Although dura mater tissue is believed to have an important role in calvarial reossification in many in vivo studies, few studies have shown the direct effect of dura mater cells on osteoblasts. In addition, no reports have yet identified the potential factor(s) responsible for various biological activities exerted by dura mater on calvarial reossification (e.g., cell proliferation). In this study, we tested the effect of dura mater on calvarial-derived osteoblasts by performing both heterotypic coculture and by culturing osteoblast cells with conditioned media harvested from dura mater cells of juvenile (3-day-old) and adult (30-day-old) mice. The results presented here demonstrate that cellular proliferation of juvenile osteoblast cells was significantly increased by juvenile dura mater either in the coculture system or when dura mater cell-conditioned medium was applied to the osteoblast cells. Moreover, high levels of FGF-2 protein were detected in juvenile dura mater cells and their conditioned medium. In contrast, low levels of FGF-2 protein were detected in adult dura mater cells, whereas FGF-2 protein was not detectable in their conditioned medium. Abrogation of the mitogenic effect induced by juvenile dura mater cell-conditioned medium was achieved by introducing a neutralizing anti-FGF-2 antibody, thus indicating that FGF-2 may be responsible for the mitogenic effect of the juvenile dura mater. Moreover, data obtained by exploring the three major FGF-2 signaling pathways further reinforced the idea that FGF-2 might be an important paracrine signaling factor in vivo supplied by the underlying dura mater to the overlying calvarial osteoblasts.
- mitogenic activity
- fibroblast growth factor-2
successful calvarial reossification is a character generally restricted to immature animals and infants younger than 2 years of age (12, 30). Glaser and Blaine (9) reported that calvariectomies in children younger than 2 years completely reossify. Similarly, in vivo animal experiments of trephination have also demonstrated that calvarial reossification is generally restricted to juvenile animals (30). Because of the paracrine relationship, most research on calvarial healing has focused on the osteoblasts in the skull and the underlying dura mater. Although studies on osteoblasts have indicated that juvenile osteoblast (JOb) cells display much higher intrinsic proliferation and differentiation than adult osteoblasts (AObs) (4), many other studies have suggested that the dura mater tissue underlying the calvarial bones also plays an important role in calvarial osteogenesis and reossification (30). In vivo studies, like the trephination in animals, have demonstrated that dura mater is an imperative component for any successful calvarial osteogenesis (13, 23). In clinical situations, complete reossification of calvarial defects in children younger than 2 yr needs an intact underlying dura mater (24). The dura mater has been hypothesized to contribute to calvarial reossification in many ways, including providing a source of osteoblastic precursors and/or supplying osteogenic cytokines (29, 35, 37). A gene expression study by Greenwald et al. (10) showed a group of upregulated osteogenic genes in juvenile rat dura mater cells, including transforming growth factor β1, fibroblast growth factor 2 (FGF-2), osteocalcin, and collagen Iα1 (10). Paracrine and autocrine patterns of the cytokine secretion by dura mater cells have been postulated (1, 31, 32). The direct influence of dura mater cells on osteoblasts has also been observed in rats with the use of a coculture model system (31). However, the specific factor(s) and signaling underlying the biological activity exerted by dura mater on osteoblasts have not yet been fully elucidated.
FGF-2, a strong mitogen for a wide variety of cell types (8), also promotes cell proliferation of osteoblasts both in vivo and in vitro (15, 20, 28). To exert biological functions, FGF-2 binds to tyrosine kinase receptors resulting in phosphorylation of downstream signaling proteins. Three signal-transduction pathways have been widely reported to be involved in FGF-2 signaling, including mitogen-activated protein kinase (MAPK), phosphatidylinositol 3-kinase (PI3-kinase), and PKC pathways (7), which all exist in osteoblasts (2, 3, 5, 6, 17–19, 36). In this study, using a combination of a heterotypic coculture model system (between juvenile dura mater and JOb-AOb) and osteoblast culture with conditioned media prepared from dura mater cells, we have gained insight into molecular mechanisms governing the mitogenic activity induced by dura mater on osteoblasts. Our data suggest that FGF-2 is an important driving force responsible for the proliferative activity elicited by dura mater on JObs. In addition, the results indicate that the dura mater's mitogenic influence is mediated through the activation of the three major FGF-2 signaling pathways.
MATERIALS AND METHODS
All experiments using animals were performed in accordance with an approved protocol reviewed by the Stanford University Animal Care and Use Committee. CD-1 Webster mice were purchased from Charles River Laboratories (Wilmington, MA). The date of birth was considered the first day of life. Animals were housed in light- and temperature-controlled rooms and were given food and water ad libitum.
Tissue harvest and primary cell culture.
JOb and AOb or juvenile and adult dura mater cells (JDM and ADM cells) were harvested from skulls of 3- and 30-day-old mice, respectively. Changes and modifications were made to the methods described previously (4, 11). The periosteum was meticulously stripped off, and the suture mesenchyme was carefully dissected away. The dura mater was gently removed and digested with 0.05% trypsin (Cellgro, Herndon, VA) at 37°C for 5 min. The digestion was neutralized with an equal volume of culture medium: α-MEM (GIBCO Life Technologies and Invitrogen, Carlsbad, CA) supplemented with 10% FCS (Gemini Bioproducts, Woodland, CA) and 100 IU/ml of penicillin and streptomycin (Invitrogen). Skulls depleted of periosteum, suture mesenchyme, and dura mater were digested with 0.2% dispase II and 0.1% collagenase A (Roche Diagnostics, Indianapolis, IN) in serum-free medium. The digestion was repeated six times for 10 min each for skulls from juvenile mice and 15 min each for skulls from adult mice at 37°C in a water bath shaker. The first two digestions were discarded. The latter four digestions were pooled, pelleted, and resuspended in the α-MEM supplemented with 10% FCS. Both dura mater and osteoblast cells were plated in 100-mm tissue culture dishes (Corning, San Mateo, CA) and maintained at 37°C in an atmosphere of 5% CO2. The media were changed every other day. Osteoblasts or dura mater cells were passaged by trypsinization. Only cells from passages 0 and 1 were used for all experiments.
Coculture experiments were carried out as follows: JOb and AOb cells were plated onto six-well Transwell cell culture chamber plates (5 × 104/well) (Corning Costar, Cambridge, MA). Equal numbers of JDM and ADM cells were plated onto the Transwell cell culture chamber polyester inserts; the inserts had a pore size of 0.4 μm. This pore size would allow the diffusion of soluble factors between the two cell populations but would prevent the transfer of any cells or organelles. The two cell populations were cultured separately for 3 days with α-MEM supplemented with 10% FCS, 100 IU/ml penicillin, and 100 IU/ml streptomycin. At day 3, the inserts containing the dura mater cells were combined with the osteoblast cells. Once the inserts were placed in the well, all cells were maintained with α-MEM supplemented with 1% FCS, 100 IU penicillin, and 100 IU streptomycin. Four different combinations of heterotypic cocultures were set up: JOb-JDM, AOb-JDM, JOb-ADM, and AOb-ADM. As control, JOb and AOb cells were cultured alone (i.e., empty Transwell inserts were combined with osteoblast cells). After 3 and 6 days, coculture cells were trypsinized and suspended in 300 μl of PBS. Cell counting was conducted by hemacytometer. All experiments were run in triplicate.
Cell growth was assessed by counting the total number of cells on days 3 and 6. For the coculture experiments designed to test the paracrine effect of FGF-2 produced by dura mater cells on JObs, JOb cells were preincubated for 12 h with 2 μg/ml anti-FGF-2 antibody in serum-free medium to abrogate their endogenous/autocrine FGF-2 activity. The same concentration of anti-FGF-2 antibody was added to the medium after the two different types of cells (JOb-JDM) were combined in the Transwells for the coculture. Coculture without anti-FGF-2 antibody was used as control. The cell growth was measured by cell counting at days 3 and 6. The coculture method was also used to investigate the signaling pathways induced by FGF-2. JOb cells were cocultured with JDM cells. Three inhibitors (Calbiochem-Novabiochem, San Diego, CA) of the major cellular signaling pathways were used as follows: 10 μM U-0126 to block the MAPK pathway, 20 μM LY-294002 to block the PI3-kinase pathway and 2 μM Gö-6983 to block the PKC pathway. After confluence was reached, cells were washed twice with 1× PBS before any treatment. Osteoblasts were starved in serum-free medium for 12 h before the treatments. Incubation with inhibitors was carried out for 12 h, and bromodeoxyuridine (BrdU) incorporation was assessed at the same time using a BrdU kit (Roche Diagnostics) according to the manufacturer's instructions. Osteoblasts without inhibitor were used as controls. All experiments were performed in triplicate.
The results are presented as the means ± SD of three independent experiments. Statistical differences between the means were examined by Student's t-test. P < 0.05 was considered statistically significant.
Preparation of cell-conditioned medium.
Cell-conditioned media were obtained from either JDM-derived cell-conditioned medium (JCM) or ADM-derived cell-conditioned medium (ACM) cells. Two types of medium were used: serum-free medium or medium supplemented with 10% FCS. Briefly, dura mater cells were plated in six multiwell plates (1.5 × 105/well), and 2 ml of either serum-free or 10% FCS-supplemented medium were added to each well. To prepare serum-free conditioned medium, cells were washed three times with sterile PBS before addition of serum-free medium. After 24 h, the cell-conditioned media were collected and concentrated 10× with Centricon filters (Centricon-3, 3000 NMWL; Millipore, Bedford, MA). Collection and concentrating of the media were carried out at 4°C. All experiments were performed three times with freshly harvested media.
Cell proliferation assays.
The growth rates of JOb and AOb cells were assessed by BrdU incorporation and immunoblotting analysis of proliferating cellular nuclear antigen (PCNA). For the BrdU labeling assay, JOb and AOb cells were plated at 1,500 cells/well, in 96-multiwell culture plates with flat bottoms (Corning). For short-term experiments, serum-free cell-conditioned media were used. Cells were washed twice with sterile PBS and starved in serum-free α-MEM overnight before the cell-conditioned media were applied. BrdU incorporation was carried out for 24 h (Roche Diagnostics) according to the manufacturer's instructions. Photometric detection was done with an ELISA reader at 370-nm wavelength. The background was subtracted when the resulting data were processed. For long-term experiments, cell-conditioned media supplemented with 10% FCS were used, and the cell proliferation assay was conducted at 4, 8, 12, 16, and 20 days. Each time point was run in triplicate.
Western blotting analysis of PCNA was performed on cell lysates prepared at different time points from JOb and AOb using RIPA buffer (Sigma-Aldrich, St. Louis, MO). Total cellular protein was quantified by bicinchoninic acid protein assay (Pierce). Fifty micrograms of total cellular protein were resolved by 10% Tris·HCl SDS-PAGE gels (Precast Nupage gels; Invitrogen Life Technologies) and transferred onto Immobilon-P membranes (Millipore). Membranes were probed with anti-PCNA antibody (FL-261 sc7970, rabbit polyclonal 1:400; Santa Cruz Biotechnology, Santa Cruz, CA). A horseradish peroxidase (HRP)-conjugated anti-rabbit antibody (1:8,000; Santa Cruz Biotechnology) was used as secondary antibody. Immunoblotted products were visualized by enhanced chemiluminescence substrate (Amersham Biosciences). Subsequently, the membranes were stripped of the antibodies by washing in stripping solution (62.5 mmol/l Tris·HCl, pH 7.5, 2% SDS, 100 mmol/l β-mercaptoethanol) for 1 h at 37°C followed by three washes with 20 mM Tris and 136 mM NaCl, pH 7.6. Membranes were then incubated with α-tubulin antibody (B-7 sc-5286, mouse monoclonal 1:600; Santa Cruz Biotechnology) to control for equal loading and transfer of the samples.
Quantification of FGF-2 protein in dura mater cells and conditioned medium.
To quantify the FGF-2 present in dura mater cells and their cell-conditioned media, a double sandwich ELISA was employed. All antibodies and detection reagents were purchased from R&D Systems (Minneapolis, MN). The assay was performed according to the manufacturer's instructions. Briefly, high-binding 96-well ELISA plates (BD Biosciences, Bedford, MA) were coated with the capture antibody, a biotinylated, monoclonal anti-human FGF-2 antibody at a concentration of 2 μg/ml. Serum-free cell-conditioned media were prepared as described before. Briefly, JDM and ADM cells were plated in six multiwell plates (1.5 × 105/well), and 2 ml of serum-free medium were added to each well. After culture for 24 h, the conditioned media were harvested from cells; cells were then counted, and the media were normalized for cell numbers and concentrated 10×. Then 100 μl of concentrated cell-conditioned media (JCM or ACM) were applied to the wells coated with the capture antibody and incubated at room temperature for 6 h. For cell lysates (JDM and ADM), protein concentrations were measured and total proteins of each lysates were calculated. Then, 50 μg of cell lysates obtained from JDM and ADM cells were applied to the wells coated as described above, and the results were also normalized to cell numbers. Each sample was run in triplicate. The optical density at 450 nm was measured with an ELISA microplate reader. The resulting optical density readings were normalized against a recombinant human FGF-2 protein (rhFGF-2) (R&D Systems) standard curve.
Neutralization of FGF-2 bioactivity.
A polyclonal goat anti-FGF-2 antibody at a concentration of 0.8 μg/ml (R&D Systems) was used for FGF-2 neutralization. Serum-free JCM and ACM were preincubated with 0.8 μg/ml of the anti-FGF-2 antibody for 3 h at 4°C before application on osteoblast cells. Osteoblast proliferation was determined by BrdU incorporation assay as described above. Normal rabbit IgG (sc2027; Santa Cruz Biotechnology) was used as negative control.
Characterization of signaling pathways activation.
Cellular signaling transduction was investigated in JOb cells treated either with JCM or 20 ng/ml of rhFGF-2 (Santa Cruz Biotechnology). Three inhibitors (Calbiochem-Novabiochem) of the major cellular signaling pathways were used as follows: 10 μM U-0126 to block the MAPK pathway, 20 μM LY-294002 to block the PI3-kinase pathway, and 2 μM Gö-6983 to block the PKC pathway. After confluence was reached, JOb cells were starved in serum-free medium overnight. Cells were washed twice with 1× PBS before any treatment. In the experiments examining FGF-2 signaling pathways, the cells were preincubated for 30 min with one of the three inhibitors described above, dissolved in 0.1% DMSO (vehicle). To determine the nonspecific contribution of DMSO to FGF-2 signal inhibition, a separate group of cells was incubated with DMSO alone. Treatments were carried out for 15 min, 30 min, and 3 h. After the treatments, the cells were washed twice with ice-cold PBS and lysed with cold lysis buffer (50 mmol/l HEPES, pH 7.5, 150 mmol/l NaCl, 1 mmol/l EDTA, 10% glycerol, 1% Triton X-100, 25 mmol/l sodium fluoride) containing 1 mmol/l sodium orthovanadate and protease inhibitor cocktail (Sigma-Aldrich). Cell lysates were assayed for protein concentration by bicinchoninic acid assay as described above. Aliquots (50–100 μg) of cell lysate were electrophoresed on 12% Tris·HCl SDS-PAGE gels (Precast Nupage gels; Invitrogen, Life Technologies) and transferred onto Immobilon-P membrane (Millipore). Immunoblotting analysis was performed with the following antibodies: anti-phospho-ERK p44/42 (Thr202/Tyr204 rabbit polyclonal antibody, 1:1,000; Cell Signaling Technology, Beverly, MA), anti-phospho-Akt (Ser473 rabbit monoclonal antibody, 1:1,000; Cell Signaling), anti-phospho-PKC-αβ (Thr638/641 rabbit polyclonal antibody, 1:1,000; Cell Signaling), anti-phospho-PKC-δ (Thr505 rabbit polyclonal antibody, 1:1,000; Cell Signaling), anti-ERK2 (C-14 mouse monoclonal antibody, 1:400; Santa Cruz Biotechnology), anti-Akt (rabbit polyclonal antibody, 1:1,000; Cell Signaling), and anti-PKC/PKD (rabbit polyclonal antibody, 1:1,000; Cell Signaling). An HRP-conjugated anti-rabbit antibody (1:8,000) and an HRP-conjugated anti-mouse antibody (1:8,000; Santa Cruz Biotechnology) were used as secondary antibodies. Immunoblotted products were visualized by use of an enhanced chemiluminescence substrate (Amersham Biosciences). For the densitometric analysis of immunoblotting for the signaling pathway experiments, bands were scanned and quantified with the use of the ImageJ program (ImageJ 1.36b; National Institutes of Health, Bethesda, MD). The densitometric results of both phosphorylated and nonphosphorylated signaling proteins were normalized to their respective loading control (α-tubulin bands) and presented as percent increase (see ⇓⇓⇓Fig. 4).
FGF-2 signaling transduction pathways were also examined with a cell proliferation assay. JOb cells were incubated with inhibitors and stimulated either with 20 ng/ml of rhFGF-2 or JCM for 12 h. Inhibitors were used at the same concentration as above. After exposure to the inhibitors, BrdU assay was performed as described above to assess the effect of these inhibitors on cellular proliferation.
Dura mater-osteoblast cell heterotypic coculture.
To determine whether dura mater affected osteoblast proliferation, we established a heterotypic coculture system in which mouse osteoblasts were cultured with mouse dura mater cells or cultured alone. The heterotypic cocultures were performed with cells obtained from juvenile and adult mice. Relative cell proliferation was determined by counting osteoblasts at 3 or 6 days of culture and comparing results with controls (same number of JOb or AOb cells alone cultured with α-MEM supplemented with 1% FCS). As shown in Fig. 1A, heterotypic coculture of JOb with JDM cells showed the highest mitogenic effect on JOb cells, However, JDM also significantly increased AOb cell proliferation. Heterotypic coculture of osteoblast cells with ADM cells did not promote proliferation of either JOb or AOb (Fig. 1B).
Effect of JCM and ACM on osteoblast proliferation.
The influence of cell-conditioned media collected from both JCM and ACM cells on cell proliferation of JOb and AOb was studied for a time period of 20 days. The media were supplemented with 10% FCS to ensure osteoblast survival. Preliminary results obtained by cell growth assay showed that, for the 2-wk treatment, JCM significantly increased the proliferation of JOb but had only a slight effect on the growth rate of AOb cells. In contrast, ACM did not induce growth of either JOb or AOb cells (data not shown). Analysis of cell proliferation performed with BrdU incorporation was used to compare the growth rates of JOb and AOb under the influence of either JCM or ACM at days 4, 8, 12, 16, and 20 of cell culture (Fig. 2, A and B). Similar to the preliminary coculture study, the ACM did not display any positive effect on the cell growth of either JOb or AOb; rather, it slightly inhibited cell growth (Fig. 2, A and B). In contrast, JCM treatment significantly increased the proliferation of JOb (P < 0.05), with the highest relative increase at day 16 (compared with control) (Fig. 2A). JCM treatment also slightly increased proliferation of AOb. However, there was a statistically significant increase only at day 16 (P < 0.01; Fig. 2B).
To rule out the serum influence on the cell growth, a short-term experiment was performed in which cell-conditioned media were produced by culturing JDM and ADM cells under serum-free conditions. Serum-free cell-conditioned media were collected and tested for their mitogenic activity. Starved JOb and AOb cells were incubated in serum-free JCM or ACM for 24 or 48 h. Cell proliferation of treated JOb and AOb cells was assessed either by BrdU incorporation or by PCNA immunoblotting analysis (Fig. 2, C and D). BrdU incorporation analysis indicated that, in JOb stimulated for 24 h with serum-free JCM, there was a sixfold increase in proliferation compared with JOb cells incubated in the presence of serum-free control medium, as well as a fourfold increase in proliferation compared with JOb cells incubated in medium supplemented with 10% FCS (Fig. 2C). JCM failed to induce proliferation on AOb cells. Similar to the data obtained with ACM supplemented with 10% FCS, (see Fig. 2, A and B), serum-free ACM did not stimulate proliferation on either JOb or AOb.
To further confirm the growth stimulation exerted by serum-free JCM, we analyzed the level of PCNA protein in JOb and AOb. As shown in Fig. 2D, immunoblotting analysis revealed the highest level of PCNA protein only in JOb cells treated for 24 and 48 h with serum-free JCM. As a loading control, α-tubulin was used. Together, these results strongly suggest the presence of mitogenic factor(s) in the cell-conditioned medium harvested from JDM cells.
Identification of biologically active FGF-2 protein in dura mater and its inactivation by neutralizing anti-FGF-2 antibody.
The mitogenic activity observed in JCM prompted us to investigate the presence of potential growth factor(s) in the serum-free medium harvested from JDM cells. It is well established that FGF-2 is a potent mitogen for osteoblasts, and previous studies performed on rat osteoblast and dura mater cells have suggested a role for FGF-2 in cell proliferation and osteogenesis (4). Therefore, to identify one of the factors responsible for the mitogenic-inducing activity present in the conditioned media harvested from dura mater cells, we determined FGF-2 levels by performing an ELISA assay on both cell lysates and serum-free cell-conditioned media of JDM and ADM cells. Considering the small amount of FGF-2 released in the medium, due to the lack of a signal peptide, freshly harvested cell-conditioned media were concentrated 10-fold by a centrifugal filter device. The final FGF-2 concentrations were calculated with the use of a protein standard as a reference and also adjusted by the concentration factors for the conditioned media or by the total cell number present at time of medium collection, The assay clearly detected FGF-2 protein in dura mater cell lysate and conditioned medium, although significant quantitative differences were observed between JDM and ADM cells. The results are presented in Fig. 3, A and B. The ELISA assay revealed 106 pg of FGF-2 per 1.0 × 105 cells in JDM cell lysates and ∼24.1 pg of FGF-2 per 1.0 × 105 cells in ADM cell lysates. Thus the endogenous level of FGF-2 was about fourfold higher in JDM than in ADM cell lysates (Fig. 3A). The difference was statistically significant (P < 0.01). Approximately 7.5 pg of FGF-2 were detected in the corresponding JCM, and no detectable FGF-2 was found in the ACM (Fig. 3B). Because ADM cells synthesize four time less FGF-2 than JDM cells and because of the poor release of FGF-2, as a result of lack of the canonical signal peptide, the amount of FGF-2 present in ACM is below the detectable threshold level. To produce the conditioned media, an equal number of cells from JDM or ADM were used (1.5 × 105/well, as described in materials and methods). As a control, to ensure that differences in FGF-2 production did not reflect differences in cell proliferation, BrdU labeling was performed on sister cell plates. No significant differences in cell proliferation were observed between JDM and ADM after 24 h (data not shown).
Next, we hypothesized that FGF-2 could be responsible for the mitogenic activity in JCM. To validate our hypothesis, we tested whether abrogation of FGF-2 activity in JCM could abolish most of the mitogenic activity exerted by JCM on osteoblasts. For this purpose, cell-conditioned media harvested from JDM and ADM cells were incubated in the presence of neutralizing anti-FGF-2 antibody and subsequently tested for mitogenic activity on osteoblast cells. As shown in Fig. 3C, JOb cells cultured with unperturbed JCM had a significant increase in proliferation, as previously demonstrated (twofold higher than cells incubated with control serum-free medium). Incubation of JCM with anti-FGF-2-neutralizing antibody (0.8 μg/ml) abolished cell proliferation in osteoblast cells by ∼60%, whereas incubation with nonimmune IgG control did not abrogate the cell proliferation activity induced by JCM. In contrast, ACM did not stimulate JOb proliferation, and anti-FGF-2 antibody treatment had no effect. We also investigated the effect of anti-FGF-2 neutralizing antibody in JOb cells cocultured with JDM cells. As shown in Fig. 3D, the effect exerted by anti-FGF-2 antibody on heterotypic cocultures, specifically on JOb cell proliferation, was similar to that observed in JOb cells treated with cell-conditioned medium derived from JDM cells. Thus the latter result further confirms the paracrine effect elicited by dura mater cells on osteoblasts.
Characterization of signaling pathway(s) activated by JCM.
Three signal-transduction pathways of FGF-2 have been widely reported in the literature: MAPK, PI3-kinase, and PKC pathways (2, 5, 18). Two subclasses of the PKC pathway, PKC-α/β and PKC-δ, have been previously studied in the context of osteogenesis, and their relevance has been highlighted (36). Therefore, we investigated which of these pathways is activated by JCM on osteoblasts. To characterize the JCM-activated pathways, we performed immunoblotting analysis on cells treated for 15 min, 30 min, and 3 h (time points were chosen based on results obtained from preliminary time course experiments) with JCM or rhFGF-2 (20 ng/ml) using specific antibodies for the phosphorylated signaling proteins: phospho-ERK1/2, phospho-Akt, and phospho-PKC-α/β or phospho-PKC-δ, the three canonical FGF-2 signal-transduction pathways. To further characterize the signaling, cells were also treated with specific inhibitors of each pathway. Figure 4 illustrates the induction of the FGF-2 signaling pathways analyzed by immunoblotting technique. For the MAPK pathway (Fig. 4A), strong bands of phosphorylated ERK1/2 were detected at 15 min, 30 min, and 3 h of treatment. After a 3-h stimulation, both JCM and 20 ng/ml exogenous rhFGF-2 (serum free + FGF-2) still induced a robust activation of ERK1/2, which was completely abrogated by the addition of 10 μM U-0126 (the MEK inhibitor). Thus, similar to the exogenous rhFGF-2, JCM also activated the ERK1/2 pathway, which was inhibited by U-0126.
PI3-kinase/Akt is another major pathway mediating most of FGF-2 biological activity. In a second set of experiments, we tested the capability of dura mater JCM to activate the PI3-kinase/Akt (Fig. 4B). Exogenous rhFGF-2 at a concentration of 20 ng/ml was used as reference. As shown in Fig. 4B, JCM induced a strong phosphorylation of Akt at 15 min, but only a faint phosphorylated Akt band was observed in cells treated for 30 min and 3 h. The specific inhibitor LY-294002 at a concentration of 20 μM effectively blocked this signaling pathway on both cells stimulated with either exogenous rhFGF-2 or dura mater JCM. Finally, we investigated the activation of the PKC pathway (Fig. 4, C and D). Therefore, an immunoblotting analysis was performed using an anti-phospho-PKC-α/β antibody and an anti-phospho-PKC-δ antibody. After treatment with dura mater JCM, phosphorylation was observed at early time points (15 and 30 min). However, at 3 h, phosphorylation of PKC-αβ was slightly decreased (Fig. 4C), and PKC-δ phosphorylated level was less than that observed at the earlier time points (Fig. 4D). Moreover, in the presence of the specific inhibitor Gö-6983 (2 μM), there was only a partial inhibition of phosphorylation induced by dura mater JCM.
Next, to identify the pathway through which JCM induces cell proliferation on osteoblasts, we performed a BrdU assay using specific inhibitors of each pathway. Starved osteoblasts were treated for 12 h with either rhFGF-2 (20 ng/ml) or JCM in the presence and absence of each specific inhibitor. BrdU incorporation was assessed as described in materials and methods. Treatment with inhibitors for 12 h did not affect the viability of the cells, as assessed by morphological analysis and Trypan blue staining (data not shown). As shown in Fig. 5A, cell proliferation increased significantly, about fivefold in rhFGF-2-treated osteoblasts and about eightfold in JCM-treated osteoblast cells compared with cells treated with serum-free medium. However, all three inhibitors abolished the dura mater JCM-induced cell proliferation, as well as the rhFGF-2-induced cell proliferation. Thus these results indicate that dura mater JCM stimulates osteoblast cell proliferation at least through three major FGF-2-related pathways.
Characterization of signaling pathway(s) in dura mater-osteoblast cell heterotypic coculture.
The above results prompted us to further investigate the biological relationship between dura mater cells and osteoblasts. Therefore, we tested the effect of the inhibitors of FGF-2 signaling pathways in the heterotypic coculture system. The concentrations of the inhibitors were the same as previously used for the experiment with the conditioned medium. Figure 5B demonstrates osteoblast cell proliferation in coculture system supplemented with or without inhibitors. Cell proliferation was assessed by BrdU incorporation during 12 h. As shown in Fig. 5B, JOb cells cocultured with JDM cells proliferated more than juvenile control cells cultured alone. These results further confirm what was previously demonstrated (see Fig. 1). Addition of each inhibitor abrogated the proliferation of cocultured osteoblast cells. These results suggest that these same signaling pathways, previously identified as major players of mitogenic activity exerted by the JCM on osteoblast cells, also mediate osteoblast proliferation in the coculture system.
A large number of studies have suggested that JDM may play important roles in calvaria osteogenesis, including cellular contributions and paracrine functions, by providing growth factors. One of the direct observations supporting this hypothesis was provided by a coculture study reported by Spector et al. (31). That study demonstrated that rat JOb cells proliferate more rapidly when cocultured with rat JDM cells. However, these data could not definitively provide evidence that the mitogenic effect observed on osteoblasts was elicited by paracrine function of dura mater cells rather than by an augmented autocrine loop of osteoblasts as result of cell interactions.
The data presented in our study provide direct evidence for a paracrine effect of JDM on osteoblasts. Using cell-conditioned media produced by JDM cells, as well as a heterotypic coculture model system, we identified FGF-2 as one of the paracrine factors provided by JDM cells that has potent mitogenic activity on osteoblasts. The abrogation of mitogenic activity obtained by neutralizing anti-FGF-2 antibody clearly indicates that this growth factor is a key inducing mitogenic molecule produced by dura mater cells.
Because FGF-2 forms do not contain a classic secretory signal sequence, their secretion from cells is very poor and uses an alternative pathway (21). Analyses of endogenous FGF-2 in JDM and ADM cells indicated that the threshold levels of endogenous FGF-2 in JDM are higher than levels in ADM cells. Therefore, differences in endogenous levels of FGF-2 may be responsible for the paracrine activity elicited by JDM cells.
Early preosteoblasts proliferate before maturing into bone-producing osteoblasts. Osteoblasts direct osteogenesis through the production of extracellular matrix, mineralization of matrix, and regulation of bone remodeling through resorption and deposition. Our previous work (4) has demonstrated the ability of JObs to proliferate and maturate into AObs. Moreover, JObs were in a less differentiated state, had less endogenous FGF-2, and had more cell surface-associated FGF receptor (FGF-R1 and FGF-R2) protein with increased phosphorylation. In contrast, AObs were more fully differentiated and produced high levels of FGF-2 and less FGF-R1 and -R2 proteins and phosphorylation. Furthermore, after rhFGF-2 stimulation, JObs organized and produced more matrix proteins and bone nodules. Thus JObs appeared to be a suitable target for FGF-2 in the context of calvarial ossification at the autocrine and paracrine levels.
FGF-2 has a two-pronged effect because it is an osteoblast mitogen and it stimulates bone formation, as demonstrated in several mammalian models (20, 28). Interestingly, FGF-2 has differential effects on osteoblast biology depending on the stage of cellular maturation (14, 15, 20, 22, 28, 34). FGF-2 elicits its biological effect on binding to its receptors (FGF-R1, -R2, -R3, -R4) for differential regulation of osteogenesis (26). These tyrosine kinase receptors become phosphorylated on FGF-2 binding and activate ERK and p38 MAPK signaling pathways (25). Because the Ras/ERK, PI3-kinase/Akt, and PKC pathways mediate most FGF-2 effects on a large variety of cells, we analyzed the extent of phosphorylation on treatment of JObs with JCM. Our results indicated that JCM strongly activated ERK1/2 signaling, similar to the effect of exogenously added rhFGF-2; however, its effect on Akt activation was more intriguing. Our results are in agreement with data previously reported by other investigators and indicate that the effect of FGF-2 on Akt activation is considerably more transient than that on ERK1/2 activation (27). FGF-2-induced PI3-kinase signaling has been reported as a mitogenic signal in neuronal progenitor cells (16). In osteoblasts, PI3-kinase activated by FGF-2 plays an important role in cell survival and apoptosis (5) and an inhibitory role in VEGF release (33, 38).
Moreover, we targeted these pathways by using specific inhibitors with the aim to characterize the signaling through which the FGF-2 released from JDM cells induces osteoblast proliferation. Our results demonstrated that all of the analyzed pathways are possibly engaged in the FGF-2 mitogenic signaling. However, because other growth factors also elicit their biological effect, mostly activating the ERK and PI3-kinase signaling pathways, we cannot rule out that the inhibitors are also targeting down-signaling pathways induced by other growth factors (e.g., PDGF, EGF). This possibility deserves further investigation.
Together, our results strongly indicate that JDM may regulate adjacent osteoblast biology by paracrine-mediated signaling. Indeed, FGF-2 could be one of the paracrine factors involved in successful calvarial ossification, an orchestrated process that may require biomolecular interactions between osteoblast and dura mater cells.
This investigation was supported by National Institute of Dental Research Grants R01 DE-014526 and R01 DE-13194 to M. T. Longaker.
We are very grateful to Dr. Deepak Gupta and Dr. Matthew Kwan for the critical reading of the manuscript.
N. Quarto is on leave of absence from the Department of Structural and Functional Biology, University of Naples Federico I, Naples, Italy.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society