More than 2 million people each year in the United States suffer from osteoporosis-related fracture, and the financial costs for the patients amounted to as much as $17 billion in 2005,1 giving a strong incentive to improve the treatment of patients with osteoporotic bone fracture and learn more about bone physiology. Bone healing process involves a variety of molecular and cellular events, which recapitulate, to a certain degree, some of the characteristics of skeletal development;2, 3 however, to date, the molecular mechanisms behind the fracture healing processes are not fully understood.
The signal transducer and activator of transcription 1 (STAT1) was originally identified as a signaling molecule involved in the interferon (IFN) pathway.4 Interestingly, recent studies revealed that STAT1 participates not only in immune regulations but also in osteoclastogenesis and osteoblast differentiation. It has been shown that both type I and type II interferons (IFN-α/β and IFN-γ) suppress osteoclastogenesis and that STAT1 has critical roles in this IFN-mediated inhibition of osteoclastogenesis.5 In accordance, Stat1−/− mice exhibited an increased osteoclast number and enhanced bone resorption, further corroborating IFN-STAT1 signaling as a critical suppressor of osteoclastogenesis.6 However, in contrast to these observations, Stat1−/− mice showed increased, not decreased, net bone mass with significantly upregulated bone formation, indicating that STAT1 also functions as a potent inhibitor of bone formation in vivo.7 As expected, the further study revealed that STAT1 suppresses bone formation by directly interacting with Runx2, an essential transcription factor for osteoblast differentiation.7 These studies have established STAT1 as a critical negative regulator in both bone resorption and bone formation, and, therefore, STAT1 is regarded as a crucial component in the maintenance of skeletal homeostasis in vivo. On the other hand, little is currently known about potential involvements of STAT1 in the bone fracture healing process.
In the current study, we examined the contribution of STAT1 in fracture healing, and found that calcified callus resorption and the subsequent remodeling, as well as membranous ossification were highly accelerated in Stat1−/− mice and that STAT1 inhibits the transcription of Osterix (Osx), another essential transcription factor for osteoblast differentiation besides Runx2. Furthermore, we found that fludarabine, a potent STAT1 inhibitor, significantly increases bone formation in a heterotopic ossification model. These results reveal previously unknown contributions of STAT1 in bone fracture healing and may have important clinical implications for the treatment of skeletal bone fracture.
MATERIALS AND METHODS
Stat1−/− mice4 were purchased from Taconic Farm Inc. (Hudson, NY, USA). All comparisons described in this study were between the littermates from the crossing between Stat+/− mice. The mice were housed in a specific-pathogen free environment and fed with sterile water and feed. All experiments were performed according to the protocol approved by the Laboratory Animal Care and Use Committee of School of Medicine, Keio University.
Bone Fracture Model and Cortical Bone Defect Model
The bone fracture model was performed essentially as previously described with some modifications.8 In short, under general anesthesia, a transverse osteotomy was performed at the middle of the tibia of 8-week-old male mice with an electric bone saw, and an inner pin of a spinal needle was inserted intramedullary to stabilize the osteotomized bones. A cortical bone defect model in the tibia was performed as previously described.9 Briefly, under general anesthesia, a round cortical defect 1 mm in diameter was made using a dental burr in the anteromedial aspect of the tibia. The tibiae were collected at the designated time points and were histologically and radiologically evaluated. X-rays and CT images (obtained with Scan Xmate-A090S Comscantecno; Yokohama, Japan) were analyzed by ImageJ software (National Institute of Health) and TRY/3D Bon (Ratoc System Engineering, Tokyo, Japan), respectively.
COS-7 cells were maintained in Dulbecco's modified Eagle's medium (DMEM; Sigma-Aldrich, St. Louis, MO, USA) containing 10% fetal bovine serum (FBS) and antibiotics. Primary osteoblasts derived from the calvaria (POBs) of newborn mice were prepared as previously described.10, 11 POBs and MC3T3-E1 cells were maintained in αMEM (Sigma-Aldrich) containing 10% FBS and antibiotics. To induce osteoblast differentiation, POBs were incubated with αMEM Medium supplemented with recombinant human BMP-2 (200 ng/mL, provided by Astellas, Tokyo, Japan), ascorbic acid (50 µg/mL) and β-glycerophosphate disodium salt hydrate (10 mM).
Expression levels for the transcripts for β-actin, Runx2, Osx, and osteocalcin were quantified by RT-PCR using LightCycler (Roche, Basel, Switzerland). cDNA was reverse-transcribed from total RNA prepared from POBs or MC3T3-E1 cells and was used as a PCR template. β-Actin was used as an internal control.
Luciferase reporter vectors bearing Osx promoter sequence (−1269/+91) and NF-κB-responsive element (RE) were generously provided by Dr. Mark S. Nanes (Emory University) and Dr. Akihiko Yoshimura (Keio University), respectively. Luciferase reporter vectors were cotransfected with STAT1 or p65 expression vector into MC3T3-E1 or COS-7 cells using Lipofectamine 2000 (Invitrogen, San Diego, CA, USA). Cells were incubated for 48 h and the luciferase activities were measured by Dual-luciferase Reporter Assay Systems (Promega, Madison, WI, USA) as instructed by the manufacturer.
Heterotopic Ossification Model
Recombinant BMP-2 (3 µg) and/or fludarabine (10 µg, Toronto Research Chemicals Inc, Tronto, ON, Canada) was diluted in 10 µL phosphate-buffered saline (PBS) and immersed in gelatin pellets (Gelfoam; Pfizer Inc, New London, CT, USA). The pellets were freeze dried and stored at −20°C until use. The pellets were subcutaneously transplanted in the back of wild-type mice, and were collected 2 weeks after surgery.
Data are presented as mean ± SD. Student's t-test for two samples assuming equal variances was used to calculate the p values. Values of p smaller than 0.05 were considered statistically signficiant. Each experiment was repeated at least three times with similar results.
Stat1−/− Mice Exhibit Accelerated Bone Remodeling in a Murine Fracture Model
To investigate potential involvements of STAT1 in skeletal fracture healing, we utilized a murine bone fracture model as described in Materials and Methods. X-rays and micro CT scanning were performed at 1, 2, 3, 4, 6, and 10 weeks after surgery (Fig. 1A). In wild-type mice, calcified area appeared at postoperative week (POW) 1, and formed a bony bridge by POW 3. The volume of the fracture callus peaked at around POW 4 and subsequently decreased in size (Fig. 1B), whereas in Stat1−/− mice, the volume of the fracture callus peaked at around POW 3 and decreased at a higher rate than in wild-type mice, indicating that the fracture healing process is accelerated in STAT1-deficient environment. The cross-sectional callus area of the fracture site was significantly smaller in Stat1−/− mice compared with wild-type animals (0.76 ± 0.19 mm2 in Stat1−/− mice and 1.72 ± 0.60 mm2 in wild-type mice at POW 6, respectively; Fig. 1B), indicating that the callus resorption and bone remodeling is accelerated in the lack of STAT1 activity.
Rapid Turnover of Calcified Callus in Stat1−/− Mice
To examine how the lack of STAT1 affects the formation and resorption of callus and subsequent remodeling process in bone fracture healing, the tibiae from wild-type and Stat1−/− mice were harvested and histologically analyzed (Fig. 1C). In both genotypes, fractured bone was bridged by cartilage callus by POW 2 with no apparent differences in the volume of the cartilaginous tissue. The cartilage callus became fully calcified by POW 4 in both wild-type and Stat1−/− mice; however, the following remodeling process (i.e., resorption of calcified callus and subsequent bone formation) was highly accelerated in Stat1−/− mice compared with wild-type mice. These findings suggest that the lack of STAT1 activity during fracture healing results in accelerated callus remodeling, but has little effect in the inflammation phase or in the subsequent cartilage callus formation.
Membranous Ossification Is Accelerated in Stat1−/− Mice during Fracture Healing
The findings in the tibial fracture model experiments indicated that the lack of STAT1 leads to activation of both calcified callus resorption by osteoclasts and bone formation by osteoblasts. Although upregulation in osteoclast activity was clearly illustrated by faster disappearance of calcified callus, it was still unclear whether bone formation was also upregulated in the remodeling phase. To further evaluate the activity of osteoblasts during fracture healing; we next utilized a cortical defect model. In normal fracture healing, both enchondral ossification (characterized by transient cartilage formation followed by the bone formation on the cartilage template) and membranous ossification (characterized by bone formation by osteoblasts without cartilage formation) are usually observed. However, in this experimental model, skeletal healing is achieved without cartilage callus formation or resorption of calcified callus by osteoclasts so that membranous ossification by osteoblasts can solely be assessed.9 As shown in Figure 2A and B, we found that bone formation in the cortical defect is highly accelerated in Stat1−/− mice compared with wild-type control when evaluated by X-rays. Histological analysis confirmed this observation (Fig. 2C). No apparent cartilage tissue was found when the sections were evaluated by alcian blue staining (data not shown). These results show that bone formation by osteoblasts (membranous ossification) is also accelerated in a STAT1-deficient environment during fracture healing.
STAT1 Negatively Regulates Osx Expression in Osteoblasts
We next sought to elucidate whether STAT1 affects the expression levels of the transcripts for Runx2 and Osx, essential transcription factors for osteoblast differentiation. POBs from wild-type and Stat1−/− newborn mice were incubated in the presence of BMP-2 and RNA was prepared at the designated time points. As shown in Figure 3A, there was no significant difference in the transcripts level of Runx2 between wild-type and Stat1−/− POBs; on the other hand, the expression levels of Osx and osteocalcin were significantly upregulated in Stat1−/− POBs, indicating that STAT1 also exerts its inhibitory effects on osteoblast differentiation via suppressing Osx transcription. To further confirm this observation, we next investigated the effects of STAT1 on the transcriptional activity of Osx by luciferase reporter assay using a reporter vector bearing Osterix promoter sequence.12 Based on the previous studies showing that STAT1 binds to p65 (one of the components of NF-κB complex) and inhibits its transcriptional activity (Ref.13–15 and data not shown), and that p65 stimulates the Osx promoter,12 we hypothesized that downregulation of Osx transcription by STAT1 is mediated by p65. As shown in Figure 3B, although STAT1 did not show any inhibitory effect on the basal Osx transcription level in either MC3T3-E1 or COS-7 cells, p65-induced Osx transcriptional activity was effectively inhibited by STAT1 in a dose-dependent manner. To confirm that STAT1 directly inhibits p65 transcriptional activity, we also did a luciferase reporter assay using a reporter vector bearing an NF-κB response element (NF-κB-RE), and confirmed that STAT1 effectively suppresses the transcriptional activity induced by p65 (Fig. 3C). In accordance, we also found that introduction of p65 in wild-type POBs leads to an increase in the transcripts levels of Osx and osteocalcin, and that this upregulation is blocked by STAT1 (Fig. 3D). Taken together, these observations suggest that, in addition to the inhibition of the nuclear translocation of Runx2 as previously reported,7 STAT1 also negatively regulates osteoblast differentiation by suppressing Osx transcription through inhibition of p65 activity.
Based on the results of the fracture model experiments and luciferase assays that STAT1 negatively regulates osteoblast differentiation, we next asked if the ossification process in vivo could be stimulated by fludarabine, a potent inhibitor of STAT1. Fludarabine is a purine analog and used as a chemotherapeutic agent for the treatment of hematopoietic malignancies; however, it has also been shown that fludarabine reduces STAT1 phosphorylation and suppresses STAT1 protein and transcript levels.16 We subcutaneously transplanted gelatin pellets containing BMP-2 with or without fludarabine and collected the pellets 2 weeks after the implantation. As expected, the pellets with BMP-2 showed a marked calcification compared to the control pellets. The pellets with fludarabine alone also showed moderated upregulation in calcification; however, when fludarabine was used in combination with BMP-2, ossification of the pellets was further accelerated (Fig. 4A and B). These observations indicate that ossification process can be enhanced by suppression of STAT1 activity in vivo.
The findings in the tibial fracture model and the cortical bone defect model showed that bone formation by osteoblasts is accelerated in a STAT1-deficient environment during fracture healing. In vitro data showed STAT1 negatively regulates osteoblast differentiation by suppressing Osx transcription through inhibition of p65 activity. These results reveal a previously unknown contribution of STAT1 in the remodeling phase of the fracture healing and show that inhibition of STAT1 with fludarabine can effectively stimulate ossification in vivo. Our observations and the recent studies by others suggest that the mechanisms behind accelerated fracture healing in Stat1−/− mice is due to upregulation in bone formation (by increased Osx expression, in addition to the enhanced nuclear localization of Runx2 in osteoblasts7) and increased bone resorption by osteoclasts.5, 6 On the other hand, because STAT1 is also implicated in the regulation of angiogenesis and chondrogenesis;17, 18 potential contributions of other factors cannot be ruled out. Nevertheless, the results of the current study further support the notion that STAT1 as a crucial negative regulator for both osteoblast differentiation and osteoclastogenesis, and also suggest that inhibition of STAT1 activity may be beneficial for the treatment of skeletal fracture.
The authors thank Shizue Tomita and Yuko Hashimoto for their excellent technical support. This study was supported by the grants of General Insurance Association of Japan.