Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Fracture healing is a special form of wound healing characterized by regeneration of the normal osseous anatomy. Bone represents the only organ in the body capable of complete repair without the presence of an intervening fibrous scar (Hulth, 1989). However, some events, including aging, metabolic alterations, and poor blood supply, may lead to the lack of inductive callus and cause delayed or nonunion results (Buckwalter et al., 1996). Especially in osteoporotic patients the mechanical strength of the fracture site is decreased due to insufficient amount of callus and calcification, therefore they are liable to delayed recovery from functional impairment and also to the risk of repeated fractures (Melton, 1995; Walsh et al., 1997). Thus, whereas most fractures heal without problem, there are several clinical cases that require enhancement of the healing to ensure the rapid restoration of skeletal function.

Fracture healing involves a complex cascade of several cellular events. These events include immediate injury response, intramembranous ossification, chondrogenesis, and endochondral ossification, resulting in the formation of a fracture callus (Simmons, 1985; Bolander, 1992). Recent studies suggest that local growth factors such as bone morphogenetic proteins (BMPs), transforming growth factor- (TGF-), insulin-like growth factors (IGFs), fibroblast growth factors (FGFs), and platelet-derived growth factor regulate cellular proliferation, differentiation, and extracellular matrix synthesis in the initiation and the development of the fracture callus (Bolander, 1992). The local application of these purified or recombinant molecules with chondrogenic and/or osteogenic capacities has been shown to induce bone regeneration and stimulate fracture healing in animal models (Lind, 1996). These findings indicate that the clinical uses of these growth factors may make possible new therapies for enhancing fracture healing. However, the safety, utility, and cost effectiveness of these growth factors must be considered. Therefore, there has been substantial interest in developing a chemical compound that safely promotes bone formation and facilitates fracture repair.

Previous studies (Notoya et al., 1994) have demonstrated that ipriflavone (7-isopropoxy-isoflavone; mw 280.32), a derivative of natural isoflavone isolated from alfalfa (Medicago setiva L.), stimulates the formation of bone-like nodules in rat bone marrow stromal cells cultured according to the method of Maniatopoulos et al. (1988). Whereas the marrow stromal population in vitro contains a diversity of cell types including osteoblastic cells, fibroblastic cells, adipocytes and monocytes/macrophages, the rich cellular environment is similar to that in the marrow cavity (Aubin et al., 1992). In addition, the nodules formed in vitro exhibits features closely resembling woven bone in vivo (Maniatopoulos et al., 1988). Hence, this model has been extensively used for studies of bone cell lineage and of the regulation of osteogenesis by various agonists (Aubin et al., 1993). These findings persuaded us to focus on the discovery of potent bone formation stimulators using this culture system. By structure-activity relationship studies, we have recently found that a novel series of 3-benzothiepin-2-carboxyamides having a 4-(dialkoxy-phosphorylmethylphenyl) moiety, derived from ipriflavone as the lead compound (Oda et al., 1999). These compounds induce potent cellular alkaline phosphatase (ALP) activity in rat bone marrow stromal cell culture. In this study, we have further evaluated the pharmacological profile of one of the most potent compounds, TAK-778 [(2R, 4S)-()-N-(4-diethoxyphosphorylmethylphenyl)-1,2,4,5-tetrahydro-4-methyl-7,8-methylenedioxy-5-oxo-3-benzothiepin-2-carboxyamide], on osteogenesis in vitro. Furthermore, to elucidate the in vivo osteogenic potential of TAK-778, we examined the effect of TAK-778 sustained-release microcapsules applied locally on skeletal regeneration and bone repair in vivo.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. TAK-778 (Fig. 1) was synthesized at Takeda Chemical Industries (Osaka, Japan) by the method of Oda et al. (1999). Ipriflavone was also supplied by Takeda Chemical Industries. All other chemicals and regents used were of the highest grade available and were purchased from regular commercial sources.

View larger version (8K): [in this window] [in a new window]   Fig. 1.   Chemical structure of TAK-778.

Cell Cultures. Rat bone marrow stromal cells were prepared and cultured according to the method of Maniatopoulos et al. (1988). Cells were obtained from femoral bone marrow of 7-week-old male Sprague-Dawley rats (Japan Charles River, Tokyo, Japan). The standard culture medium consisting of -minimum essential medium (MEM) containing 10 mM HEPES (pH, 7.0) was used. It contained 15% fetal bovine serum (FBS), 2 mM glutamine, 50 µg/ml of ascorbic acid, 10 mM -glycerophosphate, 10⁷ M dexamethasone, and antibiotics (80 µg/ml gentamicin and 100 µg/ml kanamycin). The concentration of dexamethasone used induced maximal mineralized nodule formation in our study (data not shown). The cells were plated in 100-mm culture dishes (Falcon 3003; Becton Dickinson, Lincoln Park, NJ) containing 10 ml of the culture medium. Culture medium was changed every other day, and nonadherent blood cells were removed by washing the cells with culture medium at each change. After 1 week, confluent cells in primary culture were harvested after treatment with 0.25% trypsin in calcium- and magnesium-free PBS containing 0.2% EDTA (trypsin solution) and subcultured in 35-mm culture dishes (Falcon 3046; Becton Dickinson) at a cell density of 4 × 10⁴ cells/well (day 0). All cell cultures used in this study were maintained at 37°C in a humidified atmosphere of 5% CO₂/95% air.

Biological Assays In Vitro. Bone-like nodule formation was evaluated by measuring the area stained by alizarin red using an image analyzer (PIAS LA-555, PIAS Co., Tokyo, Japan). The cell number was determined using a Coulter counter after cells were enzymatically removed with 0.25% trypsin, 1% collagenase, and 0.2% EDTA solution. The DNA content was measured by the method of Labarca and Paigen (1980). The cellular ALP activity as a marker enzyme of osteoblast phenotype was assayed by the method of Lowry et al. (1954) using the supernatant of the cell lysate as previously described (Notoya et al., 1994). To evaluate bone matrix synthesis, osteocalcin and soluble collagen released into culture media over a 48-h culture period (days 5-7) were measured by radioimmunoassays (RIAs) according to the method of Gundenberg et al. (1984) as previously described (Notoya et al., 1994), and a commercial available kit using a dye-binding method (SIRCOL; Biocolor Ltd., Belfast, N. Ireland), respectively. TGF- and IGF-I in serum-free conditioned media (CM) were measured by an enzyme-linked immunosorbent assay kit (Amersham, Tokyo, Japan) and a RIA kit (Amersham Pharmacia Biotech). For collection of the CM, the cells (days 7, 14, and 21) were washed with PBS and maintained in serum-free medium for an additional 48 h. The medium conditioned by the cells was collected and centrifuged for 30 min at 3000g to remove cellular debris. Before the assay, the CM samples were concentrated 5-fold by speed vac centrifugation.

Preparation of a Sustained-Release Microcapsule. To evaluate the in vivo osteogenic potential of TAK-778, we used sustained release microcapsules consisted of a biodegradable polymer, poly (dl-lactic/glycolic) acid (PLGA); TAK-778/PLGA-MC). An injectable sustained release microcapsule was prepared according to the method of Ogawa (1997) with some modifications. PLGA with a copolymer ratio of 85:15 (mol/mol) and an average molecular weight of 15,000 (Wako Pure Chemical Industries, Ltd., Osaka, Japan) was used. A mixture of TAK-778 and the PLGA (1:9, w/w) were dissolved in dichloromethane. This solution was poured into an aqueous 0.1% polyvinyl alcohol solution with stirring with a homogenizer to make an oil-in-water emulsion. To evaporate the dichloromethane, the emulsion was further stirred for 3 h. After removing large particles by sieving, the resulting microcapsules (30 µm in diameter) were collected by centrifugation and lyophilized into a powder. When the microcapsules suspended in 0.5 ml of a dispersing vehicle containing 2.5% D-sorbitol, 0.9% NaCl, 0.1% polysorbate 80, 0.5% caboxymethylcellulose sodium, and 0.07% Na₂HPO₄ were s.c. injected into rats, the microcapsules [TAK-778/PLGA-microcapsules (PLGA-MC)] released TAK-778 for 4 weeks after the injection (data not shown).

Animal Experiments. All animal experiments in this study were carried out in accordance with ethical guidelines established by the Experimental Animal Care and Use Committee of Takeda Chemical Industries.

Skull Defect Model in Rats. Six-week-old male Sprague-Dawley rats (Japan Charles River) were used. After achieving suitable anesthesia by i.p. injection (0.1 ml/100 g b.wt.) of pentobarbital sodium (50 mg/ml; Abbott Laboratories, North Chicago, IL), the skin overlying the calvaria was shaved and sterilized. A semilunar incision was made through the tissues of the scalp, and skin was raised to expose the right parietal bone. The pericrania were stripped off the parietal skulls. Under continuous saline irrigation, a trephine defect, 4 mm in diameter, was made in the center of the parietal bone avoiding sutures using a low-speed dental drill. Particular care was taken to preserve the structure and continuity of the underlying midsaggital blood sinus. The animals were fed with a standard rodent chow containing 1.2% calcium and 1.1% phosphorus (CE-2; Japan Crea, Tokyo, Japan). Two days later, a suspension of placebo or TAK-778/PLGA-MC was injected s.c. adjacent to the defect. In some experiments for histological study, a single injection was given 6 days after the operation. Four weeks after the craniotomy, the animals were sacrificed by carbon dioxide inhalation and the skulls were excised. The new bone formation was evaluated radiographically and histologically. The newly formed radio-opaque area in the defect was calculated using an image analyzer: [area of the initial defect created by trephination (12.6 mm²)]  [area of the remaining defect]. The new bone area in the skull defect of none of the treated rats increased with time (4 weeks after the craniotomy: 2.4 ± 0.6 mm², 8 weeks: 4.5 ± 1.4 mm², 12 weeks: 8.2 ± 1.4 mm², mean ± S.E. for six rats).

Tibial Segmental Defect Model in Rabbits. Adult (age, older than 6 months; weight, 3.8-4.2 kg) male Japanese white rabbits (Kitayama Labes, Nagano, Japan) were used. Animals were anesthetized with an i.v. injection (1 ml/kg b.wt.) of pentobarbital sodium (50 mg/ml, Abbot Laboratories) and local infiltration of lidocaine hydrochloride (1%; Fujisawa Pharmaceutical Co., Osaka, Japan). After the skin overlying the right hind limb was shaved and sterilized, a longitudinal incision was made along the right femoral axis and the tibia exposed. To excise an osteoperiosteal segment from the tibia, two transverse osteotomies were performed. First an osteotomy was made just below the tibiofibular junction using a low- speed dental drill attached to a round diamond saw under saline irrigation. A second osteotomy was made about 5 mm distal of the first osteotomy. After reposition, an external fixator (Orthofix M-100; Orthofix, Venoma, Italy) was affixed with four screws to the medial aspect of the tibia and the axial width of the defect was adjusted to 5 mm. A pellet containing placebo or TAK-778/PLGA-MC was packed into place to fill the defect. The animals were allowed full weight-bearing activity and fed with a standard rabbit chow containing 1.4% calcium and 0.6% phosphorus (RC-4; Oriental Yeast, Tokyo, Japan). Two months after the operation, the animals were sacrificed by injection of excess pentobarbital sodium and the tibiae excised. The defects were examined radiographically and histologically. In none of the treated rabbits, the 5-mm osseous defects failed to bridge the defect within 2 months (data not shown).

A Tibial Fracture Model in Rabbits. Operative procedures, including animals used and fixation, were carried out in accordance with the tibial segmental defect model. Only one transverse osteotomy was made just below the tibiofibular junction. Two days later, a suspension of placebo or TAK-778/PLGA-MC was injected s.c. into the osteotomy site of the right tibia. Thirty days after the osteotomy, the animals were sacrificed and the tibiae removed. Fracture callus formation was evaluated radiographycally and biomechanically. For biomechanical analysis, a three-point bending test was performed using a bone strength testing machine (MZ-500S; Maruto Co., Tokyo, Japan) coupled to a computer for data analysis. The crosshead speed and the distance of two supporting points in the metaphysis were established at 20 mm/min and 55 mm, respectively. The load was added on the anterior midshaft of each tibia until breakage. Breaking force was interpreted from the load-deflection curve.

Histological Evaluation. The rat skulls were fixed in 70% ethanol, immersed in 5% Villanueva mineralized bone stain in 70% methanol for 3 days, dehydrated, and embedded, without decalcification, in methyl methacrylate. A coronal section at the center of the defect was cut with a precision bone saw (Maruto Co.), ground to a thickness of 100 µm using a precision lapping machine (Maruto Co.), and microradiographed on Kodak Pelicula film for 6 min at 35 kVp, 20 mA. The 100-µm sections were mounted on plastic microscope slides, grounded to 20 µm, and covered with coverslips for light microscopic analysis.

Statistical Analysis. All statistical analyses were accomplished using the Statistical Analysis System (SAS) software (SAS Institute, Cary, NC). Data are expressed as the mean ± S.E. Statistical comparisons were performed with a one-way ANOVA and Dunnett's multiple comparison method or Student's t test. A p value of less than 0.05 was considered statistically significant.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

TAK-778 Stimulates Osteogenesis In Vitro

Mineralized Nodule Formation. Rat bone marrow stromal cells selectively differentiated into osteoblasts and subsequently formed mineralized bone-like nodules in the presence of -glycerophosphate and dexamethasone, as previous described (Notoya et al., 1994). Under the culture conditions, continuous treatment with TAK-778 (10⁵ M) for 1 to 21 days resulted in an increase in the area of mineralized nodules (Figs. 2 and 3). The effect of TAK-778 was dose dependent and more potent than that of ipriflavone (Figs. 2 and 3). TAK-778 did not accelerate the onset of mineralization (data not shown). With regard to the exposure timing for TAK-778 on the nodule formation after culturing for 21 days, a consecutive exposure during all culture period caused the most potent stimulatory effect, whereas an initial 2 days exposure (days 1-3) had a significant effect (Table 1). In addition, early exposure (days 1-7) to TAK-778 resulted in twice the nodule area induced by late exposure (days 14-21) to this compound. This indicates that TAK-778 might preferentially act on osteoblast lineage cells at early differentiation stages.

View larger version (61K): [in this window] [in a new window]   Fig. 2.   Culture dishes containing alizarin red-stained nodules after culturing for 21 days for the control (A), ipriflavone (105 M; B), and TAK-778 (105 M; C) groups.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Dose-dependent effect of TAK-778 on mineralized nodule formation in rat bone marrow stromal cell culture. Cells were cultured in medium with or without various concentrations of TAK-778 for 21 days. Mean ± S.E. (n = 5). *p < .05; **p < .01 versus control. Dunnett's multiple comparison.

View this table: [in this window] [in a new window]   TABLE 1 Effect of exposure timing for TAK-778 on mineralized nodule formation in rat bone marrow stromal cell culture Cells were exposed to TAK-778 (105 M) for various durations, and then were cultured in medium without TAK-778 during remainder of culture period. After culturing for 21 days, alizarin red-stained nodule area was measured. Mean ± S.E. (n = 5).

ALP and Other Osteoblast Products. To define further the stimulatory effect of TAK-778 on mineralized nodule formation, several markers related to osteoblast function were assessed. TAK-778 at concentrations of 10⁶ M and higher significantly stimulated the activity of cellular ALP, one of the hallmarks of osteoblast differentiation, on day 7 when the cells reached confluence accompanied by the initiation of nodule formation (Table 2). TAK-778 increased slightly but significantly the DNA content of the cells at the confluence stage (Table 2). Treatment with TAK-778 also resulted in dose-dependent increases in the amount of soluble collagen and osteocalcin secreted into culture medium from days 5 to 7, indicating a stimulatory effect on bone matrix deposition (Table 3). Because several growth factors have been shown to exert regulatory effects on growth and differentiation of osteoblasts via autocrine/paracrine mechanisms (Rodan, 1991), we quantified TGF- and IGF-I in serum-free medium conditioned by the cells on days 7, 14, and 21. As shown in Table 4, TAK-778 enhanced the secretion of both TGF- and IGF-I at every time point during the 21 days of culture. These results indicate that TAK-778 stimulates expression of the osteoblast phenotype.

View this table: [in this window] [in a new window]   TABLE 4 Effect of TAK-778 on amounts of TGF- and IGF-I in serum-free medium conditioned by rat bone marrow stromal cells Cells were cultured in medium with or without various concentrations of TAK-778 in the presence of 15% FBS for 7, 14, or 21 days. Then cells were washed with PBS and cultured in serum-free medium for an additional 48 h. CM in each well was harvested, and TGF- and IGF-I were measured by enzyme-linked immunosorbent assay and RIA, respectively. Mean ± S.E. (n = 3-5).

Effect on C3H10T1/2 Cells and Rat Calvarial Cells. We also examined the effect of TAK-778 on cells at different stages during the differentiation process using an uncommitted mesenchymal pluripotential line C3H10T1/2 and committed osteoblastic cells isolated from newborn rat calvaria. C3H10T1/2 cells did not exhibit any features characteristic of the osteoblast phenotype. Treatment of the cells with TAK-778 did not induce ALP activity, but did result in a dose-dependent increase in the saturated cell density (Table 5). On the other hand, TAK-778 stimulated ALP activity of rat calvarial cells that spontaneously exhibited osteoblast phenotype without dexamethasone (Table 5). TAK-778 at a concentration of 10⁵ M significantly reduced the saturated cell density (Table 5). Continuous treatment of the rat calvarial cells with this compound for 1 to 21 days resulted in an increase in mineralized nodule formation (data not shown). These results suggest that TAK-778 stimulates the phenotype expression of committed osteoblast lineage cells, as opposed to inducing the differentiation of uncommitted mesenchymal cells into osteoblasts.

TAK-778/PLGA-MC Enhances New Bone Formation During Bone Repair In Vivo

Skull Defect Model in Rats. To elucidate the in vivo osteogenic potential of TAK-778, we used a rat skull defect model treated with sustained release microcapsules (TAK-778/PLGA-MC) injected s.c. adjacent to the defect. The 4-mm trephine defects made after stripping off the pericrania did not completely heal within 12 weeks and the spontaneous new bone area that formed in the defect by 4 weeks was less than 20 to 30% of the defect area (see Experimental Procedures). Four weeks after the operation, treatment with a single local application of TAK-778/PLGA-MC (0.2-5 mg/site) resulted in a dose-dependent increase in the radio-opaque area formed in the defect (Fig. 4, A and B and 5). Histological studies showed the defect area was occupied by a bony bridge (Fig. 4D) and the newly-formed radio-opaque area corresponded to a calcified bone containing bone marrow cavities surrounded by thick osteoid seams with cuboidal osteoblasts (Fig. 4F). In contrast, the placebo-treated defects were filled with a dense fibrous connective tissue and an only small amount of bony callus at the periphery (Fig. 4, C and E). No chondrocytes or cartilage tissues were seen in either the placebo- or TAK-778/PLGA-MC-treated defects during this period (data not shown).

View larger version (81K): [in this window] [in a new window]   Fig. 4.   Rat skull defects 4 weeks after the operation. Soft X-ray radiographs (A and B), contact microradiographs (C and D; original magnification 10×), and microphotographs (E and F, undecalcified ground section, Villanueva bone stain; original magnification 100×) of the defects treated with placebo (A, C, and E) and TAK-778/PLGA-MC (1 mg/site; B, D, and F). Approximately two-thirds of the defects were occupied by a radio-opaque area (B) and a bony bridge across the defect treated with TAK-778/PLGA-MC (D). Histologic examination revealed the bony bridge consisted of a calcified bone containing bone marrow cavities surrounded by thick osteoid seams with cuboidal osteoblasts (F). In contrast, the placebo-treated defect was filled with a dense fibrous connective tissue and an only small amount of bony callus at the periphery (A, C, and E).

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Dose-dependent effect of TAK-778/PLGA-MC on new bone formation in the rat skull defects 4 weeks after the operation. Mean ± S.E. (n = 6). *p < .01 versus control. Dunnett's multiple comparison.

Tibial Segmental Defect Model in Rabbits. Anatomically, there are two types of bones in the skeleton, flat bones and long bones. These are distinguished by their histogenesis, that is, intramembranous ossification and endochondral ossification, respectively, although the development and growth of long bones actually involves both types of ossification. For this reason, we also evaluated the effect of TAK-778/PLGA-MC on skeletal regeneration of long bones using a tibial segmental defect model in rabbits. A pellet containing TAK-778/PLGA-MC (4 mg/pellet) was immediately packed into place to fill the 5-mm osteoperiosteal defect created by the operation. Two months after the operation, the TAK-778/PLGA-MC pellets induced radiological osseous union across the defects (Fig. 6, B and D). The number of rabbits with osseous union was four of four rabbits treated with TAK-778/PLGA-MC. The beginning of the union was observed within 1 month (data not shown). The histological examination revealed that radiological osseous union consistent with abundant cancellous woven bone filled the segmental defect (Fig. 6F). Newly formed lamellar bone containing some osteons was also observed along the callus-cortex junction (Fig. 7A). This indicates that the bony callus converted to mature bone through remodeling and is tightly connected to the existing cortical bone. Furthermore, spotted cartilaginous matrix was noticed in the central region of the osseous union (Fig. 7B), indicating traces of endochondral ossification. In the defects that had been filled with placebo pellet, small callus formed at the free bone ends, but they failed to bridge the defect (Fig. 6, A, C, and E). The number of placebo-treated rabbits with osseous union was zero of four. When the TAK-778/PLGA-MC pellets were implanted s.c. or i.m. in rabbits, no ectopic bone formation was observed (data not shown). These results suggest that TAK-778/PLGA-MC enhances skeletal repair in long bone as well as flat bone defects.

View larger version (68K): [in this window] [in a new window]   Fig. 6.   The rabbit tibial segmental defects 2 months after the operation. Soft X-ray radiographs (A, B, C, and D) and microphotographs (E and F, H&E; original magnification 8×) of the defect treated with placebo (A, C, and E) and TAK-778/PLGA-MC (4 mg/site; B, D, and F). TAK-778/PLGA-MC pellets induced radiological osseous union across the defects (B and D), whereas placebo pellets did not (A and C). Histologic observation showed that the TAK-778/PLGA-MC-treated bone segments were bridged by a bony callus (F).

View larger version (116K): [in this window] [in a new window]   Fig. 7.   Microphotographs of the rabbit tibial segmental defects treated with TAK-778/PLGA-MC 2 months after the operation. A, at the callus-cortex junction (arrowheads indicate fracture line), the bony callus (upper half) and the existed cortical bone (lower half) were partly remodeled and replaces by some osteons (Os). H&E, magnification 100×. B, scattered foci of metachromatic cartilaginous matrix (arrows) were noted at the center region of the bony callus. Toluisin blue. Original magnification 200×.

Tibial Fracture Model in Rabbits. To assess whether TAK-778/PLGA-MC-induced callus contributed to the advantage of early biomechanical stability after the fracture, we performed an osteotomy on adult rabbit tibia and injected placebo or TAK-778/PLGA-MC (10 mg/site) adjacent to the fracture. Thirty days after the operation, the radiological fracture line of TAK-778/PLGA-MC group partially disappeared, due to an increase in the callus formation at the osteotomy site, whereas the fracture line of the placebo group still clearly appeared (Fig. 8). The three-point bending test showed the breaking force of TAK-778/PLGA-MC-treated bone was 1.4 times higher than that of the placebo-treated bone (placebo, 78.2 ± 6.3 N; TAK-778/PLGA-MC, 112.1 ± 6.5 N, mean ± S.E. for 7 rabbits, p < 0.05, Student's t test).

View larger version (97K): [in this window] [in a new window]   Fig. 8.   Soft X-ray radiographs of the rabbit tibial osteotomy sites 30 days after the operation in the placebo (A) and TAK-778/PLGA-MC (10 mg/site; B) groups. The radiological fracture line of TAK-778/PLGA-MC group partially disappeared, whereas that of the placebo group still clearly appeared.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

This report provides the first evidence that a nonendogenous chemical compound stimulates potently bone-like nodule formation in vitro and enhances new bone formation during skeletal regeneration and bone repair in vivo. The new bone formed in response to TAK-778 was woven and was normally mineralized. The histological features were similar to those of callus formed in natural fracture healing (Frame, 1980; Ashhurst, 1986; Buckwalter et al., 1996). In addition, TAK-778-treated callus also consisted of newly formed lamellar bone that contained osteon in the rabbit tibial defect model, suggesting that the callus changes to mature bone through remodeling. Thus, this compound may increase the quantity of callus without qualitative alterations.

TAK-778 stimulated the expression of the osteoblast phenotype, that is, ALP activity, soluble collagen release, and osteocalcin secretion, with a slight increase in cellular proliferation in the presence of dexamethasone in rat bone marrow stromal cell culture. The process of bone-like nodule formation in vitro resembles that of intramedullary bone formation in repair of bone marrow tissue after local injury: pluripotential osteoprogenitor cells in the marrow stroma proliferate and differentiate to functional osteogenic cells, resulting in the formation of woven bone in the bone marrow (Amsel et al., 1969). Hence, the stimulatory effect of TAK-778 on cellular growth and differentiation during bone-like nodule formation in vitro may reflect the effect of this compound on cellular events during bone marrow regeneration, especially intramembranous direct bone formation. Furthermore, a recent study using the chondrogenic cell line ATDC5 showed that TAK-778 stimulates the expression of type II collagen mRNA and the accumulation of Alcian blue-stainable cartilage-specific proteoglycan in ATDC5 cells, suggesting that TAK-778 also stimulates chondrogenesis as well as osteogenesis (Akiyama et al., 1999). The molecular mechanism of action is presently under investigation.

TAK-778-induced growth stimulation appears to be differentiation stage specific and independent of its differentiation promotion in the mixed populations of rat bone marrow stromal cell culture. The idea is supported by results showing that TAK-778 stimulates proliferation of uncommitted mesenchymal C3H10T1/2 cells without induction of differentiation, whereas it inhibits proliferation of osteoblast-enriched rat calvarial cells accompanied by stimulation of cellular ALP activity. Treatment of C3H10T1/2 cells with TAK-778 in the presence of serum doses lower than 1% resulted in no or negligible growth stimulation (our unpublished observation). Therefore, growth stimulation may require a certain amount of serum, suggesting that TAK-778 modulates the actions of serum growth factors on cells at early differentiation stages along osteogenic lineage cells. Cell division indicated by [³H]thymidine incorporation is usually observed only in cells in the morphologically unspecialized "osteoprogenitor" state, and mature osteoblasts or osteocytes are considered nonmitotic cells (Robey et al., 1992). Consequently, the difference between bone marrow stromal cells and calvarial cells concerning the effect of TAK-778 on cellular division seems to be related to the difference in the maturity of their cell populations (Turksen and Aubin, 1991).

Hughes and McCulloch (1991) suggested that differentiation of osteoprogenitor cells might be regulated by the production of autocrine/paracrine factor(s) based on the following data: conditioned media from calvaria-derived cells enhances mineralized nodule formation in rat bone marrow stromal cell culture. In this study, TAK-778 promoted the secretion of TGF- and IGF-I during bone-like nodule formation. Relating to these findings, Demetriou et al. (1996) reported that neutralizing antibody against TGF- suppressed mineralized nodule formation in rat bone marrow stromal cell culture under basal conditions. Furthermore, neutralizing anti-IGF-I antibody blocked stimulation of osteoblast phenotype expression induced by intermittent exposure of parathyroid hormone, which stimulates IGF-I secretion and bone-like nodule formation in rat calvarial cell culture (Ishizuya et al., 1997). These lines of evidence suggest that the secretion of these endogenous growth factors plays a crucial role in the formation of bone-like nodules in vitro. In addition, the increase in the expression of these growth factors in vivo may contribute to the development of the callus formation, because exogenous administration of TGF- or IGF-I to bone in vivo increases new bone formation (Noda and Camillere, 1989; Aspenberg et al., 1989). Further work is needed to elucidate the expression of these growth factors in the bony callus treated with TAK-778.

There are several similarities and differences between the pharmacological profiles of TAK-778 and those of other endogenous osteogenic agents, such as BMPs and basic fibroblast growth factor (bFGF). These agents are also synthesized and localized in the callus during fracture healing (Bolander 1992; Onishi et al., 1998). Both BMPs and bFGF stimulate bone-like nodule formation in vitro (Hanada et al., 1997) and the local application of these agents enhances callus formation during bone repair in animal models (Kawaguchi et al., 1994; Riley et al., 1996). In rat bone marrow stromal cell culture, when TAK-778 was added in combination with BMP-4/7 or bFGF mutein CS23 at submaximal concentrations, an additive or synergistic stimulatory effect on cellular ALP activity and bone-like nodule formation was observed (M.G., K.N., and H.M., unpublished observations). It is expected that TAK-778 enhances osteogenesis induced by endogenous or exogenous application of these osteogenic agents. On the other hand, BMPs, unlike TAK-778, have been shown to induce the development and expression of chondrogenic and osteogenic lineages in uncommitted mesenchymal C3H10T1/2 cell cultures (Katagiri et al., 1990; Wang et al., 1993). The difference between BMPs and TAK-778 in vitro may reflect the in vivo phenomenon that BMPs induce ectopic bone formation in vivo, whereas TAK-778 dose not. Basic FGF has been reported to suppress the features of differentiated osteoblasts (Rodan et al., 1989; Hurley et al., 1993). This effect, which TAK-778 does not have, may lead to timing limitations for exogenous application of bFGF.

Local application of the sustained-release microcapsules, TAK-778/PLGA-MC, resulted in an increase in new bone formation in bony defect and fracture animal models. The sustained-release formulation was designed to release TAK-778 over 4 to 6 weeks for the following reasons. First, in general, callus formation occurs during 4 to 40 days postfracture, after the immediate inflammatory reaction-hematoma stage (Simmons, 1985), although the timetable for these events during fracture healing depends on the intrinsic tissue capacity for healing and variables that influence healing, such as type of injury (Buckwalter et al., 1996). Second, with regard to the exposure timing for this compound, the in vitro results show that the effect of TAK-778 on bone-like nodule formation is dependent on the duration of exposure of this compound (Table 1). When the PLGA microcapsules containing a contrast agent, iodoform, were injected into sites adjacent to the tibia in rabbits, radiographical finding showed microcapsules clumped around the injection sites without broad dispersion (our unpublished observation). Assuming that the local distribution volume of this sustained-release formulation (5 mg/site) is 5 ml, the estimated local concentration of this compound corresponded to the in vitro concentration of TAK-778, in which this compound stimulated osteogenesis in vitro (T.H., K. Saito, H.M. and Y. Ogawa, unpublished observations).

In the rabbit tibial fracture model with external fixation, TAK-778/PLGA-MC enhanced callus formation at the fracture site, accompanied by an increase in the breaking force. Clinical application of this may lead to early removal of the fixation device followed by earlier weight bearing, leading to the formation of stress-generated potentials in the callus (Bringhton, 1984) and an increase in the release of growth factors (Canalis et al., 1988; Klein-Nulend et al., 1995).

TAK-778/PLGA-MC led to no significant changes in either mineralized bone or osteoid width of the intact bone, whereas this compound promoted new bone formation in the defect site of a rat skull defect model. This difference may be due to differences in cellular populations and local environment between the intact and defect bone. In fact, in an autoradiographic study using [³H]thymidine by Tonna and Cronkite (1961), the labeling index at the peak during the fracture healing is 10 times higher than that of control. In addition, if new bone is induced by TAK-778/PLGA-MC on intact bone, the bone might be rapidly resolved and return to the original state, possibly in accordance with Wolff's hypothesis: the structure of bone adapts to changes in its stress environment (Wolff, 1892). Actually, 1 week after the injection of TAK-778/PLGA-MC or PLGA-MC alone, focal pericranial woven bone formation was observed on the intact bone just beneath the aggregation of the injected microcapsules, possibly due to the irritation effects of PLGA itself (our unpublished observation). On the other hand, long-term oral administration of TAK-778 induced no observable effects on intact bones in normal rats and dogs (our unpublished observation).

In summary, a novel chemical compound, TAK-778, stimulated potent bone-like nodule formation accompanied by an increase in the expression of osteoblast phenotype in vitro. Local application of TAK-778 sustained-release microcapsules enhanced callus formation in animal bony-defect and fracture models. These results suggest possible new strategies for treatment of skeletal injuries and diseases and for oral and maxillofacial applications. Results of further experimental and clinical studies are needed to demonstrate whether TAK-778/PLGA-MC has clinical utility for the enhancement of fracture healing.