Activations of ERK1/2 and JNK by transforming growth factor beta negatively regulate Smad3-induced alkaline phosphatase activity and mineralization in mouse osteoblastic cells.

Transforming growth factor (TGF) beta inhibits alkaline phosphatase (ALP) activity and mineralization in mouse osteoblastic MC3T3-E1 cells, whereas local administration of TGF-beta stimulates bone formation in vivo. We recently demonstrated that Smad3, a TGF-beta signaling molecule, promotes ALP activity and mineralization in MC3T3-E1 cells. Moreover, the target disruption of Smad3 in mouse is reported to cause a decrease in bone mineral density. These findings indicate that Smad3 plays an important role in the regulation of bone formation. However, why the effects of TGF-beta and Smad3 on ALP activity and mineralization are different remains unknown. The purpose of the present study is to clarify the role of mitogen-activated protein kinase (MAPK) in TGF-beta and Smad3 pathways in osteoblast. TGF-beta activated extracellular signal-regulated kinases/p42/p44 (ERK1/2), p38 MAPK, and c-Jun N-terminal kinase (JNK) in mouse osteoblastic MC3T3-E1 cells. The expression of dominant negative type Smad3, Smad3DeltaC, affected neither TGF-beta-activated MAPKs nor TGF-beta-inhibited ALP activity. Specific inhibitors of ERK1/2 activation (PD98059 and U0126), as well as JNK inhibitors (curcumin and dicumarol) antagonized the inhibitory effects of TGF-beta on ALP activity and mineralization, whereas the specific inhibitor of p38 MAPK (SB203580) did not affect them. PD98059 and curcumin enhanced Smad3-induced ALP activity and mineralization, whereas SB203580 inhibited them. In the luciferase reporter assay using 3TP-lux with the specific Smad3-responsive element, PD98059, and curcumin enhanced TGF-beta- and Smad3-induced transcriptional activity in MC3T3-E1 cells. On the other hand, TGF-beta-induced production of type I collagen was antagonized by curcumin but not by PD98059. The present study indicated that TGF-beta-responsive ERK1/2 and JNK cascades negatively regulate Smad3-induced transcriptional activity as well as ALP activity and mineralization in osteoblasts.


Transforming growth factor (TGF) ␤ inhibits alkaline phosphatase (ALP) activity and mineralization in mouse osteoblastic MC3T3-E1 cells, whereas local administration of TGF-␤ stimulates bone formation in vivo.
We recently demonstrated that Smad3, a TGF-␤ signaling molecule, promotes ALP activity and mineralization in MC3T3-E1 cells. Moreover, the target disruption of Smad3 in mouse is reported to cause a decrease in bone mineral density. These findings indicate that Smad3 plays an important role in the regulation of bone formation. However, why the effects of TGF-␤ and Smad3 on ALP activity and mineralization are different remains unknown. The purpose of the present study is to clarify the role of mitogen-activated protein kinase (MAPK) in TGF-␤ and Smad3 pathways in osteoblast. TGF-␤ activated extracellular signal-regulated kinases/p42/p44 (ERK1/2), p38 MAPK, and c-Jun N-terminal kinase (JNK) in mouse osteoblastic MC3T3-E1 cells. The expression of dominant negative type Smad3, Smad3⌬C, affected neither TGF-␤-activated MAPKs nor TGF-␤-inhibited ALP activity. Specific inhibitors of ERK1/2 activation (PD98059 and U0126), as well as JNK inhibitors (curcumin and dicumarol) antagonized the inhibitory effects of TGF-␤ on ALP activity and mineralization, whereas the specific inhibitor of p38 MAPK (SB203580) did not affect them. PD98059 and curcumin enhanced Smad3induced ALP activity and mineralization, whereas SB203580 inhibited them. In the luciferase reporter assay using 3TP-lux with the specific Smad3-responsive element, PD98059, and curcumin enhanced TGF-␤-and Smad3-induced transcriptional activity in MC3T3-E1 cells. On the other hand, TGF-␤-induced production of type I collagen was antagonized by curcumin but not by PD98059. The present study indicated that TGF-␤-responsive ERK1/2 and JNK cascades negatively regulate Smad3-induced transcriptional activity as well as ALP activity and mineralization in osteoblasts.
Transforming growth factor ␤ (TGF-␤) 1 is most abundant in bone matrix compared with other tissues (1). TGF-␤ is stored in an inactive form, released from the bone matrix, and activated in the bone microenvironment (2). It is produced by osteoblasts and appears to regulate bone metabolism in various ways, including skeletal development and bone remodeling (3). Several reports demonstrated that TGF-␤ induced bone formation when it was locally administered into bone tissues in rat (4 -7). However, it is disputable whether TGF-␤ would possess bone anabolic effects in vitro (8 -10), and the mechanism by which TGF-␤ stimulates bone formation in vivo remains unknown.
The Smad family proteins are critical components of the TGF-␤ signaling pathways (11). TGF-␤ exerts growth inhibitory and transcriptional response through the two receptorregulated Smads, Smad2 and Smad3 (11). Receptor-mediated phosphorylation of Smad2 or Smad3 induces their association with the common partner Smad4, followed by translocation into the nucleus where these complexes activate transcription of specific genes (12). As for osteoblasts, Li et al. (13) reported that overexpression of Smad2 suppressed Runx2(cbfa1) and osteocalcin mRNA expression in primary rat calvaria cells and ROS17/2.8 cells. Moreover, TGF-␤ stimulated the ␤ V -integrin subunit expression and p57 kip2 proteolysis via Smad signaling in osteoblastic cells (14). Although Alliston et al. (15) reported that Smad3 decreased Runx2 and osteocalcin gene expressions in MC3T3-E1 cells, our recent study revealed that Smad3 inhibited the proliferation and enhanced the levels of bone matrix proteins, such as type I collagen, osteopontin, and matrix Gla protein in a manner similar to TGF-␤ in these cells (16). On the other hand, unlike TGF-␤, Smad3 enhanced ALP activity and mineralization of MC3T3-E1 cells in that study. Our findings suggested that Smad3 plays an important role in the regulation of bone formation. Indeed, Borton et al. (17) recently reported that mice with targeted deletion of Smad3 were osteopenic, compared with wild type littermates, because of a lower rate of bone formation. The increased synthesis of type I collagen was common effect of TGF-␤ and Smad3 on osteoblasts. However, Smad3 greatly increased ALP activity and mineralization, whereas TGF-␤ inhibited them in these cells. The reason for the discrepant effects of TGF-␤ and Smad3 on ALP activity and mineralization remained unknown in our previous study (16). We therefore hypothesized that TGF-␤ might inhibit ALP activity and mineralization of osteoblasts through some pathways other than the Smad3 pathway. Alternatively, it is also possible that some kinds of TGF-␤-responsive intracellular signalings that are independent of Smad3 pathway might alter the activity of Smad3 signaling.
There are actually three distinct MAPKs that have been identified in mammalian cells, referred to as extracellular signal-regulated kinases/p42/44 MAPK (ERK1/2), p38 MAPK (P38), and c-Jun N-terminal kinases (JNK)/stress-activated protein kinases (18). These MAPKs are all proline-directed, serine-threonine kinases that are activated on threonine and tyrosine residues in response to a wide variety of extracellular stimuli. TGF-␤ also stimulates ERK1/2, P38, and JNK in a variety of cell lines (18). Numerous reports suggested that MAPK pathways cross-talk with Smad pathway and modulate the transcriptional regulation of the target genes (19 -25). Our aim of this study is to clarify the role of MAPKs in TGF-␤ and Smad3 pathways in osteoblastic cells.

EXPERIMENTAL PROCEDURES
Materials-MC3T3-E1 cells were kindly provided by Dr. H. Kodama (Ohu Dental College, Ohu, Japan). Human recombinant TGF-␤ 1 , mouse anti-c-Myc antibody, and mouse anti-␤-actin monoclonal antibody were purchased from Sigma. Anti-Smad3 antibody was purchased from Zymed Laboratories (San Francisco, CA). PD98059, U0126, SB203580, curcumin, and dicumarol were purchased from Sigma. All of the other chemicals used were of analytical grade.
Construct and Transient or Stable Transfection-Myc-tagged Smad3 was prepared as described previously (26). Smad3 DNA was derived from rat. A mutant form of Myc-tagged Smad3 (Smad3⌬C), in which the MH2 domain corresponding to amino acid residues 278 -425 was removed, was kindly provided by Dr. Y. Chen. Myc-Smad3, Myc-Smad3⌬C, and empty vector (pcDNA3.1ϩ) (each 3 g) were transfected to MC3T3-E1 cells with LipofectAMINE (Invitrogen). 6 h later, the cells were fed with fresh ␣-MEM containing 10% FBS. 48 h later, the cells were employed as transiently transfected ones for the experiments. For a stable transfection, after incubation in ␣-MEM containing 10% FBS for 48 h, the cells were passaged, and the clones were selected in ␣-MEM supplemented with G418 (0.3 mg/ml) (Invitrogen) and 10% FBS. To rule out the possibility of clonal variation, we characterized at least three independent clones for each stable transfection. Empty vector (V)-transfected cells were used as the control.
Luciferase Assay-MC3T3-E1 cells were seeded at a density of 2 ϫ 10 5 /6-well plate. 24 h later, the cells were transfected with 3 g of the reporter plasmid (p3TP-Lux) and the pCH110 plasmid expressing ␤-galactosidase (1 g) using LipofectAMINE (Invitrogen). 15 h later, the medium was changed to ␣-MEM containing 4% FBS, and the cells were incubated for an additional 9 h. Thereafter, the cells were cultured for 24 h in the absence or presence of TGF-␤ in ␣-MEM containing 0.2% FBS. The cells were lysed, and the luciferase activity was measured and normalized to the relative ␤-galactosidase activity as described (26).
Protein Extraction and Western Analysis-The cells were lysed with radioimmunoprecipitation buffer with 0.5 mM phenylmethylsulfonyl fluoride, complete protease inhibitor mixture, 1% Triton X-100, and 1 mM sodium orthovanadate. The cell lysates were centrifuged at 12,000 ϫ g for 20min at 4°C, and the supernatants were stored at Ϫ80°C. Protein quantitation was performed with BCA protein assay reagent (Pierce). 20 g of protein was denatured in SDS sample buffer and separated on 10% polyacrylamide-SDS gel. The protein was transferred in 25 mM Tris, 192 mM glycine, and 20% methanol to polyvinylidene difluoride. The blots were blocked with Tris-buffered saline (20 mM Tris-HCl, pH 7.5, and 137 mM NaCl) plus 0.1% Tween 20 containing 3%

FIG. 2. Dominant negative effects of Smad3⌬C.
A, cells expressing V and Myc-Smad3⌬C were cultured for 48 h after transfection. Then protein extraction and Western analysis were performed using anti c-Myc antibody. B, cells were transfected with 3 g of the reporter plasmid (p3TP-Lux), the pCH110 plasmid expressing ␤-galactosidase (1 g), and 3 g of V/Smad3⌬C. 48 h later, the cells were stimulated with 5 ng/ml TGF-␤. Then, 24 h later, the cells were harvested, and the relative luciferase activity was measured as described under "Experimental Procedures." The values of relative luciferase activity represent the means Ϯ S.E. *, p Ͻ 0.01, compared with the TGF-␤-untreated group. C, confluent cells were fed with fresh serum-free medium with or without 2.5 ng/ml TGF-␤ for 48 h. The ALP activity was measured as described under "Experimental Procedures." Each bar is expressed as the mean Ϯ S.E. (n mol/min/mg protein) of four determinations. *, p Ͻ 0.01, compared with the TGF-␤-untreated group. dried milk powder. We used anti-Myc antibody as the first antibody to confirm the high expression. Anti-Myc antibody was immunized against the sequence of amino acid residues 410 -419 in the epitope of human c-Myc. Anti-␤-actin antibody was used to confirm the equal supplement of equal protein on each lane. The antigen-antibody complexes were visualized using the appropriate secondary antibodies (Sigma), and the enhanced chemiluminescence detection system, as recommended by the manufacturer (Amersham Biosciences).
RNA Extraction and Northern Analysis-Total RNA was prepared from MC3T3-E1 cells using the acid guanidinium-thiocyanate-phenolchloroform extraction method (27). 20 g of total RNA was denatured, run on a 1% agarose gel containing 2% formaldehyde, then transferred to a nitrocellulose membrane, and fixed with ultraviolet light (Funa-UV-Linker, Funakoshi, Tokyo, Japan). The membrane was hybridized to a 32 P-labeled (Amersham Biosciences) DNA probe overnight at 42°C. The hybridization probes were the 2.8-kb fragment of the ␣1 gene of type I procollagen (COLI) (a gift from Dr. T. Kimura, Osaka University, Osaka, Japan). After hybridization, the filter was washed twice with 2ϫ standard saline citrate containing 0.5% SDS and subsequently washed twice with 0.1ϫ standard saline citrate containing 0.5% SDS at 58°C for 1 h. The filter was exposed to x-ray film using an intensifying screen at Ϫ80°C. All of the values were normalized for RNA loading by probing blots with human ␤-actin cDNA (Wako Pure Chemical Industries, Ltd., Osaka, Japan).
Assay of ALP Activity-Confluent cells in 24-well plates were rinsed three times with phosphate-buffered saline, and 600 l of distilled water was added to each well. ALP activity was assayed at 37°C by a method modified from that of Lowry et al. (28). In brief, the assay mixtures contained 0.1 M 2-amino-2-methyl-1-propanol (Sigma), 1 mM MgCl 2 , 8 mM p-nitrophenyl phosphate disodium, and cell homogenates. After 3 min of incubation, the reaction was stopped with 0.1 N NaOH, and the absorbance was read at 405 nm. A standard curve was prepared with p-nitrophenol (Sigma). Each value was normalized with the value in protein content. ALP staining was performed as described previously by Harlow and Lane (29). In brief, cultured cells were rinsed in phosphate-buffered saline, fixed in 100% methanol, rinsed with phosphatebuffered saline, and then overlaid with 1.5 ml of 0.15 mg/ml 5-bromo-4-chloro-3-indolylphosphate (Invitrogen) plus 0.3 mg/ml nitro blue tetrazolium chloride (Invitrogen) in 0.1 M Tris-HCl, pH 9.5, 0.01 N NaOH, 0.05 M MgCl 2 , followed by incubation at room temperature for 2 h in the dark.
Assay of Mineralization-The mineralization of MC3T3-E1 cells was determined in 6-and 12-well plates using von Kossa staining and Alizarin Red staining, respectively. After the confluent cells were grown in ␣-MEM supplemented with 10% FBS, 1% penicillin/streptomycin, and 10 mM ␤-glycerophosphate for 2 weeks, the cells were fixed with 95% ethanol and stained with AgNO 3 by the Von Kossa method to detect phosphate deposits in bone nodules (30). At the same time, the other plates were fixed with ice-cold 70% ethanol and stained with Alizarin Red S (Sigma) to detect calcification. For quantitation, the cells stained with Alizarin Red were destained with ethylpyridinium chloride (Wako Pure Chemical Industries, Ltd.), and then the extracted stain was transferred to a 96-well plate, and the absorbance at 562 nm was measured using a microplate reader, as described previously (31).
Statistics-The data are expressed as the means Ϯ S.E. The statistical analysis was performed using an unpaired t test or analysis of variance.
TGF-␤-induced Activations of MAPKs Were Independent of Smad3 Signaling Pathway-We investigated whether TGF-␤induced activations of MAPKs would be independent of Smad3 signaling pathway. The MH2-region of Smad3 is indispensable for protein-protein interaction and the transcriptional regulation of the target genes (11,12). In several studies, C-terminally truncated Smad3 was used to inactivate endogenous Smad3 in a dominant negative manner (20). We therefore used the Smad3⌬C, which lacks the MH2-region. We confirmed that the Myc signal was detected in Myc-Smad3⌬C-transfected MC3T3-E1 cells but not in V-transfected cells ( Fig. 2A). To investigate whether Smad3⌬C has a dominant negative effect on TGF-␤-induced transcriptional activity, we employed luciferase assay using 3TP-Lux containing the promoter of plasminogen inhibitor 1 with a Smad3-specific responsive element. Although TGF-␤ promoted luciferase activity in V-transfected MC3T3-E1 cells, Smad3⌬C suppressed TGF-␤-induced luciferase activity (Fig. 2B). These findings suggested that Smad3⌬C exhibited dominant negative effects on TGF-␤-Smad3 signaling in MC3T3-E1 cells. We examined whether activations of MAPKs by TGF-␤ would be dependent or independent of Smad3. TGF-␤ increased the phosphorylation of ERK1/2, P38, and JNK, and Smad3⌬C did not affect the phosphorylation of these MAPKs by TGF-␤ in MC3T3-E1 cells (Fig. 3). These results indicated that TGF-␤-induced activations of ERK1/2, P38, and JNK were independent of TGF-␤-Smad3 signaling pathway.

Requirement of ERK1/2 and JNK in the Inhibitory Effects of TGF-␤ on ALP Activity and Mineralization-Although TGF-␤ inhibits ALP activity and mineralization in MC3T3-E1 cells in
our study and previous studies (4 -7, 16), our recent study revealed that Smad3 promoted them in MC3T3-E1 cells (16). These findings raised the hypothesis that TGF-␤ inhibits ALP activity and mineralization of osteoblasts through a pathway other than Smad3 signaling in MC3T3-E1 cells. As shown in Fig. 2C, Smad3⌬C did not affect ALP activity inhibited by TGF-␤, suggesting that TGF-␤ inhibited ALP activity through pathways other than Smad3. We therefore investigated the effects of MAPK inhibitors on TGF-␤-inhibited ALP activity and mineralization. PD98059 and U0126 (each 10 M) rescued the reduction of ALP activity and mineralization by TGF-␤ (Figs. 4 and 5). Curcumin and dicumarol (each 10 M) also rescued them (Figs. 4 and 5). On the other hand, 10 M SB203580 did not affect the inhibitory effects of TGF-␤ (Figs. 4 and 5). These results indicated that TGF-␤ inhibited ALP activity and mineralization through ERK1/2 and JNK pathways in osteoblasts.

Inhibitors of ERK1/2 and JNK Enhanced Smad3-induced ALP Activity and Mineralization-To test the hypothesis that
MAPKs activated by TGF-␤ negatively regulates Smad3induced ALP activity and mineralization, we investigated the effects of MAPK inhibitors on Smad3-induced ALP activity and mineralization by using stably Smad3-overexpressed MC3T3-E1 cells. Smad3 overexpression promoted ALP activity and mineralization in MC3T3-E1 cells (Figs. 6 and 7), as described in our previous study (16). PD98059, U0126, curcumin, and dicumarol enhanced Smad3-induced ALP activity and mineralization, whereas SB203580 suppressed Smad3-induced ALP activity and mineralization (Figs. 6 and 7), indicating that inhibitors of ERK1/2 and JNK augmented Smad3-induced ALP activity and mineralization in osteoblasts. ERK1/2 and JNK pathways might negatively regulate Smad3-induced ALP activity and mineralization in osteoblasts.
Inhibitors of ERK1/2 and JNK Enhanced Transcriptional Activity of Smad3-To investigate whether MAPKs activated by TGF-␤ would negatively regulate the transcriptional activity of Smad3, we employed luciferase assay using 3TP-lux. Without MAPK inhibitors, TGF-␤ promoted transcriptional activity at a level of about three times that of the basal line in MC3T3-E1 cells (Fig. 8A). However, PD98059 and curcumin significantly enhanced TGF-␤-induced transcriptional activity more effectively than that without these MAPK inhibitors (Fig.  8A). Moreover, Smad3-induced transcriptional activity was also significantly increased by PD98059 and curcumin, compared with that without these inhibitors (Fig. 8B). These results indicated that ERK and JNK pathways negatively regulated transcriptional activity of the TGF-␤-Smad3 signaling pathway in osteoblasts.
Inhibitor of ERK1/2 but Not JNK Enhanced TGF-␤-induced Expression of COLI-Although the effects of TGF-␤ and Smad3 on ALP activity and mineralization were contrary, both TGF-␤ and Smad3 promoted the expression of COLI in MC3T3-E1 cells (16). We investigated the effects of PD98059 and curcumin on TGF-␤-induced COLI mRNA expression in MC3T3-E1 cells. PD98059, but not curcumin, enhanced TGF-␤-induced COLI mRNA expression (Fig. 9). These findings suggested that inhi-bition of ERK, but not JNK, enhanced TGF-␤-induced COLI expression in osteoblasts.

DISCUSSION
In the present study, the inhibition of the Smad3 signaling pathway by the expression of dominant negative type Smad3, Smad3⌬C, did not affect TGF-␤-induced activations of MAPKs, ERK1/2, P38, and JNK. These findings indicated that activations of these MAPKs by TGF-␤ were independent of Smad3 signaling pathways. Indeed, Engel et al. (20) reported that JNK was rapidly and transiently activated by TGF-␤ receptor type I in a Rho-GTPase-dependent and Smad-independent manner in mink lung epithelial (Mv1Lu) cells and a breast carcinoma cell line (MDA-MB-468). These findings suggested that the coincident activation of the Smad and JNK/AP-1 pathways is necessary for full transcriptional activation in response to TGF-␤. Moreover, TGF-␤ induced proliferation in colon carcinoma cells by Ras-dependent but Smad-independent down-regulation of p21 cip1 (32). On the other hand, Lai and Cheng (33) reported that TGF-␤ activated Ras/MAPK/AP-1 signaling and that the stimulation of AP-1 by TGF-␤ was dependent on Smad signaling in human osteoblastic cells. In addition, several studies suggested the cross-talk of Ras-ERK/MAPKs and TGF-␤-Smad pathway (19,(22)(23)(24)(25). Taken into account with our findings, TGF-␤-Smad signaling, MAPK pathways, and AP-1 might cross-talk in a complex manner, although there are TGF-␤responsive and Smad3-independent MAPK pathways in mouse osteoblastic cells.
Several studies and our previous study indicated that TGF-␤ suppressed ALP activity and mineralization in osteoblasts including MC3T3-E1 cells (8,9,34). In contrast, Smad3 promoted them in MC3T3-E1 cells (16). In the present study, inactivation of ERK1/2 and JNK with their specific inhibitors antagonized the inhibitory effects of TGF-␤ on ALP activity and mineralization in MC3T3-E1 cells (Figs. 4 and 5). These findings suggested that the activation of ERK1/2 and JNK contributes to the inhibitory regulation of ALP activity and mineralization by TGF-␤. In the present study, the specific inhibitors of ERK1/2 and JNK enhanced the transcriptional activity (Fig. 8), as well as ALP activity and mineralization (Figs. 6 and 7) induced by Smad3. Taken together, these findings suggested that ERK1/2 and JNK signaling pathways negatively regulate Smad3 signaling pathway, resulting in the suppression of Smad3-induced ALP activity and mineralization in osteoblasts. In support of this speculation, several studies have shown that JNK, as well as c-Jun and JunB, represses Smad3-mediated transcriptional activity in the human hepatoma cell line (HepG2), mouse fibroblasts, and keratinocytes (35,36). Moreover, several studies indicated that oncogene Ras or epidermal growth factor-induced Ras repressed TGF-␤/Smad signaling in cancer cells or cell lines other than osteoblasts (19,(22)(23)(24)(25). The present study indicated that TGF-␤ activates ERK1/2 as well as JNK and inhibits ALP activity in a manner independent of Smad3 in MC3T3-E1 cells. These findings therefore suggested that TGF-␤-responsive ERK1/2 and JNK cascades negatively regulate Smad3-induced transcriptional activity as well as ALP activity and mineralization in osteoblasts. The negative signal of TGF- ␤-responsive ERK1/2 and JNK for the Smad3 signaling pathway might explain the discrepant effects of TGF-␤ and Smad3 on ALP activity and mineralization in MC3T3-E1 cells. However, we cannot rule out the possibility that some other mechanisms might be responsible for these discrepant effects of TGF-␤ and Smad3. First, there might be some intracellular signaling pathways by which Smad3 but not TGF-␤ enhances ALP activity. Takeuchi et al. (37) reported that the type I collagen and ␣ 2 ␤ 1 -integrin interaction up-regulates ALP activity and down-regulates TGF-␤ receptor activity, which allows the cells to escape the inhibitory effects of TGF-␤ in MC3T3-E1 cells. Therefore, ␣ 2 ␤ 1 -integrin may play some role in Smad3stimulated ALP activity. Smad3 overexpression may enhance the interaction between type I collagen and integrin by upregulating the expression of the integrin in MC3T3-E1 cells. The second mechanism concerns the transformed state of osteoblasts. TGF-␤ promotes the production of COLI in ROS 17/2.8 (10) and MC3T3-E1 cells (2,37). However, TGF-␤ stimulates and inhibits ALP activity in ROS 17/2.8 and MC3T3-E1 cells, respectively (10,38). Although the effects of Smad3 on ROS 17/2.8 cells are unknown, the intracellular signals that modulate the effects of TGF-␤ and Smad3 might be different, depending on the cell lines, species, and how the cells have been transformed.
The present study could not clarify the exact molecular mechanism by which TGF-␤ negatively regulates Smad3-induced transcriptional activity as well as ALP activity and mineralization through ERK1/2 and JNK. Several studies revealed the mechanisms by which RAS/ERK1/2 or JNK cascades negatively regulate TGF-␤ signaling pathway. Kretzschmar et al. (22) reported that oncogenic Ras inhibited the TGF-␤-induced nuclear accumulation of Smad2 as well as Smad3 and Smaddependent transcription by phosphorylation of Smad2 and Smad3 via ERK1/2, whereas Saha et al. (24) reported that oncogenic Ras repressed TGF-␤ signaling by ERK1/2-dependent down-regulation of Smad4. The AP-1 family protein c-Jun, which is a substrate for JNK, directly suppressed Smad/DNA interaction (36). It is also possible that the target step in which TGF-␤-responsive ERK1/2 and JNK negatively regulate Smad3 signaling pathway might be the recruitment of Smad3 to intracellular membranes that contains TGF-␤-receptor type I by Smad anchor for receptor activation (39,40), phosphorylation/activation of a motif SSXS in the C terminus of Smad3 by serine/threonine kinase activity of TGF-␤-receptor type I (41), the association of Smad3 and the common partner, Smad4, the translocation of the Smad3-Smad4 complex into the nucleus, and its DNA binding or interaction with other transcriptional regulators. In our preliminary study, both ERK1/2 and JNK inhibitors did not promote the TGF-␤-responsive nuclear translocation of Smad3 (data not shown). These findings suggested that the TGF-␤-responsive ERK1/2 and JNK cascade might not affect the nuclear translocation of Smad3 in MC3T3-E1 cells. Furthermore, there might be the autoinduction system that TGF-␤ up-regulates the production of TGF-␤ itself (42). It is possible that TGF-␤ increases the production of Smad3 and that ERK1/2 and JNK negatively regulate Smad3 expression by the TGF-␤-responsive autoinduction system. Further studies are in progress in our laboratory. Smad3 enhanced ALP activity and mineralization in MC3T3-E1 cells (16), suggesting that Smad3 is involved in the transcriptional mechanism leading to bone formation. In support of this, Borton et al. (17) recently reported that mice with targeted deletion of Smad3 were osteopenic compared with wild type littermates, because of a lower rate of bone formation. In the present study, inhibitors of ERK1/2 and JNK rescued TGF-␤-inhibited ALP activity and mineralization in MC3T3-E1 cells. Moreover, these inhibitors enhanced Smad3-induced ALP activity and mineralization in these cells. The negative effects of TGF-␤ on ALP activity and mineralization in osteoblasts negatively influence bone formation. If in vivo ERK1/2 inhibitor and/or JNK inhibitor FIG. 9. Effects of MAPK inhibitors on TGF-␤-induced expression of COLI mRNA. Confluent MC3T3-E1 cells were treated with TGF-␤ (2.5 ng/ml) in serum free ␣-MEM in the presence or absence of MAPK inhibitors (each 10 M) after pretreatment with these inhibitors for 1 h. 24 h later, RNA extraction and Northern analysis were performed as described under "Experimental Procedures." FIG. 8. Effects of MAPK inhibitors on TGF-␤-and Smad3-induced transcriptional activity. A, MC3T3-E1 cells were transfected with 3 g of the reporter plasmid (p3TP-Lux) and the pCH110 plasmid expressing ␤-galactosidase (1 g). 48 h later, the cells were stimulated with 5 ng/ml TGF-␤ in the presence or absence of MAPK inhibitors (each 10 M) after pretreatment with MAPK inhibitors for 1 h. 24 h later, the cells were harvested, and relative luciferase activity was measured, as described under "Experimental Procedures." The values of relative luciferase activity represent the means Ϯ S.E. *, p Ͻ 0.01, compared with inhibitors-untreated group. B, the cells were transfected with 3 g of the p3TP-Lux, the pCH110 plasmid expressing ␤-galactosidase (1 g), and 3 g of V/Smad3. The treatment and measurement of luciferase activity were performed, as in A.
antagonize the negative effects of TGF-␤ on bone formation and enhance the positive effects of TGF-␤-responsive Smad3 on bone formation, the combination of TGF-␤ and inhibitors of ERK1/2 and/or JNK may be a novel therapeutic strategy for bone disease or fracture healing. Type I collagen is the abundant protein in bone matrix and plays an important role in bone formation, mineralization, and maintenance of bone strength (37). As shown in Fig. 9, TGF-␤-induced expression of COLI was enhanced by the ERK1/2 inhibitor but not by the JNK inhibitor. We therefore speculated that the combination of TGF-␤ with ERK1/2 inhibitors might be better than with JNK inhibitor for inducing bone anabolic action.
In conclusion, the present study indicated that TGF-␤-activated ERK1/2 and JNK cascades negatively regulated the transcriptional activity as well as ALP activity and mineralization induced by Smad3 in mouse osteoblastic cells. We propose that Smad3 is an important molecule in the regulation of bone formation and that the local combined administration TGF-␤ with inhibitors of ERK1/2 might be a novel therapeutic strategy for the stimulation of bone formation.