Signaling involved in osteoblastic cell differentiation remains largely unknown. This study further investigates mechanisms involved in BMP-2-induced osteoblastic cell differentiation. We report that BMP-2 can activate JNK and p38 in osteoblastic cells and provide evidences that these MAP kinases have distinct roles in regulating alkaline phosphatase and osteocalcin expression.
Introduction: Bone morphogenetic protein (BMP)-2 exerts many of its biological effects through activation of the Smad pathway. Cooperative interactions between the Smads and the stress-activated protein kinase (SAPK) p38 and c-Jun-NH2-terminal kinase (JNK) pathways have recently been observed in TGF-β signaling.
Materials and Methods: Activation of mitogen-activated protein (MAP) kinases by BMP-2 and the role of these signaling pathways for cell differentiation induced by BMP-2 was investigated in mouse MC3T3-E1 and primary cultured calvaria-derived osteoblastic cells using immunoprecipitation, in vitro kinase assay and Western blot analysis, as well as specific MAP kinase inhibitors.
Results: Associated with the rapid activation of Smads, BMP-2 barely affected extracellular-signal regulated kinase (ERK) activity, whereas it induced a transient activation of p38 and JNK. The role of p38 and JNK in mediating BMP-2-induced stimulation of osteoblastic cell differentiation was evaluated using the respective specific inhibitors SB203580 and SP600125. Inhibition of p38 by SB203580 was mainly associated with decreased alkaline phosphatase (ALP) activity, whereas inhibition of JNK by SP600125 was associated with a marked reduction in osteocalcin (OC) production induced by BMP-2. Corresponding alterations in ALP and OC mRNA levels were found in cells treated with BMP-2 and inhibitors, suggesting an implication of p38 and JNK pathways in BMP-2-induced osteoblastic cell differentiation at a transcriptional level.
Conclusion: Data presented in this study describe p38 and JNK as new signaling pathways involved in BMP-2-induced osteoblastic cell differentiation with evidences for a distinct role of each MAP kinase in the control of alkaline phosphatase and osteocalcin expression.
Bone morphogenetic proteins (BMPs) are members of the transforming growth factor-β (TGF-β) superfamily and exert a wide range of biological effects in different tissues. In particular, they contribute to the formation of bone and connective tissues(1) by inducing the differentiation of mesenchymal cells into bone forming cells.(2) This factor is one of the most potent stimulators of osteoblastic cell differentiation, which is mainly characterized by expression of alkaline phosphatase (ALP), type I collagen, and osteocalcin (OC).(3–5) Members of the TGF-β superfamily exert their biological activities by binding to cell surface type I and II serine/threonine kinase receptors. The type II receptor phosphorylates the type I receptor, which in turn phosphorylates and activates intracellular substrates such as proteins of the Smad family. Smad 1(6) and the closely-related Smads 5 and 8, specifically mediates BMP-2 responses, such as, for example, the osteoblastic differentiation of precursor cell lines.(7,8) On phosphorylation, Smad 1/5/8 proteins interact with a common partner, Smad 4, and the complex Smad 1/5/8-Smad 4 translocates to the nucleus, where it exerts transcriptional activity either through direct binding to DNA or through association with other DNA-binding proteins.(9) Among signal transducers that have been recently reported to participate in TGF-β signaling, the mitogen-activated protein kinases (MAPKs) probably play a significant role in cooperating with Smads. MAPKs are a group of well-described serine/threonine kinases implicated in the transmission of extracellular signals to intracellular targets.(10) Three structurally related MAPK pathways have been characterized in mammalian cells. The extracellular-signal regulated kinases (ERKs) are activated downstream of cell membrane receptors and regulate cell proliferation and differentiation.(11) The c-Jun-NH2-terminal kinases (JNKs) and the p38 MAPK (p38) pathways are generally activated by treatment of cells with inflammatory cytokines or by environmental stress leading to apoptosis and have therefore been named stress-activated protein kinases (SAPKs).(12) SAPK pathways also contribute to multiple cellular processes such as proliferation, differentiation, and survival in response to extracellular stimulation.(13) Some MAPK substrates are transcription factors that are activated through phosphorylation. Cooperative interaction between Smads and transcription factors activated by MAPKs has been recently described for TGF-β signaling. Several TGF-β-responsive elements containing AP1-binding sites are activated by c-Jun/c-Fos heterodimers.(14) It has also been reported that TGFβ and Smad 3 only weakly induce AP1-containing promoters in absence of c-Jun or c-Fos binding, suggesting that Smads require active c-Jun/c-Fos dimers as DNA binding partners.(15) Interestingly, Smads were found to preferentially interact with the phosphorylated form of c-Jun,(15) which is generated by the activity of JNK. As for c-Jun/c-Fos, Smads can also cooperate with activated ATF2, which is generated on phosphorylation by p38 MAPK.(16)
Recently, two studies described that BMP-2 activates MAP kinases in the pluripotent C2C12 cell line,(17,18) with controversial issues regarding their role in myoblastic cell transdifferentiation. The first study described that BMP-2 induces an osteoblastic phenotype through activation of ERK and p38.(17) The second study also found that BMP-2 activates p38, but in contrast with the first report, did not confirm the stimulation of ERK by BMP-2 and found that the p38 pathway has a negative effect on the transdifferentiation of C2C12 cells into osteoblastic cells.(18) Very recently, a role of ERK and p38 in BMP-2-induced human osteoblastic cell differentiation has also been reported.(19) All these studies point to a possible role of MAP kinases in the effects of BMP-2 in osteoblastic cells, but their role in mediating cell differentiation induced by BMP-2 remains unclear.
In this study, we investigated the effect of BMP-2 on the activation of the three MAP kinases, ERK, p38, and JNK, in MC3T3-E1 and primary cultured mouse calvaria-derived osteoblastic cells. We found that BMP-2 has a very small effect on activation of ERK in these cells compared with epidermal growth factor (EGF) and serum growth factors. However, BMP-2 strongly activated both p38 and JNK. Interestingly, we found that these MAP kinases are involved in osteoblastic differentiation induced by BMP-2. These observations indicate that p38 and JNK MAP kinases are novel signaling pathways involved in BMP-2 regulation of osteoblastic cells.
MATERIALS AND METHODS
Antibodies against ERK1/2, phospho(p)-JNK (agarose-conjugated), JNK, p-c-Jun, c-Jun, p38, and Smad1/5 were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Antibodies against p-ERK1/2 (agarose-conjugated) and p-p38 (agarose-conjugated) were purchased from New England Biolabs (Beverly, MA, USA). The antibody against MAPKAP-K2 was from StressGen (Victoria, British Columbia, Canada). Antibodies against p-Smad1/5 were generously provided by New England Biolabs.
Primary mouse calvaria-derived osteoblastic cells, isolated as previously described,(20) and mouse MC3T3-E1 cells(21) were cultured in either 75-cm2 flasks or 6-well plates and grown in α-MEM (Amimed, Allschwill, Switzerland) supplemented with 10% fetal calf serum (FCS; Life Technologies, Basel, Switzerland) and 0.5% non-essential amino acids (Amimed). After 8–10 days, cells were cultured with α-MEM containing 1% FCS for 24 h before addition of test agents. In some experiments, cells were treated with either 4-(4-fluorophenyl)-2-(4-methylsulfinylphenyl)-5-(4-pyridinyl)imidazole (SB203580) or anthra[1,9-cd]pyrazol-6(2H)-one 1,9-pyrazoloanthrone (SP600125; Calbiochem-Novabiochem Corp., San Diego, CA, USA) or their respective vehicles for 1 h before and during exposure to BMP-2.
Western blotting and immunocomplex in vitro kinase assays
Cells treated with different agents were rapidly frozen in liquid nitrogen. Cells were lysed in buffer A for 10 minutes. Buffer A contained 50 mM Tris (pH 7.4), 150 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 2 mM Na3VO4, 0.01 μM calyculin A, 0.1 μM mycrocystin LR, 1% NP-40, 1% sodium deoxycholate, and 0.1% SDS. Lysates were cleared by centrifugation at 6000 rpm, 4°C, for 30 minutes. For Western blotting analysis, an aliquot of the supernatant was diluted with an equal amount of 2× reducing sample buffer consisting of 125 mM Tris (pH 6.8), 20% glycerol, 4% SDS, 200 mM dithiothreitol, and 0.025% bromophenol blue, and heated at 70°C for 30 minutes. Proteins were then fractionated by reducing SDS/PAGE on 6–15% acrylamide gradient gels and transferred to Immobilon P membranes (Millipore Corp., Bedford, MA, USA). Immunoblotting was performed as described previously with appropriate primary antibodies and horseradish peroxidase-conjugated secondary antibodies.(22) Immunoreactive bands were visualized by ECL (Amersham International, Little Chalfont, UK). For in vitro kinase assays, equal amounts of protein lysate were incubated with the appropriate agarose conjugated antibody for 18 h at 4°C. Immunoprecipitated complexes were washed twice in buffer A and twice in kinase buffer. Kinase buffer contained 50 mM Tris (pH 7.4), 25 mM β-glycerophosphate, 20 mM MgCl2, 1 mM dithiotreitol, 10 μM32P-ATP (50 μCi/ml), and 2.5 μg of GST-ATF2 (Santa Cruz Biotechnology) for assaying p38 activity, GST-c-Jun (Biomol Research Laboratories, Campus Drive, PA, USA) for assaying JNK activity, or recombinant HSP27 (StressGen) for MAPKAP-K2 assay. After 30 minutes of incubation at 30°C, the reaction was stopped by addition of 2× reducing buffer, and proteins were separated on SDS/PAGE and transferred to Immobilon P membranes as described above for Western blot analysis. Membranes were exposed for autoradiography at −80°C.
Measurement of biochemical parameters in osteoblastic cells
As previously reported,(23) ALP activity was assessed using P-nitrophenyl phosphate as a chromogenic substrate, and osteocalcin released into the medium was measured by radioimmunoassay using a goat anti-mouse osteocalcin antibody and a donkey anti-goat secondary antibody (Biomedical Technologies Inc., Stoughton, MA, USA). Protein content was determined using the Pierce Coomassie Plus assay reagent (Pierce, Rockford, IL, USA).
RNA isolation and Northern blotting analysis
Total cellular RNA was extracted from MC3T3-E1 cells using the Tripure reagent (Roche Molecular Biochemicals, Rotkreuz, Switzerland) according to the manufacturer's instructions. Equal amounts (10 μg) of RNA were electrophoretically resolved on an 1.5% agarose gel in phosphate buffer and transferred onto GeneScreen Plus nylon membranes (DuPont de Nemours, Brussels, Belgium) by overnight capillary transfer. Membranes were prehybridized at 68°C for 2 h in Quickhyb (Stratagene) complemented with 5 μg/ml ssDNA before hybridization at 68°C for 3 h with the desired cDNA probe labeled with 50 μCi [α-32P] deoxy-CTP by random priming (Megaprime DNA Labeling System; Amersham Life Science, Dubendorf, Switzerland). The filters were finally exposed for autoradiography at −80°C. A 2.5-kb EcoRI restriction fragment corresponding to the rat ALP cDNA(24) was provided by Dr Rodan (Merck Sharp and Dohme Research Laboratories, Westpoint, PA, USA). The mouse osteocalcin cDNA probe was obtained by reverse transcriptase-polymerase chain reaction (RT-PCR) using specific primers as previously described.(25) The RT-PCR product was then purified using a gel extraction kit (Quiagen II; Quiagen, Basel, Switzerland).
Levels of mRNA were quantified using an ImageQuant application for McIntosch (IQMacv1.2) and normalized to the amount of cyclophilin transcripts detected in the same conditions, with a cDNA probe for human cyclophilin.(26)
When not specified, all experiments were carried out independently at least three times. Results are expressed as mean ± SE. Comparative studies of means were performed using one-way ANOVA followed by a post hoc test (projected least significant difference Fisher) with a significance value of p < 0.05.
As previously described, BMP-2 increased ALP activity in MC3T3-E1 cells.(27,28) In the MC3T3-E1 subclone used in this study, this effect was maximal after 48 h (Fig. 1A) and was associated with increased ALP mRNA expression (Fig. 1B), an effect that was completely blocked by 2 μg/ml actinomycin D (data not shown), confirming the implication of a transcriptional process in this cellular response.(29)
Activation of MAP kinases by BMP-2 in MC3T3-E1 cells
Recent information indicates that BMP-2 can activate the MAP kinases ERK and p38 in both human osteoblastic cells(19) and the pluripotent C2C12 cell line.(17,18) In MC3T3-E1 cells, we found that BMP-2 (100 ng/ml) enhances the phosphorylation of ERK, but this effect was very small compared with those of either EGF (10−8 M) or FCS (10%) (Fig. 2A). We also determined the effect of BMP-2 on ERK activation by in vitro kinase assay using Elk-1 as the substrate. In experiments showing a strong activation of p38 and JNK by BMP-2 (Figs. 2C and 2D), we could not find any activation of ERK by this method (data not shown). The striking activation of p38 and JNK by BMP-2 was transient and less rapid than the phosphorylation of Smad1/5 proteins (Fig. 2B). Activation of Smads by BMP-2 was detectable after 15 minutes, maximal after 1 h, and maintained for at least 6 h. Activation of both JNK and p38 was only detected after a lag time of approximately 1 h. Both kinases were then transiently activated with maximal stimulation after 3 h, and their activity returned to baseline values within 6 h. Activation of p38 and JNK was not associated with any change in the respective amount of each MAP kinase, suggesting that their activation results from the stimulation of regulatory kinases.
Effect of the p38 MAP kinase inhibitor, SB203580, on BMP-2-induced MC3T3-E1 signaling and cell differentiation
We recently used SB203580, a specific p38 MAP kinase inhibitor,(30) for analysis of the role of p38 in MC3T3-E1 cell differentiation induced by serum growth factors, and we found that 10 μM SB203580 was an optimal dose for selectively inhibiting p38 in these cells.(23) This concentration was therefore used for analysis of the role of p38 in BMP-2-induced osteoblastic cell differentiation. As shown in Fig. 3A, 10 μM SB203580 did not interfere with activation of Smad1/5. As expected, it completely blocked activation of p38 induced by BMP-2 assessed by the determination of MAPKAP-K2 activity, a selective and physiological cellular p38 substrate.(31) In our experimental conditions, BMP-2 enhanced both ALP activity (Fig. 3B; Table 1) and OC production (Fig. 3C) but had no effect on collagen deposition (data not shown). SB203580 time-dependently (Table 1) reduced both the stimulation of ALP (Fig. 3B), and to a lesser extent, OC production induced by BMP-2 (Fig. 3C). Associated with these effects on BMP-2-induced osteoblastic cell differentiation, this inhibitor also blunted ALP and OC mRNA expressions (Figs. 6A and 6B).
Table Table 1. Effect of p38 and JNK Inhibitors on Time-Dependent Changes in ALP Activity Induced by BMP-2 in MC3T3-E1 Cells
Effect of the JNK inhibitor, SP600125, on BMP-2-induced MC3T3-E1 osteoblastic cell differentiation
SP600125 has been reported to selectively inhibit JNK activity in Jurkat T cells(32) and also recently in MC3T3-E1 cells.(33) As described in Fig. 4 (bottom panel), SP600125 dose-dependently inhibited the cellular phosphorylation of c-Jun, a specific substrate of JNK kinases. In our experimental conditions, a significant effect was observed at the dose of 15 μM with a complete inhibition at 25 μM. The same range of effective inhibitory concentrations have been reported in Jurkat T cells.(32) The dose of 25 μM SP600125, which maximally inhibited JNK activity, only slightly influenced the stimulation of the Smad1/5 proteins and of the p38 MAP kinase induced by BMP-2, confirming that this compound is a specific inhibitor of the JNK pathway inMC3T3-E1 osteoblastic cells.(33) The role of the JNK pathway in mediating osteoblastic differentiation induced by BMP-2 was therefore investigated using 15 and 25 μM SP600125. Effect of this compound on the stimulation of ALP induced by BMP-2 was variable. It either slightly inhibited (Table 1) or enhanced (Fig. 5A) BMP-2-induced ALP activity. In contrast to this observation on ALP, SP600125 dose-dependently and markedly reduced the stimulation of OC production induced by BMP-2 (Fig. 5B). Associated with these biochemical changes induced by SP600125, corresponding changes in mRNA for ALP and OC were observed with a slight increase in ALP mRNA (Fig. 6C) and a marked dose-dependent inhibition in OC mRNA expression (Fig. 6D).
Effect of BMP-2 on activation of MAP kinases and their roles in the differentiation of primary cultured calvaria-derived osteoblastic cells
As found in MC3T3-E1 cells, BMP-2 induced a slight and transient increase of ERK activity in osteoblasts derived from mouse calvaria. Time-dependent stimulation of Smads by BMP-2 was very similar to that found in MC3T3-E1 cells, with a slight increase after a 15-minute incubation, maximal activation between 1 and 2 h, and sustained activation for at least 6 h (Fig. 7A, top panels). In calvaria-derived osteoblasts, we found that BMP-2 also induced activation of p38 and JNK. The onset of activation for these two kinases was delayed compared with activation of Smads and appeared only after a 1-h exposure. The duration of p38 and JNK activation by BMP-2 was, however, different compared with that recorded in MC3T3-E1 cells. The transient activation of p38 was shorter, whereas that of JNK lasted much longer; it was still present after 6 h of exposure to BMP-2 (Figs. 2C, 2D, and 7A).
Except for a marked effect of the p38 inhibitor on basal OC production (Fig. 7C), p38 and JNK inhibitors had similar effects on the stimulation of osteoblastic cell differentiation induced by BMP-2 in calvaria-derived compared with MC3T3-E1 osteoblastic cells. SB203580 significantly reduced the stimulation of ALP, whereas SP600125 had no effect (Fig. 7B). Concerning OC, the JNK inhibitor completely blunted the stimulation of this protein induced by BMP-2 (Fig. 7C). In calvaria cells treated with the p38 inhibitor and having a high basal level of OC production, the effect of BMP-2 was slightly, but not significantly, reduced (Fig. 7C).
BMP-2 is a well-described stimulator of osteoblastic cell differentiation characterized mainly by increased expression of ALP, type I collagen, and OC.(3–5) The results of this study indicate that, in addition to the well-documented Smad pathway and the more recently described PI3-K/Akt pathway,(34) BMP-2 also activates both JNK and p38 in osteoblastic cells. This response was detected in MC3T3-E1 and calvaria-derived osteoblastic cells, suggesting a role of these pathways in mediating BMP-2-induced osteoblastic cell differentiation. Whereas previous studies have already reported that BMP-2 can stimulate the MAP kinases ERK and p38, our data indicate that BMP-2 can also induce activation of JNK in osteoblastic cells. In MC3T3-E1 cells, BMP-4 was reported to induce a stimulation of p38, but activation of JNK could not be detected.(35) Whether this difference in MAP kinase activation between BMP-2 and BMP-4 in MC3T3-E1 cells is related to differences in receptor signaling remains to be investigated. As mentioned above, BMP-2 has also been reported to activate ERK and p38 MAP kinases, but not JNK, in the pluripotent C2C12 myoblastic cell line.(17,18) The role of these MAP kinases in mediating BMP-2-induced C2C12 osteoblastic transdifferentiation is, however, controversial and remains to be further investigated. Recently, a role of ERK and p38 in BMP-2-induced human osteoblastic cell differentiation has also been reported.(19) This information points to a possible role of MAP kinases in BMP-2 effects on osteoblastic cells.
In this study, we found that BMP-2 mainly activated the MAP kinases p38 and JNK in MC3T3-E1 and calvaria-derived osteoblastic cells, whereas it barely affected activation of ERK (Figs. 2 and 7). This later observation is concordant with the previous observation that BMP-2 does not influence the proliferation of MC3T3-E1 cells.(27) In addition, two recent studies also reported that ERK is a negative regulator of BMP-induced differentiation of C2C12 and MC3T3-E1 cells,(35,36) suggesting that ERK is not a pathway involved in BMP-2-induced osteoblastic cell differentiation.
Analysis of the role of p38 and JNK activated by BMP-2 suggests that each MAP kinase has differential effects on osteoblastic cell differentiation. As recently found for serum growth factors and epinephrine,(23,37) data obtained in the present study with SB203580 suggest that p38 is mainly involved in mediating activation of ALP induced by BMP-2 (Figs. 3B and 7B; Table 1). In addition to its consistent effect on ALP, our data suggest that p38 might also play a role in modulating OC production by growth factors. In MC3T3-E1 cells, we previously reported that SB203580 slightly reduced the production of OC by cells cultured in presence of serum growth factors.(23) In this study, a similar effect was found in these cells in response to BMP-2 (Fig. 3C). In contrast to our previously mentioned observation that the p38 inhibitor slightly reduces the production of OC in MC3T3-E1 cells, this inhibitor markedly enhanced the basal production of this protein in calvaria-derived osteoblastic cells (Fig. 7C). This observation was reproduced in two separate experiments and was completely unexpected. It suggests that the p38 pathway either negatively regulates or delays OC expression in calvaria cells. When these cells were treated with the p38 inhibitor and had an increased OC expression, we could not detect any BMP-2 response on this parameter (Fig. 7C). Interpretation of these data is difficult and requires further investigation to clarify the mechanism by which inhibition of the p38 pathway leads to increased expression of OC. Data obtained with the JNK inhibitor, SP600125, suggest that JNK also differentially influences the regulation of ALP and OC. Indeed, the complete inhibition of JNK by SP600125 was associated with variable and small effects on ALP (Figs. 5A and 7B; Table 1), and a marked and consistent decrease in OC production in response to BMP-2 in both MC3T3-E1 and calvaria cells (Figs. 5B and 7C). This finding suggests a particular role of JNK in regulating late osteoblastic differentiation, an interesting issue now investigated in our laboratory. Taken together, these observations indicate that the SAP kinases, p38 and JNK, are implicated in BMP-2-induced osteoblastic cell differentiation with distinct regulatory effects on ALP and OC expression.
The molecular mechanism by which p38 and JNK mediate the effects of BMP-2 on ALP and OC in osteoblastic cells are not known. Data shown in Figs. 1 and 6, however, suggest the implication of transcriptional processes. A major group of DNA binding proteins that could potentially be involved is activating protein-1 (AP1) dimers.(38) The AP-1 family of transcription factors are homodimers and heterodimers composed of basic region-leucine zipper proteins that belong to the Jun, Fos, and the closely related activating transcription factors (ATFs). Among the mechanisms that account for the stimulation of AP-1, MAP kinases are the most important mediators.(38) Interestingly, AP-1 has been described to be a DNA-binding partner for Smads. Several TGFβ-responsive elements have been found to contain AP-1 binding sites, which are activated by c-Jun and c-Fos heterodimers. In addition, Smads were found to preferentially interact with the phosphorylated form of c-Jun,(15) suggesting an implication of the JNK pathway. Whether Smads interact with AP-1 to directly activate the expression of osteoblastic gene markers or whether these responses are mediated by induction of master osteoblastic genes such as Runx2/Cbfa1(39,40) or osterix(41) is an interesting issue that remains to be investigated.
In conclusion, data presented in this study describe p38 and JNK as new signaling pathways involved in BMP-2-induced osteoblastic cell differentiation, with evidence for distinct roles of each MAP kinase in the transcriptional control of ALP and/or OC expression.
We thank P Apostolides and S Troccaz for their expert technical assistance. We thank Genetics Institute for providing recombinant human BMP-2 and New England Biolabs for providing the antibodies against p-Smad1/5. This study was supported by the Swiss National Science Foundation (Grant 32–58937.99) and the Novartis Foundation.