Search

The molecular mechanisms by which bone morphogenetic proteins (BMPs) promote skeletal cell differentiation were investigated in the murine mesenchymal stem cell line C3H10T1/2. Both BMP-7 and BMP-2 induced C3H10T1/2 cells to undergo a sequential pattern of chondrogenic followed by osteogenic differentiation that was dependent on both the concentration and the continuous presence of BMP in the growth media. Differentiation was determined by the expression of chondrogenesis and osteogenesis associated matrix genes. Subsequent experiments using BMP-7 demonstrated that withdrawal of BMP from the growth media led to a complete loss of skeletal cell differentiation accompanied by adipogenic differentiation of these cells. Continuous treatment with BMP-7 increased the expression of Sox9, Msx 2, and c-fos during the periods of chondrogenic differentiation after which point their expression decreased. In contrast, Dlx 5 expression was induced by BMP-7 treatment and remained elevated throughout the time-course of skeletal cell differentiation. Runx2/Cbfa1 was not detected by ribonuclease protection assay (RPA) and did not appear to be induced by BMP-7. The sequential nature of differentiation of chondrocytic and osteoblastic cells and the necessity for continuous BMP treatment to maintain skeletal cell differentiation suggests that the maintenance of selective differentiation of the two skeletal cell lineages might be dependent on BMP-7-regulated expression of other morphogenetic factors. An examination of the expression of Wnt, transforming growth factor-beta (TGF-beta), and the hedgehog family of morphogens showed that Wnt 5b, Wnt 11, BMP-4, growth and differentiation factor-1 (GDF-1), Sonic hedgehog (Shh), and Indian hedgehog (Ihh) were endogenously expressed by C3H10T1/2 cells. Wnt 11, BMP-4, and GDF-1 expression were inhibited by BMP-7 treatment in a dose-dependent manner while Wnt 5b and Shh were selectively induced by BMP-7 during the period of chondrogenic differentiation. Ihh expression also showed induction by BMP-7 treatment, however, the period of maximal expression was during the later time-points, corresponding to osteogenic differentiation. An interesting phenomenon was that BMP-7 activity could be further enhanced twofold by growing the cells in a more nutrient-rich media. In summary, the murine mesenchymal stem cell line C3H10T1/2 was induced to follow an endochondral sequence of chondrogenic and osteogenic differentiation dependent on both dose and continual presence of BMP-7 and enhanced by a nutrient-rich media. Our preliminary results suggest that the induction of osteogenesis is dependent on the secondary regulation of factors that control osteogenesis through an autocrine mechanism.

Fracture healing is a specialized post-natal repair process that recapitulates aspects of embryological skeletal development. While many of the molecular mechanisms that control cellular differentiation and growth during embryogenesis recur during fracture healing, these processes take place in a post-natal environment that is unique and distinct from those which exist during embryogenesis. This Prospect Article will highlight a number of central biological processes that are believed to be crucial in the embryonic differentiation and growth of skeletal tissues and review the functional role of these processes during fracture healing. Specific aspects of fracture healing that will be considered in relation to embryological development are: (1) the anatomic structure of the fracture callus as it evolves during healing; (2) the origins of stem cells and morphogenetic signals that facilitate the repair process; (3) the role of the biomechanical environment in controlling cellular differentiation during repair; (4) the role of three key groups of soluble factors, pro-inflammatory cytokines, the TGF-beta superfamily, and angiogenic factors, during repair; and (5) the relationship of the genetic components that control bone mass and remodeling to the mechanisms that control skeletal tissue repair in response to fracture.

Previous studies have demonstrated that both mechanical perturbation and cell adhesion induced the expression of osteopontin (opn) by osteoblasts (Carvalho et al. [1998] J. Cell. Biochem. 70:376-390). The present study examined if these same stimuli on osteoblasts would induce the expression of other integrin binding proteins, specifically fibronectin (fn) and bone sialoprotein (bsp). All three genes showed three- to four-fold maximal induction in response to both cell adhesion and a single 2-h period of an applied spatially uniform, dynamic biaxial strain of 1.3% at 0.25 Hz. Each gene, however, responded with a different time course of induction to mechanical strain, with bsp, fn, and opn showing their maximal response at 1, 3, and 9 h, respectively, after the perturbation period. In contrast, peak induction to cell adhesion was observed at 24 h for bsp and opn, while fn levels peaked at 8 h. Interestingly, while both opn and fn mRNA expression returned to base line after cell adhesion, bsp mRNA levels remained elevated. Examination of collagen type I and osteocalcin mRNAs showed unaltered levels of expression in response to either type of perturbation. A common feature of the signal transduction pathways, which mediate the gene expression in response to both cell adhesion and mechanical perturbation, was the activation of specific tyrosine kinases based on the ablation of the induction of these genes by the tyrosine kinase inhibitor genistein. While cycloheximide blocked the induction of all three mRNAs in response cell adhesion, it failed to block the induction of any of these genes in response to mechanical perturbation. Such results suggest that the induction of these genes after mechanical perturbation was mediated by an immediate response to signal transduction, while cell adhesion mediated effects secondary to signal transduction. Depolymerization of microfilaments with cytochalasin D had no effect on the overall expression of any of these genes in response to cell adhesion and only blocked the induction of opn expression in response to mechanical perturbation. These results suggest that cytoskeletal integrity is only selectively important in the signal transduction of certain types of stimuli and for the regulation of certain genes. In summary, both mechanical perturbation and cell adhesion stimulated the expression of integrin binding proteins. Furthermore, while there are common features in the signal transduction processes that mediate the induction of these genes in response to both stimuli, specific genes are separately regulated by precise mechanisms that are unique to both forms of stimuli.

This study identifies a cis-acting element that confers tissue-restricted expression to the bone sialoprotein (BSP) gene. Using both gain of function and loss-of function studies, we demonstrate that this element acts as a tissue specific enhancer of BSP expression in osteoblasts and hypertrophic chondrocytes but does not function in non-hypertrophic chondrocytes or fibroblasts. Furthermore, our data demonstrate that binding of this element occurs in correlation with active BSP expression. While Dlx5 has been implicated as the tissue-specific regulator of BSP expression through direct DNA binding at an element with homology to the one under study here, our results demonstrate that Dlx5 does not act as a positive regulator of BSP expression. Finally, mutational analyses of this element demonstrate that while there is homology to putative homeodomain binding elements, this site is unlikely to bind homeodomain factors including Dlx5. Thus, these studies identify an important cis-acting element in the BSP promoter that acts as a tissue-specific enhancer of BSP expression in both osteoblasts and hypertrophic chondrocytes. As such this is the first demonstration of a common regulatory mechanism utilized by both chondrocytes and osteoblasts for the tissue-restricted expression of the BSP gene.

The expression of neuropilin-1 (NRP1), a recently described VEGF and semaphorin receptor expressed by endothelial cells (EC) but some non-EC types as well, was analyzed in osteoblasts in vitro and in vivo. Cultured MC3T3-E1 osteoblasts expressed NRP1 mRNA and bound VEGF165 but not VEGF121, characteristic of the VEGF isoform-specific binding of NRP1. These cells did not express VEGFR-1 or VEGFR-2 so that VEGF binding to osteoblasts was strictly NRP1-dependent. In a chick osteocyte differentiation system, NRP1 was expressed by osteoblasts but its expression was absent as the cells matured into osteocytes. Immunohistochemical localization of NRP1 within the developing bones of 36-day-old mice and embryonic Day 17 chicks demonstrated that NRP1 was expressed by osteoblasts migrating alongside invading blood vessels within the metaphysis of the growth plate, as well as by osteoblasts at the developing edge of trabeculae within the marrow cavity. On the other hand, NRP1 was not expressed by osteocytes in either species, consistent with the in vitro results. In addition to osteogenic cells, NRP1 expression by EC was observed throughout the bone. Together these results suggest that NRP1 might have a dual function in bone by mediating osteoblast function directly as well as angiogenesis. J. Cell. Biochem. 81:82–92, 2001. © 2001 Wiley-Liss, Inc.

Osteopontin is a predominant integrin binding protein of bone and its expression has been shown to be induced by mechanical stimuli within osteoblasts (Toma et al. [1997] J. Bone Miner. Res. 12:1626-1636). The present studies examined if the cell adhesion would mimic the mechano-transduction that stimulated opn mRNA expression and whether integrin receptors were involved in these processes. Osteopontin mRNA expression was induced three- to four-fold, 24 hours after embryonic chicken calvaria osteoblast attachment to fibronectin (FN), however no induction was observed if the cells were plated on tissue culture plastic alone. Osteopontin mRNA induction in response to cell attachment on FN was dependent on new protein synthesis and the activation of a tyrosine protein kinase(s) but unlike mechano-induction was independent of the maintenance of the cell's microfilament structure. Integrin receptor(s) were shown to be involved in mediating the signal transduction processes of both cell attachment and mechanical stimulation since incubation of osteoblasts with the integrin binding peptide RGDS partially blocked the induction of opn expression in response to both stimuli. Interestingly, incubation of the osteoblasts that were adherent on tissue culture plastic alone with the RGDS peptide lead to an induction in opn expression suggesting that integrin occupancy by itself was sufficient to initiate the signal transduction process that induced opn expression. In order to assess the role of integrin occupancy vs. focal adhesion complex formation that accompanies cell attachment, in the signal transduction process that induces opn expression, receptor clustering was stimulated pharmacologically with bombesin or lysophosphatidic acid in osteoblasts attached to tissue culture plastic. Neither compound in the absence of occupancy of the integrin receptors was capable of stimulating opn expression in attached cells, however if the cells were placed in suspension pharmacological mediation of receptor clustering and integrin occupancy were additive in their effect of inducing opn expression. These data demonstrate that induction of opn expression by mechanical stimuli and cell attachment are commonly mediated through integrin receptor(s). However, when cells are attached receptor clustering alone which accompanies focal adhesion formation was incapable of mediating signal transduction suggesting that receptor occupancy was a prerequisite to the signal transduction process that leads to the induction of opn mRNA expression.

Bone sialoprotein (BSP) is an extracellular matrix protein that has a highly restricted expression to mineralized skeletal tissues. The chicken bone sialoprotein-encoding gene (bsp) was isolated and shown to contain two less exons than similar mammalian genes, with the absence of an untranslated 5' exon and the fusion of the first two exons that encode the signal peptide and amino terminal end of the mature BSP peptide. Primer extension analysis showed one strong transcriptional start point (tsp) in mRNA prepared from embryonic bone. Comparison of the avian bsp promoter sequence to those of other genes expressed in vertebrate skeletal tissues, identified the presence of homeobox protein binding sequence motifs for engrailed (en-1) and Msx 2 (Hox 8.1), and two collagen type II gene silencer elements. Two TATA sequences one at -21 bp and the second at -172 bp to the tsp were identified. For the first TATA element no CCAAT sequence was observed at an appropriate cis position however two Sp1 sequences (GGGCGG) were identified at -66 and -85 bp. A CCAAT element was seen in an appropriate cis position in relationship to the second upstream TATA, but transient expression analysis in embryonic chicken calvaria osteoblasts using two separate promoter/reporter constructs (+24 to -1244 bp or -121 to -1244 bp), confirmed that only the proximal TATA and Sp1 elements were functional. The +24 to -1244 bp promoter sequence demonstrated 33.6, 13.2, and 3.2 fold activity above base line respectively, within cells prepared from embryonic chicken calvaria bone, cephalic sterna, a cartilage that undergoes mineralization and caudal sterna, a cartilage that does not mineralize during embryogenesis. Only base line activity was observed within cells prepared from embryonic dermal fibroblasts a non-skeletal tissue, which does not express BSP. These same cells demonstrated comparable steady state mRNA levels, corroborating that this segment of promoter DNA had tissue specific activity. A series of nested deletions from the 5' end of the -1244 construct demonstrated that a portion of the tissue specific regulation was controlled by the presence of a silencer element(s) between -1244 and -620 bp since deletion of this segment of DNA resulted in a 6 fold increase in the promoter activity in dermal skin fibroblasts. The -1244-(+)24 nt promoter construct was shown to be stimulated by dexamethasome approximately 1.5 fold over control, inhibited by 1,25(OH)2D3 approximately 60% of control and was strongly stimulated approximately 5.0 fold by parathyroid hormone (PTH) in embryonic calvaria osteoblasts. These data define the proximal promoter of the avian bsp gene and identify several potential regulatory elements that have been observed in the promoters of other genes expressed in skeletal tissues. These elements imparted both tissue and hormone specific promoter activity to bsp expression within skeletal cells.

Endochondral bone formation is one of the most extensively examined developmental sequences within vertebrates. This process involves the coordinated temporal/spatial differentiation of three separate tissues (cartilage, bone, and the vasculature) into a variety of complex structures. The differentiation of chondrocytes during this process is characterized by a progressive morphological change associated with the eventual hypertrophy of these cells. These cellular morphological changes are coordinated with proliferation, a columnar orientation of the cells, and the expression of unique phenotypic properties including type X collagen, high levels of bone, liver, and kidney alkaline phosphatase, and mineralization of the cartilage matrix. Several studies indicate that hypertrophic chondrocytes also express osteocalcin, osteopontin, and bone sialoprotein, three proteins which until very recently were widely believed to be restricted in their expression to osteoblasts. Recent studies suggest that the hypertrophic chondrocytes are regulated by the calcitropic hormones, morphogenic steroids, and local tissue factors. These considerations are based on the regulation by 1,25 (OH)2D3 and retinoids of the cartilage specific genes as well as osteopontin and osteocalcin expression in hypertrophic chondrocytes. They are also based on the effects on growth plate development caused by 1) transgenic ablation of autocrine/paracrine regulators such as PTHrP and of the transcriptional regulator c-fos and 2) naturally occurring genetic mutations of the FGF receptor. These studies further suggest that specific transcriptional factors mediate exogenous regulatory signals in a coordinated manner with the development of bone. While it has been widely demonstrated that the majority of hypertrophic chondrocytes undergo apoptosis during terminal stages of the developmental sequence, their response to specific exogenous regulatory signals and their expression of bone-specific proteins give rise to questions about whether all growth chondrocytes have the same developmental fates and have identical functions. Furthermore, specific questions arise as to whether there are similar mechanisms of regulation for commonly expressed genes found in both cartilage and bone or whether these genes have unique regulatory mechanisms in these different tissues. These recent findings suggest that hypertrophic chondrocytes are functionally coupled during endochondral bone formation to the recruitment of osteoblasts, vascular cells, and osteoclasts.

Seventeen day chicken embryonic osteoblasts treated over a 30-day period with 1,25(OH)2 D3 showed a 2-10-fold decrease in collagen, osteopontin and osteocalcin protein accumulation, alkaline phosphatase enzyme activity, and mineral deposition. Comparable inhibition in the steady state mRNA levels for alpha 1(I) and alpha 2(I) collagen, osteocalcin, and osteopontin were observed, and the inhibitory action of the hormone was shown to be specific for only the late release populations of cells from sequential enzyme digestions of the chick calvaria. In order to determine whether the continuous hormone treatment blocked osteoblast differentiation, the cells were acutely treated for 24 h with 1,25(OH)2 D3 at culture periods when the cells proliferate (day 5), a culture period when the cells cease further cell division and are increasing in the expression of their differentiated functions (day 17), and a culture period when the cells are encapsulated within a mineralized extracellular matrix (day 30). Inhibition of the expression of collagen, osteocalcin, and osteopontin were observed at days 17 and 30, while no effect could be detected for the 5-day cultures. To further define whether the inhibitory effect was specific for cells expressing their differentiated phenotype, 1,25(OH)2 D3 treatment was initiated at day 17 and continued to day 30 after the cells have established their collagenous matrix. In these experiments further collagenous matrix deposition, mineral deposition, alkaline phosphatase activity, and osteocalcin synthesis were also inhibited after the hormone treatment was initiated. These results, in summary, show that 1,25(OH)2 D3 in primary avian osteoblast cultures derived from 17-day embryonic calvaria inhibits the expression of several genes associated with differentiated osteoblast function and inhibit extracellular matrix mineral deposition.

Endochondral bone formation occurs by a series of developmentally regulated cellular events from initial formation of cartilage tissue to stages of calcified cartilage, resorption, and replacement by bone tissue. Several studies have raised the question of the possibility that the hypertrophic chondrocytes associated with the calcifying cartilage matrix can acquire properties similar to osteoblasts. We have addressed this possibility by measuring synthesis within hypertrophic chondrocytes in vitro of two bone-related proteins, osteopontin and osteocalcin. Chondrocytes derived from chick embryo ventral vertebral tissue were cultured under conditions that promoted extracellular matrix mineralization and differentiation towards the hypertrophic phenotype as indicated by the induction of Type X collagen, alkaline phosphatase, and diminished expression of Type II collagen and the core protein of large proteoglycan. In these cultures, osteopontin synthesis was detected in early cultures in the absence of a calcified matrix; in contrast, an absence of the bone-specific protein osteocalcin was observed. However, with onset of development of the hypertrophic phenotype an induction of protein expression for osteocalcin was observed with a significant (twofold) increase in osteopontin. Maximal levels of osteocalcin synthesis occurred with the peak of alkaline phosphatase activity and Type X collagen mRNA levels. The levels of osteocalcin synthesis were induced fiftyfold from the earliest level of detection but this level was only one one-hundredth of that observed for mature chick osteoblast cultures. Osteocalcin and osteopontin were characterized by several criteria (electrophoresis, immunoblotting, chromatographic characteristics, and response to 1,25(OH)2D3) which confirmed their molecular properties as being identical to osteoblast synthesized proteins. The coordinate change in the cellular phenotype to the hypertrophic chondrocyte was shown to be concurrent with ultrastructural maturation of the cells and the accumulation of osteocalcin and osteopontin in the extracellular matrix associated with hydroxyapatite at sites of mineralization. Since the ultrastructural features of the cells in vitro and the extracellular matrix surrounding the lacunae have features of the hypertrophic chondrocyte and associated matrix in vivo, the induction of the bone-specific protein osteocalcin suggests that at least a population of these cells may develop osteoblastic phenotypic markers in association with mineralizing matrix. The detection of osteocalcin and the high level of synthesis of osteopontin may represent an advanced stage of chondrocyte hypertrophy or the possibility of a trans-differentiation of the chondrocytes to an osteoblastic-like cell.