Integrins α6Aβ1 and α6Bβ1 Promote Different Stages of Chondrogenic Cell Differentiation*

The differentiation of chondrocytes and of several other cell types is associated with a switch from the α6B to the α6A isoform of the laminin α6β1 integrin receptor. To define whether this event plays a functional role in cell differentiation, we used anin vitro model system that allows chick chondrogenic cells to remain undifferentiated when cultured in monolayer and to differentiate into chondrocytes when grown in suspension culture. We report that: (i) upon over-expression of the human α6B, adherent chondrogenic cells differentiate to stage I chondrocytes (i.e. increased type II collagen, reduced type I collagen, fibronectin, α5β1 and growth rate, loss of fibroblast morphology); (ii) the expression of type II collagen requires the activation of p38 MAP kinase; (iii) the over-expression of α6A induces an incomplete differentiation to stage I chondrocytes, whereas no differentiation was observed in α5 and mock-transfected control cells; (iv) a prevalence of the α6A subunit is necessary to stabilize the differentiated phenotype when cells are transferred to suspension culture. Altogether, these results indicate a functional role for the α6B to α6A switch in chondrocyte differentiation; the former promotes chondrocyte differentiation, and the latter is necessary in stabilizing the differentiated phenotype.

The onset of chondrogenesis in developing long bones is characterized by the reduction of intercellular spaces and establishment of extensive cell-cell contacts between mesenchymal chondrogenic cells, i.e. cell condensation (21). Several factors have been shown to play a role in this process, including cell-cell interactions (22)(23)(24)(25)(26)(27), composition of the ECM (25, 28 -30), changes in cell shape (31), and response to cytokines (32). Following cell condensation, chondrogenic cells that produce type I collagen, fibronectin (FN), and its integrin receptor ␣ 5 ␤ 1 , differentiate to stage I chondrocytes that express type II collagen and eventually to stage II hypertrophic chondrocytes, characterized by type II and X collagen production (12).
Most of these events can be reproduced in vitro in a tissue culture model system that allows condensation and differentiation of chick embryo tibiae chondrogenic cells (12,(33)(34)(35). These cells, which adhere to tissue culture dishes and display a fibroblast-like phenotype (pre-chondrogenic cells), proliferate and secrete type I collagen and FN. When transferred to suspension culture, the cells rapidly aggregate and secrete type II collagen and other stage I chondrocyte-specific ECM molecules (12). In addition, differentiation is associated with: (i) downregulation of type I collagen, FN, and its receptor, ␣ 5 ␤ 1 integrin; (ii) switch from the ␣ 6B to the ␣ 6A isoform of the integrin ␣ 6 ␤ 1 receptor for laminin (LN); (iii) reduced growth rate; and (iv) acquisition of a cobblestone morphology when re-plated in monolayer culture (12,25). In ϳ3 weeks of culture, cell clusters are completely separated into single hypertrophic stage II chondrocytes that express type X collagen (12,36).
A switch from ␣ 6B to ␣ 6A has been associated with the progression of cell differentiation of several cell types and tissues (13,(37)(38)(39). The two alternatively spliced isoforms of ␣ 6 differ in their cytoplasmic domain (40 -43) and, upon binding to LN, induce different levels of phosphorylation of the signaling proteins paxillin and MAP kinases (44,45). Furthermore, in cultured myoblasts, the ectopic expression of ␣ 6A ␤ 1 (19,20) modulates functions that are associated with differentiation and mitogenic responses, via MAP kinases.
The family of MAP kinases comprises several subtypes, including ERK-1 and -2 (extracellular signal-regulated kinases 1/2), JNK (c-Jun N-terminal kinase or stress-activated protein kinase), and p38. Each kinase subtype is activated by similar, but distinct, upstream kinases. Extracellular stimuli appear to activate either single or multiple MAP kinase pathways (46 -48). In chondrocytes, inhibition of ERK1/2 and activation of p38 are associated with differentiation in vitro (49 -51).
Here we have studied the role of the ␣ 6B to ␣ 6A switch, at the onset of chondrocyte differentiation, by over-expressing ␣ 6A and ␣ 6B integrin subunits in chondrogenic cells derived from chick embryo tibiae. We have evaluated the effect on the expression of several differentiation markers (cell morphology, growth rate, collagen expression pattern) in cells grown under culture conditions that are either permissive or nonpermissive for chondrocyte differentiation, i.e. cells cultured in suspension or in monolayers.
Our results indicate a functional role for ␣ 6B ␤ 1 in the induction of chondrocyte differentiation and for ␣ 6A ␤ 1 in its progression. Furthermore, we suggest that these events depend on p38 MAP kinase activity.
Cell Culture-Primary cultures of chondrocytes were isolated from chick embryo tibiae developed in virus-free eggs (Lohmann GMBH, Cuxhaven, Germany) at stages 28 -30 (53) as described previously (34 -36). Briefly, embryo tibiae were rinsed in phosphate-buffered saline (PBS), pH 7.4, and digested with 400 units/ml collagenase I and 0.25% trypsin for 15 min at 37°C to remove the perichondrium, including the putative perichondral ossification sleeve. The tissue debris was removed by centrifugation and the pellet digested for an additional 45 min at 37°C with the above mixture supplemented with 1000 units/ml collagenase II. Released chondrocytes were pooled, harvested by low speed centrifugation, and resuspended in culture medium (35).
Cells were grown in monolayer culture in Coon's modified F-12 medium (Sigma) supplemented with 10% fetal calf serum, 50 units/ml penicillinstreptomycin and 2 mM L-glutamine. The medium was changed every other day, and confluent cells were split at a ratio of 1:3. After 12 days, cells were transfected with the appropriate vector and then cultured in monolayer for a maximum of 15-17 days. When required, the transfected cells were released by trypsin treatment and transferred to anchorageindependent conditions on a 1% agarose layer (35).
Construction of ␣ 6A , ␣ 6B , and ␣ 5 cDNAs-All constructs were assembled in the avian retroviral expression vector pSFCV-LE (54), a kind gift from B. Vennstrom (Department of Cell and Molecular Biology, Karolinska Institute, Stockholm). Human wild type ␣ 6A (3177 bp) and ␣ 6B (3207 bp) cDNAs cloned into the pRC-CMV vector (Invitrogen, Carlsbad, CA) were a kind gift from Dr. A. Sonnenberg (Division of Cell Biology, The Netherlands Cancer Institute, Amsterdam). The fulllength cDNAs were isolated by digestion with HindIII and subcloned into pSFCV-LE.
The human wild type ␣ 5 (3147 bp) cDNA, cloned in the pECE vector, was a kind gift from Dr. E. Ruoslahti (Cancer Research Center, The Burnham Institute, La Jolla, CA). The full-length cDNA was isolated by digestion with SalI and XbaI and, after Klenow treatment, inserted in-frame into the retroviral vector linearized with EcoRV.
Transfection-The empty pSFCV-LE vector or, alternatively, the vector containing the full-length human cDNA of either ␣ 5 , ␣ 6A , or ␣ 6B , was transfected into avian chondrogenic cells using polybrene and Me 2 SO. 24 h before transfection, 8 ϫ 10 5 -1 ϫ 10 6 chondrogenic cells were seeded in culture dishes. After 24 h, cells were rinsed with PBS and resuspended in 3 ml of F-12 medium supplemented with 1% fetal calf serum, 1% chicken serum, and 30 g of polybrene. Ten g of circular plasmid DNA were admixed with the cells and incubated at 37°C in 5% CO 2 for 6 h with occasional agitation. Cells were then incubated at room temperature with fresh medium and 30% Me 2 SO for 4 min, washed twice with PBS, and covered with complete medium (Coon's modified F-12). Neomycin-resistant cells were isolated by use of a selection medium containing G418 (500 g/ml, Invitrogen) for 1 week. When indicated, cells were transferred to suspension culture at day 10 of adherent culture to promote their differentiation. Using this method, we obtained a much higher transfection efficiency than with other protocols tested (not shown). For the flow cytometric analysis of the endogenous levels of avian integrin subunits ␣2 or ␣3, transfected cells were permeabilized with PBS, 0.1% saponin for 10 min at room temperature. Cells were then incubated with antibodies to ␣2 (1:20) or ␣3 (1:5) diluted in PBS, 0.1% saponin and processed as described above. As a negative control, transfected cells were incubated with the secondary antibody alone. Each analysis was performed twice on two different batches of transfected cells.
To determine the number of human ␣5 and ␣6 expressing cells at early time points after transfection (day 2 and day 4), cells were immunostained with primary antibodies to either human ␣6 (1:50) or human ␣5 (1:25) and processed as described above. Analyses were performed with an epifluorescence inverted microscope (Olympus Optical Co.); images stored using a chilled Hamamatsu CCD black and white camera and processed with IP-LAB and Adobe Photoshop software. Ten epifluorescence microscope fields, obtained using a 40ϫ lens, were collected for each sample.
Evaluation of Apoptosis-Cells were seeded on 10-mm glass coverslips, transfected, and grown in the presence or absence of 10 M p38 inhibitor, SB203580, as described below. After 2 and 4 days, cells were fixed with 4% formaldehyde at 4°C, washed several times with PBS, and permeabilized with 0.2% Triton X-100 for 5 min at room temperature. After several washes with PBS, apoptosis was determined using the TUNEL method (TdT-mediated dUTP nick-end labeling) (Promega) according to the manufacturer's instructions. Briefly, cells were incubated at room temperature for 10 min with equilibration buffer followed by a 1 h-incubation at 37°C in a humidified chamber with TdT diluted in the reaction buffer. The TdT reaction was stopped with 2ϫ SSC for 15 min at room temperature, and cells were washed with PBS and deionized water. Coverslips were mounted in Moviol, and apoptosis was assessed and quantified by nuclear staining observed with an epifluo-rescence inverted microscope. Ten fields, obtained using the 40ϫ lens, were counted for each sample.
Type X Collagen and FN Metabolic Labeling-Cells, cultured either in monolayer or in suspension, were starved for 2 h at 37°C in serum and methionine-free F-12 medium containing 100 g/ml ascorbic acid. The medium was then supplemented with 100 Ci of [ 35 S]methionine (Amersham Biosciences) for 2 h at 37°C. The culture medium was harvested and clarified by low speed centrifugation. To evaluate type X collagen secretion, aliquots containing 4 ϫ 10 5 cpm of labeled medium were dialyzed against 0.5 N acetic acid for 24 h at 4°C, and proteins were digested with 100 g of pepsin at 4°C for 12-15 h. The digested proteins were lyophilized, dissolved in Laemmli reducing buffer, and separated by a 12.5% SDS-PAGE (57). The radiolabeled proteins were detected by autoradiography with Hyperfilm TM MP (Amersham Biosciences).
Immunoblots-Cells obtained after different times of culture under adherent or suspension conditions were pooled, washed with PBS, lysed for 1 h in ice-cold buffer (50 mM Tris-HCl, 150 mM NaCl, 1% Nonidet P-40) containing 2 mM PMSF and 2% aprotinin as proteinase inhibitors, and centrifuged. The protein concentration was determined using the Bradford dye colorimetric assay (Bio-Rad, Hercules, CA). Samples containing 50 -100 g of total protein were separated by SDS-PAGE, blotted onto nitrocellulose membranes, and analyzed by immunostaining. Membranes were blocked with 3% nonfat dry milk in TBS, 0.05% Tween for 1 h at room temperature and incubated with antibodies to type I (1:500) or type II (1:250) collagen diluted in TBS, 0.05% Tween, 3% nonfat dry milk for 1 h at room temperature. HRP-conjugated swine anti-rabbit Ig (Dako) diluted at 1:2000 was used for detection. Reactivity was evaluated using a chemiluminescence emission assay (ECL, Amersham Biosciences).
p38 Kinase Assay-The activity of p38 was determined by an immunocomplex kinase assay. Cell lysates from transfected cells were cultured in the presence or absence of the p38 inhibitor SB203580 (10 M). The culture medium, containing SB203580, was changed every 12 h. Cells were collected on days 2 and 4 after transfection (e.g. during the selection period). Experiments were performed in parallel, with cells treated with the same amount of Me 2 SO used to dilute the SB203580 inhibitor. Cells were lysed in buffer containing 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 ␤-glycerolphosphate, 1 mM Na 3 VO 4 , 1 g/ml leupeptin, and 1 mM PMSF. The protein concentration was determined using the Bradford dye colorimetric assay (Bio-Rad).
Samples containing 200 g of total cell extract were precipitated with immobilized phospho-p38 MAP kinase monoclonal antibody (Cell Signaling Technology, New England Biolabs) and incubated with rotating overnight at 4°C. After two rinses with lysis buffer and kinase buffer (25 mM Tris-HCl, pH 7.5, 5 mM ␤-glycerolphosphate, 2 mM dithiothreitol, 0.1 mM Na 3 VO 4 , 10 mM MgCl 2 ), beads were resuspended in 50 l of kinase buffer supplemented with 200 M ATP and 2 g of ATF-2 fusion protein and incubated at 30°C for 30 min. The reaction was stopped by the addition of 25 l of 3ϫ Laemmli sample buffer followed by boiling. Samples were analyzed by SDS-PAGE and Western blot, by probing the membranes with the phospho-ATF-2 antibody diluted 1:1000 overnight at 4°C. Bound antibodies were detected as described above.
Furthermore, samples containing 200 g of total cell extract were separated by SDS-PAGE and analyzed by Western blot using antibodies to either actin or p38 and type II collagen. The reaction was developed using the ECL luminescence reagent.
Cells were solubilized in ice-cold lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 5 mM EDTA, 2% Nonidet P-40, 0.2% SDS) containing prote-ase inhibitors (2 mM PMSF, 2 mM leupeptin, 2 mM pepstatin) and centrifuged at 12,000 ϫ g for 10 min at 4°C. The supernatant was pre-cleared with protein A-Sepharose beads (Sigma) and normal rabbit serum (Sigma) in lysis buffer for 1 h at 4°C, centrifuged at 2,000 ϫ g for 5 min at 4°C, and incubated with the specific antibody to chicken ␤ 1 integrin subunit (V2E9) overnight at 4°C with gentle agitation. Soluble immunocomplexes were bound to protein A-Sepharose beads (Sigma) and recovered by centrifugation. After being washed with lysis buffer and Pastan buffer (20 mM Tris-HCl, 100 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40, 2.5 M KCl), bound proteins were eluted with Laemmli buffer, boiled, separated by 6% SDS-PAGE, and blotted onto nitrocellulose membranes. Blots were washed extensively with TBS-T (50 mM Tris-HCl, 150 mM NaCl, 0.05% Tween 20), blocked at room temperature for 1 h with 5% BSA in TBS-T, and incubated for 2 h at room temperature with the primary antibody specific for avian ␣ 5 integrin subunits diluted in TBS-T, 5% BSA. After three washes, membranes were incubated with the secondary antibodies (Dako) diluted with TBS-T, 5%BSA (1:2000), washed extensively with TBS-T and water, and developed using the ECL detection system.
Densitometric Analysis-Autoradiographic films from Western blot experiments were analyzed by calculating the total pixel value in the area of interest using Kodak Digital Science 1D image analysis software (Kodak, Rochester, New York).

RESULTS
Expression of ␣ 5 , ␣ 6A , and ␣ 6B Integrin Subunits in Chondrogenic Cells-We have previously reported that chondrocyte differentiation in vitro is associated with a down-regulation of the ␣ 5 ␤ 1 fibronectin receptor and a switch from the ␣ 6B to the ␣ 6A isoform of the ␣ 6 ␤ 1 laminin receptor (25). These results suggest that changes in the integrin expression pattern could be relevant for the induction and/or progression of chondrocyte differentiation. If this is the case, ectopic expression of either ␣ 6A or ␣ 6B should affect the differentiation pattern of cells cultured either in monolayer (nonpermissive for differentiation) or in suspension (permissive for differentiation) (36).
The expression of the human ␣ subunits could deplete the monomeric ␤ 1 pool available for endogenous integrins. To rule out this possibility, we surface-labeled transfected cells with biotin and analyzed the cell lysates by immunoprecipitation with an antibody to avian ␤ 1 followed by SDS-PAGE of precipitates, electrophoretic transfer to nitrocellulose, and detection of biotinylated integrins with streptavidin/HRP. ␤ 1 expression was 2-3-fold higher in cells transfected with vectors expressing ␣ subunits than in control cells (pSFCV-LE) ( Fig. 2A), suggesting that ectopic expression of ␣ subunits induced a parallel over-expression of endogenous ␤ 1 as described previously in other cell model systems (20).
To exclude possible negative effects of the transfected subunits on the expression of endogenous ␣ subunits, we evaluated by flow cytometry the surface levels of chick ␣ 2 and ␣ 3 and found that their levels were similar in all transfected cells (Fig.  2B). These results indicate that ectopic expression does not affect endogenous integrins. Therefore, the effects that follow human ␣ 5 or ␣ 6 transfection are not due to a reduced surface expression of endogenous integrins.
To assess whether transfection of ectopic ␣ subunits affects the levels and distribution of FN and LN, immunofluorescence experiments were performed after 10 days of adherent culture, using a polyclonal antibody against the globular domains G3-G5 of laminin-1 or the B3/D6 monoclonal antibody against avian FN.
Staining for laminin was similar in all transfected chondrocytes, showing a weak punctate distribution around the nucleus, which is associated with a strong positive staining of focal contacts (Fig. 3A).
In contrast, we found that pSFCV-LE and pSFCV-LE/␣ 5 cells displayed an extensive extracellular fibrillar network of FN underneath and between the cells (Fig. 3B), whereas pSFCV-LE/␣ 6A and pSFCV-LE/␣ 6B cells showed a faint staining with a few short fibrils (Fig. 3B).
To evaluate whether the reduced staining for FN was a consequence of a lower level of synthesis and/or secretion, we metabolically labeled transfected cells on day 10 and immunoprecipitated equal amounts of culture medium with an antibody to FN. The results (Fig. 3C) demonstrate a reduction in the levels of FN in all transfected cell populations in comparison with pSFCV-LE cells. However, although pSFCV-LE/␣ 5 cells still secreted appreciable amounts of FN, almost no signal was detected in pSFCV-LE/␣ 6 A and B. These results suggest that ectopic ␣ 6 inhibits FN synthesis and secretion.
Because reduced levels of FN during chondrogenesis are associated with a parallel down-regulation of the cognate ␣ 5 ␤ 1 receptor (25), the levels of endogenous ␣ 5 subunit were analyzed in the same cells. Western blot analysis using an antibody specific for the chick ␣ 5 subunit demonstrated a significant reduction in the expression of this subunit in cells expressing either of the two ␣ 6 isoforms but not in pSFCV-LE and pSFCV-LE/␣ 5 cells (Fig. 3D), which instead gave a higher signal.
pSFCV-LE/␣ 6A and pSFCV-LE/␣ 6B Cells Display Different Levels of Differentiation under Nonpermissive Culture Conditions-During chondrogenic differentiation, the down-regulation of FN and ␣ 5 is associated with the loss of other markers of undifferentiated cells (i.e. fibroblast morphology, high proliferation rate, and type I collagen expression) and the expression of stage I chondrocyte markers (i.e. cobblestone morphology, low proliferation rate, and type II collagen expression) (12,25).
Altogether, these data suggest that over-expression of ␣ 6B induces full differentiation of chondrogenic cells to stage I chondrocytes under nonpermissive culture conditions. In con- trast, ␣ 6A -transfected cells synthesized type II collagen and down-regulated FN and ␣ 5 . However, they were unable to modify other undifferentiated chondrogenic cell markers (i.e. type I collagen and cell morphology). Finally, control ␣ 5 -transfected cells did not display any change when compared with pSFCV-LE, with the exception of a very low level of type II collagen expression.
Prevalence of ␣ 6A and ␣ 6B Is Necessary to Stabilize the Differentiated Phenotype in Chondrocytes Suspension Cultures-The expression of type X collagen was determined to evaluate whether ectopic expression of ␣ 6A and ␣ 6B is able to induce faster differentiation to stage II chondrocytes under nonpermissive culture conditions. Lysates of metabolically labeled cells were pepsin-digested and analyzed by SDS-PAGE. Under these culture conditions neither the ␣ 6 nor the control transfected cells expressed type X collagen (Fig. 5, Adh., Col. X).
We next evaluated whether ␣ 6A and ␣ 6B subunits play any role in the progression from stage I to stage II chondrocytes in culture conditions permissive for differentiation (i.e. suspension cultures). Although in such cultures type X collagen is produced first after several days, pSFCV-LE/␣ 6A and pSFCV-LE/␣ 6B cells expressed type X collagen after only 24 h (Fig. 5, Susp., Col. X). This was due specifically to the over-expression of the ␣ 6 subunits, demonstrated by the absence of type X collagen after 24 h of suspension culture in pSFCV-LE and pSFCV-LE/␣ 5 (Fig. 5, Susp., Col. X). Levels of type II collagen expression were also analyzed and, as anticipated, were found to be increased in pSFCV-LE/␣ 6A , pSFCV-LE/␣ 5 , and pSFCV-LE. In contrast, no type II collagen was expressed by pSFCV-LE/␣ 6B (Fig. 5, Susp., Col. II), and type I collagen was almost undetectable in all samples (Fig. 5, Susp., Col. I). These results suggest that a prevalence of ␣ 6A versus ␣ 6B is necessary to FIG. 3. Fibronectin and ␣ 5 ␤ 1 , but not laminin, are down-regulated upon ectopic ␣ 6 expression. A, pSFCV-LE, pSFCV-LE/␣ 5 , pSFCV-LE/␣ 6A , and pSFCV-LE/␣ 6B cells from day 7 were plated on glass coverslips and further cultured adherent for 3 days. Cells were subsequently fixed, permeabilized, and immunostained with an antibody to avian LN-1 followed by a Cy3-conjugated secondary antibody. B, pSFCV-LE, pSFCV-LE/␣ 5 , pSFCV-LE/␣ 6A , and pSFCV-LE/␣ 6B cells from day 7 were plated on glass coverslips and cultured adherent for three additional days. Cells were subsequently fixed, permeabilized, and immunostained with an antibody to avian FN followed by a donkey anti-mouse FITC-conjugated antibody. Nuclear staining by 4Ј,6-diamidino-2-phenylindole (asterisks) identifies single cells. C, transfected cells were metabolically labeled with 35 S on day 10. FN was immunoprecipitated from labeled medium using the B3/D6 antibody to FN. Immunoprecipitates were analyzed by SDS-PAGE under reducing conditions. D, surface expression of the endogenous FN receptor, ␣ 5 ␤ 1 , was determined by surface biotinylation of transfected cells. Cell lysates were immunoprecipitated with the avian ␤ 1 antibody (V2E9), separated by SDS-PAGE, and blotted onto nitrocellulose membranes. Blots were detected with ␣ 5 antibody and developed using the ECL system. p38 Kinase Activation Is Necessary for ␣ 6 -induced Chondrocyte Differentiation-p38 MAP kinase activation is associated with differentiation of chondrogenic cells (50,51,58,59) and myoblasts (19,20) in culture. The activity of p38 was determined in early phases of chondrocyte differentiation induced by ␣ 6B transfection, with the aim of identifying the earliest events associated with differentiation to stage I chondrocytes. At early time points after chondrocyte transfection (days 2 and 4, before selection was completed), p38 activity was evaluated by measuring the levels of phosphorylated ATF-2, a major phospho-p38 kinase substrate. Equal amounts of total lysate proteins were immunoprecipitated with an antibody to p-p38. Immunoprecipitates were then incubated with ATF-2 fusion protein in the presence of ATP, thus allowing immunoprecipitated active p38 MAP kinase to phosphorylate ATF-2. Phosphorylation of ATF-2 at Thr-71 was detected by Western blotting using a phospho-ATF-2 (Thr-71) antibody. In addition, the total amount of p38 was also measured in each cell population by Western blot of total lysates. A densitometric analysis was performed, and the data for each sample were normalized to the actin and p38 contents. Values were further corrected for the number of cells expressing the ectopic ␣ subunits as determined by immunofluorescence. The results indicate that p38 activity on day 2 is 2.56 Ϯ 0.251 and 4.07 Ϯ 0.266 times higher in pSFCV-LE/␣ 6A and pSFCV-LE/␣ 6B cells, respectively, in comparison with control cells (pSFCV-LE) ( Table I).
On the contrary, on day 2 the pSFCV-LE/␣ 5 cells displayed 50% of the p38 activity detected in control cells (pSFCV-LE) ( Table I, Fig. 6A). These results demonstrate that high levels of p38 activation are dependent on the expression of ectopic ␣ 6B .
To address a possible role for p38 in the differentiation events associated with ectopic ␣ 6 expression, we first inhibited p38 activity using the specific inhibitor SB203580 (69). As expected, addition of 10 M SB203580 to the culture medium of transfected cells had no effect on the expression of total p38, but it drastically reduced the level of p-ATF-2 (Fig. 6B). In addition, the number of apoptotic cells was determined by analyzing the nuclear TUNEL staining using epifluorescence microscopy. No significant differences were found between the treated and nontreated cells (data not shown).
Next, the amount of type II collagen synthesized on days 2 and 4 (Fig. 6C) in the presence or absence of SB203580 was determined. Western blot analyses of cell lysates from all of the transfected cell populations (pSFCV-LE, pSFCV-LE/␣ 5 , pSFCV-LE/␣ 6A , and pSFCV-LE/␣ 6B ) showed that on day 4, in the absence of the p38 inhibitor, only pSFCV-LE/␣ 6B synthesized type II collagen. The addition of SB203580 to the culture medium abolished such expression. It is notable that type II collagen is detected on day 10 in ␣ 6A -transfected cells only (day 10, Fig. 4). These results suggest that a functional correlation exists between p38 activation and expression of a stage I chondrocyte differentiation marker induced by ␣ 6B over-expression. curve) and grown in monolayer for 10 additional days (day 10 of the growth curve). Every other day, triplicate dishes of transfected cells were trypsinized and counted using a Neubauer chamber. The graph shows the mean number of transfected cells at each time point including standard deviations. The drop in cell number, observed between day 0 and day 2 of the growth curve, is due to a loss of cells observed in most cell lines after seeding (67). C, pSFCV-LE, pSFCV-LE/␣ 5 , pSFCV-LE/␣ 6A , and pSFCV-LE/␣ 6B cells from day 7 were plated on glass coverslips and cultured in monolayer for 3 additional days. Immunofluorescence was performed on permeabilized cells using antibodies to type I (Col. I) or type II (Col. II) collagens followed by Cy3-conjugated secondary antibodies.

DISCUSSION
A switch in expression from the ␣ 6B to the ␣ 6A isoform of the ␣ 6 ␤ 1 laminin receptor is associated with differentiation in several cell types (37)(38)(39). The two alternatively spliced isoforms of ␣ 6 differ in their cytoplasmic domains, which consist of 36 and 54 amino acids in ␣ 6A and ␣ 6B (40 -43), respectively, and display similar binding affinities and ligand specificities. Moreover, upon phorbol ester treatment, only the ␣ 6A isoform is phosphorylated (41,60). Differential interactions with cytoskeletal proteins have been suggested to explain the greater ability of ␣ 6A to induce pseudopodia formation and to promote migration in a macrophage cell line plated on a laminin-1 (19,44,61). However, little is known about the roles played by ␣ 6A and ␣ 6B in the modulation of cell differentiation. We have addressed this question using an in vitro model system that allows differentiation of chondrogenic cells transferred from nonpermissive (adherent) to permissive (suspension) culture conditions. The in vitro chondrocyte differentiation from prechondrogenic cells to stage I chondrocytes (33,35) is associated with a switch from the ␣ 6B to the ␣ 6A (25). Our approach has been to over-express the two alternative isoforms, evaluating the effect on chondrocyte differentiation under both nonpermissive and permissive culture conditions. ␣ 6B Expression Promotes Full Differentiation to Stage I Chondrocytes-Under nonpermissive culture conditions, ectopic expression of the ␣ 6B isoform is sufficient to specifically promote chondrocyte differentiation to stage I, as indicated by activation of the expression of type II collagen, reduction of the growth rate, reduction of type I collagen, fibronectin and ␣ 5 ␤ 1 expression, and changes in cell morphology from a fibroblastlike to a cobblestone shape.
The specificity of the ␣ 6B effects listed above is supported by the observation that transfection with human ␣ 6A or ␣ 5 subunits only prompts a low level and delayed expression of type II collagen. Furthermore, there is no loss of markers of the undifferentiated phenotype.
The full differentiation to stage I chondrocyte following ␣ 6B transfection is not due to a nonspecific imbalance in the expression of endogenous chick integrins. This is demonstrated by the unchanged levels of chick ␣ 2 and ␣ 3 after transfection of either human ␣ 6 or ␣ 5 .
On the contrary, the specific ␣ 6 -dependent reduction of chick p38 activity was determined using an in vitro kinase assay and ATF-2 as a substrate. The levels of ATF-2 were determined by Western blot using an antibody to pATF-2. Equal amounts of total cell extracts from day 2 and 4 were analyzed by Western blot using an antibody to p38. For actin, equal amounts of total cell extracts from days 2 and 4 were analyzed by Western blot using an antibody to actin. For p38/ cells, the total amount of p38/cell was determined. The results were evaluated by dividing the p38/actin ratio normalized to the number of cells expressing ectopic ␣ subunits as determined by immunofluorescence. This value ranged between 52 and 69% of the whole cell population. ␣ 5 may contribute to the differentiation process by removing the inhibitory activity exerted by ␣ 5 ␤ 1 (62). The reduction in endogenous ␣ 5 is a direct and specific effect of ␣ 6 transfection and is not observed when cells are transfected with ␣ 5 . In addition, it occurs despite the increased levels of endogenous ␤ 1 , which provide sufficient protein to assemble chick ␣ 5 ␤ 1 heterodimers.
␣ 6A Prevalence Is Necessary to Fully Maintain the Stage I Differentiation State-Upon differentiation to stage I chondrocytes, the switch from expression of ␣ 6B to ␣ 6A leads to a prevalence of the latter on the cell surface. Experiments to impair this were performed by over-expressing ␣ 6B .
␣ 6B -transfected chondrocytes, when differentiated in adherent cultures, are unable to maintain their phenotype once transferred in suspension, as shown by loss of expression of type II collagen. In contrast, ␣ 6A , ␣ 5 , and mock-transfected cells differentiate normally to stage I.
These results suggest that the prevalence in ␣ 6A surface expression during chondrogenesis may contribute to sustaining the differentiated phenotype. In addition, upon transfection of ␣ 6B differentiation to stage II chondrocyte progresses normally in vitro. These observations suggest that other events may contribute to support the chondrocyte phenotype in vivo. Therefore, it is likely that transition to stage II chondrocytes implies several independent events, one of which may be the role of ␣ 6A in sustaining type II collagen expression.
Stage I to Stage II Transition Requires Suspension Culture and ␣ 6 to Occur-To assess whether ␣ 6B and/or ␣ 6A are involved in the progression from stage I to stage II chondrocyte differentiation, the expression of type X collagen was evaluated in transfected cells cultured either in monolayer or in suspension.
We found that type X collagen is exclusively expressed in suspension cultures but that it appears earlier (24 h) in both types of ␣ 6 -transfected cells as compared with control cells where several days are required. These results further suggest that neither of the ␣ 6 subunits is sufficient to promote differentiation to stage II chondrocytes and that other conditions associated with culture in suspension are necessary. In addition, this further step of differentiation is likely to be independent from the expression of type II collagen, as shown by the behavior of ␣ 6B -transfected cells.
Early Activation of the p38 MAP Kinase Pathway Is Functionally Associated with ␣ 6B -induced Type II Collagen Expression-p38 MAP kinase has been shown to modulate chondrogenic differentiation. Growth/differentiation factor-5 (GDF-5) induces p38 activation and chondrocyte differentiation in the mouse chondrogenic cell line ATDC5 (49). Chun and co-workers (50,51,58,59), using a micromass culture system, have shown that chondrocyte differentiation associates with a protein Kinase C (PKC)-independent activation of p38 in chondrogenesis. In addition, treatment of micromass cultures with epidermal growth factor, a limb bud cartilage development modulator, decreases the rate of chondrogenesis by reducing p38 activity (51).
Furthermore, adhesion to substrates and actin cytoskeletal organization have been shown to be important in integrin/ growth factor receptor signaling via MAP kinases (2,(63)(64)(65)(66). These pathways may be relevant to the induction of chondrocyte differentiation in anchorage-independent cultures (12) and following cytochalasin D treatment (31).
Here we show that inhibition of p38 activation by SB203580 inhibits induction of type II collagen expression by ␣ 6B . Therefore it appears that this signaling pathway is involved in chondrocyte differentiation. In addition, by determining the level and time of p38 activity we found a peak at day 2 in ␣ 6B transfected cells, which was about 2-fold higher than the levels measured on days 2 and 4 in ␣ 6A -and ␣ 5 -transfected cells. These results correlate with the early expression of type II collagen in ␣ 6B but not in ␣ 6A -and ␣ 5 -transfected cells. Although these data do not allow us to conclude that p38 activation is directly dependent on ␣ 6B , they suggest that p38 activation is functionally related with type II collagen expression.
The presence of LN in chondrogenic area of the condensing mesenchyme suggests that the interaction of ␣ 6B ␤ 1 receptor may be relevant in vivo, as well as in vitro. In addition, in in vitro model systems more closely reproducing the limb bud chondrogenic area, i.e. micromass cultures, the inhibition of p38 activity has been proven to block type II collagen expression (50), further supporting this hypothesis.
The interpretation of the relevance of ␣ 6A role in vivo is more complex. The presence of LN in the ECM surrounding stage II chondrocytes in cartilage tissue (67), suggests that its interaction with ␣ 6A ␤ 1 may play a role in chondrocyte differentiation in vivo, as we have highlighted in vitro. However, mice deficient in ␣ 6A do not display defective endochondral bone formation (68), suggesting that the potential role of ␣ 6B ␤ 1 -LN interactions is not sufficient to promote the progression of chondrocyte differentiation and that other events may be relevant to this process. In particular, further studies are needed to address the role of other cell-ECM interactions and soluble factor effects.
In conclusion, our results demonstrate that ectopic expression of the ␣ 6B ␤ 1 laminin receptor is sufficient to promote differentiation of chondrogenic cells even under nonpermissive culture conditions. Furthermore the expression of ␣ 6A contributes to stabilizing the differentiated phenotype and to sustaining the progression to further steps of chondrocyte differentiation in vitro.