Hepatocyte growth factor (HGF) is activated and the expression of BMP receptors (BMPRs) is induced around the fracture site during the early phase of fracture repair. HGF facilitates the expression of BMPRs in mesenchymal cells. This study suggests that HGF contributes to fracture repair by inducing the expression of BMPRs.
Introduction: The precise mechanisms that control the upregulation of BMP, BMPRs, and other molecules involved in bone repair are not completely understood. In this study, we hypothesized that HGF, activated through the action of thrombin on the HGF activator, may enhance BMP action through the local induction of BMP or BMPRs.
Materials and Methods: Callus samples from tibial fractures in mice were harvested for immunohistochemical analysis of HGF and phosphorylated c-Met, for in situ hybridization of BMPRs, and for real-time RT-PCR analysis for the expression of HGF, c-Met, and BMPRs. To study the changes in gene expression of BMPRs in response to HGF, C3H10T1/2 cells were cultured with or without HGF and harvested for real-time RT-PCR and for Western blot analysis. To evaluate the contribution of HGF to the biological action of BMP2, C3H10T1/2 cells and primary muscle-derived mesenchymal cells were precultured with HGF and cultured with BMP2. In addition, the expression of the luciferase gene linked to the Id1 promoter containing the BMP responsive element and alkaline phosphatase (ALP) activity were assayed.
Results: Positive immunostaining of HGF and phosphorylated c-Met was detected around the fracture site at 1 day after the fracture was made. mRNA expression of BMPRs was increased 1 day after fracture and localized in mesenchymal cells at the fracture site. From an in vitro study, the expression of mRNA for BMPRs was elevated by treatment with HGF, but the expression of BMP4 did not change. Western blot analysis also showed the upregulation of BMPR2 by HGF treatment. The results from the luciferase and ALP assays indicated increased responsiveness to BMPs by treating with HGF.
Conclusions: This study indicates that HGF is activated and expressed at the fracture site and that HGF induces the upregulation of BMPRs in mesenchymal cells. Furthermore, HGF may facilitate BMP signaling without altering the expression of BMP molecules.
THE REGULATION OR promotion of repair in fractures or damaged bone is one of the most important subjects in the basic research and clinical practice areas of orthopedic surgery. (1) Classically, it has been recognized that fracture repair is achieved by local new bone or callus formation, which is attributed to the regenerating potential inherent to skeletal tissues. Attempts have been made to enhance the regeneration potential of bone to promote bone repair. To devise a more effective method to achieve this goal, it is important to gain a more precise understanding of the biological mechanisms underlying the repair reaction in skeletal tissues. For many years, the precise molecular and cellular events involved in bone repair have remained a mystery. Fortunately, recent advances in molecular biology and related technologies have provided new approaches and insights into our understanding of how bone is repaired. Previous studies had linked fracture repair with the molecule(s) responsible for the regenerating potential of bone. (1, 2) Because the discovery of bone-inducing activity in organic bone matrix, the sources have been identified (BMP-2, -4, and -7) and are now currently produced by the use of recombinant DNA technology. (3) Thereafter, receptors for the BMPs (type I and type II serine/threonine kinase receptors) and Smad-dependent pathways involved in the cascade of BMP-related intracellular signaling have also been identified. (4) In terms of fracture repair, it has been shown that the expression of BMPs and BMP receptors (BMPR1 and 2) is upregulated in cells surrounding the fracture site during the initial phase of the repair process. (5) However, the trigger for the regulation of BMP expression elicited by the onset of fracture remains unknown.
Hepatocyte growth factor (HGF) was originally cloned as a potent growth factor for hepatocytes in the regenerating liver. (6) Subsequent extensive studies of HGF revealed a variety of biological activities (motogenic, mitogenic, or morphogenetic potential) that could contribute to the regenerating reaction in a broad range of damaged organs such as liver, (7) heart, (8) lung, (9) kidney, (10) and blood vessels(11) or other organs(12, 13) by binding with its receptor, c-Met, and by phosphorylation of a tyrosine residue in its intracellular domain with kinase activity. However, the involvement of HGF in the bone repair reaction has, to date, not been explored. This report describes the results of a study into the role of HGF in the fracture healing process and, specifically, the effects of this growth factor on the biological activities and signaling mechanism of BMPs in in vivo and in vitro experimental systems.
MATERIALS AND METHODS
The right tibias of 65 ICR male mice were fractured by manual bending under anesthesia by methyl ether inhalation. The mice were maintained and monitored in cages with free access to water and food. Thirteen mice were killed by anesthesia at scheduled intervals (0, 1, 3, 7, and 14 days after the onset of the fracture), and nine of the fractured tibias with soft tissue around the fracture site were harvested: six were fixed in 10% neutral buffered formalin for in situ hybridization and immunohistochemical analysis of phosphorylated c-Met, and three were fixed in 70% ethanol for immunohistochemical analysis of HGF. The four remaining tibias were harvested for RNA extraction. This experimental protocol was approved by the Institutional Committee of Animal Care and Experiments of Osaka City University.
The specimens were decalcified in 0.5 M EDTA, dehydrated through a graded ethanol series, and embedded in paraffin. Sections of 5 μm thickness were prepared using a microtome and processed for routine hematoxylin/eosin staining, immunohistochemistry, and in situ hybridization.
Polyclonal antibody against rat HGF (cross-reacts with murine HGF) and anti-c-Met (pYpYpY1230/1234/1235) phospho-specific antibody were obtained from the Institute of Immunology (Tokyo, Japan) and BioSource International (Camarillo, CA, USA), respectively, and were diluted to 10 μg/ml and 1:50, respectively, as described previously. (14) Sections were deparaffinized and treated with 0.3% hydrogen peroxidase for 30 minutes at room temperature to block endogenous peroxidase. The sections were treated with 3% defatted dried milk in PBS for 30 minutes. After blocking with the dried milk solution, the sections were incubated with or without the primary antibody overnight at 4°C, washed three times with PBS for 5 minutes, and incubated with biotinylated horseradish peroxidase-conjugated anti-rabbit second antibody (DakoCytomation, Glostrup, Denmark) for 30 minutes at room temperature. After washing three times with PBS for 5 minutes, the sections were incubated with the Vectastain Elite ABC visualization system (Vector Laboratories, Burlingame, CA, USA), and the color reaction was developed by diaminobenzidine (DAB) followed by washing with distilled water. Finally, sections were counterstained with methyl green for 5 minutes at room temperature.
In situ hybridization
A 0.50-kb fragment of mouse BMPR1A cDNA and a 0.55-kb fragment of mouse BMPR2 cDNA were used as templates to synthesize RNA probes. They were subcloned into pBluescript SK(−) plasmid (Stratagene, La Jolla, CA, USA). The cDNA encoding the mouse BMPRs were obtained by RT-PCR, and the primers of BMPRs for PCR were selected as described previously. (15) In situ hybridization was carried out as described previously. (16)
The mouse fibroblastic cell line C3H10T1/2 was obtained from the RIKEN Cell Bank (Tsukuba, Japan). Primary muscle-derived mesenchymal cells were prepared from the hindlimb of mice embryo (E15.5) as described previously without using collagenase treatment. (17) Cells were seeded at a cell density of 3 × 105 cells per 100-mm plastic dish and cultured in α-MEM (Sigma, St Louis, MO, USA) containing 10% (vol/vol) heat-inactivated FBS (Gibco, Grand Island, NY, USA) for growth or 2.5% FBS for examination at 37°C in 5% CO2 humidified air. On reaching confluency, the cells were used in the subsequent experiments. C3H10T1/2 cells maintained between passages 7 and 12 were used for the in vitro experiments.
Cell proliferation was evaluated with an assay kit as described previously(18) (Promega). Fifteen microliters of dye solution was added to the cells on each well of a 96-well tissue culture plate and incubated at 37°C for 4 h. One hundred microliters of Solubilization/Stop solution was added to each well, and the plate was incubated at 37°C for 1 h and mixed thoroughly. The plates were read on a microplate reader at a wavelength of 595 nm.
Total RNA was prepared from cells treated with or without 5 ng/ml of recombinant human HGF (rhHGF) for 0, 1, 3, 6, and 12 h on a 10-cm dish or from homogenized fractured tibias using ISOGEN (Wako, Osaka, Japan). One microgram of total RNA was reverse-transcribed into first-strand cDNA with oligo dT primer using Superscript II reverse transcriptase (Invitrogen). Real-time RT-PCR was performed according to the manufacturer's instructions. Sequences for primers and TaqMan fluorogenic probes (Applied Biosystems, Foster City, CA, USA) were as follows: BMPR1A, forward primer, 5′-GGATCTCTCTATGACTTCCTGAAATGT-3′, reverse primer, 5′-CAGCAGAATAAGCTAACTTGAGTAGGG-3′, TaqMan probe, 5′(FAM)-CCACACTAGACACCAGAG-(TAMRA)3′; BMPR2, forward primer, 5′-GCCAAGATGAATACAATCAATGCA-3′, reverse primer, 5′-CTTCTACCTGCCACACCATTCATA-3′, TaqMan probe, 5′(FAM)-AGAGCCTCATGTGGTGAC-(TAMRA)3′. TaqMan probes for HGF, BMP-4, and GAPDH were purchased from Applied Biosystems. Real-time RT-PCR for c-Met was performed using SYBR Green Supermix (Bio-Rad Laboratories). Experimental samples were matched to a standard curve generated by amplifying serially diluted products using the same PCR protocol. To correct for variability in RNA recovery and efficiency of reverse transcription, GAPDH cDNA was amplified and quantified in each cDNA preparation. Normalization and calculation steps were performed as described previously. (19) For the in vitro study, experiments were performed on three separate test occasions with an n of 3 for each test occasion.
Immunoprecipitation and Western blot analysis
Cells were plated at a density of 1–2 × 104 cells/cm2 on 100-mm plates and cultured for 2–3 days until a confluence of 80–90% was reached. rhHGF (5 ng/ml) was added to the media, and the cells were cultured for 12, 24, 48, or 72 h to examine the time dependency in response to HGF. To examine the dose-dependent response to HGF, different concentrations of HGF (0, 1, 5, 10, and 20 ng/ml) were added to the plates, and the cells were cultured for 72 h. After HGF treatment, immunoprecipitation was performed using the immunoprecipitation kit, Immunoprecipitation Starter Pack (Amersham Bioscience) according to the manufacturer's instructions. The polyclonal goat antibodies against mouse BMPR2 (1 μg for each sample; Santa Cruz Biotechnology, Santa Cruz, CA, USA) were used. The resultant sample was boiled for 5 minutes in 20 μl of sample buffer for SDS-PAGE as described previously. (15) Equal amounts of protein samples were applied and run on each lane of an SDS 10% acrylamide gel (40 mA, low voltage, 90 minutes), and ultimately blotted to an enhanced chemiluminescence (ECL) nitrocellulose membrane (Amersham). The membranes were developed with ECL plus reagent (Amersham). To examine the possible contribution of platelet growth factors to modulate or upregulate expression of the BMPRs, the heterodimer of platelet-derived growth factor AB (PDGF-AB; R&D Systems) was used at a dose of 5 ng/ml using the same protocol described for rhHGF. We selected PDGF-AB instead of PDGF-BB based on a previous report indicating the negative effect of PDGF-BB on osteogenic differentiation. (20) The intensity of staining of each band was calculated using a digital densitometer (Bio-Rad).
Relative luciferase activity assay
To detect the changes in intensity of BMP signaling after treatment with HGF, the C3H10T1/2 cells and primary muscle-derived mesenchymal cells were transfected with 1 μg of the luciferase gene linked to the Id1 promoter containing the BMP responsive element or a mutated form without the promoter function (kindly provided by Dr T Katagiri, Saitama Medical School Research Center for Genomic Medicine, Saitama, Japan). Each construct was used together with 250 ng of the control luciferase vector (pRL) that was used as an internal control to calculate transfection efficiency by the calcium phosphate/DNA precipitation method. For the transfection, 5 × 103 cells/well were plated on 48-well plates for 1 day, and medium was changed 2 h before transfection. Twelve hours later, cells were washed with Hanks' balanced salt solution (Sigma) twice, and cells were treated with or without 5 ng/ml of rhHGF for 0, 12, 24, 48, or 72 h. The other group was treated with medium containing 1 μg/ml of the anti-c-Met antibody (R&D Systems) before pretreatment with HGF for 72 h. The culturing of the cells was continued with or without the addition of 100 ng/ml of rhBMP-2 (Yamanouchi Pharmaceutical Co., Tokyo, Japan) for an additional 24 h. Luciferase activity was determined by a Dual-Glo Luciferase Reporter Assay System (Promega, Madison, WI, USA) as previously described. (21)
Protocols for assay for alkaline phosphatase activity
Alkaline phosphatase (ALP) levels in C3H10T1/2cells and primary muscle-derived mesenchymal cells were assayed to check for the effects of HGF on ALP induction by BMP signaling at the translational level. Both types of cells were seeded at a density of 1 × 104 cells/well in 48-well plates (n = 8 per group). On achieving confluency, the cells were pretreated with 5 ng/ml of rhHGF in culture medium for 0, 12, 24, 48, and 72 h, washed twice with PBS, and treated with medium plus 100 ng/ml of rhBMP-2 for 2 days. The other group was treated with medium containing 1 μg/ml of the anti-c-Met antibody as described above. The cells were washed twice with normal saline, and ALP activity was assayed as described previously using p-nitrophenylphosphate as the substrate. (22) The effects of HGF on BMP-induced ALP activity were normalized by protein. Experiments were performed in triplicate independently.
Data are expressed as the mean ± SD for each group. Statistical differences among treatment groups were analyzed using Fisher's protected least significant difference (PLSD) test. Values of p < 0.05 were considered significant.
Immunohistochemical detection of endogenous HGF and phosphorylated c-Met at the early phase of fracture repair
Immunohistochemical analyses using anti-rodent HGF and phosphorylated c-Met revealed positive signals especially around the fracture sites 24 h after the onset of the fracture. The HGF+ cells were predominantly localized in stromal cells around the fracture site (Figs. 1B and 1C), whereas c-Met phosphorylation was noted mainly in parenchymal areas nearby the fracture site (Figs. 1F and 1G). In contrast, there was no apparent signal in the negative controls stained without primary antibodies (Figs. 1D and 1H). Of note, immunopositive signals for HGF and phosphorylated c-Met were not detected around the fracture site immediately after the onset of the fracture and were only weakly visible 3 days after fracture (data not shown). This result indicates a specific activation of HGF/c-Met signals that is caused by the injury of the bone. Based on the histological data, we hypothesized that the paracrine delivery system of HGF toward mesenchymal cells may be critically involved in the initial phase of fracture repair. Furthermore, the upregulation of HGF and c-Met mRNA expression was detected in the early phase of fracture repair by real-time RT-PCR (Fig. 2). Based on previous results indicating upregulated HGF and c-Met expression by activated HGF, (23) these results also provide evidence of immediate HGF activation and activity around the fracture site.
Upregulation of BMPR mRNA expression at the fracture site
In situ hybridization at the fracture site showed mRNA expression of BMPRs in the mesenchymal cells around the fracture site in the early phase of fracture repair, chondrocyte-like cells in the callus at day 7, and osteoblastic cells in newly formed bone at day 14 (Fig. 3). We could not detect the positive staining cells in the specimens using sense probes at any time-point (data not shown). Real-time RT-PCR for BMPRs using extracted total RNA from the fracture site showed a statistically significant 4-fold increase in BMPR1A mRNA expression at day 1 that was maintained up to day 14. The significantly increased BMPR2 mRNA expression was also detected at day 1, and it increased a further 6-fold by day 14 (Fig. 4).
Upregulated expression of c-Met, BMPR1A, and BMPR2 by HGF in an in vitro system
The effects of HGF on cell proliferation under in vitro conditions were examined by MTT assay. There was no significant proliferation effect of HGF on the cells at the concentrations used in this experiment (data not shown). All in vitro experimental results were normalized by values calculated from the MTT assay.
Based on the above results observed in an in vivo system, the effects of HGF in relation to BMP expression and the BMP signaling system were analyzed in an in vitro system using a cell line of mesenchymal origin (C3H10T1/2 cells) and various molecular biological methods.
Real-time RT-PCR analysis showed expression of c-Met in the C3H10T1/2 cells stimulated by exposure to exogenous rhHGF for >1 h (Fig. 5A), thereby mimicking the in vivo result described above.
Although the qualitative RT-PCR analysis indicated the constitutive expression of BMPR1A and BMPR2 in the cells (data not shown), the real-time RT-PCR analysis showed a significant elevation in the expression of the BMP-receptors at 6 h. Thereafter, a decrease in BMPR1A and BMPR2 expression was noted after the addition of rhHGF relative to the control where there was no HGF treatment (Fig. 5B). However, no change in the level of BMP4 expression was noted after treatment with HGF (Fig. 5B).
To verify that the upregulated translation resulted from upregulation of transcription of the BMP receptors, we ran a Western blot analysis for BMPR2. The results confirmed the upregulated translational expression of BMPR2 after addition of rhHGF in a time- (Fig. 5C) and dose- (Fig. 5D) dependent manner. The BMPR2 synthesis was not affected by PDGF-AB at the concentration of 5 ng/ml that we used in this study (Fig. 5C).
Effects of HGF on signaling of BMP
The relative luciferase expression assay in C3H10T1/2 cells (data not shown) and primary muscle-derived mesenchymal cells using the Id1 promoter containing the BMP responsive element showed elevation of transcriptional activity of luciferase by pretreatment with HGF for >48 h followed by treatment with BMP compared with the group that did not receive either HGF or BMP (Fig. 6A). Furthermore, increased transcriptional activity obtained by HGF pretreatment was abolished by addition of the anti-c-Met neutralization antibody. ALP activities in primary muscle-derived mesenchymal cells also were significantly elevated in the group treated with rhHGF for 24 h or more in comparison with the group treated with BMP only. The enhanced BMP-2-induced ALP activity by HGF was also blocked by addition of the anti-c-Met antibody (Fig. 6B). In C3H10T1/2 cells, almost the same results were obtained (data not shown). These results indicate that HGF and its receptor are involved in a mechanism to regulate BMP signaling through the transcriptional regulation of BMPRs.
In this experimental study, we investigated the contribution of HGF to the healing reaction in fracture. The results indicate that localization and activation of HGF occur around the fracture site during the early phase of the healing process and that the expression of HGF is upregulated at the fracture site during the same time. This early upregulation of expression of HGF mRNA resembles that of pulmonary ischemia-reperfusion injury. (24) The observations in healing bone are consistent with previous studies that have shown that HGF is linked with the regenerative processes in a wide range of organs and tissues. As already reported, HGF is excreted from cells in a latent form and is converted to the active form by HGF activation factor (HGF-AF), which in turn is activated by thrombin, a blood coagulation factor. (25) In this type of pathological condition, HGF has multiple roles in the repair of injured tissues, for example, by inducing angiogenesis. The active form of HGF was not detected directly at the fracture sites in this study. However, we could confirm the phosphorylation of c-Met (Fig. 1), which is essential for the biological activity of HGF and mRNA expression of HGF itself and of c-Met (Fig. 2), which could be induced by HGF as previously reported, (23) instead of detection of the active form of HGF, (23) although it could not be easily compared at each time-point because of the heterogeneity of cells in the fracture site. In terms of activation of the locally produced or circulating HGF, the presence of thrombin in the hematoma formed at the fracture site just after the onset of the fracture might contribute through the activation of HGF-AF and HGF. Thus, a nonspecific circulatory disruption at the onset of injury to any tissue might enable the local activation of HGF and initiate the tissue repair reaction. The results of in situ hybridization of BMPRs (Fig. 3) indicate that the fracture healing process begins as a result of the immediate mRNA expression of BMPRs in the cells around the fracture site. It is possible that the immediate reactions mentioned above do not occur by newly induced molecules, but instead by activated molecules at the fracture site. From this perspective, we tried to study the relationship between HGF, the activation of which is related to the hemorrhage, and the expression of BMPRs.
In this study, the upregulated expression of BMPRs at the transcriptional and translational levels was confirmed by real-time RT-PCR and Western blotting, respectively. These results have revealed the potential contribution of HGF to the fracture healing process. The HGF receptor, c-Met, has a tyrosine kinase activity at its intracellular domain and acts downstream of the MAPK cascade. Furthermore, the activating protein-1 binding element has been located in the promoter region of both BMPR1A and BMPR2. These facts suggest that mRNA expression of BMPRs is able to be induced by HGF. However, we have not shown that the transcriptional regulation of BMPRs occurs as a consequence of intracellular signaling by c-Met.
Although BMP molecules were reported to be upregulated in adjacent periosteal cells during the early phase of fracture repair at the fracture site, (1) this type of change in BMP expression was not noted after stimulation by HGF in the in vitro system. However, BMP receptors in the mesenchymal cells were upregulated by HGF and could potentially contribute to healing of the fracture by amplifying BMP signal transduction and promoting fracture healing reactions during the initial phase of fracture repair. This response seems to be specific to HGF, because PDGF, the receptors of which also have tyrosine kinase domains, did not significantly induce expression of BMPRs for C3H10T1/2 in our study. However, the in vitro studies are limited by the clonal nature of the cell types involved in these studies and the absence of circulating hormones.
From our in vivo study, the expression levels of BMPRs were elevated for several days during fracture repair (Fig. 4); however, from our in vitro study, HGF transiently induced mRNA expression of BMPRs (Fig. 5B). This discrepancy can be explained by the following two points: (1) the stimulation to the multipotent cells by HGF may continue during the early phase of fracture repair in vivo, although the stimulation was transient in vitro, and (2) because the in vivo studies take place in an environment different from the in vitro studies, there may be molecules, for example, BMPs, (15) which can induce the expression of BMPRs around the fracture site. Therefore, our results suggest that the effect of HGF in the expression of BMPRs has a significant role as the trigger of fracture repair by BMP signaling despite the effect being transient and mild.
To further understand the interaction between the activation of HGF and the expression of BMPRs, it will be necessary to form a system for the local administration of HGF into the fracture site and use a null animal model. Unfortunately, the null mouse of c-Met(26) and HGF(27, 28) are embryonic lethal as previously reported, so siRNA intervention or a conditional knockout animal model will be required. On the other hand, to clarify the precise mechanism of fracture repair, further study will be necessary to determine the identity and molecular mechanisms by which other factors regulate expression of BMP molecules during the early phase of fracture healing.
This work was supported in part by a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan (Project Grant 12137203) and by grants from the Ministry of Health, Labour and Welfare of Japan. The authors thank the members of Central Laboratory of Osaka City University Graduate School of Medicine for cooperation in this work.