Fetal bovine bone cells synthesize bone-specific matrix proteins.

We isolated cells from both calvaria and the outer cortices of long bones from 3- to 5-mo bovine fetuses. The cells were identified as functional osteoblasts by indirect immunofluorescence using antibodies against three bone-specific, noncollagenous matrix proteins (osteonectin, the bone proteoglycan, and the bone sialoprotein) and against type 1 collagen. In separate experiments, confluent cultures of the cells were radiolabeled and shown to synthesize and secrete osteonectin, the bone proteoglycan and the bone sialoprotein by immunoprecipitation and fluorography of SDS polyacrylamide gels. Analysis of the radiolabeled collagens synthesized by the cultures showed that they produced predominantly (approximately 94%) type I collagen, with small amounts of types III and V collagens. In agreement with previous investigators who have employed the rodent bone cell system, we confirmed in bovine bone cells that (a) there was a typical cyclic AMP response to parathyroid hormone, (b) freshly isolated cells possessed high levels of alkaline phosphatase, which diminished during culture but returned to normal levels in mineralizing cultures, and (c) cells grown in the presence of ascorbic acid and beta-glycerophosphate rapidly produced and mineralized an extracellular matrix containing largely type I collagen. These results show that antibodies directed against bone-specific, noncollagenous proteins can be used to clearly identify bone cells in vitro.

Osteoblasts synthesize and respond to transforming growth factor-type beta (TGF-beta) in vitro.

Transforming growth factor-type beta (TGF-beta) has been identified as a constituent of bone matrix (Seyedin, S. M., A. Y. Thompson, H. Bentz, D. M. Rosen, J. M. McPherson, A. Conti, N. R. Siegel, G. R. Gallupi, and K. A. Piez, 1986, J. Biol. Chem. 261:5693-5695). We used both developing bone and bone-forming cells in vitro to demonstrate the cellular origin of this peptide. TGF-beta mRNA was detected by Northern analysis in both developing bone tissue and fetal bovine bone-forming cells using human cDNA probes. TGF-beta was shown to be synthesized and secreted by metabolically labeled bone cell cultures by immunoprecipitation from the medium. Further, TGF-beta activity was demonstrated in conditioned media from these cultures by competitive radioreceptor and growth promotion assays. Fetal bovine bone cells (FBBC) were found to have relatively few TGF-beta receptors (5,800/cell) with an extremely low Kd of 2.2 pM (high binding affinity). In contrast to its inhibitory effects on the growth of many cell types including osteosarcoma cell lines, TGF-beta stimulated the growth of subconfluent cultures of FBBC; it had little effect on the production of collagen by these cells. We conclude that bone-forming cells are a source for the TGF-beta that is found in bone, and that these cells may be modulated by this factor in an autocrine fashion.

Infrared analysis of rat bone: age dependency of amorphous and crystalline mineral fractions.

Quantitative infrared spectrophotometric analysis of whole femurs from male rats demonstrates that anorphous calcium phosphate is a major component of bone mineral. The amount of amorphous calcium phosphate in whole bone decreases while the crystalline bone apatite increases during early stages of bone formation. Mature rat bone contains constant levels of both amorphous and crystalline calcium phosphate.

Experimental renal osteodystrophy. The response to 25-hydroxycholecalciferol and dicholomethylene diphosphate therapy.

Bone mineral and matrix maturation in chronically uremic, nonacidotic rats were investigated after 25-hydroxcholecalciferol (25OHD) and/or dichloromethylene diphosphonate (C12MDP) therapy utilizing bromoform-toluene density gradient fractionation and X-ray diffraction analyses. The bromoform-toluene density gradient analyses demonstrated that the progressive accumulation of less dense, more immature bone characteristic of progressive uremia was reversed by 25OHD and/or C12MDP therapy for a 2-wk period, and that after 4 wk of therapy the maturational profile of bones from chronically uremic animals treated with 250HD and/or C12MDP was comparable to that from nonuremic littermates. X-ray diffraction analysis revealed that by the 4th wk of therapy with 25OHD and C12MDP both the degree of crystallinity and the crystal size/perfection parameters in the uremic bones were comparable to those of nonuremic, pair-fed control littermates. Treatment for 4 wk with 25OHD resulted in enlarged and/or more perfect apatite crystallites, while C12MDP alone slightly inhibited crystal growth and/or perfection after 2 wk of treatment. Soft tissue calcification was diminished in uremic animals treated for 4 wk with C12MDP or a combined C2MDP/25OHD regimen, the latter being much more effective in this regard. The accumulated data in this study support the premise that the attendant accelerated bone resorption, soft tissue calcification, and abnormal mineralization and maturation of the skeletal tissue, well documented to characterize experimental ranal insufficiency, may be alleviated with therapeutic dosages of 25OHD and/or C12MDP.

Abnormal bone mineral maturation in the chronic uremic state.

X-ray diffraction analysis of bone from chronically uremic but nonacidotic rats with normocalcemia and hyperphosphatemia revealed smaller apatite crystals and an increase in the X-ray amorphous mineral fraction when compared to age-matched, pair-fed control animals, indicating less advanced mineral maturation in the uremic animals. Studies in animals with varied degrees of chronic renal insufficiency revealed a progression of the bone crystal maturational defect with advancing uremia.

The effects of phosphoproteins on collagen self-assembly in tail tendon and incision dentin from rats.

Phosphoproteins retard the rate at which collagen molecules undergo self-assembly into fibrils. The inhibition appears to be dependent on the amount of phosphoprotein present, with increasing phosphoprotein concentrations resulting in greater inhibition. Prior treatment of the phosphoprotein with calcium markedly increases the resultant inhibitory effect. Dentin phosphoproteins are considerably more effective than phosvitin in retarding collagen self-assembly, with retardation times for these hard tissue extracellular matrix proteins being 25-30 times greater than control values.

Bone matrix mRNA expression in differentiating fetal bovine osteoblasts.

In the accompanying study, we report an in vitro culture system from bovine bone cells that can be applied to investigate bone cell growth and differentiation. In this system, bovine bone cells placed in mineralization medium formed multilayers (days 2-3), began deposition of mineral (days 5-6), and eventually acquired a mineralized matrix sheet (days 14-20) through the stages of mineralizing nodules and trabecular-like structure. In the current study we used this system to investigate the relative expression of bone matrix genes that may play an important role in bone development and metabolism. alpha 1(I)-collagen, alkaline phosphatase, osteonectin, biglycan (PgI), decorin (PgII), osteopontin, and bone sialoprotein mRNA gene expression were measured on days 0, 2, 6, 10, and 20 (date when the cells were placed in mineralization medium as day 0). Total RNA was purified and analyzed by northern blot using radiolabeled cDNA encoding these genes. To comprehend the relationship between gene expression and mineralization, total calcium content in the cultures was also measured. During the culture period we observed several very different gene expression profiles. The expression of both alpha 1(I)-collagen and biglycan increased 3- to 4-fold by day 6 and then returned to basal levels by day 20. The osteonectin gene was highly expressed throughout the culture, with no significant increase in induction found during any time of culture. A significant induction of alkaline phosphatase (13.8-fold) gene expression was observed by day 6. Osteopontin showed a similar profile to that of alkaline phosphatase but had a much greater level of relative expression (26-fold) compared to day 0. Interestingly, downregulation during mineral accumulation seemed a common occurrence among many of the genes measured. In contrast, the bone sialoprotein gene showed a significant and distinct expression pattern, increasing rapidly after the onset of mineralization on day 6 and ultimately reaching 140-fold that of day 0. Decorin (Pg II) showed an increasing pattern, with the final relative level of induction 5-fold on day 20. These data suggest that the development of the mature osteoblastic phenotype, complete with the ability to produce a thick mineralized matrix, requires the differential regulation of a series of genes and their gene products over the culture period.

Expression of the osteonectin gene potentially controlled by multiple cis- and trans-acting factors in cultured bone cells.

The cis-acting regulatory elements of the osteonectin gene have been studied using a chloramphenicol acetyltransferase (CAT) promoter assay in osteonectin-expressing and nonexpressing cultured cells. When various stretches of the promoter were transiently transfected into fetal bovine bone cells, a positive element was detected in the DNA located between bases -504 and 11 (1 being the start of transcription) and a negative element between bases -900 and -504. The positive element of the promoter also conferred preferential expression of the gene, showing more activity in cells with higher levels of osteonectin mRNA expression. A 1.2 kb fragment of intron 1 displayed a negative effect on CAT expression when inserted 5' to the promoter. An additional regulatory element was found in DNA encoding exon 1, which significantly influenced expression of the gene in fetal bovine bone cells. Gel shift analysis using positive genomic elements located 5' to the start of transcription indicated that one of the nuclear proteins that interacts with the osteonectin promoter may be related to the transcription factor AP2.

Osteogenesis imperfecta: changes in noncollagenous proteins in bone.

The noncollagenous proteins osteonectin, bone sialoprotein, osteocalcin, the small proteoglycan decorin (PG II), and alpha 2-HS glycoprotein (which is synthesized in the liver but highly concentrated in bone) were measured in extracts of cortical bone from 3 type I, 2 type II, 8 type III and 13 type IV patients with osteogenesis imperfecta (OI) and from 7 control subjects. Osteonectin was found to be reduced in the bone of all OI patients. The bone from severely affected type III OI patients contained the lowest levels of osteonectin. In contrast, bone sialoprotein was found to be elevated in the bones of OI patients. The highest levels were found in individuals classified as type IV patients. Osteocalcin and alpha 2-HS glycoprotein concentrations were increased in all OI patients. Decorin levels were not significantly altered in OI bones compared to controls. These changes in the concentrations of the noncollagenous proteins may contribute to the fragility of the OI bone by interfering with complete mineralization and/or normal tissue architecture.

Purification and fragmentation of nondenatured bone sialoprotein: evidence for a cryptic, RGD-resistant cell attachment domain.

Bone sialoprotein (BSP), a small (approximately 80,000 M(r)) integrin binding, RGD-containing bone matrix glycoprotein, has been purified in milligram quantities from the serum-free medium of the rat osteosarcoma cell line UMR-106-BSP using nondenaturing conditions. Routine protein purification without serine protease inhibitors or reducing agents consistently resulted in three major fragments. The largest fragment (E1) started at amino acid 117 and did not bind to antibodies made to the RGD region of the protein. Furthermore, the smallest fragment (E3), was shown by sequencing to contain the RGD region of the protein. Digestion of intact BSP with highly purified chymotrypsin also resulted in a large fragment (C1) with properties nearly identical to those of E1. The large, non-RGD-containing fragments, E1 and C1, as well as the intact BSP, supported attachment by normal human bone cells and human skin fibroblasts in vitro. Attachment to the intact BSP was totally blocked by 0.4 mM GRGDS peptide. Both preparations of skin fibroblasts and approximately half of the preparations of normal human bone cells, however, also would not attach to the E1 and C1 fragments in the presence of 0.4 mM GRGDS peptide. In contrast, half of the bone cell preparations had significant attachment activity to E1 (> 50%) and C1 (> 25%) in the presence of 0.4 mM GRGDS peptide. These data suggest that cleavage of the BSP results in either (1) the exposure of a previously unavailable or cryptic cell attachment site or (2) a conformational change that increases the affinity of the complex between a non-RGD-encoded binding region of the E1 and C1 fragments and at least one receptor. The possible homology of the second, non-RGD-suppressible site of BSP with the second cell attachment site on the gamma chain of fibrinogen is discussed.

Estrogen and raloxifene stimulate transforming growth factor-beta 3 gene expression in rat bone: a potential mechanism for estrogen- or raloxifene-mediated bone maintenance.

Estrogen or raloxifene (LY156758) prevent estrogen deficiency-induced bone loss in animals and humans. We demonstrated in the rat that a 22% reduction in bone mineral density generated by ovariectomy was associated with a 2-fold reduction of transforming growth factor-beta 3 (TGF beta 3) messenger RNA expression in the femur. Administration of 17 beta-estradiol or raloxifene to ovariectomized rats restored both bone mineral density and TGF beta 3 messenger RNA expression in the femur to levels measured in intact animals. In transient transfection assays, the promoter sequence from -38 to + 110 of the human TGF beta 3 gene, which contains no palindromic estrogen response element, was sufficient to mediate 17 beta-estradiol or raloxifene induced-reporter gene expression in presence of the estrogen receptor. Raloxifene activated TGF beta 3 promoter as a full agonist at nanomolar concentrations. In the same cellular system, raloxifene inhibited the estrogen response element-containing vitellogenin promoter expression as a pure estrogen antagonist. In two well characterized osteoclast differentiation models, TGF beta 3 significantly inhibited the differentiation and bone-resorptive activities of murine and avian osteoclasts. These findings suggest that regulation of TGF beta 3 gene expression by raloxifene or estrogen in bone may be an important target to mediate bone maintenance.

The cDNA cloning and RNA distribution of bovine osteopontin.

We have isolated and sequenced the bovine cDNA (OPN) counterpart of osteopontin. The cDNA is 1356 nucleotides (nt) in length with an open reading frame of 834 nt, encoding a 278-amino acid (aa) protein. Cell-free transcription and translation of OPN RNA resulted in a major species of approx. 40 kDa in size, in agreement with the predicted size of the deduced aa sequence. Northern analysis of bovine OPN RNA indicated the presence of the message in mineralized, as well as soft tissues. A comparison of the deduced aa sequence among various species indicates both regions of similarity and divergence. One prominent region of dissimilarity in bovine OPN compared to all other species is a 22-aa gap which may represent a loss of a potential Ca(2+)-binding loop. Despite the variability among the species, several regions of conservation are apparent, including a hydrophobic leader sequence, a potential site for Asn-linked glycosylation, a stretch of polyaspartic acid residues, and the cell attachment Arg-Gly-Asp tripeptide. Whether bovine OPN enhances cell attachment is unknown. Furthermore, whether the loss of a potential Ca(2+)-binding loop alters the function of OPN would be interesting to determine.

Matrix protein profiles in calf bone development.

The major noncollagenous proteins (NCP) of the mineralized extracellular matrix of bovine bone were present from early fetal life through adulthood. The alpha 2-HS-glycoprotein was the dominant bone NCP at 2 months in utero and then diminished with advancing fetal age. Osteocalcin levels were quite low in early fetal bone, became elevated near term, reached a maximum shortly after birth, and dropped again in adult bone. Osteonectin levels were relatively constant at all stages of bone development, but somewhat lower values were seen in woven bone than in lamellar bone matrix. For a given fetus, NCP profiles were identical in both flat and long bones. The 24,000 dalton phosphoprotein was significantly elevated in fetal secondary spongiosa, bony areas known to be undergoing extensive resorption. Adult bone matrix contained greater quantities of lower and intermediate molecular weight NCPs compared to actively forming fetal bone.

The selective estrogen receptor modulator raloxifene regulates osteoclast and osteoblast activity in vitro.

Raloxifene is a selective estrogen receptor modulator (SERM) that prevents bone loss. Although it is largely used for the treatment of osteoporosis, the mechanisms by which this compound modulates the activity of bone cells are still poorly understood. In this study we investigate whether raloxifene affects osteoclast and osteoblast activity in vitro. Bone marrow cultures were established from neonatal mice and treated with 1,25(OH)(2) vitamin D(3) (VitD(3), 10(-8) mol/L) to induce osteoclast generation. Similar to 17beta-estradiol, raloxifene significantly reduced the number of osteoclasts in a concentration-dependent manner, with maximal inhibition at 10(-11) mol/L (-48%). However, as for 17beta-estradiol, at a high concentration (10(-7) mol/L), the inhibitory effect of raloxifene was abolished. In a pit assay, raloxifene inhibited bone resorption. A maximal effect was observed at 10(-9) mol/L, and maintained at a high concentration, indicating that inhibition of osteoclast formation and inhibition of bone resorption may be due to activation of, at least in part, different pathways. Osteoblasts from neonatal mice calvariae were also exposed to raloxifene. In these cells, this compound induced a concentration-dependent increase of proliferation, which was blocked by the estrogen-receptor antagonist ICI 164,384. Raloxifene also increased the osteoblast-specific transcription factor Cbfa1/Runx2 and alpha2 procollagen type I chain mRNAs, with a pattern that only partially coincided with that of 17beta-estradiol. Consistent with decreased osteoclastogenesis, raloxifene inhibited the mRNA expression of interleukin (IL)-1beta and IL-6 at a low concentration, but not at a high concentration, whereas 17beta-estradiol had similar effects on IL-6 and inhibited IL-1beta at both concentrations. Furthermore, both compounds were able to inhibit tumor necrosis factor (TNF)-alpha-induced IL-1beta, but not IL-6, increase. In conclusion, these data show that raloxifene negatively modulates osteoclasts, and positively affects osteoblasts, suggesting not only an antiresorptive role, but also an osteoblast stimulatory role.

Expression and localization of the two small proteoglycans biglycan and decorin in developing human skeletal and non-skeletal tissues.

The messenger RNAs and core proteins of the two small chondroitin/dermatan sulfate proteoglycans, biglycan and decorin, were localized in developing human bone and other tissues by both 35S-labeled RNA probes and antibodies directed against synthetic peptides corresponding to nonhomologous regions of the two core proteins. Biglycan and decorin expression and localization were substantially divergent and sometimes mutually exclusive. In developing bones, spatially restricted patterns of gene expression and/or matrix localization of the two proteoglycans were identified in articular regions, epiphyseal cartilage, vascular canals, subperichondral regions, and periosteum, and indicated the association of each molecule with specific developmental events at specific sites. Study of non-skeletal tissues revealed that decorin was associated with all major type I (and type II) collagen-rich connective tissues. Conversely, biglycan was expressed and localized in a range of specialized cell types, including connective tissue (skeletal myofibers, endothelial cells) and epithelial cells (differentiating keratinocytes, renal tubular epithelia). Biglycan core protein was localized at the cell surface of certain cell types (e.g., keratinocytes). Whereas the distribution of decorin was consistent with matrix-centered functions, possibly related to regulation of growth of collagen fibers, the distribution of biglycan pointed to other function(s), perhaps related to cell regulation.

Bone sialoprotein (BSP) secretion and osteoblast differentiation: relationship to bromodeoxyuridine incorporation, alkaline phosphatase, and matrix deposition.

We defined two distinct maturational compartments (proliferative and secretory) of osteogenic cells in vivo on the basis of ALP activity, BrdU incorporation, cell shape, and BSP production. BSP immunoreactivity was found to mark cells in the secretory but not in the proliferative compartment. We established the phenotypic similarity of primitive marrow stromal cells with proliferating perichondral cells (fibroblast-like, ALP+, BrdU+, BSP-). This suggests the potential functional equivalence of the two cell types as committed non-secretory osteogenic cells and points to the duality of osteogenic cell compartments as a generalized feature of bone formation. We further showed that although BSP secretion is a hallmark of the onset of osteogenesis, BSP antigenicity is lost both in osteoid and in a large proportion of mature osteoblasts during subsequent phases of bone deposition. This suggests that bone formation may not be a uniform event, as bone cells actually deposit antigenically, and likely biochemically, distinct matrices at specific times.

Localization of bone sialoprotein (BSP) to Golgi and post-Golgi secretory structures in osteoblasts and to discrete sites in early bone matrix.

Bone sialoprotein (BSP), a bone matrix-enriched glycoprotein containing the Arg-Gly-Asp (RGD) motif and endowed with cell binding properties, was localized in osteoblasts and early bone matrix of developing rat bone at the ultrastructural level. Preliminary light microscopic observations indicated that intracellular labelling was restricted to a paranuclear dot corresponding to the "negative Golgi image" of classical histology. The same pattern was observed whether antisera against the fully glycosylated protein or a peptide antiserum to a stretch of amino acids in human BSP sequence were employed. At the EM level, we obtained labeling over the Golgi area of osteoblasts but not over the rER. The labeling was concentrated over distensions of the trans Golgi and over pro-secretory granules. In the matrix, BSP was distributed in a non-random manner. The label was concentrated over spherical aggregates of finely fibrillar material corresponding to the sites of early mineral deposition (so-called "mineralization nodules"). Such BSP-positive foci were seen both close to and away from the cell surface. The predominant association of BSP with Golgi and post-Golgi secretory structures and its absence from rER, as well as the reproducibility of the same pattern of localization with different antisera, might indicate a slow transit of the protein through the Golgi, not necessarily associated with protein glycosylation.

Cellular activity, matrix proteins, and aging bone.

The extracellular matrix of bone is composed of proteins from both local and exogenous sources. Many of these are known growth factors (e.g., transforming growth factors-beta; insulin-like growth factors; and fibroblast growth factors) which concentrate in mineralized bone and probably contribute to the ability of bone to regenerate itself on injury. With advancing age, human osteoblasts have reduced bone formative properties and in vitro, osteoblasts from fetuses generally have greatly increased proliferative and biosynthetic capacities compared to cells from adult donors. Aside from type I collagen, many noncollagen components are synthesized by osteoblasts and secreted to the bone matrix space. Several of these are cell attachment proteins (e.g., fibronectin, thrombospondin, osteopontin) which greatly influence cytodevelopment and differentiation. They are degraded to lower molecular weight fragments with advancing age, probably deactivating their true bioactivities. It is not know whether the age-related degradation of these proteins affects bone cell function in aged individuals. Several of the bone matrix proteins are also found in platelets and have been implicated in the wound-repair process. One of these, osteonectin, is found in a wide variety of nonbone cell systems, but only in periods of rapid growth and proliferation. Osteonectin production is the highest and is maintained the longest in bone compared to all other tissues of the body. Thus, reduced osteonectin production in aged bone cells may be an important parameter for further study.