Apert (Ap) syndrome is characterized by premature cranial suture ossification caused by fibroblast growth factor receptor 2 (FGFR-2) mutations. We studied the role of cadherins and signaling events in the phenotypic alterations induced by the Ap FGFR-2 S252W mutation in mutant immortalized fetal human calvaria osteoblasts. The FGFR-2 mutation caused increased expression of the osteoblast markers alkaline phosphatase (ALP), type 1 collagen (COLIA1), and osteocalcin (OC) in long-term culture. The mutation also increased cell-cell aggregation, which was suppressed by specific neutralizing anti-N- and anti-E-cadherin antibodies. Mutant osteoblasts showed increased N- and E-cadherin, but not N-cell adhesion molecule (N-CAM) messenger RNA (mRNA) and protein levels. This was confirmed in vivo by the abundant immunoreactive N- and E-cadherins in preosteoblasts in the Ap suture whereas N-CAM and α- and β-catenins were unaffected. Neutralizing anti-N-cadherin antibody or N-cadherin antisense (AS) oligonucleotides but not anti-E-cadherin antibody or AS reduced ALP activity as well as ALP, COLIA1, and OC mRNA overexpression in mutant osteoblasts. Analysis of signal transduction revealed increased phospholipase Cγ (PLCγ) and protein kinase Cα (PKCα) phosphorylation and increased PKC activity in mutant cells in basal conditions. Inhibition of PKC by calphostin C or the PKCα-specific inhibitor Gö6976 suppressed the increased N-cadherin mRNA and protein levels as well as the overexpression of ALP, COLIA1, and OC mRNA in mutant cells. Thus, N-cadherin plays a role in the activation of osteoblast differentiation marker genes in mutant osteoblasts and PKCα signaling appears to be involved in the increased N-cadherin and osteoblast gene expression induced by the S252W FGFR-2 mutation in human osteoblasts.
FIBROBLAST GROWTH factor receptors (FGFRs) are members of a family of tyrosine kinase receptors that are involved in the control of cell proliferation, migration, differentiation, and survival.(1–3) Binding of FGFs to FGFRs leads to receptor dimerization, intrinsic tyrosine phosphorylation, and intracellular signaling events that trigger gene expression and subsequent activation of proliferation or differentiation.(2) Although the mechanisms involved in FGF-induced signal transduction are not fully established, multiple signal transduction pathways have been implicated in FGF signaling.(3, 4) Tyrosine phosphorylation after FGFR activation stimulates a kinase cascade involving Raf, mitogen-activated protein (MAP) kinase kinase (MEK), and MAPKs,(4–6) a family composed of extracellular signal-regulated kinases (ERKs), stress-activated protein kinase/c-jun N-terminal kinase (SAPK/JNK), and p38 proteins.(7) In some cells, FGF-2 induces tyrosine phosphorylation of src,(8, 9) activation of p38 MAPK(10) or phospholipase Cγ (PLCγ),(11) a physiological activator of protein kinase C (PKC).(12) The PKC family is composed of serine/threonine kinases and more than 10 PKC isoenzymes have been identified in mammalian tissues.(13, 14) PKC isoforms have been classified into three subclasses according to their ability to be activated by diacylglycerol (DAG), phosphatidylserine, or calcium. Classical PKCs include the calcium- and DAG-dependent α, βI, βII, and γ isoforms. Novel PKCs comprise the calcium-independent and DAG-dependent ζ, λ, μ, and ϵ isoforms and the atypical PKCs are composed of the calcium- and DAG-independent ζ, λ, and μ isoforms.(13, 14) The various PKC isoenzymes involved in cellular signaling pathways play important and distinct roles in proliferation and differentiation in different cell systems.(4, 15–17)
Recent data indicate that FGFs play important roles in the regulation of osteoblastic cell growth and differentiation.(18–20) In humans, mutations in FGFR-1 and -2 were shown to induce abnormal ossification of cranial sutures (craniosynostosis), implying a major role of FGFR signaling in intramembranous bone formation.(21, 22) Most syndromic craniosynostosis associated with FGFR-2 mutations in humans are associated with point mutations in the immunoglobulin (Ig) III domain or in the linker between the Ig II and the Ig III domain, leading to complex functional alterations of the mutant receptor. In Apert (Ap) syndrome, most cases are associated with S252W or P253R point mutations in the linker between the second and third Ig-like domain of FGFR-2.(23, 24) Experimental studies suggest that FGFR-2 mutations engender gain of function through different signal alterations.(25–27) In Ap syndrome, FGFR-2 mutations were found to enhance receptor occupancy by FGF ligands and/or to induce prolongation of the duration of receptor signaling.(28) Using ex vivo analysis of osteoblast cultures,(29) we previously showed that the Ap FGFR-2 S252W mutation activates the differentiation pathway in preosteoblasts in vitro but does not affect cell proliferation or the response to FGF-2.(30) This alteration of osteoblast differentiation induced by the mutation also was confirmed in vivo.(31) However, the molecular events and signaling pathways linking the FGFR-2 mutation to activation of the osteoblast differentiation pathway remained unclear.
Cell-cell adhesion plays an essential role in the control of cell growth and differentiation and tissue development.(32) The cell adhesion molecule (CAM) cadherins play an important role in cell-cell adhesion in various cell types.(33) Bone cells express several cadherins, including the classic E- and N-cadherins and N-CAM.(34–38) Although the role of cadherins in osteoblast function is not clear, recent studies indicate that N-cadherin is involved in the expression of osteoblast markers and differentiation.(36–38) In the present study, we investigated the role of specific cadherins and the transduction signaling events that link the S252W FGFR-2 mutation to the activation of osteoblast differentiation in Ap craniosynostosis. The data support a role for PKCα and N-cadherin in the overexpression of osteoblast markers induced by the activating S252W FGFR-2 mutation in human calvaria osteoblasts.
MATERIALS AND METHODS
Antibodies and inhibitors
Rabbit antibody against FGFR-2 (residues 805-821 mapping within the COOH-terminus of the human bek receptor), the corresponding FGFR-2 blocking peptide, and goat polyclonal antibodies against α- or β-catenins were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Mouse monoclonal antibodies against N-cadherin or N-CAM, rat monoclonal antibody against E-cadherin, and polyclonal antibody against actin were from Sigma (St. Louis, MO, USA). A nonspecific IgG used as control of neutralizing cadherin antibodies was from DAKO (Glostrup, Denmark). Antisense (AS) oligonucleotides protected from endonucleases by phosphorothioate modification at the 3′ end were from Eurobio (Les Uris, France). Calphostin C, a potent and selective inhibitor of PKC,(39) and PD-98059, a specific inhibitor of MEK activation,(40) were from Biomol Research Laboratories (Plymouth, PA, USA). SB 203580, a highly specific inhibitor of p38 kinase,(41) Gö6976, a potent and specific PKCα inhibitor,(15, 42) and Gö6983, a selective inhibitor of PKCα, -β, and -γ(43) were from Calbiochem (San Diego, CA, USA). Transduction antibodies specific to phosphotyrosine (4G10, agarose conjugated), phosphorylated PKCα, and PKCϵ were from Upstate Biotechnology, Inc. (Lake Placid, NY, USA). The antiphosphorylated PKCδ and the antiphospho-p38 antibodies were from New England Biolab (Beverly, MA, USA), whereas antibodies to phosphorylated ERK-1 and -2 were from Santa Cruz Laboratories (Santa Cruz, CA, USA).
Calvaria samples (supplied by Dr. A-L. Delezoide, Hopital R. Debré, AP-HP, Paris, France) were obtained from two aborted normal and two Apert 26-week-old fetuses in accordance with the French Ethical Committee recommendations. Mutation analyses (performed by Dr. J. Bonaventure, INSERM U 393, Hôpital Necker, Paris, France) revealed the S252W mutation, the most frequent mutation in Apert syndrome, in the two Ap fetuses, as shown by single-strand conformation polymorphism and restriction analyses of the coding sequence of the FGFR-2 gene.
Cell cultures and differentiation studies
Cells were obtained by collagenase digestion from the coronal sutures in one Ap and one control (Co) fetus as previously described.(29) The cells were immortalized by transfection with the SV40 oncogene as described(44) and called Ap and Co fetal cells.(30) Previous analysis showed that this immortalization does not change the properties of the cells.(30, 31) The cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with glutamine (292 mg/liter), 10% heat inactivated fetal calf serum (FCS), and antibiotics (100 IU/ml penicillin and 100 μg/ml streptomycin). For cell differentiation studies, Ap and Co cells were cultured in aggregates to induce in vitro osteogenesis.(45) We used this aggregation system because Ap or Co cells do not form nodules spontaneously. In this system, cells cultured at high density in aggregates allow optimal cell-cell contacts and differentiation.(38) Ap and Co cells were cultured in aggregates by pooling 107 cells on bacteriological grade culture dishes (Becton-Dickinson, Plymouth, UK) and incubated in DMEM plus 10% FCS for 24 h and then in the same medium with 50 μg/ml ascorbic acid and 50 μg/ml vitamin K for up to 21 days. In some experiments, 3 mM Pi were added to induce mineralization of the matrix.(45) Calcium content in the matrix, alkaline phosphatase (ALP) activity, and type 1 collagen (Col 1) synthesis evaluated by collagen type 1 carboxyterminal peptide (P1CP) levels were determined as previously described.(29)
For immunohistochemistry, coronal sutures obtained from one Ap and one Co fetus were fixed in 10% formaldehyde, decalcified in 4.1% ethylenediamine tetracetic acid for 3 weeks at 37°C, and embedded in paraffin. Sections (5 μm thick) were deparaffinized in xylene, rehydrated in ethanol, placed in 10% EDTA, trypsin diluted in calcium-free phosphate-buffered saline (PBS), washed in PBS/0.01% Triton X-100, incubated with 0.1% bovine serum albumin (BSA)/3% goat serum to block unspecific binding, and then exposed for 1 h at room temperature to anti—N-cadherin, anti-E-cadherin, anti-N-CAM, anti-FGFR-2, or anti-α- or -β-catenin antibodies, all diluted 1:100. Control sections were incubated with the appropriate solution (PBS/0.001% Na azide for cadherins, anti-FGFR-2 antibody neutralized with the blocking peptide for FGFR-2, or nonimmune serum for catenins). After 1 h of exposure at room temperature, sections were washed three times for 10 minutes in PBS and exposed to a second appropriate antibodies (1:50) linked to colloidal gold particles (IntenSETMM; Amersham, Arlington Heights, IL, USA) for 1 h at room temperature. The gold particle staining was enhanced by precipitation of metallic silver (ImmunoGold Silver Staining) and then washed before visualization. For immunocytochemistry, Ap and Co cells were cultured at confluence, fixed in acetone/methanol (1:1), and immunostained for N-cadherin, E-cadherin, or N-CAM as described previously.
Cell aggregation analysis
Cell-cell adhesion was assayed as described(46) with minor modification. Ap and Co cells suspended in DMEM plus 10% FCS were plated in bacteriological grade tissue culture wells and incubated to allow aggregation for 60 minutes at 37°C on a gyratory shaker in the presence or absence of neutralizing antibodies against N- and E-cadherins or a nonspecific IgG. Single cells and cell clusters (two cells or more) were counted and cell aggregation was evaluated by the ratio of the total cell number plated to the number of total particles (i.e., single cells and cell clusters) per well after 60 minutes of incubation.(38)
Neutralizing antibodies and AS oligonucleotides
To determine the role of cadherins in the alteration of osteoblast gene expression induced by the mutation, subconfluent Ap and Co cells were cultured in the presence or absence of neutralizing anti-N- or -E-cadherin antibodies (15 μg/ml), their combination, or a nonspecific IgG for 5 days. ALP activity was measured biochemically, and ALP and pro-α1(I) collagen (COLIA1) messenger RNA (mRNA) levels were determined by reverse-transcription polymerase chain reaction (RT-PCR) analysis. We also used an AS strategy to suppress part of the expression of mRNA in osteoblasts.(47) The N-cadherin AS oligomer used was complementary to the first 21 bases of the translation start site of human N-cadherin mRNA and had the following sequence: CGCTCCCGCTATCCGGCACAT. A random (R) oligonucleotide composed of the 21 randomly arranged bases of the N-cadherin AS was used as Co and had the following sequence: ACTCTGCTGTCACCACGGCCC. The E-cadherin AS oligomer complementary to the first 20 bases of the translation start site of human E-cadherin mRNA had the following sequence: CCTTAAACCTGTACAGTCAT. The R oligonucleotide for the E-cadherin AS had the following sequence: TGATAACTTTATGCACACCC. The R sequences were checked on computed database not to be complementary to any known sequence. The oligonucleotides were protected from endonucleases by phosphorothioate modification at the 3′ end. Preliminary dose-response experiments revealed that only high doses of oligos (50 μM) induced toxic cellular effects determined by cell viability (measured by trypan blue exclusion) and cell detachment (measured by the number of cells attached) and we therefore used an effective nontoxic lower dose of AS oligos (10 μM). Ap cells were cultured for 5 days in the presence or absence of AS or R oligonucleotides (10 μM) and the medium was replaced every 2 days. The cells were then washed and the effect of AS and R oligonucleotides on ALP activity was determined biochemically. In parallel, the effects of AS and R oligonucleotides on N- and E-cadherin protein levels were analyzed by Western blot analysis.
Western blot analysis
Ap and Co cell extracts were prepared by adding 500 μl of (ice-cold) lysis buffer containing 10 mM Tris-HCl, 5 mM EDTA, 150 mM NaCl, 30 mM sodium pyrophosphate, 50 mM NaF, 1 mM Na3VO4, 10% glycerol, and protease inhibitors (Boehringer, Mannheim, Germany). Lysates were clarified by centrifugation at 12,000g for 10 minutes at 4°C and the protein content of supernatants was determined using the DC Protein Assay (Bio-Rad Laboratories, Hercules, CA, USA). Proteins were subjected to on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and in some experiments actin was used as internal Co for protein loading. Proteins were transferred onto polyvinylidene difluoride (PVDF) membranes (Hybond-P; Amersham) in buffer containing 20% methanol. The membranes were incubated overnight with 1% blocking buffer Boehringer) in Tris-buffered saline (TBS; 50 mM Tris-HCl and 150 mM NaCl) containing 0.1% Tween 20, and then for 1 h at room temperature with appropriate antibodies in 0.5% blocking buffer, washed twice with TBS/0.1% Tween 20, and 0.5% blocking buffer, incubated for 1 h with horseradish peroxidase (HRP)-conjugated secondary antibody. After washing the membranes three times in TBS/0.1% Tween 20, the signal was visualized using the enhanced chemiluminescence (ECL) + chemiluminescent kit (Amersham). The specific bands on the autoradiograms were quantitated by densitometry.
RT-PCR analysis and Southern blot analyses
The levels of N- and E-cadherins mRNA were analyzed by RT-PCR. Optimization of RT-PCR results was carried out by generating saturations curves of RT-PCR products of each gene and of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) against cycle number (0-35 cycles) to allow semiquantitative variation of product levels.(38) The same cycle number (30 cycles) was used for all genes, except for N-cadherin and GAPDH (25 cycles), in which the amplification was linear. After 1-5 days of culture in aggregates, Ap and Co cells were washed with PBS and lysed with Extract-All (Eurobio) reagent according to the manufacturer's instructions. Three micrograms of total cellular RNA from each sample were reverse transcribed and the complementary DNA (cDNA) samples were then divided and amplified by PCR using specific primers for ALP, pro-α1(I) collagen (COLIA1), osteocalcin (OC), or GAPDH.(38) The quantities of the cDNAs for starting template in PCR reaction were 0.45 μg for GAPDH and 0.75 μg for other genes. The following were the primer sequences used for E-cadherin: sense, 5′-TGGCTGAAGGTGACAGAGC-3′; AS, 5′-CGTTAGCCTCGTTCTCAGG-3′; and internal, 5′-GAGGTTCCTGGAAGAGCACC-3′. The primer sequences used for N-cadherin were as follows: sense, 5′-CAAGTGCCATTAGCCAAGG-3′; AS, 5′-TTAAGCCG AGTGATGGTCC-3′; and internal, 5′-GTTGTCAACAT GGTACCGGC-3′.(38) Southern blots were performed by running aliquots of amplified cDNA on 1.2% agarose gel followed by transfer onto nylon membrane according to the manufacturer's protocol. Hybridization of blots was carried out overnight at 55°C with (γ32P)adenosine triphosphate (ATP)-labeled internal primers. Membranes were washed twice in 2× SSC/0.1% SDS at room temperature for 15 minutes, once in 0.1× SSC/0.1% SDS at 55°C for 15 minutes, and the filters were exposed to X-ray films. Autoradiographic signals were quantified using a scanner densitometer (Transydine General Corp., Ann Arbor, MI, USA), and the signal for each gene was related to that of GAPDH.(38)
Signal transduction analyses
Subconfluent Ap and Co cells were cultured in DMEM without serum for various incubation times and proteins were extracted in lysis buffer as described previously. For analysis of activated PLCγ, equal amounts of lysate proteins were incubated with agarose-conjugated antiphosphotyrosine 4G10 antibody for 18 h at 4°C. Immunoprecipitates were washed three times in lysis buffer and immunoprecipitated proteins were solubilized in reducing sample buffer containing 125 mM Tris, pH 6.8, 100 mM dithiothreitol, 10% glycerol, 0.025% bromophenol blue, and 2% SDS. Proteins were fractionated by reducing SDS/PAGE acrylamide gradient gels and transferred to Immobilon P membranes according to the manufacturer's instructions (Millipore Corp., Bedford, MA, USA) and Western blot analysis was performed as described previously using a specific anti-PLCγ antibody. For analysis of phosphorylated signaling proteins, Western blot analysis was performed as described previously using specific antibodies against phosphorylated PKCα, PKCϵ, PKCδ, p38, or ERK-1 and -2.(48) The levels of phosphorylated PKCα were measured by scanning densitometry. For direct analysis of PKC activation, lysates of Ap and Co cells were prepared as described previously and PKC activity was determined by measuring the transfer of [32P]-labeled phosphate to a biotinylated peptide substrate (AAKIQASFRGHMARKK) that is specific for PKC activity(49) using the Signa TECT PKC Assay System (Promega, Charbonne, France).
To determine the signal transduction pathway involved in the alterations of N-cadherin and osteoblast marker genes, Ap cells were cultured in the presence of the specific inhibitors calphostin C (0.5-2 μM), SB 203580 (20 μM), PD-98059 (20 μM), Gö6976 (0.1-1 μM), or Gö6983 (0.25-1 μM), and N-cadherin, ALP, COLIA1, or OC mRNA levels were determined at 24 h by RT-PCR analysis as described previously. At the protein level, the effects of three different PKC inhibitors (calphostin C, 2 μM; Gö6976, 1 μM; and Gö6983, 1 μM) on N-cadherin were determined at 48 h in Ap cells by Western blot analysis as described previously.
The data were repeated 2- to 3-fold and the results presented are representative of all experiments. The results are expressed as the mean ± SEM and were analyzed using the statistical package superanalysis of variance (ANOVA; Macintosh, Abacus Concepts, Inc., Berkeley, CA, USA). Differences between the mean values were evaluated with a minimal significance of p < 0.05.
Mutant and Co cells were cultured at high density in aggregates, allowing optimal cell-cell contacts and rapid differentiation.(30, 45) In these conditions, osteoblast markers are maximal in early time points and then decrease with time in culture. The time-course study showed that ALP activity decreased with time. In addition, type 1 collagen synthesis determined by P1CP levels released in the medium decreased with time and then remained stable (Fig. 1A), consistent with the evolution of these genes in human calvaria osteoblasts cultured in similar conditions.(38) In parallel, calcium content in the extracellular matrix deposited by Ap cells increased with time (Fig. 1A). Throughout the culture period, we found that ALP in Ap cells was 2-3 times higher than in Co cells. Type 1 collagen synthesis was also 2-fold increased in Ap cells at all time points. Moreover, calcium content in the extracellular matrix was higher than normal at 4-8 days of culture (Fig. 1A). RT-PCR analysis also showed that ALP, COLIA1, and OC mRNA levels were higher in Ap cells compared with Co cells during the first 7 days of culture in aggregates (Fig. 1B). These results show that the S252W FGFR-2 mutation increases the expression of these osteoblast differentiation markers during early and late stages of in vitro osteogenesis, confirming the increased ALP and OC gene expression previously observed in Ap mutant cells compared with Co cells in monolayers.(30)
Mutant osteoblasts overexpress N- and E-cadherins
The expression of N- and E-cadherins was determined by RT-PCR analysis in mutant cells cultured in aggregates. Because cell-cell interactions occur when cells establish contacts at early time points, cadherin expression was studied at 1-5 days of culture. In Co cells, maximal N-cadherin and E-cadherin mRNA levels were found at 1 day and the levels decreased with time in culture. N-cadherin mRNA levels were higher in Ap cells compared with Co cells on 1, 3, and 5 days of culture. E-cadherin mRNA levels also were increased in mutant cells compared with Co cells on 3 days and 5 days of culture (Fig. 2A).
To confirm this finding at the protein level, cadherin levels in mutant cells were determined at the same time points by Western blot analysis. As for mRNA levels, N-cadherin and E-cadherin protein levels were high on 1 day and decreased with time. N-cadherin and E-cadherin protein levels were higher in Ap cells compared with Co cells on 1 day and 3 days of culture and this was confirmed by densitometric analysis of the expected size bands (Fig. 2B). The slight difference in N- and E-cadherin expression at the mRNA and protein levels in cells cultured in aggregates may reflect a complex control of expression at the transcriptional and posttranscriptional levels, which remains to be identified. In contrast, the levels of α- and β-catenins, proteins that are associated with cadherins and are components of signal transduction pathways(50) did not differ in Co and Ap cells (Fig. 2C). These results indicate that the FGFR-2 mutation induces overexpression of N- and E-cadherin mRNA and protein levels at early time points in mutant osteoblasts in vitro.
To further document this point at the cellular level, the expression of cadherins was investigated by immunocytochemistry in confluent Ap cells in comparison with Co cells. Immunoreactivity for N-cadherin (Figs. 3C and 3D) and E-cadherin (Figs. 3E and 3F) was localized at the cellular membrane, as expected. Co cells showed no specific staining (Figs. 3A and 3B). The immunostaining showed that N-cadherin was much stronger in Ap cells (Fig. 3D) compared with Co cells (Fig. 3C). Similarly, E-cadherin immunostaining was higher in Ap cells (Fig. 3F) compared with Co cells (Fig. 3E), confirming the Western blot analysis (Fig. 2). In contrast, immunostaining with an anti-N-CAM antibody gave a similar signal in Ap (Fig. 3H) and Co cells (Fig. 3G). In addition, α- and β-catenin immunostaining did not differ in Ap and Co cells (not shown), indicating that the mutation altered mainly N- and E-cadherin synthesis. When Ap cells were cultured in aggregates, a similar increase in N-cadherin and E-cadherin immunostaining was found in mutant cells compared with Co cells (data not shown). These data indicate that mutant cells overexpress N- and E-cadherins in vitro.
The S252W FGFR-2 mutation increases N- and E-cadherin expression in vivo
To confirm the alteration of cadherin expression induced by the mutation in vivo, the expression of cadherins was analyzed by immunohistochemistry in coronal sutures from an Ap fetus affected by the S252W FGFR-2 mutation and an age-matched normal fetus. The normal suture consisted of a thin bone margin (Fig. 4B) whereas the fused fetal Ap coronal suture consisted of large bony trabeculae (Fig. 4C), reflecting the premature ossification induced by the mutation. In the normal suture, N-cadherin (Fig. 4B) and E-cadherin immunostaining (Fig. 4F) was faint in mesenchymal cells and was abundant in preosteoblasts and osteoblasts along the bone matrix. In the Ap suture, strong immunostaining for N-cadherin (Fig. 4C) and E-cadherin (Fig. 4G) was present in preosteoblasts and osteoblasts. In contrast, N-CAM immunostaining in preosteoblasts and osteoblasts was similar in the normal (Fig. 4J) and Ap (Fig. 4K) sutures. Immunostaining for α- and β-catenins in the Ap suture did not differ from normal (not shown). Control sections of the normal suture (Figs. 4A, 4E, and 4I) and Ap suture (Figs. 4D, 4H, and 4L) showed no specific staining. We also analyzed FGFR-2 immunostaining in Ap suture, using an anti-FGFR-2 that recognizes the COOH-terminus of the human receptor. FGFR-2 immunostaining was present in preosteoblasts and osteoblasts (Fig. 4N) that overexpress N-cadherin (Fig. 4C) and E-cadherin (Fig. 4G), indicating that FGFR-2 is present where the cadherins are overexpressed. The control section incubated with anti-FGFR-2 antibody neutralized with the blocking peptide showed no specific staining (Fig. 4M). In separate immunostaining experiments, we found that preosteoblasts and osteoblasts in the Ap suture also overexpress type 1 collagen, OC, and osteopontin compared with the normal suture.(31) These data indicate that the mutation induces overexpression of N- and E-cadherins in vivo and this is associated with increased osteoblast gene expression in the Ap suture.
N- and E-cadherins are involved in mutant osteoblastic cell-cell aggregation
To evaluate the functional role of N- and E-cadherins in mutant cells, cell-cell adhesion was determined in Ap cells using a cell aggregation assay(46) previously validated for human calvaria osteoblasts.(38) As shown in Fig. 5A, Ap cells formed more compact aggregates at 2 days of culture compared with Co cells. Further quantification of single cells and clusters in the whole cultures showed that cell-cell aggregation was greater in Ap cells than in Co cells (Fig. 5B). To determine the role of N- and E-cadherins in this effect, we tested the effects of specific neutralizing anti-N- and anti-E-cadherin antibodies. The neutralizing anti-N- or anti-E-cadherins either alone or in combination reduced cell-cell adhesion in Ap cells whereas a nonspecific IgG had no significant effect (Fig. 5C). Inhibition of N- or E-cadherin function by these specific antibodies in Ap cells reduced cell-cell aggregation to the Co levels in Co cells (Fig. 5C). These data support the hypothesis that N- and E-cadherins are involved in the increased cell-cell adhesion induced by the mutation.
Increased N-cadherin expression is involved in osteoblast gene activation in mutant cells
To determine the role of N- and E-cadherin overexpression in the activation of osteoblast gene expression induced by the S252W mutation, we used specific anti-N- and anti-E-cadherin neutralizing antibodies and measured ALP activity and osteoblast marker mRNA levels in Ap cells cultured in monolayers. The addition of anti-N-cadherin antibody reduced by 30% the mutation-induced increase in ALP activity in Ap cells whereas a nonspecific IgG had no significant effect (Fig. 6A). In contrast to the anti-N-cadherin antibody, the anti-E-cadherin antibody alone had no effect and did not affect the inhibitory effect of anti-N-cadherin antibody on ALP activity (Fig. 6A). To further document the role of cadherins in the activation of ALP activity in mutant cells, we used an AS strategy to reduce the synthesis of N- or E-cadherins. Western blot analysis showed that N-cadherin and E-cadherin AS oligonucleotides reduced N- and E-cadherin protein levels by about 20% in Ap cells, compared with R oligos (Fig. 6B). Treatment of Ap cells with N-cadherin AS, but not with R oligos, reduced the mutation-induced increase in ALP activity in Ap cells by about 20% compared with Co cells (Fig. 6C). This was not caused by a toxic effect of oligonucleotides because AS-treated Ap cells did not show alteration of adherence or viability, measured by trypan blue exclusion and cell detachment, which were only affected at high dosage levels of oligos (50 μM). In contrast to the N-cadherin AS, E-cadherin AS had no significant effect on ALP activity in mutant cells. Furthermore, the combination of N- and E-cadherin AS oligos reduced ALP activity as much as with N-cadherin AS alone (Fig. 6C), confirming the inhibitory effect of the anti-N-cadherin neutralizing antibody (Fig. 6A). These data indicate that inhibition of N-cadherin but not E-cadherin synthesis reduces the increased ALP activity in mutant cells.
To further assess the involvement of N- or E-cadherin in the activation of osteoblast gene expression induced by the mutation, the expression of ALP and COLIA1 mRNA was determined in Ap cells cultured in the presence of specific neutralizing anti-N- or anti-E-cadherin antibodies. The anti-N-cadherin antibody decreased ALP mRNA levels compared with nonspecific IgG, whereas anti-E-cadherin antibody had a minor effect (Fig. 6D). The combination of anti-N- and anti-E-cadherin antibodies had the same effect as anti-N-cadherin alone on ALP mRNA (Fig. 6D), confirming the results on ALP activity (Fig. 6A). Similarly, the anti-N-cadherin antibody reduced COLIA1 mRNA levels in Ap cells, whereas the anti-E-cadherin antibody alone had no effect (Fig. 6D). These results indicate that inhibition of N-cadherin but not E-cadherin reduces the ALP and COLIA1 mRNA overexpression in mutant osteoblasts.
S252W FGFR-2 mutation increases PKCα phosphorylation and PKC activity
To determine the signaling pathways stimulated by the mutation that may be involved in the activation of N-cadherin and osteoblast genes in mutant cells, phosphorylation of signaling molecules was investigated in Ap cells in basal culture conditions using immunoprecipitation and Western blotting analysis. In the absence of serum, Ap cells showed an increased amount of tyrosine phosphorylated PLCγ occurring at 0-6 h. In contrast, basal levels of phosphorylated p38 MAPK and phosphorylated ERK-1 and -2 did not differ in Ap and Co cells (Fig. 7A). Because the signaling pathway of PLCγ may involve PKC,(12) we looked for possible phosphorylation of PKC isoforms in mutant cells using available specific anti-phospho-PKCα, -PKCδ, and -PKCϵ antibodies. PKCδ was not detected in the two cell lines (not shown). Phosphorylated PKCα levels were increased in Ap cells compared with Co cells in basal conditions (Fig. 7A). Scanning densitometry revealed that phospho-PKCα was increased about 2-fold in Ap cells compared with Co cells. In contrast, basal phosphorylated PKCϵ levels were similar in Ap and Co cells. Although phosphorylated PKCϵ tended to increase with time in Co cells, this was not the case in Ap cells (Fig. 7A). To determine the activation of PKC in Ap cells, total PKC activity was measured by the transfer of [32P]-labeled phosphate to a biotinylated peptide substrate (AAKIQASFRGHMARKK) that is specific for PKC activity.(49) As shown in Fig. 7B, PKC activity was higher in Ap cells compared with Co cells. These results show that the S252W mutation increases PLCγ and PKCα phosphorylation and increases PKC activity in mutant osteoblasts.
Involvement of PKC in N-cadherin and osteoblast gene overexpression in mutant cells
To identify more specifically the role of PKC in the overexpression of N-cadherin in mutant cells, Ap cells were treated with agents that are specific inhibitors of signaling pathway molecules,(15, 39-43, 51) and N-cadherin mRNA levels were then determined. We used calphostin C, a potent and selective inhibitor of PKC,(39) whose action may be restricted specifically to the membrane compartment where it interacts with the PKC regulatory domain.(51) We also used PD-98059, a specific MEK inhibitor,(40) and SB 203580, a specific inhibitor of p38 MAPK.(41) N-cadherin mRNA levels were reduced by calphostin C (1 μM) and, to a lesser extent, by SB 203580 (20 μM) in Ap cells. Both compounds reduced N-cadherin mRNA to levels lower than the levels in Co cells. In contrast, PD-98059 (20 μM) had no effect on N-cadherin mRNA levels (Fig. 8A). The Western blot analysis showed that calphostin C and the p38 MAPK inhibitor reduced N-cadherin protein levels almost to the levels in Co cells (Fig. 8B). This indicates that inhibition of PKC and, to a lesser extent p38 MAPK, reduces the N-cadherin overexpression in mutant cells.
To evaluate the role of signaling molecules in osteoblast gene overexpression in mutant cells, we tested the effect of the different specific inhibitors on ALP, COLIA1, and OC mRNA expression in Ap cells. We found that calphostin C (1 μM) reduced ALP and COLIA1 mRNA overexpression and restored OC mRNA in Ap cells to the levels in Co cells (Fig. 8A). The p38 MAPK inhibitor SB 203580 had distinct effects because it reduced ALP and OC, but not COLIA1 mRNA levels (Fig. 8A). In contrast, PD-98059 had no effect on ALP and COLIA1 but reduced OC mRNA levels (Fig. 8A). Thus, the three signaling inhibitors induced distinct effects on ALP, COLIA1, and OC expression in Ap cells, as expected from their distinct activity on the signaling pathways. However, only calphostin C suppressed the overexpression of N-cadherin, ALP, COLIA1, and OC in mutant osteoblasts, suggesting an important role of PKC in the control of these genes.
To further assess the possible implication of PKC and the PKCα isoform in the increased N-cadherin and osteoblast markers induced by the mutation, we studied the effect of the staurosporine-related PKC inhibitor Gö6976, which is the most potent inhibitor available for inhibition of PKCα activity,(42) and the less specific PKCα inhibitor Gö6983,(43) in comparison with calphostin C. As shown in Fig. 9A, treatment of Ap cells with calphostin C (0.5-2 μM for 24 h) induced a dose-dependent inhibition of N-cadherin mRNA levels, which were reduced below the levels in Co cells. The PKCα-specific inhibitor Gö6976 (0.1-1 μM) also reduced N-cadherin mRNA levels, whereas Gö6983, a less sensitive PKCα inhibitor, was less effective (Fig. 9A). The effects of the three PKC inhibitors also were determined on N-cadherin protein levels in Ap cells at 48 h. Western blot analysis showed that calphostin C (2 μM), Gö6976 (1 μM), and Gö6983 (1 μM) reduced N-cadherin protein levels in Ap cells to or below the levels in Co cells (Fig. 9B). Furthermore, the three PKC inhibitors inhibited osteoblast gene expression in Ap cells. As shown in Fig. 9C, both calphostin C and Gö6976 reduced ALP mRNA levels in Ap cells to the levels in Co cells at 24 h, whereas the less sensitive PKCα inhibitor Gö6983 was less effective. In addition, the three PKC inhibitors reduced COLIA1 mRNA levels to the levels in Co cells (Fig. 9D). These results show that calphostin C and the PKCα-specific inhibitor Gö6976 suppressed the overexpression of N-cadherin, ALP, and COLIA1 induced by the S252W FGFR-2 mutation in osteoblasts, which supports a role for PKCα in the overexpression of these genes in Ap cells.
In the present study, we used mutant osteoblasts to investigate the role of specific cadherins and FGFR signaling pathways that may be involved in the premature osteoblast differentiation induced by the Ap S252W FGFR-2 mutation. Our studies indicate that N-cadherin, a molecule playing an essential role in cell-cell adhesion, is involved in the activation of osteoblast marker genes by the S252W mutation in vitro and in vivo. Furthermore, our data support the hypothesis that PKCα plays a role in the increased N-cadherin and osteoblast gene expression induced by the FGFR-2 mutation in mutant osteoblasts.
We used immortalized mutant osteoblastic cells cultured in aggregates to allow osteogenesis in long-term culture.(30, 38, 45) In this system, ALP and type 1 collagen level decreased whereas calcium content in the matrix increased, reflecting the changes observed during human calvaria osteogenesis in vitro.(38) We showed that the S252W FGFR-2 mutation increases osteoblast marker gene expression such as ALP, COLIA1, and OC. These phenotypic properties of the Ap cell line are the same as those of primary culture calvaria cultures.(30) Moreover, the in vitro alteration of the osteoblast phenotype induced by the mutation is relevant to the in vivo situation because we also found premature expression of ALP, type 1 collagen, and OC expression in preosteoblasts in the fused Ap suture.(31) This increased osteoblast differentiation phenotype is not peculiar to the FGFR-2 S252W mutation because we(30) and others(52) found that mutation of the adjacent amino acid P253R in FGFR-2, the other prevalent mutation found in Ap syndrome, also activates osteoblast differentiation markers in mutant calvaria cells. To investigate the cellular mechanisms that link the S252W mutation to the observed alterations of osteoblast differentiation, we evaluated the changes in cell-cell adhesion molecules that play essential roles in the control of morphogenesis and mesenchyme condensation.(32, 33, 50) Our results show that E- and N-cadherins were expressed at the early time points of aggregate cultures and their expression decreased with time in culture, which is consistent with an important role in early cell-cell interactions. In addition, cultured mutant osteoblasts constitutively displayed increased N- and E-cadherin mRNA and protein levels in basal culture conditions. This was not related to artifactual cell culture conditions, because more abundant immunoreactivity for N- and E-cadherins also were found in nonimmortalized Ap calvaria cells (data not shown) as well as in preosteoblasts and osteoblasts in the fused Ap suture in vivo. We could not detect changes in other molecules involved in cell-cell adhesion, such as N-CAM or α- and β-catenins, indicating a relative specificity of effect of the mutation on N- and E-cadherins in mutant cells. Furthermore, we found that the mutation increased cell-cell aggregation and that specific neutralizing anti-N- and anti-E-cadherin antibodies restored cell-cell aggregation to control levels. This supports the hypothesis that the increased N- and E-cadherin expression is involved in the increased cell-cell adhesion in mutant cells. To identify the possible role of N- and E-cadherins in the alterations of the osteoblast phenotype induced by the FGFR-2 mutation, we used specific antibodies to inhibit the function of each molecule. Neutralizing anti-N-cadherin antibody reduced ALP, COLIA1, and OC mRNA expression in mutant osteoblasts, whereas the anti-E-cadherin antibody had no effect. Thus, N-cadherin but not E-cadherin appears to be important in the activation of these osteoblast marker genes in mutant cells. Because N- and E-cadherins are expressed at relatively high levels in these cells, it was not possible to suppress completely the expression of these proteins using AS oligonucleotides. Nevertheless, we found that N-cadherin AS oligos that reduced the expression of N-cadherin protein levels decreased significantly the mutation-induced increase in ALP activity in Ap cells, whereas reduction of E-cadherin levels by specific AS oligos had no effect. These data confirm that N-cadherin appears to be important in the activation of osteoblast markers by the S252W FGFR-2 mutation. This is consistent with recent data indicating that N-cadherin plays a role in the expression of ALP activity(36) and osteoblast differentiation marker gene expression in human osteoblasts.(37, 38) The mechanisms by which N-cadherin overexpression may control osteoblast marker genes in mutant cells are unclear at the present time. One possibility is that the increased cell-cell adhesion may lead to increased cell-cell communication and to coordinated gene expression in osteoblasts.(53) Alternatively, N-cadherin itself may participate in some signal transduction events(50) resulting in the observed increased osteoblast differentiation in mutant Ap cells, and this hypothesis awaits further investigation.
We also analyzed the signaling pathways linking the FGFR-2 mutation to the activation of N-cadherin and osteoblast marker genes in mutant cells. Our data indicate that the S252W FGFR-2 mutation increased PKC activity constitutively in mutant osteoblasts, which is consistent with the observed PLCγ phosphorylation. PKC proteins are composed of several kinase isoforms that exhibit distinct tissue expression pattern(13, 14) and some of them were found previously in osteoblastic cells.(54) Using the presently available specific antibodies against phosphorylated PKC isoforms, we found expression of PKCα and PKCϵ but not PKCδ, in Co and Ap calvaria osteoblasts. Interestingly, only PKCα phosphorylation was found to be increased in mutant cells compared with Co cells. Phosphorylation of at least two specific sites of the catalytic domain of PKCα is believed to contribute to the activity of the enzyme.(55, 56) However, the activation of PKC isotypes involves several complex mechanisms,(55–57) making the determination of PKCα activation difficult.(58) Although DAG-activated PKC can be restricted to membrane compartments, membrane localization may not be appropriate to monitor PKCα activation because PKC isotypes can associate with cell membrane before PKC stimulation.(58) A slight increase in membrane-associated PKCα was reported recently in calvaria cells with another (P253R) Ap FGFR-2 mutation. However, no data were available on the direct implication of this PKC isoenzyme in the osteoblast phenotype and it is unknown whether this represents the cause or the consequence of this particular FGFR-2 mutation.(52) In the present study, we examined the role of PKCα in the phenotypic alterations of osteoblasts with the S252W FGFR-2 mutation using two different PKC inhibitors calphostin C and Gö6976, which recognize distinct functional domains of PKCα. Calphostin C is targeted to the phorbol ester binding site of the PKCα regulatory domain and may inhibit only PKCα that is localized specifically in the membrane compartment,(51) whereas the potent PKCα-specific inhibitor Gö6976 recognizes a target site in the PKCα catalytic domain.(42) Our data show that the two inhibitors reduced the increased N-cadherin, ALP, and COLIA1 gene expression in mutant osteoblasts to control levels. Gö6983 was less effective than Gö6976, which is consistent with its less potent activity on PKCα.(43) In parallel experiments, we found that PKC inhibitors do not significantly affect N-cadherin expression in Co cells, as assessed by Western blot analysis (data not shown). This supports the hypothesis that PKCα may be implicated in the overexpression of N-cadherin and osteoblast genes induced by the mutation in Ap cells. Although our data are consistent with a role of PKCα in the N-cadherin and osteoblast gene overexpression induced by the activating S252W FGFR-2 mutation, we cannot rule out that other PKC isoforms, not yet identified, also may be involved in the increased osteoblast phenotype in mutant cells, because several PKC isoforms may play critical roles in the control of cell differentiation in different cell systems.(59) Recent data indicate that PKC may play a key role in chondrogenic differentiation and that PKC signaling is coupled with the regulation of N-cadherin during chondrogenesis.(60) However, in this study inhibition of PKC resulted in sustained expression of N-cadherin in chondrocytes.(60) In contrast, we found that specific inhibition of PKC reduced N-cadherin and osteoblast marker genes overexpression in mutant calvaria cells. Thus, distinct regulatory mechanisms and signaling pathways appear to be involved in the expression of this cell adhesion molecule in chondroblasts and osteoblasts.
Several FGF/FGFR signaling pathways are involved in the control of cell proliferation and differentiation.(3, 4, 10–12) In osteoblastic cells, recent data indicate that FGF-2 activates ERK(61) and p38 MAPKs.(48, 62) However, neither ERK nor p38 MAPK phosphorylation was found to be increased by the activating S252W FGFR-2 mutation in mutant osteoblasts in basal culture conditions, suggesting that these pathways are not involved in the phenotype induced by the mutation. However, ALP, COLIA1, and OC gene expression may be controlled by ERK, p38 MAPK and PKC in osteoblasts,(48, 61–64) and inhibition of these signaling pathways may affect these genes. Indeed, we found that inhibition of ERK by PD-98259 reduced OC mRNA in Ap cells but not N-cadherin mRNA or protein levels, suggesting that this MAPK may play a role in the activation of OC gene expression independently of N-cadherin. This is supported by recent data showing the direct involvement of the MAPK pathway in OC gene transcription in murine preosteoblastic cells.(64) We also found that the p38 MAPK inhibitor SB 203580 reduced ALP and OC expression in Ap cells, suggesting a role for p38 MAPK in the control of these genes. This is consistent with the involvement of p38 MAPK in the up-regulation of ALP activity in murine osteoblasts.(48, 62) Although more than one signaling pathway is likely to be involved in the regulation of osteoblast marker genes, the finding that N-cadherin, ALP, COLIA1, and OC gene expressions were reduced to control levels only by specific PKC inhibitors supports the hypothesis that PKC plays a role in the overexpression of N-cadherin and osteoblast genes in Ap mutant osteoblasts.
In summary, the present study indicates that N-cadherin plays a role in the increased human calvaria osteoblast differentiation induced by the S252W FGFR-2 mutation. Moreover, the results suggest that PKCα is involved in the activation of N-cadherin and osteoblast marker gene expression in Ap mutant osteoblasts. The identification of this PKC isoform will allow us to investigate the downstream signaling pathways and the molecular events that control the alterations of N-cadherin and osteoblast markers in mutant osteoblasts.