2 as substrate. Active ATF-2 was detected with an antiphospho-Thr-71 ATF-2 antibody. FGF1 was reduced in the presence of the MEK1/2 inhibitor, indicating that MEK1/2-ERK1/2 signaling can contribute to p38 MAPK activation or that the U0126 compound is not completely specific. Similar results were obtained with the PD98059 MEK1/2 inhibitor (data not shown). This result suggests that by using our experimental conditions, the ability of MEK1/2 inhibitors to prevent the FGF-induced growth arrest may depend in part on their effect on p38 MAPK activity.
The ERK1/2 and p38 MAPK Pathways Are Involved in p107 and p130 Dephosphorylation-Laplantine et al. (10) have shown that FGF treatment of RCS cells causes dephosphorylation of Rb family proteins and that p107 and p130 function is required for FGF-induced growth inhibition of chondrocytes in vitro and in vivo. Further data (12) indicated that dephosphorylation of p107 and p130 is generated by different mechanisms and that p107 dephosphorylation is one of the earliest events in FGF-induced growth arrest of chondrocytes. We have studied the effect of the MEK1/2 and p38 MAPK inhibitors on the dephosphorylation of p107 and p130 induced by FGF in RCS cells. As shown in Fig. 4A, RCS cells contain a mixture of hyper-and hypophosphorylated p107 molecules that upon FGF treatment are rapidly converted to the hypophosphorylated form. This event is significantly prevented by the U0126 MEK1/2 inhibitor but not by the p38 inhibitor SB202190. The effect of the MEK1/2 inhibitor is clearly detectable at 1 h after FGF addition and persists up to 20 h. On the other hand, growing RCS cells contain predominantly hyperphosphorylated forms of p130, and p130 dephosphorylation is clearly detectable only after several hours of treatment with FGF1 (Fig. 4B). Treatment of RCS cells with either the MEK1/2 or the p38 MAPK inhibitors caused a partial but clear reduction of the extent of p130 dephosphorylation. Thus these results indicate that the ERK1/2 and p38 MAPK pathways affect FGFinduced p107 and p130 dephosphorylation in a distinct way. ERK1/2 signaling is required for p107 dephosphorylation, whereas p130 dephosphorylation appears to be dependent on the activation of the ERK1/2 as well as the p38 MAPK pathways.
Expression of TrkA and ErbB1 in RCS Cells-To better understand the role of p38 MAPK and ERK1/2 in the response to FGF of RCS cells, we created RCS cells expressing the NGF receptor TrkA or the EGF receptor ErbB1, two tyrosine kinase receptors that are known to activate both pathways. Signaling through these two receptors has been shown to elicit different responses in PC12 cells where EGF stimulation induces cellular proliferation, whereas NGF and FGF stimulation results in differentiation into a sympathetic neuron-like phenotype (41)(42)(43). RCS cells, which do not express either of these receptors and do not respond to NGF or EGF, were infected with a retrovirus carrying the cDNA encoding for the rat TrkA or human ErbB1. After selection, the expression and activation of each of the two receptors were analyzed by immunoprecipitation and Western blot. TrkA is synthesized as a 110-kDa Nglycosylated precursor, which is converted into the mature 140-kDa isoform of the receptor that is then translocated to the cell surface (44). Fig. 5A shows that RCS/TrkA cells express the two isoforms of TrkA, gp140 and gp110, but only gp140 is activated upon NGF stimulation, as shown by its ability to be recognized by an anti-phosphotyrosine antibody. Fig. 5B shows the expression level of ErbB1 (gp170) protein in RCS/ErbB1 and its activation, detected with anti-phosphotyrosine antibody, upon EGF stimulation.

NGF Induces Growth Inhibition in RCS/TrkA Cells and Causes Sustained ERK1/2 and p38 MAPK Activation-NGF
signaling has been extensively studied in PC12 cells where it causes cell differentiation and growth arrest (41,43). However, it has been also shown that NGF is mitogenic and transforming in NIH3T3 fibroblasts expressing TrkA (45). In order to determine the effect of NGF on RCS/TrkA proliferation, a BrdU incorporation assay was performed. Fig. 5C shows that NGF was able to induce a dose-dependent inhibition of cellular proliferation, although the degree of inhibition was not as strong as observed with FGF1. FACScan TM analysis reveals that after 20 h of treatment, NGF (10 ng/ml) also causes an accumulation of RCS/TrkA cells in the G 0 /G 1 phase of cell cycle (not shown). We analyzed the signal transduction pathways activated by NGF through TrkA activation in RCS/TrkA cells that may contribute to the growth-inhibitory response. Fig. 6A shows that NGF stimulation led to phosphorylation of ERK1 and ERK2 that peaked within 5 min and remained high for at least 1 h. No phosphorylation of JNK1 (p46) and JNK2 (p54) was detected. Moreover, NGF did not increase the levels of AKT phosphorylation (Fig. 6A), similarly to what was observed with FGF treatment. To determine p38 MAPK activity, the in vitro p38 kinase assay was performed using ATF-2 as substrate. As shown in Fig. 6B NGF induced p38 MAPK activity, which peaked at 15 min. The inhibitors U0126 and SB202190 were used to determine whether ERK1/2 and p38 MAPK played a role in NGF-mediated growth arrest in RCS/TrkA. Both the MEK1/2 and p38 MAPK inhibitors were able to counteract the NGF-mediated growth inhibition, with SB202190 having the stronger effect (Fig. 6C). These results indicate that NGF, as FGF, requires the ERK1/2 and p38 MAPK pathways to mediate its inhibitory effect on RCS cell proliferation.

EGF Does Not Inhibit RCS/ErbB1 Cells Proliferation and Transiently Activates ERK1/2, p38 MAPK, and AKT-The
ErbB1 receptor is widely expressed in epithelial, mesenchymal, and neuronal tissues where it plays a pivotal role in regulating cell proliferation, survival, and differentiation. Overexpression of wild-type ErbB1 or expression of constitutively activated receptor mutants transforms fibroblasts in vitro (46,47). The response to EGF of RCS/ErbB1 was determined by examining cells incorporating BrdU in response to EGF. Fig. 5D shows that, unlike FGF1, EGF does not cause inhibition of cellular proliferation over a wide range of concentrations (0.5-20 ng/ ml). To compare tyrosine kinase receptor signaling in an inhibitory versus non-inhibitory response, the signal transduction pathways activated by EGF in RCS/ErbB1 cells were investigated. Interestingly, the addition of EGF (100 ng/ml) leads to a transient ERK1/2 and p38 MAPK activation compared with FGF1 and NGF. Although a strong ERK1/2 phosphorylation was detected after 1 min of stimulation, it declined to lower levels after 15 min (Fig. 7A). p38 MAPK activity reached its maximum in 15 min and declined to basal levels within 1 h (Fig. 7B). EGF was unable to activate JNK1/2 in RCS/ErB1 cells but caused AKT phosphorylation (Fig. 7A). Together these results show that unlike FGF1 and NGF, EGF signaling does not inhibit RCS proliferation and only transiently activates ERK1/2, p38 MAPK, while increasing phosphorylation of AKT.
Time Course of ERK1/2 and p38 MAPK Activation-It has been shown (48) that both the magnitude and duration of ERK1/2 activation can determine the nature of the cellular response to growth factors. In PC12 cells, EGF transiently stimulates ERK1/2 and p38 MAPK and induces cellular proliferation. In contrast, NGF and FGF cause sustained activation of both pathways and neuronal differentiation (43,49). To compare the kinetics and magnitude of ERK1/2 activation in response to FGF1, NGF, and EGF stimulation of RCS/TrkA and RCS/ErbB1 cells, ERK1/2 phosphorylation was monitored at various time points up to 5 h. As shown in Fig. 8, FGF1 induced a strong and prolonged ERK1/2 activation that peaked in 5 min and persisted up to 5 h. NGF stimulation of RCS/TrkA also led to a strong and prolonged ERK1/2 activation that peaked in 5 min and was still detectable after 5 h of stimulation. However, EGF stimulation of RCS/ErbB1 led to a very rapid but transient phosphorylation of ERK1/2 that peaked at 1 min and decreased to the basal level by 3 h.
As shown in Figs. 1C, 6B, and 7B, the activity of p38 MAPK upon FGF1 and NGF stimulation is stronger, reaches its max- Parental RCS cells as well as cells stably expressing the TrkA (RCS/TrkA) or ErbB1 (RCS/ErbB1) receptors were serum-starved overnight and treated with 100 ng/ml NGF or 100 ng/ml EGF for 5 min, respectively. The expression of the receptors was tested by immunoprecipitation (IP) with anti-TrkA or anti-ErbB1 antibody followed by immunoblotting as indicated. The activation of receptors was detected using the anti-phosphotyrosine antibody 4G10 (P-Tyr). C and D, NGF inhibits proliferation of RCS/TrkA cells, whereas EGF does not inhibit proliferation of RCS/ErbB1. Cells were seeded on coverslips at a density of 2 ϫ 10 4 cells/well (24-well plates) in DMEM containing 10% FCS. After 24 h, the medium was replaced, and the cells were treated with indicated concentrations of NGF or EGF. Cells were labeled with BrdU (5 g/ml) for the last 4 h, fixed, and processed for immunofluorescent detection of BrdU.
imum at 15 min, and is still high after 1 h, while upon EGF stimulation, p38 activity reaches its maximum at 15 min but then decreases to basal levels. Thus a clear correlation between the magnitude and particularly the duration of ERK1/2 and p38 MAPK activation and growth arrest can be detected in RCS cells upon stimulation of the NGF, EGF, or of the endogenous FGF receptors. NGF and FGF receptor stimulation leads to sustained activation of p38 MAPK and ERK1/2 and to growth arrest, although stimulation of the EGF receptor does not cause growth arrest and is accompanied by a significant but shortlived activation of p38 MAPK and ERK1/2.

DISCUSSION
The rate of proliferation and differentiation of chondrocytes determines the overall growth in the length and shape of long bones. Evidence from human and mouse genetics indicates that longitudinal bone growth is negatively regulated by FGF sig-naling (4), and accordingly, FGF signaling inhibits chondrocytes proliferation in vitro and in vivo (8), an unusual response for FGFs that usually act as mitogens in other cell types. Although genetic evidence clearly identified STAT1 as an essential mediator of the FGF-induced growth arrest of chondrocytes both in vitro and in vivo (8,9), the mechanism by which STAT1 participates in this process is still unclear and may not require that it function as a transcriptional regulator. Indeed data from our laboratory 2 show that interferon ␥, a potent inducer of STAT1 transcriptional activation, does not induce 2 A. Raucci, E. Laplantine, and C. Basilico, unpublished data.
FIG. 6. ERK1/2 and p38 MAPK are involved in NGF-induced growth inhibition of RCS/TrkA cells. A, RCS/TrkA cells were serum-starved in DMEM overnight and treated with 100 ng/ml NGF for the indicated times, and cellular extracts were prepared. 30 g of total lysates was analyzed by SDS-PAGE and immunoblotting. Specific antibodies against phospho-ERK1/2, phospho-JNK1/2, and phospho-AKT were used. To ensure equal loading of total cellular extracts, the membranes were reprobed with antibodies detecting total ERK2, JNK1, and AKT as indicated. B, in vitro p38 MAPK assay. C, cells were seeded on coverslips at a density of 2 ϫ 10 4 cells/well (24-well plates) in DMEM containing 10% FCS. After 24 h, the medium was replaced, and the cells were treated with the indicated concentrations of MEK1/2 inhibitor (U0126) or p38 MAPK inhibitor (SB202190). After 1 h NGF (10 ng/ml) was added to the medium for 20 h. Cells were labeled with BrdU (5 g/ml) for the last 4 h, fixed, and processed for immunofluorescent detection of BrdU. IP, immunoprecipitation.

FIG. 7. EGF signal transduction in RCS/ErbB1 cells.
Cells were serum-starved in DMEM overnight and treated with 100 ng/ml EGF for the indicated times, and cellular extracts were prepared. A, 30 g of total lysates was analyzed by SDS-PAGE and immunoblotting. Specific antibodies against phospho-ERK1/2, phospho-JNK1/2, and phospho-AKT were used. To ensure equal loading of total cellular extracts, the membranes were reprobed with antibodies detecting total ERK2, JNK1, and AKT as indicated. B, in vitro p38 MAPK assay. IP, immunoprecipitation.
FIG. 8. Kinetics of ERK1/2 activation. Cells were serum-starved in DMEM overnight and treated with 100 ng/ml FGF1, NGF, or EGF for the indicated times; cellular extracts were prepared, and 30 g of total lysate was analyzed by SDS-PAGE and immunoblotting with a specific antibody for phospho-ERK1/2. growth arrest in RCS cells, indicating that either STAT1 function is necessary but not sufficient to promote growth arrest or that its phosphorylation is not required. In this report we show that two components of the MAPK family, ERK1/2 and p38 MAPK, also play an important role in FGF-mediated growth arrest in chondrocytes.
Blocking the ERK or p38 MAPK Pathways Prevents the Growth Arrest Induced by FGF in Chondrocytes-We have used the RCS cell line to show that FGF causes a strong and prolonged activation of the ERK1/2 and p38 MAPK pathways in chondrocytes. By exposing RCS cells to specific inhibitors of MEK-ERK1/2 or p38 signaling, we showed that inhibition of the function or the activation of these enzymes partially or completely prevented FGF-induced growth arrest, in that cells continued to proliferate and enter the S phase similarly to untreated controls. On the other hand, inhibition of PLC␥ activity had no significant effect. Thus PLC␥ does not appear to play an important role in FGF-mediated growth inhibition, similar to what had been shown previously (50) for FGF induction of proliferation in fibroblasts.
Interestingly, inhibition of ERK1/2 or p38 MAPK differentially affects the FGF-mediated growth arrest of RCS. Whereas SB202190 and SB203580 (p38 MAPK inhibitors) almost completely restore cell proliferation, PD98059 and U0126 (MEK1/2 inhibitors) cause a more partial block of the FGF effect. We have also shown that the p38 MAPK inhibitor, SB202190, does not influence the activation of ERK1/2 by FGF1, whereas the MEK1/2 inhibitor, U0126, partially blocks the p38 MAPK activity (detected by phosphorylation of the substrate ATF-2).
Together with the observation that the ability of p38 MAPK inhibitors to prevent the FGF-induced growth arrest is stronger than that of the ERK inhibitors, this result could be interpreted to suggest that the effect of the ERK inhibitors depends at least in part on the inhibition of the p38 MAPK pathway. Although U0126 has generally been considered specific for MEK1/2, it is possible that at the concentrations used, it could also directly affect p38 MAPK activity or activation (36,37). Alternatively, it is possible that the MEK1/2-ERK1/2 pathway contributes to p38 MAPK activation, as suggested by Morooka and Nishida (49). We have found that the phosphorylation of p38 MAPK induced by FGF in RCS cells is partially inhibited in the presence of the U0126 compound (data not shown). We have attempted to investigate this point further by using dominant-negative forms of MEK1 and ERK2, but we found that in our system the inhibition of ERK signaling obtained by the expression of these mutants was very slight, and thus the results obtained are not conclusive. However, irrespective of the mechanisms by which the MEK1/2 inhibitor may affect p38 MAPK activity, the data on p107 and p130 dephosphorylation indicate that these two pathways produce signals differentially affecting the ability of FGFs to induce dephosphorylation of Rb proteins. ERK1/2 activity appears to be important for p107 dephosphorylation, which is not affected by the p38 MAPK inhibitors, whereas p130 dephosphorylation is reduced by inhibition of both pathways. Other FGF responses are also differentially sensitive to inhibition of p38 MAPK or ERK1/2 signaling. We have shown previously that FGF treatment of RCS induces the expression of several differentiation-related genes, such as osteopontin, MMP13, and TIMP1 (12). This induction is completely blocked by the MEK1/2 inhibitors and not by the p38 inhibitors. 3 Thus, although the effect of the MEK1/2 inhibitors could depend in part on inhibition of p38 MAPK activity, this cannot be the sole mechanism by which these compounds prevent many FGF responses in RCS cells. ERK1/2 are usually activated by growth factors and are implicated in cell proliferation, differentiation, and survival (22)(23)(24), whereas p38 MAPK is preferentially activated by inflammatory cytokines, cellular stresses, withdrawal of growth factors, and proapoptotic stimuli (25). It is interesting to note that two recent reports (51,52) on the growth arrest caused by stress responses have detected a rapid dephosphorylation of p107 in response to UV or oxidative stress stimuli, which are known to cause strong activation of p38 MAPK (53). In our case, however, the rapid dephosphorylation of p107 induced by FGF in chondrocytes appears to depend on the activation of the ERK pathway, rather than on p38 MAPK. We had suggested previously that the rapid dephosphorylation of p107 induced by FGF in RCS cells, which occurs in the absence of new protein synthesis, could be due to "activation" of a specific phosphatase (10,12,13). In view of the results presented in this report, it is likely that activation of this phosphatase, which may be due to phosphorylation, requires the direct or indirect action of ERK1/2 but that the p38 MAPK may also contribute to its activation.
In contrast to p107, p130 dephosphorylation is a late event that we had defined as a characteristic of the "maintenance" of growth arrest and probably results from inactivation of cyclindependent kinases (12). p130 dephosphorylation is partially blocked by both types of inhibitors. ERK1/2 and p38 MAPK can regulate the expression and the stability of cyclin-dependent kinase inhibitors such as p21 (54 -57), which is induced by FGF in RCS cells. A block in p21 induction could affect p130 dephosphorylation. Together with previous experiments aimed at identifying the key molecules and events that determine the growth-inhibitory response of chondrocytes to FGF signaling, the data presented here highlight the concept that in these cells FGF induces a complex network of signaling and transcriptional events, each of which is necessary, but not sufficient, to establish a growth-inhibitory response.
The Duration of MAP Kinase Activation Correlates with the Biological Response of RCS Cells to Growth Factors-We created new RCS cell lines expressing two different tyrosine kinase receptors, TrkA and ErbB1, the receptors for NGF and EGF, respectively. Stimulation with NGF of RCS cells expressing TrkA induces growth inhibition and G 1 arrest, whereas stimulation of RCS cells expressing ErbB1 with EGF does not. In both cases the receptors are activated upon ligand binding. Notably, like FGF, NGF induces sustained activation of ERK1/2 and p38 MAPK, and ERK1/2 and p38 MAP kinase inhibitors prevent the NGF-mediated growth arrest. In contrast, EGF induces a transient activation of ERK1/2 and p38 MAPK. These findings are in accordance with numerous reports indicating that the duration of MAPKs signal may hold the key to very different cellular outcomes.
In PC12 cells, NGF and FGF induce sustained activation of MAPKs, which is required for cell cycle arrest and terminal differentiation, whereas EGF, which only elicits a transient activation of MAPKs, does not promote differentiation but induces proliferation (43,49). In addition, overexpression of EGFR in PC12 cells leads to a prolonged MAPK response resulting in cell differentiation (58). Sustained activation of MAPK has been associated with nuclear accumulation of these enzymes (58) where it may modulate gene expression via the phosphorylation of transcription factors, so that the quantitative differences in the duration of MAPK activation may be reflected in qualitative changes in gene expression. For example, the activation of MAPKs regulates AP-1 activity through the phosphorylation of distinct substrates (59, 60). The AP-1 complexes, which include the c-Jun and c-Fos family, regulate the transcription of a number of genes, and depending on the dimer combination they can be either positive or negative regulators of gene expression (61). It has been demonstrated that Fra-1, Fra-2, c-Jun, and JunB are substrates of sustained ERK1/2 activity (62), and we have shown previously that FGF treatment causes a strong up-regulation of the expression of Fra-1, c-Jun, and JunB in RCS cells (12). Thus different modifications of the activity of proteins that make up the AP1 complex due to sustained or transient MAPK activation could result in different programs of gene expression that play a role in directing the response of RCS cells toward growth-inhibitory or growth-stimulating pathways.
The mechanisms by which these different tyrosine kinase receptors induce distinct patterns of MAPK activation and distinct biological responses in the same cell type are still unclear. Considerable attention has been focused on the signaling molecules linking the receptors to the ERK1/2 pathway. The ErbB1, TrkA, and FGF receptors all recruit a variety of signaling molecules upon growth factor stimulation. Most of these effectors are shared by the three receptors including PLC␥, Shp-2, Shc, and Grb2 (15), but TrkA and FGFRs associate with an additional signaling molecule, FRS2 (16,63). Activated FRS2 has been shown to form a multiprotein complex with Shp2 and Grb2, which has been postulated to play an important role in the sustained activation of ERK1/2 elicited by NGF and FGF in PC12 cells (16,17).
The PI3K/AKT Pathway-In contrast with finding that FGF protects several types of cells from apoptosis by activating the anti-apoptotic AKT pathway, FGF promotes apoptosis in chondrocytes (9). We show here that FGF1 does not activate AKT in RCS cells, and the basal level of AKT phosphorylation decreases after FGF stimulation. Similarly, NGF does not induce AKT phosphorylation in RCS cells expressing TrkA, whereas EGF stimulates AKT phosphorylation in cells expressing its cognate receptor. Although RCS cells do not show increased apoptosis in response to FGF, a phenomenon that instead can be observed in primary chondrocytes (9), it is possible that the down-regulation of AKT activity reflects the proapoptotic function of FGF in these cells and that the actual apoptotic response is blocked further downstream in these tumor cells. It has recently been reported that AKT negatively regulates the Ras/Raf/MEK/ERK pathways via phosphorylation of Raf at Ser-259 (64). In vascular smooth muscle cells, platelet-derived growth factor induced proliferation and a sustained activation of PI3K/AKT, but only a transient activation of Ras/Raf/MEK/ ERK, because activated AKT was able to phosphorylate and block Raf1 kinase activity (65). In addition, the existence of cross-talk between p38 MAPK and PI3K/AKT pathways has been shown by Gratton et al. (66), who showed that blockade of AKT signaling increases p38 activity in endothelial cells, enhancing apoptosis.
Our results show that in RCS cells there is an evident correlation between sustained activation of ERK1/2 and p38 MAPK, lack of activation of AKT, and the induction of growth arrest. FGF1 stimulation induces a strong activation of ERK1/2 and p38, dephosphorylation of AKT, and growth arrest, whereas EGF stimulation activates both pathways and does not induce growth arrest. In accordance with our results, it has been shown that PC12 cells overexpressing wild-type AKT are not susceptible to the growth-arresting effect of NGF (67). Thus the inability of FGF to induce AKT activation in RCS cells may contribute to sustained MAPK activation and growth arrest.
The inhibition of proliferation that FGFs cause in chondrocytes is a complex event that requires the intervention of several proteins including STAT1, p21, p107, and p130 (8 -10, 12, 13). In this report we show that among the signaling pathways downstream of the FGFRs, ERK1/2 and p38 MAPK play a role in the inhibition of the proliferation of chondrocytes induced by FGF and that the intensity and the duration of the activation of the two pathways may be necessary for the growth-inhibitory response. Microarray experiments ongoing in our laboratory should shed light on the complex interplay between the genes and their products that are responsible for growth arrest and initiation of differentiation induced by FGF in chondrocytes, and on how the expression or the activity of these genes is regulated by the MAPKs cascades.