Yoshinori Kuboki

To elucidate the biochemical mechanism of osteogenesis, the effect of matrix geometry upon the osteogenesis induced by bone morphogenetic protein (BMP) was studied. A series of five porous hydroxyapatites with different pore sizes, 106-212, 212-300, 300-400, 400-500, and 500-600 microns, was prepared. A block (approximately 5 x 5 x 1 mm, 40.0 mg) of each hydroxyapatite ceramics was combined with 4 micrograms of recombinant human BMP-2 and implanted subcutaneously into the back skin of rat. Osteoinductive ability of each implant was estimated by quantifying osteocalcin content and alkaline phosphatase activity in the implant up to 4 wk after implantation. In the ceramics of 106-212 microns, the highest alkaline phosphatase activity was found 2 wk after implantation, and the highest osteocalcin content 4 wk after implantation, consistent with the results observed with particulate porous hydroxyapatite [Kuboki, Y. et al. (1995) Connect. Tissue Res. 32: 219-226]. Comparison of the alkaline phosphatase activities at 2 wk and the osteocalcin contents at 4 wk after implantation revealed that the highest amount of bone was produced in the ceramics implants with pore size of 300-400 microns. In the ceramics with smaller or larger pore sizes, the amount of bone formation decreased as the pore size deviated from 300-400 microns. The results indicated that the optimal pore size for attachment, differentiation and growth of osteoblasts and vascularization is approximately 300-400 microns. This study using chemically identical but geometrically different cell substrata is the first demonstration that a matrix with a certain geometrical size is most favorable for cell differentiation.

Bone morphogenetic protein (BMP) is known to require a suitable carrier to induce ectopic bone formation in vivo. Hydroxyapatite ceramics have been reported to be effective in some forms but ineffective in others as a carrier of BMP-induced bone formation. In this study we compare three geometrically different forms of hydroxyapatite to examine their functions as carriers of BMP-induced bone formation. A fraction containing all the active BMPs (BMP cocktail) was partially purified from a 4M guanidine extract from bovine bone by a three-step chromatographic procedure. The BMP cocktail was combined with each of three forms of hydroxyapatite--solid particles (SPHAP), porous particles (PPHAP), and coral-replicated porous tablets (coral-HAP)--and implanted subcutaneously into rats. Both the PPHAP and coral-HAP systems induced osteogenesis 2 weeks after implantation, as evidenced by morphological and biochemical observations. Details of the osteogenetic process were followed by double-fluorescence labeling in the coral-HAP system to confirm bone formation on the surface of hydroxyapatite. However, there was no evidence of osteogenesis or chondrogenesis in the SPHAP system. The results indicate that the geometry of the interconnected porous structure in PPHAP and coral-HAP create spaces for vasculature that lead to osteogenesis while the smooth structure and close contact of particles in SPHAP inhibit vascular formation and proliferation of mesenchymal cells, preventing bone and cartilage formation. It was concluded that the geometrical structure in hydroxyapatite ceramics that induces vasculature is crucial as a carrier for BMP-induced bone formation.

BACKGROUND: The effect of the geometry of extracellular matrices on bone morphogenetic protein (BMP)-induced osteogenesis has not been systematically studied. Geometry is crucially important for the scaffold in bone and joint tissue engineering. The purpose of this study was to elucidate principles of geometry of matrices in designing new scaffolds and matrices for use in reconstruction of bone and joints. METHODS: More than ten biomaterials with different geometries, including a unique device of honeycomb-shaped hydroxyapatite, were combined with BMPs of recombinant (rhBMP-2) or natural bovine origin (S300 BMP cocktail) and implanted subcutaneously into 4-week-old Wistar-King rats. The implanted pellets were removed at 1-4 weeks and analyzed for bone and cartilage formation by histological and biochemical methods. RESULTS: BMP-induced bone and cartilage induction was highly dependent on the geometric properties of the carrier. Some carriers such as porous particles or blocks of hydroxyapatite induced osteogenesis directly, without detectable chondrogenesis, whereas other carriers such as fibrous glass membrane induced cartilage exclusively. Still other carriers induced mostly cartilage followed by bone formation. Solid particles of hydroxyapatite and fibrous glass membrane with too tight a meshwork did not induce bone or cartilage. The optimal pore size for bone-forming efficacy in porous blocks of hydroxyapatite was a diameter of 300-400 microm. In straight tunnel structures with various diameters in honeycomb-shaped hydroxyapatite, tunnels with smaller diameters (90-120 microm) induced cartilage followed by bone formation, whereas those with larger diameters (350 microm) induced bone formation directly within the tunnels. CONCLUSIONS: BMP carriers were classified into three types: bone-inducing, cartilage-inducing, and cartilage-bone-inducing. From the analysis of causative factors inducing osteogenesis and chondrogenesis in the BMP system, we concluded that the geometry of the carrier is crucially important and vasculature-inducing geometry should be considered in designing effective scaffolds for bone formation. We propose a classification of geometry of the artificial extracellular matrices that is useful for designing a scaffold for tissue engineering of bone and related tissues. CLINICAL RELEVANCE: Conventional requisites of the BMP carriers for clinical use have mainly concerned the affinities of carriers with cells and biomolecules and their mechanical strength. The vasculature-inducing geometry of carriers adds a new criterion in designing systems for effective bone and joint reconstruction. The geometries of porous structures-their sizes, continuity, and straightness as verified by hydroxyapatite in this study-will be applicable for other biomaterials for clinical reconstruction therapy.

Bone marrow cells are multipotent cells. When bone marrow cells were cultured with type I collagen matrix gels, they showed high alkaline phosphatase activity, collagen synthesis, and formed mineralized tissues. Furthermore, cells expressed osteocalcin and bone sialoprotein genes, which are osteoblast-specific genes. These findings indicate that type I collagen matrix gels induce osteoblastic differentiation of bone marrow cells. Type I collagen interacts with the alpha 2 beta 1 integrin receptor on the cell membrane and mediates extracellular signals into cells. DGEA peptide is a cell-binding domain of type I collagen molecule. When collagen-integrin interaction was interrupted by the addition of Asp-Gly-Glu-Ala (DGEA) peptide to the culture, the expression of osteoblastic phenotypes of bone marrow cells was inhibited. Furthermore, anti-alpha 2 integrin antibody, which interacts with alpha subunit of integrin and blocks the binding of integrin with collagen, suppressed the expression of osteoblastic phenotypes. These findings imply that collagen-alpha 2 beta 1 integrin interaction is an important signal for the osteoblastic differentiation of bone marrow cells.

Bone marrow contains multipotent cells that differentiate into fibroblasts, adipocytes, and osteoblasts. Recently we found that type I collagen matrix induced the osteoblastic differentiation of bone marrow cells. Three weeks after cells were cultured with type I collagen, they formed mineralized tissues. In this study, we investigated the expression of osteoblast-related genes (alkaline phosphatase, osteocalcin, bone sialoprotein, osteopontin, and cbfa-1) during the osteoblastic differentiation. The expression of alkaline phosphatase and osteopontin genes increased time-dependently during the osteoblastic differentiation. Osteocalcin and bone sialoprotein genes were expressed in cells that formed mineralized tissues, and both were expressed only after cells reached the mineralized tissue-formation stage. On the other hand, the cbfa-1 gene was expressed from the early differentiation stage. The Asp-Gly-Glu-Ala (DGEA) amino acid domain of type I collagen interacts with the alpha2beta1 integrin receptor on the cell membrane and mediates extracellular signals into cells. When the collagen-integrin interaction was interrupted by the addition of DGEA peptide to the culture, the expression of osteoblastic phenotypes of bone marrow cells was inhibited. These findings imply that the collagen-alpha2beta1 integrin interaction is an important signal for the osteoblastic differentiation of bone marrow cells.